Effect and safety of salt iodization to prevent iodine deficiency disorders: a systematic review with meta-analyses

Abstract

Background: Iodine deficiency, one of the most prevalent micronutrient deficiencies globally, is
the main cause of potentially preventable mental retardation in childhood, as well as a spectrum of
morbidities referred to as iodine deficiency disorders. Iodization of salt is recommended to prevent
and treat many of these disorders.
Objective: To assess the effects and safety of consumption of iodized salt in the prevention of
iodine deficiency disorders.
Data sources: The following databases were searched: China National Knowledge Infrastructure,
the Cochrane Library, EMBASE, MEDLINE, VIP (the register of Chinese trials developed by the
Chinese Cochrane Centre), The Virtual Health Library of the Pan-American Health Organization,
WANFANG, the World Health Organization (WHO) International Clinical Trials Platform search
portal, and the WHO Global Health Library regional databases. Relevant agencies were contacted,
and the reference lists were reviewed.
Study appraisal and synthesis methods: Randomized controlled trials (RCTs), non-RCTs, quasiexperimental,
cohort, and multiple cross-sectional studies were included. All studies compared
a group of individuals with exposure to iodized salt to a group without exposure to iodized salt.
Two or more reviewers independently screened potential studies, extracted study characteristics
and outcomes and, when possible, conducted meta-analyses to estimate the effect of iodized
salt relative to non-iodized salt. Results are presented as mean differences (MDs), risk ratios (RRs),
prevalence ratios (PRs), or odds ratios (ORs), with 95% confidence intervals (CIs).
Results: Two RCTs, 6 non-RCTs, 20 quasi-experimental studies, 16 cohort studies, 42 multiple crosssectional
studies, and 3 studies with mixed designs were included. The numbers of participants
ranged from 30 in a cohort study to over 5 000 000 in a multiple cross-sectional registry study.
Iodized salt significantly reduced the risk of goitre (non-RCTs RR = 0.59 [95% CI = 0.36 to 0.95];
cohort RR = 0.30 [95% CI = 0.23 to 0.41]; multiple cross-sectional PR = 0.18 [95% CI = 0.14 to 0.22]),
cretinism (multiple cross-sectional OR = 0.13 [95% CI = 0.08 to 0.20]), low intelligence (quasiexperimental
RR = 0.28 [95% CI = 0.21 to 0.36]; multiple cross-sectional PR = 0.24 [95% CI = 0.07 to
0.82]), and low urinary iodine excretion (multiple cross-sectional PR = 0.45 [95% CI = 0.34 to 0.59]).
Iodized salt significantly increased intelligence quotient (quasi-experimental MD = 8∙18 [95%
CI = 6∙71 to 9∙65]; multiple cross-sectional MD = 10.45 [95% CI = 4∙79 to 16∙11]) and urinary iodine
excretion (cohort MD = 59.22 [95% CI = 50.40 to 68.04]; multiple cross-sectional MD = 72∙35 [95%
CI = 44∙54 to 100.17]). The results regarding the potential adverse effect of hypothyroidism showed
no relationship (cohort OR = 1.14 [95% CI = 0.84 to 1.53]; multiple cross-sectional OR = 1.13 [95%
CI = 0.94 to 1.36]), and the results for hyperthyroidism were inconsistent and dependent on study
design (cohort OR = 1.36 [95% CI = 1.12 to 1.66]; multiple cross-sectional OR = 0.96 [95% CI = 0.92
to 1.00]). The quality of evidence varied from very low to moderate, depending on outcome and
study design.
Conclusions: This review showed that iodized salt has a large effect on reducing the risk of goitre,
cretinism, low cognitive function and iodine deficiency. Robust monitoring of salt iodization
programmes is important, to ensure safe and effective levels of iodine consumption, especially as
countries implement programmes to reduce population salt intake

Full text

Eect and safety of salt iodization
to prevent iodine deciency
disorders: a systematic review
with meta-analyses
Dr Nancy J Aburto1
Dr Minawaer Abudou2
Ms Vanessa Candeias3
Professor Tiaxiang Wu4
1 United Nations World Food Programme, Via Cesare Giulio Viola 68/70,
00148 Rome, Italy
Email: nancy.aburto@wfp.org
2 The Eye Department of the First Aliated Hospital, Xinjiang Medical
University, Xinjiang, China
3 Healthy Living, World Economic Forum, 91–93 route de la Capite,
1223 Cologny, Switzerland
4 Professor, Administrator, CEO Head of Chinese Clinical Trial Registry,
West China Hospital, Sichuan University, Chengdu, China

WHO Library Cataloguing-in-Publication Data
Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review
with meta-analyses.
1.Sodium Chloride, Dietary. 2.Iodine. 3.Deciency Diseases – prevention and control. 4.Food,
Fortied. 5.Review. 6.Meta-Analysis. I.Aburto, Nancy J. II.Abudou, Minawaer. III.Candeias, Vanessa.
IV.Wu, Tiaxiang. V.World Health Organization.
ISBN 978 92 4 150828 5 (NLM classication: QV 283)
© World Health Organization 2014
All rights reserved. Publications of the World Health Organization are available on the WHO web site
(www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211
Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: bookorders@who.int).
Requests for permission to reproduce or translate WHO publications – whether for sale or for non-
commercial distribution – should be addressed to WHO Press through the WHO web site
(www.who.int/about/licensing/copyright_form/en/index.html).
The designations employed and the presentation of the material in this publication do not imply the
expression of any opinion whatsoever on the part of the World Health Organization concerning the
legal status of any country, territory, city or area or of its authorities, or concerning the delimitation
of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines
for which there may not yet be full agreement.
The mention of specic companies or of certain manufacturers’ products does not imply that
they are endorsed or recommended by the World Health Organization in preference to others of
a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary
products are distinguished by initial capital letters.
All reasonable precautions have been taken by the World Health Organization to verify the
information contained in this publication. However, the published material is being distributed
without warranty of any kind, either expressed or implied. The responsibility for the interpretation
and use of the material lies with the reader. In no event shall the World Health Organization be
liable for damages arising from its use.
The named authors alone are responsible for the views expressed in this publication.
Design and layout: Alberto March
Suggested citation
Aburto N, Abudou M, Candeias V, Wu T. Eect and safety of salt iodization to prevent iodine
deciency disorders: a systematic review with meta-analyses. WHO eLibrary of Evidence for
Nutrition Actions (eLENA). Geneva: World Health Organization; 2014.

3
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Contents
Abstract 4
Plain language summary 5
Background 6
Methods 7
Study characteristics 7
Search methods 7
Data collection and analysis 7
Selection of studies 7
Data extraction and management 8
Statistical analysis 8
Results 9
Eect estimates 9
Goitre prevalence 9
Cretinism 10
Cognitive function 10
Urinary iodine excretion 11
Thyroid hormones 11
Other adverse outcomes 12
Other outcomes 12
Quality of the body of evidence 12
Discussion 13
Limitations of the study 15
Conclusion 15
Contributions 16
Acknowledgements 16
Declaration of competing interests 16
References 24
Annex 1. Electronic search strategy 33
Aneex 2. Characteristics of included studies 34
Annex 3. Summary table: goitre prevalence with and without iodization of salt 129
Annex 4. Summary tables: urinary iodine excretion with and without iodized salt 130
Annex 5. Summaries of the risk of bias of included studies 134
Annex 6. GRADE summary of evidence 142

4
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Abstract
Background: Iodine deciency, one of the most prevalent micronutrient deciencies globally, is
the main cause of potentially preventable mental retardation in childhood, as well as a spectrum of
morbidities referred to as iodine deciency disorders. Iodization of salt is recommended to prevent
and treat many of these disorders.
Objective: To assess the eects and safety of consumption of iodized salt in the prevention of
iodine deciency disorders.
Data sources: The following databases were searched: China National Knowledge Infrastructure,
the Cochrane Library, EMBASE, MEDLINE, VIP (the register of Chinese trials developed by the
Chinese Cochrane Centre), The Virtual Health Library of the Pan-American Health Organization,
WANFANG, the World Health Organization (WHO) International Clinical Trials Platform search
portal, and the WHO Global Health Library regional databases. Relevant agencies were contacted,
and the reference lists were reviewed.
Study appraisal and synthesis methods: Randomized controlled trials (RCTs), non-RCTs, quasi-
experimental, cohort, and multiple cross-sectional studies were included. All studies compared
a group of individuals with exposure to iodized salt to a group without exposure to iodized salt.
Two or more reviewers independently screened potential studies, extracted study characteristics
and outcomes and, when possible, conducted meta-analyses to estimate the eect of iodized
salt relative to non-iodized salt. Results are presented as mean dierences (MDs), risk ratios (RRs),
prevalence ratios (PRs), or odds ratios (ORs), with 95% condence intervals (CIs).
Results: Two RCTs, 6 non-RCTs, 20 quasi-experimental studies, 16 cohort studies, 42 multiple cross-
sectional studies, and 3 studies with mixed designs were included. The numbers of participants
ranged from 30 in a cohort study to over 5000000 in a multiple cross-sectional registry study.
Iodized salt signicantly reduced the risk of goitre (non-RCTs RR = 0.59 [95% CI= 0.36 to 0.95];
cohort RR = 0.30 [95% CI=0.23 to 0.41]; multiple cross-sectional PR = 0.18 [95% CI=0.14 to 0.22]),
cretinism (multiple cross-sectional OR = 0.13 [95% CI = 0.08 to 0.20]), low intelligence (quasi-
experimental RR=0.28 [95% CI=0.21 to 0.36]; multiple cross-sectional PR = 0.24 [95% CI=0.07 to
0.82]), and low urinary iodine excretion (multiple cross-sectional PR = 0.45 [95% CI=0.34 to 0.59]).
Iodized salt signicantly increased intelligence quotient (quasi-experimental MD = 818 [95%
CI=671 to 965]; multiple cross-sectional MD = 10.45 [95% CI=479 to 1611]) and urinary iodine
excretion (cohort MD = 59.22 [95% CI=50.40 to 68.04]; multiple cross-sectional MD = 7235 [95%
CI=4454 to 100.17]). The results regarding the potential adverse eect of hypothyroidism showed
no relationship (cohort OR = 1.14 [95% CI=0.84 to 1.53]; multiple cross-sectional OR = 1.13 [95%
CI=0.94 to 1.36]), and the results for hyperthyroidism were inconsistent and dependent on study
design (cohort OR = 1.36 [95% CI=1.12 to 1.66]; multiple cross-sectional OR = 0.96 [95% CI=0.92
to 1.00]). The quality of evidence varied from very low to moderate, depending on outcome and
study design.
Conclusions: This review showed that iodized salt has a large eect on reducing the risk of goitre,
cretinism, low cognitive function and iodine deciency. Robust monitoring of salt iodization
programmes is important, to ensure safe and eective levels of iodine consumption, especially as
countries implement programmes to reduce population salt intake.

5
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Plain language summary
Iodine deciency causes a spectrum of disorders, from poor growth, retarded development, low
urinary iodine excretion and poor cognitive function and goitre, to severe cognitive disability
and death. Consistent consumption of small quantities of iodine can prevent these disorders and
reverse some, but not all, of the negative eects of iodine deciency. In 1993, the World Health
Organization (WHO), the United Nations Children’s Fund (UNICEF) and the International Council
for the Control of Iodine Deciency Disorders (ICCIDD) recommended universal iodization of
salt to prevent and treat iodine deciency disorders, and this recommendation remains today.
Salt is a good vehicle for iodization because it is consumed almost universally without seasonal
variation; there are relatively few production facilities, simplifying quality control; technology for
iodization is well established;consumer acceptability of iodized salt is high; and iodizationis very
inexpensive. However, elevated salt intake results in elevated blood pressure and is associated with
cardiovascular disease, and globally most populations consume salt at levels far exceeding needs.
Therefore, public health policies based on WHO recommendations seek to reduce population salt
intakes.
This systematic review and the corresponding meta-analyses provide the rst comprehensive
synthesis of available data comparing consumption of, or exposure to, iodized salt on an entire
array of iodine deciency disorders, including goitre, urinary iodine excretion, cretinism, cognitive
function and potential adverse eects such as hypothyroidism and hyperthyroidism. Because this
review comprised diverse study designs, including randomized controlled trials, non-randomized
controlled trials, quasi-experimental, cohort observational, and multiple cross-sectional studies,
the study was large enough to explore the eectiveness of iodized salt for preventing iodine
deciency disorders by a number of subgroups, such as age, physiological status, concentration of
iodine in the salt, and risk of iodine deciency disorders at baseline. The additional information on
the large number of outcomes and on potential eect modiers provides important information
for assessing and informing an update of the current WHO guidelines on salt iodization, especially
in the context of renewed eorts to reduce salt consumption globally.

6
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Background
Iodine deciency, one of the most prevalent micronutrient deciencies globally (1), is the main
cause of potentially preventable cognitive disability in childhood (2). Iodine is a trace element that
is an integral part of the thyroid hormones essential for fetal development, regulation of metabolic
activities of cells, and proper growth and development (3). When iodine requirements go unmet,
synthesis of thyroid hormones is impaired, resulting in a spectrum of growth, developmental and
functional abnormalities referred to as iodine deciency disorders (4). Goitre, swelling of the thyroid
gland, is the most common manifestation of iodine deciency in both children and adults (4, 5).
Severe iodine deciency during pregnancy results in fetal death or cretinism, marked by severe
mental and physical growth retardation. Mild and moderate deciency during pregnancy hinders
fetal development, and ospring are at high risk of speech and hearing defects and impaired
motor and physical growth. Iodine deciency during infancy and early childhood can also cause
irreversible decits in cognitive development.
The World Health Organization (WHO) estimates that approximately 37% of school-age
children, 285 million, and nearly 2 billion individuals worldwide, have insucient iodine intake
(6, 7). Mountainous areas are particularly susceptible to iodine deciency, owing to low levels of
iodine in the soil and, thus, the locally grown and raised food (8). However, iodine deciency is
common in many contexts globally and is considered a public health problem in more than 50
countries (6).
WHO, the United Nations Children’s Fund (UNICEF) and the International Council for the
Control of Iodine Deciency Disorders Global Network (ICCIDD) promote salt iodization for
the control of iodine deciency disorders because: (i) salt is widely consumed by virtually all
population groups in all countries, with little seasonal variation in consumption; (ii) salt production
is generally limited to a few centres, facilitating quality control; (iii) technology for salt iodization
is well established and relatively easy to transfer to less developed countries; (iv)iodization does
not aect the organoleptic properties of salt and, therefore, consumer acceptability is high; and
(v)iodization is very inexpensive. To date, more than 120 countries with all levels of deciency
risk are implementing large-scale salt iodization (9). Despite the known benets of appropriate
consumption of iodine, some concern exists that widespread salt iodization could potentially lead
to excess iodine intake (10). Furthermore, though policies to support salt iodization and reduce
sodium intake are compatible, there is concern that, as populations reduce salt intake to reduce
the risk of elevated blood pressure, hypertension and stroke, there may be an increase in iodine
deciency disorders (11–13).
A systematic review of the eectiveness of iodization of salt in preventing iodine deciency
disorders, published in 2002, included data from populations at high risk of iodine deciency, and
controlled trials, and concluded that salt iodization was an eective means of improving iodine
status (14). Because it included only six studies, it could not address many outcomes of interest,
nor explore potential eect modiers. The objective of the current systematic review and meta-
analyses is to assess the eect and safety of iodized salt compared to non-iodized salt, to prevent
the numerous outcomes comprising iodine deciency disorders. Furthermore, it aims to explore
whether results are aected by age, physiological status, concentration of iodine in salt, population
salt consumption, or underlying risk of iodine deciency.

7
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Methods
Study characteristics
Randomized controlled trials (RCTs), non-RCTs, quasi-experimental studies, cohort observational
studies and multiple cross-sectional observational studies were included. The participants could
be adults (>15 years of age) and children of any age (0–15 years) and of either sex. Populations
could be in the general population (free living) or part of specic populations, such as refugees.
Studies in apparently healthy populations that may or may not have been at risk or suered from
iodine deciency were considered. The review also included studies that compared outcomes
between groups consuming iodized salt and groups consuming non-iodized salt. Salt could
have been fortied with other micronutrients if the only dierence between groups was the
inclusion of iodine in the forticant. An attempt was made to collect the following outcomes: all-
cause mortality, goitre, mental or physical development, cretinism, cognitive function, urinary
iodine concentration, thyroid-stimulating hormone (TSH) concentration, serum thyroglobulin
concentration, hypothyroidism, hyperthyroidism, and adverse eects reported by study authors.
Search methods
The following sources were searched: The China National Knowledge Infrastructure (May 2011);
the Cochrane Library (Issue 5, 2011), including the Cochrane Controlled Trials Register; EMBASE
(1966 to 18 June, 2011); MEDLINE (PubMed 1966 to 31 May, 2011); The Virtual Health Library of
the Pan-American Health Organization (May 2011); VIP (the register of Chinese trials developed by
the Chinese Cochrane Centre); WANFANG; the WHO International Clinical Trials Registry Platform
(18 June 2011); and the WHO Global Health Library regional databases (June 2011). For the
detailed search strategy used for the electronic search, see Annex 1. The following websites were
also searched: Google, ICCIDD, Thyroid Disease Manager, and the WHO Department of Nutrition
for Health and Development. The following organizations were contacted: the Sprinkles Global
Health Initiative, the Home Fortication Technical Advisory Group, the nutrition section of UNICEF,
the United Nations World Food Programme (WFP), the Micronutrient Initiative (MI), the Global
Alliance for Improved Nutrition (GAIN), Helen Keller International (HKI), Sight and Life Foundation,
the United States (US) Centers for Disease Control and Prevention (CDC), the Iodine Network, and
ICCIDD. The following journals were hand searched: Chinese Journal of Control of Endemic Diseases,
Chinese Journal of Epidemiology, Chinese Journal of Preventive Medicine and Studies of Trace Elements
and Health. The reference lists of identied papers were also scanned.
Data collection and analysis
Selection of studies
Two reviewers independently assessed references for potential relevance. The full manuscript was
retrieved when the title, abstract and keywords suggested that the study: (i) compared a group
consuming iodized salt to a group consuming non-iodized salt; (ii) reported on an outcome of
interest; or (iii) was unclear regarding these criteria. Two reviewers independently assessed all
potentially eligible studies, according to the above prespecied inclusion criteria, while two other
reviewers assessed the relevance of one third of those selected for full review. An adapted preferred
reporting items for systematic reviews and meta-analyses (PRISMA) owchart was used to depict
the selection of studies for inclusion (15).

8
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Data extraction and management
Two reviewers independently extracted relevant characteristics of the populations and
interventions of each study. A third reviewer checked the data, and disagreements were resolved
through consensus. In the case of articles originally published in the Chinese language, studies
were screened and data extracted in duplicate before translation, and then again after translation
to English. Any relevant missing information was requested from the original study authors. In the
case of companion papers of a primary study, all available data were simultaneously evaluated,
to maximize the yield of information. For RCTs, the risk of bias associated with the method of
sequence generation, allocation concealment, blinding, selective reporting, loss to follow-up, and
completeness of outcome data was assessed. For observational studies, the risk of bias associated
with the method of measuring exposure, collecting outcome data, and selecting participants was
also evaluated. The risk of bias was rated as low, unclear or high, according to established criteria
(16, 17). A risk of bias graph and a risk of bias summary were generated. The quality of the entire
body of evidence was assessed, using the grading of recommendations assessment, development
and evaluation (GRADE) methodology (18) and GRADEProler software (version 3.6, 2011).
Statistical analysis
An overall eect estimate for dichotomous data was calculated as a risk ratio (RR) with 95% condence
interval (CI), or an odds ratio (OR), in the case of rare events (19). For multiple cross-sectional studies,
prevalence ratios (PRs) were reported for the overall eect estimate from dichotomous outcomes. For
continuous variables, the overall eect estimate was calculated as the dierence in means, with 95%
CI, between the group consuming iodized salt (intervention) and the group consuming non-iodized
salt (control/comparison). To combine data and generate overall eect estimates, RRs, PRs, and mean
dierences (MDs) were calculated, using the inverse variance method, random eects, and, for ORs, the
Peto method, xed eects meta-analysis in Review Manager software (Copenhagen, 2011) (17, 19) was
used. When a study reported results for multiple intervention groups with multiple control groups, all
comparisons were included in the pooled analysis. When outcomes were assessed at more than one
time point, data from the latest time point were used for the pooled analysis, and all relevant time points
were used for subgrouping by study duration.
Results were considered to be statistically signicant at α = 0.05. Evidence was considered
conclusive and the estimate precise if the point estimate suggested a benet or harm and the 95% CI
did not overlap a threshold of relevance. If the point estimate was near the null value and the CI did not
overlap a threshold of relevance, evidence was considered conclusive of no eect. Conversely, evidence
was considered to be inconclusive, and the point estimate imprecise, if the point estimate suggested a
benet or harm but the CI crossed a threshold of relevance. Heterogeneity was assessed through visual
inspection of the forest plots and with the I2 statistic quantifying inconsistency across studies (20, 21). An
I2 statistic of 75% or greater was considered an important level of heterogeneity, and, when found, the
data in meta-analyses were combined, while taking note of the heterogeneity and attempting to explain
the heterogeneity by examining individual study and subgroup characteristics.
Specic objectives were tested and potential reasons for heterogeneity explored, using the
following subgroups: age group (child versus adult versus all ages); concentration of iodine in salt
(<20parts per million [ppm] versus 20–40 ppm versus >40 ppm); mean salt intake (<2g/day versus
2–4g/day versus >4 g/day); physiological status (pregnant versus non-pregnant); iodine deciency
in the population at baseline (adequate iodine status versus mild versus moderate versus severe
iodine deciency); and duration of population exposure to iodized salt (<1 year versus 1–5 years versus
5–15 years versus >15 years).

9
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Results
As shown in Fig. 1 (see p.17), the search of electronic databases yielded 9720 publications and
other sources identied 90 publications. After scanning the titles, 7113 publications were excluded
because of being duplicates or obviously not relevant. Of the 2697 remaining publications, 2416
were excluded after review of the abstract, for not meeting basic inclusion criteria. A total of
281 full-text publications were reviewed and 142 excluded for lack of a comparison between a
group consuming iodized salt and a group not consuming iodized salt via intervention, quasi-
experiment, cohort or multiple cross-sectional, pre/post study design. Upon review of complete
text, another 26 publications were identied as duplicates or companion papers to original studies.
Of the 113 potential studies for inclusion, 24 are awaiting classication, as additional information
is being sought from the original authors (22–52). Therefore, the current review includes 89
studies. Two studies are RCTs (53, 54). Six are non-RCTs (55–62). Twenty are quasi-experimental
studies (63–92). Sixteen are cohort observational studies (93–111). Forty-two are multiple cross-
sectional studies (112–156). Three studies with mixed designs were identied: one with cohort,
multiple cross-sectional and quasi-experimental components (157); one with cohort and multiple
cross-sectional components (158); and one with multiple cross-sectional and quasi-experimental
components (159). One study originally reported a non-RCT, and reported subsequent data from
a cross-sectional observational follow-up (60). Two ongoing studies were also identied (160, 161).
Studies were published in English, Chinese, French, German and Spanish and were conducted
in Argentina, Azerbaijan, Bangladesh, China, Colombia, Cote d’Ivoire, Denmark, Germany,
Guatemala, India, the Islamic Republic of Iran, Italy, Kazakhstan, Malaysia, Mongolia, Morocco,
Pakistan, Poland, South Africa, Spain, Switzerland, Tajikistan, Thailand, Uganda, the United States
of America and Uzbekistan. A summary of the characteristics of included studies is presented in
Table 1 (see p. 18) and Annex 2.
Eect estimates
Goitre prevalence
The results for goitre prevalence for all study designs can be found in Table 2 (see pp. 19 and 20).
In one RCT study in school-aged children, 4years of availability of iodized salt in the marketplace,
combined with the instruction in the intervention group to consume iodized salt, did not result in
a signicant change in goitre prevalence relative to control (RR = 1.06 [95% CI = 0.69 to 1.62]). The
other RCT qualitatively noted that thyroid volume did not change in a group of pregnant women
consuming iodized salt; whereas, in the control group, thyroid volume increased signicantly
during pregnancy (P<0.001).
Three non-RCTs with four comparisons contributed to a meta-analysis and found iodized salt
reduced the risk of goitre relative to control (RR = 0.59 [95% CI = 0.36 to 0.95]). Eleven cohort
studies with 13 comparisons contributed to the meta-analysis, which found iodized salt was associated
with decreased risk of goitre (RR =0.30 [95% CI =0.23 to 0.41]). The meta-analysis of 34 multiple
cross-sectional studies with 44 comparisons also associated exposure to iodized salt with decreased
prevalence of goitre (PR = 0.18 [95% CI = 0.14 to 0.22]). Additional studies of all designs that could not
be included in the meta-analyses supported the ndings that iodized salt decreases the risk or prevalence
of goitre (see Annex 3).

10
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Subgroup analyses
Subgroup analyses (see Table 2, pp. 19 and 20) of cohort studies and multiple cross-sectional
studies suggest that the consumption of iodized salt reduces the risk of goitre regardless of age
group, concentration of iodine in salt, or underlying risk of iodine deciency at baseline, with no
signicant dierences among groups. The risk reduction or prevalence reduction (depending on
study design) was signicantly larger as the duration of exposure to iodized salt increased. There
were insucient studies to conduct meaningful subgroup meta-analyses by estimated mean salt
consumption or physiological status.
Cretinism
There were no RCTs, non-RCTs, or quasi-experimental or cohort studies measuring cretinism that
reached the inclusion criteria. The meta-analysis of two multiple cross-sectional studies with two
comparisons detected a reduced odds of cretinism with availability of iodized salt (OR = 0.13 [95%
CI = 0.08 to 0.20]). One other study reported that, at the survey before iodization of salt of an entire
village in China, there were 88 individuals with cretinism and at the follow-up survey of the village
21 years after the introduction of iodized salt, there were only four incident cases of cretinism.
Cognitive function
There were 18 quasi-experimental studies with 31 comparisons that contributed to the meta-
analysis of cognitive function and found improved cognitive function, measured by intelligence
quotient (IQ), with iodized salt compared to non-iodized salt (see Table 3, p. 21; MD = 8.18 [95%
CI = 6.71 to 9.65]). The meta-analysis of 16 studies with 25 comparisons found a reduced risk of
low intelligence (IQ <70) with iodized salt (RR = 0.28 [95% CI = 0.22 to 0.36]). The meta-analysis of
two multiple cross-sectional studies with three comparisons found an increase in IQ with iodized
salt (MD = 10.45 [95% CI = 4.79 to 16.11]). The results of the one multiple cross-sectional study
that measured prevalence of low intelligence were consistent with an improvement in cognitive
function with iodized salt (PR = 0.24 [95% CI=0.07 to 0.82]).
Subgroup analyses
There was a signicant increase in IQ when the concentration of iodine in the salt was 20–40 ppm
(MD = 7.59 [95% CI = 5.19 to 9.99]); >40 ppm (MD = 14.24 [95% CI = 12.30 to 16.18]); or unknown
(MD = 6.51 [95% CI = 4.53 to 8.49]), with a greater improvement at the higher concentration of
iodine compared to the other two groups (P<0.01). Iodized salt was also associated with a reduced
risk of having low intelligence, regardless of the concentration of iodine in the salt: 20–40 ppm
(RR = 0.16 [95% CI = 0.10 to 0.25]); >40 ppm (RR = 0.20 [95% CI = 0.12 to 0.32]); or unknown
(RR = 0.40 [95% CI = 0.29 to 0.54]).
There was a signicant increase in IQ with iodized salt in those studies in which there was a
population risk of mild (MD = 8.31 [95% CI = 5.77 to 10.84]), moderate (MD = 4.01 [95% CI = 1.28 to 6.74]),
or severe (MD = 9.05 [95% CI= 6.84 to 11.26]) iodine deciency, and there was a corresponding
significant reduction in risk of low intelligence (mild RR = 0.24 [95% CI = 0.14 to 0.40];
moderate RR = 0.39 [95% CI=0.16 to 0.95]; severe RR = 0.29 [95% CI = 0.20 to 0.42]). There was no
dierence in results based on the duration of exposure to iodized salt.

11
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Urinary iodine excretion
The meta-analysis of three cohort studies with nine comparisons that reported urinary iodine
excretion as µg/L, and the meta-analysis of two studies with two comparisons that reported µg/g
creatinine, detected a signicant increase in urinary iodine excretion with iodized salt relative to
non-iodized salt (MD = 59.22 [95% CI = 50.40 to 68.04] and MD = 87.35 [95% CI = 49.74 to
124.97]) (see Table 4, p. 22). The meta-analysis of ve multiple cross-sectional studies with ve
comparisons that reported mean urinary iodine excretion as µg/L, and the three studies with
four comparisons that reported µg/g creatinine, detected a signicant increase in urinary iodine
excretion with consumption of iodized salt (MD = 72.35 [95% CI = 44.54 to 100.17] and MD = 104.11
[95% CI=55.28 to 152.94]).
The meta-analysis of the two non-RCTs with two comparisons reporting risk of iodine
deciency suggested a potential benet of iodized salt (RR = 0.42 [95% CI = 0.14 to 1.28]). However,
there were only 233 participants and 57 events. The meta-analysis of the two cohort studies with
six comparisons detected a significant reduction of risk in iodine deficiency with iodized salt
(RR = 0.40 [95% CI = 0.26 to 0.60]). The same was found in the meta-analysis of 10 multiple cross-
sectional studies with 10 comparisons (PR = 0.45 [95% CI = 0.34 to 0.59]). The RCTs, non-RCTs,
cohort studies and multiple cross-sectional studies that could not be incorporated into meta-
analyses all reported increased urinary iodine excretion with iodized salt (see Annex 4).
Subgroup analyses
Exposure to iodized salt increased urinary iodine excretion in children, adults and all ages combined.
There was no statistical dierence in the increase in urinary iodine excretion by concentration of
iodine in salt or underlying risk of iodine deciency. There was no consistent relationship between
duration of exposure and urinary iodine excretion. No studies reported salt intake and no studies
were conducted in pregnant women (data not shown).
Thyroid hormones
One cohort study reported and detected a signicant decrease in TSH (MD = –9.80 [95% CI =–10.85 to
–8.75]) in children exposure to iodized salt. The meta-analysis of the two multiple cross-sectional studies
with four comparisons reporting TSH detected no change with iodized salt (MD = 0.18 [95% CI =–0.28
to 0.64]). The meta-analysis of one multiple cross-sectional study with three comparisons detected a
signicant decrease in thyroglobulin with iodized salt (MD = –8.73 [95% CI=–9.31 to –8.14]) (147).
Hypothyroidism and hyperthyroidism
The meta-analyses of four cohort studies with four comparisons and of four multiple cross-sectional
studies with four comparisons found no relationship between iodized salt and hypothyroidism (see
Table 5, p. 23) (cohort: OR = 1.14 [95% CI = 0.84 to 1.53]; cross-sectional: OR = 1.13 [95% CI = 0.94 to
1.36]). The meta-analysis of two cohort studies with two comparisons showed an increased odds of
hyperthyroidism with iodized salt (OR = 1.36 [95% CI = 1.12 to 1.66]). However, the meta-analysis of
ve multiple cross-sectional studies with ve comparisons detected no relationship between iodized
salt and hyperthyroidism (OR = 0.96 [95% CI = 0.92 to 1.00]). One study measured the incidence of
hyperthyroidism indicated by incident use of anti-thyroid medication in the population of Denmark, and
included more than 5000000 individuals (117), whereas, all other studies measured low circulating TSH
in samples of individuals. Without the study from Denmark, the meta-analysis found that iodized salt
reduced the odds of hyperthyroidism (OR = 0.50 [95% CI = 0.40 to 0.63]). There were too few studies to
conduct meaningful subgroup analyses of hypothyroidism or hyperthyroidism.

12
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Other adverse outcomes
There was one cohort study that reported consumption of iodized salt signicantly increased the
odds of elevated positive anti-thyroid microsomal antibody (ATMA) (OR = 2.51 [95% CI = 1.93 to 3.27]).
There were two multiple cross-sectional studies with contradictory results. The meta-analysis of
those studies detected no increased odds of elevated ATMA with iodized salt (OR = 1.27 [95%
CI = 0.94 to 1.71]). The same was true for anti-thyroglobin antibody (anti-Tg) (OR = 1.43 [95% CI = 1.08
to 1.89]). One multiple cross-sectional study of school-aged children reported an increased risk of
elevated urinary iodine excretion with 10 years of exposure to iodized salt (RR = 1.35 [95% CI = 1.06
to 1.74]) (see Table 5, p. 23).
Other outcomes
There were no studies reaching inclusion criteria that reported all-cause mortality, mental
development or physical development.
Quality of the body of evidence
The risk of bias of included studies is summarized in the risk of bias graph summary table
(see Annex 5). The two RCTs were of poor methodological quality and neither followed the
consolidated standards of reporting trials (CONSORT) recommendations (162). They did not
report details on power calculation, randomization method, ow of participants, or blinding of
outcome assessment. The six non-RCTs were at high risk of bias, owing to inadequate allocation
concealment, unclear blinding of participants and outcome assessors, unclear risk of attrition bias,
and lack of comparability between the intervention and control group in half of all studies). The risk
of bias of quasi-experimental, cohort and multiple cross-sectional studies was unclear. Most quasi-
experimental and multiple cross-sectional studies had complete outcome data; however, blinding
of personnel and outcome assessors was unclear. Four cohort studies did not report outcome
data completely and four were unclear. Five cohort studies had unclear sampling methodology
or unclear methods in general. All other cohort studies suggested no other form of bias. Of the
multiple cross-sectional studies, seven were at high risk of bias because of sampling methods that
did not ensure comparability between the participants of surveys, or dierences in the methods for
measuring outcomes between surveys. Ten had unclear sampling or unclear methods in general.
The quality of evidence (see Annex 6) for goitre ranged from low to moderate. Many studies
lacked the detail needed to clearly understand bias; but the large eect sizes detected in quasi-
experimental, cohort and multiple cross sectional studies led to the upgrade of that evidence.
For cretinism, the quality of the body of evidence was moderate. All studies were multiple
cross-sectional; however, the quality of evidence was upgraded from low to moderate because of
the very large eect size.
The quality of evidence for cognitive function was low. All studies were either quasi-
experimental or multiple cross-sectional and were conducted in China. It was not possible to
assess the risk of bias in most studies because of the limited information in published reports;
therefore, the evidence was conservatively downgraded because of the potential of high risk of
bias. Nonetheless, because the actual risk of bias was unclear, the evidence was then upgraded,
owing to the large eect on both mean dierence in IQ and risk of low intelligence. The net eect
was therefore that the evidence was graded as low.

13
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
The evidence for urinary iodine excretion was low to moderate quality. Data were from cohort
and multiple cross-sectional studies; however, all studies reported large eect sizes.
The body of evidence for hypothyroidism was very low quality and that for hyperthyroidism
was very low and low quality. The studies reporting hypothyroidism had cohort or multiple cross-
sectional designs. The evidence was downgraded, owing to inconsistency. Hyperthyroidism was
reported in cohort and multiple cross-sectional studies. The evidence was not downgraded for
any reason in the cohort studies; however, the evidence from multiple cross-sectional studies was
downgraded because of imprecision.
The evidence for thyroid hormones or other adverse eects was very limited and the quality
was either low or very low.
Discussion
This review found that exposure to iodized salt was associated with reduced risk or prevalence of
goitre, cretinism, low intelligence or low urinary iodine excretion, as well as increased mean IQ and
increased median and mean urinary iodine excretion. There was no clear relationship between
iodized salt and any adverse eects, including increased hypothyroidism or hyperthyroidism.
Because of the limited number of RCTs and non-RCTs found in the previous systematic
review (14), a broad search strategy was intentionally used for the current review, to include quasi-
experimental, cohort observational and multiple cross-sectional studies with pre/post designs.
The review also beneted from the ability to search Chinese databases, which included numerous
reports from monitoring and evaluation of salt iodization programmes in that country. Though the
current review included only two RCTs, the large number of studies, 89, allowed exploration of the
many outcomes comprising iodine deciency disorders; potential adverse eects; and potential
eect modication by age, concentration of iodine in salt, underlying risk of iodine deciency, and
duration of exposure.
Data consistently showed a correlation between exposure to iodized salt and reduced risk
of goitre. The studies spanned eight decades and multiple countries and continents, including
Africa, Asia, Europe, North America and South America. The results were consistent, regardless
of age group, concentration of iodine in salt or underlying risk of iodine deciency. Some results
suggested that the longer the population exposure to iodized salt was, the lower the prevalence
of goitre would be. The present results are similar to those found in a recent systematic review of
the eect of iodine supplementation, mainly in the form of iodized oil, which reported reduced
risk of goitre in children receiving iodine (163). Though both interventions appear to be eective in
reducing goitre, salt iodization provides an inexpensive platform for unparalleled population reach
(9). For this reason, WHO, UNICEF and the ICCIDD Global Network appeal to all countries to iodize
salt, regardless of the underlying population risk of iodine deciency disorder, in order to protect
everyone, including the most vulnerable.
The current review found convincing results of the benecial eect of iodized salt on cretinism. This
nding is consistent with a narrative review published in 1989, which concluded that “iodine deciency
correction” during pregnancy reduced cases of cretinism (164). Another review found consumption of
iodized oil during pregnancy to be highly eective at reducing incident cretinism (165). A more recent

14
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
review reinforced the ndings by showing that increased intake of iodine from any source during
pregnancy reduced cretinism (166). The latter review also summarized the experimental evidence,
which showed that, in animals, iodine deciency causes fetal brain damage, and adequate iodine
consumption during pregnancy is protective. The body of knowledge around the importance of
the eect of iodine intake during pregnancy on the brain is undisputed, and the results of the
current review support the contention that salt iodization is an eective strategy to increase iodine
intake during pregnancy and reduce cretinism and brain damage (164, 166–169).
The current review found that children exposed to iodized salt during gestation, infancy
and early childhood had higher IQ and reduced risk of low intelligence compared to unexposed
children. A previous meta-analysis of 18 studies of children living in areas with iodine suciency
versus those living in iodine-decient areas also concluded that exposure to iodized salt was
positively associated with cognition in children (170). Children exposed to iodine had an average
IQ score of more than 13 points higher than children not exposed to iodine. The present review has
reported a similar increase of approximately 8 to 10 IQ points with exposure to iodized salt. The
results were also consistent with a review that found iodine supplementation during pregnancy
and infancy improved developmental scores in infants (166), and a separate review that found
improved cognitive function with iodine supplementation (163). The evidence suggests that
exposure to iodine improves cognitive development in infants and children and iodization of salt
is an eective strategy to increase exposure during pregnancy, infancy and early childhood (171).
However, the current review did not compare delivery mechanisms for consumption of iodine and
cannot determine whether salt iodization is the most eective intervention for reaching pregnant
and lactating women and infants, and thus conclusions in this regard cannot be drawn from the
current review alone.
Because of the large number of studies in the current review, it was possible to investigate
potential adverse eects such as increased hypothyroidism and hyperthyroidism. The meta-analysis
of inconsistent individual studies showed no overall eect of iodized salt on hypothyroidism. A
previous review without meta-analysis cited four studies that suggested an association between
increased population iodine intake and prevalence of hypothyroidism (172). Of these four studies,
three were observational reports and the one experimental study was included in the current
review. Across all studies, there were few events of hyperthyroidism and the results varied by
study design. There have been several studies that correlated iodine-induced hyperthyroidism to
the introduction of iodized salt (173–175). In many of these cases, iodization levels were much
higher than those recommended (174, 175), or iodization occurred in an area of iodine suciency
(173). Nonetheless, long-term studies suggest that an increase in the prevalence of iodine-induced
hyperthyroidism is temporary and that within a few years of population exposure to iodized salt,
the prevalence of hyperthyroidism reverts to baseline levels or even lower (117, 171). The few
studies reporting anti-thyroid antibodies had contradictory ndings and more data are needed to
formulate conclusions regarding these outcomes. Currently, there is little evidence of increasing
health risks with population-wide salt iodization programmes. However, potentially serious
adverse eects such as hypothyroidism or hyperthyroidism may occur in some individuals with
the introduction of iodized salt at the population level. Therefore, salt iodization programmes
require robust monitoring and evaluation systems to ensure the population consumes safe levels
of iodine (176), and to inform adjustments to iodization programmes, as public health eorts to
reduce sodium intake mature.

15
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Many of the analyses suered from heterogeneity stemming from known (age group, region of
world, decade of study implementation, sampling methodology, measurement techniques, iodine
status at baseline, iodine concentration in salt, salt consumption patterns, other dietary patterns)
and unknown factors. This limitation notwithstanding, the results were very consistent for most
outcomes measured. Iodized salt was associated with a lower prevalence of goitre in children,
adults and the entire population; in populations consuming iodized salt at <20ppm, 20–40ppm,
or >40ppm; and in populations with risk of severe, moderate or mild iodine deciency at baseline.
The other outcomes with sucient data to make such comparisons (i.e. cognitive function, urinary
iodine excretion) showed similar results. Additionally, studies that could not be combined in the
meta-analyses for these outcomes were also consistent with a benet of consuming iodized salt.
The heterogeneity of the studies may make the numerical value of the estimated eect sizes less
important, but the large eect sizes and consistent results show the value of the meta-analyses
(19), despite the heterogeneity, to conclude that iodized salt is eective at reducing the risk of
many iodine deciency disorders.
Limitations of the study
A limitation of the review was an insucient number of studies reporting the level of salt intake or
the level of overall iodine intake to test the potential eect modication of these variables on the
ability of iodized salt to reduce iodine deciency disorders. Additionally, many studies reported
the recommended concentration of iodine in fortication programmes as the iodine level in salt,
and did not consider the actual level consumed. Though potassium iodate in salt is stable (10),
programme monitoring data have shown the presence of suboptimal fortication and some
pre-consumption losses during transport, handling and storage. Future studies and monitoring
reports should attempt to report the mean salt intake of children, adults and pregnant women,
as well as the iodine concentration in salt, in order to test the potential modifying eect of these
factors. Furthermore, lack of methodological detail in some studies included in the review led
to their exclusion from the meta-analyses and, in some cases, diculty in assessing the bias of
studies or the body of evidence. All studies reporting intelligence outcomes were conducted in
China. The body of evidence could be strengthened if these results were reproduced in other
countries. Multiple cross-sectional studies without a baseline before iodization were not included
in the current review. Including such studies would provide more information on the eectiveness
of varying levels of iodine in salt. This information would be particularly important as countries
strengthen eorts to decrease overall population salt intake.
Conclusion
The current review is the largest compilation of studies regarding the eect of iodized salt on
health outcomes to date. The results suggest that iodized salt is an eective means of improving
iodine status and preventing many iodine deciency disorders. Though the quality of the body of
evidence for some outcomes could be strengthened, data were compiled from all over the world;
spanning more than eight decades; from rural and urban areas; from low-, middle- and high-income
countries; and from countries and regions traditionally plagued by severe iodine deciency. The
results were mostly consistent with a benet of salt iodization. Moreover, there was little evidence
of adverse eects from salt iodization. Continued monitoring of programmes should provide more
high-quality evidence on outcomes such as mental and physical development, and on potential
adverse eects such as hypothyroidism, hyperthyroidism and elevated urinary iodine excretion.
These data are needed to ensure iodization programmes are not inadvertently causing harm.

16
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Additionally, the focus on operational research and monitoring data is particularly important as
countries strive to reach global targets for reduced population sodium intake. As sodium intake
decreases, iodization levels in salt may need to be adjusted, to ensure that the most vulnerable
pregnant women, infants and young children receive the iodine they need.
Contributions
The priority questions to be addressed for the review were discussed and developed by the WHO
guideline development group – micronutrients, and the protocol was subsequently developed to
address the priority areas of interest for the review of the WHO guideline on salt iodization. Searches
were run by NJA and TW. Assessment of inclusion, data extraction and validity assessment were
carried out by all authors. Data analyses were completed by NJA, with input from MA, VC and TW.
The GRADE evidence proles were developed by NJA. The rst draft of the manuscript submitted for
review by the guideline development group – micronutrients was prepared by NJA, with support
from MA, VC and TW. Substantial intellectual input on research methods and interpretation of
results was provided by all authors. All authors read, provided input to, and agreed the nal draft
of the manuscript.
Acknowledgements
Funding for this review came from various sources, including WHO Department of Nutrition for
Health and Development, the Government of Luxembourg, the Bill & Melinda Gates Foundation
and the Micronutrient Initiative. Their support is gratefully acknowledged, along with the support
provided by members of the WHO guideline development group – micronutrients in providing
guidance on document scope.
Declaration of competing interests
MA and TW received funding from the Evidence and Programme Guidance Unit, Department of
Nutrition for Health and Development, WHO, to attend technical meetings in Geneva to work
in collaboration on the review; there was no further nancial support from any organization for
the submitted work that might have an interest in the submitted work in the previous 3 years;
and there were no other relationships or activities that could appear to have inuenced the
submitted work.
NJA and VC were sta members of WHO when this work was initiated. NJA is currently a sta
member at the United Nations World Food Programme. The authors alone are responsible for the
views expressed in this publication and they do not necessarily represent the views, decisions or
policies of WHO or the United Nations World Food Programme.

17
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Fig. 1 Flowchart of screening, inclusion and exclusion
9720 hits from electronic searches 90 publications identied through
additional searches
9810 publications identied
2697 potential studies from
publications
281 full-text publications
assessed for eligibility
Total 113 potential studies
89 studies included:
2 RCTs
6 non-RCTs
20 quasi-experimental studies
16 cohort observational studies
42 multiple cross-sectional studies
1 with cohort component
1 with cohort/cross-sectional/quasi
-experimental component
1 with a cross-sectional and a
quasi-experimental component
24 awaiting classication
7113 duplications and
obviously irrelevant studies
2416 studies excluded after
scanning titles and abstracts
142 full-text articles excluded
for lack of comparison of a
group consuming iodized salt
versus a group consuming
non-iodized salt
26 subsequently identied
as companion papers or
duplicates
RCT: randomized controlled trial.

18
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Table 1. Summary of studies reaching inclusion criteria for systematic review on the eect of salt iodization on iodine deciency disorder
Study design
Number of
studies Sample size Participants Duration Allocation InterventionaControl/comparison
Randomized
controlled trials
(53, 54)
2 35 to 334
(total=369)
Children/
pregnant
women
6 months to
4years
Random group assignment Participants instructed to only
purchase and consume iodized
salt, or provided iodized salt
Randomly assigned participants
instructed to only purchase and
consume non-iodized salt
Non-randomized
controlled trials
(55–62)
6 102 to 21616
(total >32350)
Children/
women all
ages
4months to
7years
Clusters (cities) or
households non-randomly
allocated to receive iodized
salt
3 studies: participants were
random sample of residents of
cities where iodized salt was
made available
Random sample of population
that did not have iodized salt
available
 
  
3 studies: participants received
iodized salt directly
Participants received non-
iodized salt
Quasi-
experimental
(63–92)
20b91 to 2592
total >14000)
Children 9 months’
gestation plus
1–4yearsc
Divided population by birth
date based on being born
before or after USI in village
17 studies: children exposed to
iodized salt during gestation,
infancy and early childhood
through USI
Children from same villages not
exposed to iodized salt during
gestation, infancy and early
childhood
   10–15years Compared individuals before
and after USI from villages
with USI to those without USI
3 studies: children exposed to
iodized salt during gestation,
infancy and early childhood
through USI
Children from similar villages
without availability of iodized
salt and therefore not exposed
to iodized salt
Cohort
observational
(93–111)
16b30–500000
(total
>910000)
Children/
women/all
ages
6months to
10years
Entire population of village
or random sample of
population followed from
pre to post availability of
iodized salt
4 studies: households provided
iodized salt
Same sample of participants
measured before iodized salt
was available



8 studies: population exposed
to iodized salt through USI
implementation


 
4 studies: sample exposed to
iodized salt through availability
in market
Multiple
cross-sectional
(112–156)
42b40–5000000
(total
>7000000)d
Children/
adults/all ages
1–36years Representative cross-
sectional surveys
Sample of population measured
after iodized salt was available or
the implementation of USI
Sample of population measured
before iodized salt was available
USI: universal salt iodization.
a All studies compared groups that consumed iodized salt to groups that consumed non-iodized salt. There were no studies comparing multiple micronutrient fortied salt with and without iodine.
b Three studies had mixed designs. One study measured goitre in both a cohort of individuals and a quasi-experimental sample of individuals, and also measured urinary iodine excretion in multiple cross-
sectional studies of children of the same age over time (157). One study measured cretinism in a multiple cross-sectional design and measured a cohort of individuals for all other outcomes (158). One study
measured intelligence in two study samples: one was a multiple cross-sectional design and the other was a quasi-experimental design (159).
c Most quasi-experimental studies were unclear on the precise period of exposure to iodized salt in the intervention group and merely noted that the intervention group was exposed to iodized salt
during the period of gestation and the rst years of life.
d 41 of 42 studies had a total sample size of 2000000. One study was a registry study in Denmark that included >5000000 individuals (117).

19
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Table 2. Eect of iodized salt on goitre prevalence, by study design and by subgroup
Study designaGroup Number of studies Number of comparisons I2RR/PR 95% CI
RCTbAll 1 1 — 1.06 0.69 to 1.62
Non-RCT All 3 4 99 0.59 0.36 to 0.95
Quasi-experimental All 1 1 — 0.10 0.08 to 0.13
Cohort All 11 13 99 0.30 0.23 to 0.41
Age group     
 Children 2 2 100 0.09 0.08 to 0.11
 Adults 1 1 — 0.45 0.27 to 0.75
 All ages 8 10 99 0.32 0.24 to 0.43
Concentration of iodine in salt, ppm     
 <20 4 6 98 0.39 0.31 to 0.50
 20–40 3 3 99 0.24 0.08 to 0.74
 >40 3 3 96 0.37 0.15 to 0.92
Underlying risk of iodine deciencyc    
 Adequate iodine status 0 — — — —
 Mild iodine deciency 4 4 100 0.29 0.16 to 0.54
 Moderate iodine deciency 1 1 — 0.43 0.21 to 0.87
 Severe iodine deciency 7 8 99 0.30 0.18 to 0.48
Duration of exposure, years     
 <1 3 3 34 0.64 0.57 to 0.73
 1–5 7 9 99 0.28 0.20 to 0.30
 5–15 1 1 — 0.09 0.08 to 0.10
 >15 0 — — — —

20
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Study designaGroup Number of studies Number of comparisons I2RR/PR 95% CI
Multiple cross-sectional All 34 44 100 0.18 0.14 to 0.22
Age group     
 Children 23 32 100 0.22 0.17 to 0.29
 Adults 2 2 99 0.17 0.04 to 0.85
 All ages 11 11 100 0.10 0.06 to 0.16
Concentration of iodine in salt, ppm     
 <20 2 2 54 0.11 0.07 to 0.17
 20–40 17 24 99 0.21 0.15 to 0.28
 >40 8 9 100 0.13 0.06 to 0.25
 Underlying risk of iodine deciencyc    
 Adequate iodine status 3 4 92 0.20 0.12 to 0.34
 Mild iodine deciency 9 11 99 0.11 0.07 to 0.18
 Moderate iodine deciency 7 11 99 0.35 0.25 to 0.49
 Severe iodine deciency 15 16 100 0.15 0.10 to 0.22
Duration of exposure, years     
 <1 1 1 — 1.07 0.88 to 1.30
 1–5 9 10 100 0.29 0.18 to 0.46
 5–15 16 25 100 0.17 0.12 to 0.24
 >15 6 6 100 0.08 0.04 to 0.17
95% CI: 95% condence interval; ppm: parts per million; PR: prevalence ratio; RR: risk ratio; RCT: randomized controlled trial.
a Studies that could not be combined in the appropriate meta-analysis by study design all showed reduced prevalence of goitre with iodized salt (1 non-RCT, 1 quasi-experimental,
1 cohort and 5 multiple cross-sectional studies).
b One RCT in pregnant women reported no increase in thyroid volume during pregnancy with iodized salt versus increase in thyroid volume in control.
c Risk estimated by prevalence of goitre or median urinary iodine excretion in the population.

21
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Table 3. Eect of iodized salt on change in intelligence quotient, by study design and by subgroup
Study designaGroup Number of studies Number of comparisons I2MD 95% CI
Quasi-experimental All 18 31 89 8.18 6.71 to 9.65
Concentration of iodine in salt, ppm     
 <20 0 — — — —
 20–40 5 11 86 7.59 5.19 to 5.19
 >40 3 3 0 14.24 12.30 to 16.18
Underlying risk of iodine deciencyb    
 Adequate iodine status 3 3 93 5.35 -4.06 to 14.76
 Mild iodine deciency 6 8 91 8.31 5.77 to 10.84
 Moderate iodine deciency 4 4 52 4.01 1.28 to 6.74
 Severe iodine deciency 12 13 88 9.05 6.84 to 11.26
Duration of exposure, years     
 <1 0 — — — —
 1–5 10 19 87 8.17 6.17 to 10.18
 5–15 5 5 87 9.44 5.95 to 12.94
 >15 0 — — — —
Multiple cross-sectional All 2 3 92 10.45 4.79 to 16.11
95% CI: 95% condence interval; MD: mean dierence; ppm: parts per million.
a No randomized controlled trials (RCTs), non-RCTs or cohort observational studies reaching inclusion criteria reported on the eect of iodized salt on change in intelligence.
b Risk estimated by prevalence of goitre or median urinary iodine excretion in the population.

22
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Table 4. Eect of iodized salt on change in urinary iodine excretion
Study designaGroup Units Number of studies Number of comparisons I2MDb95% CI
Cohort All µg/L 3 9 100 59.22 50.40 to 68.04
All µg/g creatinine 2 2 92 87.35 49.74 to 124.97
Multiple cross-sectional All µg/L 5 5 99 72.35 44.54 to 100.17
All µg/g creatinine 3 4 99 104.11 55.28 to 152.84
95% CI: 95% condence interval; MD: mean dierence; ppm: parts per million.
a No randomized controlled trials (RCTs), non-RCTs or quasi-experimental studies reaching inclusion criteria reported on the eect of iodized salt on change in urinary iodine excretion in
such a way as could be combined in meta-analyses.
b Studies that reported median urinary iodine excretion and could not be combined in the meta-analyses reported increased urinary iodine excretion with iodized salt (1 RCT, 3 non-RCTs,
6 cohort and 16 multiple cross-sectional studies).

23
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Table 5. Eect of iodized salt on hyperthyroidism, hypothyroidism and other adverse eects
Outcomes Study design Group Number of studies Number of comparisons I2Peto OR 95% CI
Hypothyroidism     
Cohort All 4 4 91 1.14 0.84 to 1.53
Multiple cross-sectional All 4 4 90 1.13 0.94 to 1.36
Hyperthyroidism     
Cohort All 2 2 0 1.36 1.12 to 1.66
Multiple cross-sectionala,b All 5 5 89 0.96 0.92 to 1.00
Elevated anti-thyroid microsomal antibody     
Cohort Adults 1 1 — 2.51 1.93 to 3.27
 Multiple cross-sectional All 2 2 96 1.27 0.94 to 1.71
Elevated anti-thyroglobin antibody     
Multiple cross-sectional All 2 2 97 1.43 1.08 to 1.89
Elevated urinary iodine excretion     
Multiple cross-sectional Children 1 1 — 1.35 1.06 to 1.74
95% CI: 95% condence interval; OR: odds ratio.
a Removing the one study from Denmark that included more than 5000 000 individuals from medical registries, the meta-analysis of three studies with three comparisons reported Peto
OR = 0.50 (95% CI =0.40 to 0.63) (117).
b There was one multiple cross-sectional study that reported qualitatively no dierence in the prevalence of hypothyroidism or hyperthyroidism before and after 12 years of exposure to
iodized salt.

24
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
References
1. Global health risks: mortality and burden of disease attributable to selected major risks.Geneva: World Health
Organization; 2009 (http://www.who.int/healthinfo/global_burden_disease/GlobalHealthRisks_report_full.pdf,
accessed 23 October 2014).
2. WHO/UNICEF/ICCIDD. Indicators for assessment of iodine deciency disorders and the control programme report
of a joint WHO/UNICEF/ICCIDD consultation 3–5 November 1992. Geneva: World Health Organization; 1993 (WHO/
NUT/93.1; http://whqlibdoc.who.int/hq/1993/WHO_NUT_93.1.pdf?ua=1, accessed 23 October 2014).
3. Larsen PR, Silva JE, Kaplan MM. Relationships between circulating and intracellular thyroid hormones: physiological
and clinical implications. Endocr Rev. 1981;2:87–102.
4. Delong F. The disorders induced by iodine deciency. Thyroid. 1994;4(1):107–28.
5. Beaufrere B, Bresson JL, Briend A, Ghisol J, Goulet O, Navarro J et al. Iodine nutrition in the infant. Committee on
Nutrition of the French Society of Pediatrics. Arch Pediatr. 2000;7(1):66–74.
6. Iodine status worldwide. WHO global database on iodine deciency. Geneva: World Health Organization; 2004
(http://whqlibdoc.who.int/publications/2004/9241592001.pdf?q=iodine-status-worldwide, accessed 23 October 2014).
7. de Benoist B, McLean E, Andersson M, Rogers L. Iodine deciency in 2007: global progress since 2003. Food Nutr
Bull. 2008;29(3):195–202.
8. Dunn JT. Iodine deciency disorders. UpToDate (http://www.uptodate.com/contents/iodine-deciency-disorders,
accessed 26 November 2014).
9. Sustainable elimination of iodine deciency. New York: United Nations Children’s Fund; 2008 (http://www.unicef.
org/iran/Sustainable_Elimination_of_Iodine_Deciency_053008%281%29.pdf, accessed 23 October 2014).
10. Salt as a vehicle for fortification. Report of a WHO expert consultation. Geneva: World Health Organization; 2008
(http://www.who.int/dietphysicalactivity/LUXsaltreport2008.pdf, accessed 23 October 2014).
11. Aburto NJ, Ziolkovska A, Hooper L, Elliott P, Cappuccio FP, Meerpohl JJ. Eect of lower sodium intake on health:
systematic review and meta-analyses. BMJ. 2013;346:f1378. doi:10.1136/bmj.f1326.
12. Guideline: Sodium intake for adults and children. Geneva: World Health Organization; 2012 (http://www.who.int/
nutrition/publications/guidelines/sodium_intake_printversion.pdf, accessed 23 October 2014).
13. Salt reduction and iodine fortication strategies in public health: report of a joint technical meeting convened
by World Health Organization and The George Institute for Global Health in collaboration with the International
Council for the Control of Iodine Deciency Disorders Global Network, Australia, March 2013. Geneva: World Health
Organization; 2014 (http://apps.who.int/iris/bitstream/10665/101509/1/9789241506694_eng.pdf?ua=1, accessed
23 October 2014).
14. Wu T, Liu GJ, Li P, Clar C. Iodised salt for preventing iodine deciency disorders. Cochrane Database Syst Rev.
2002;(3):CD003204. doi:10.1002/14651858.CD003204.
15. Liberati A, Altman DG, Tetzla J, Mulrow C, Gøtzsche PC, Ioannidis JPA et al. The PRISMA statement for reporting
systematic and meta-analyses of studies that evaluate interventions: explanation and elaboration. PLoS Med.
1999;6(7):1–28. doi:10.1016/j.jclinepi.2009.06.006.
16. Deeks JJ, Dinnes J, D’Amico R, Sowden AJ, Sakarovitch C, Song F et al. Evaluating non-randomised intervention
studies. Health Technol Assess. 2003;7(27) (http://www.journalslibrary.nihr.ac.uk/__data/assets/pdf_
le/0007/64933/FullReport-hta7270.pdf, accessed 23 October 2014).
17. Higgins JP, Thompson SG, Spiegelhalter DJ. A re-evaluation of random-eects meta-analysis. J R Stat Soc Ser A Stat
Soc. 2009;172:137–59.
18. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P et al. GRADE: an emerging consensus on rating
quality of evidence and strength of recommendations. BMJ. 2008;336:924–6. doi:10.1136/bmj.39489.470347.AD.
19. Higgins JPT, Green S, editors. Cochrane handbook for systematic reviews of interventions Version 5.1.0. Oxford: The
Cochrane Collaboration; 2011 (http://www.cochrane.org/handbook, accessed 23 October 2014).

25
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
20. Higgins J, Thompson S, Deeks J, Altman D. Statistical heterogeneity in systematic reviews of clinical trials: a critical
appraisal of guidelines and practice. J Health Serv Res Policy. 2002;7:51–61.
21. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539–58.
22. Chen XY, Chen JJ, Reng GW, Li ZZ. [Eective evaluation on iodised salt intervention in iodine deciency disorders in
Lanzhou City.] Bulletin of Disease Control and Prevention. [Chinese] 2011;26(2):49–51.
23. Crncević-Orlić Z, Ruzić A, Rajković K, Kapović M. The eectiveness of a 40-year long iodine prophylaxis in endemic
goitre region of Grobnik, Croatia. Coll Antropol. 2005;29(2):509–13.
24. Dai H, Shan ZY, Teng XC, Teng D, Guan HX, Li YS et al. [A ve-year follow-up study of hypothyroidism in areas with
dierent iodine intakes.] Chinese Journal of Endocrinology and Metabolism. [Chinese] 2006;22(6):528–31.
25. Shan ZY, Teng WP, Li YZ, Wang ZY, Jin Y, Guan HX et al. [Comparative epidemiologic study on the prevalence of
iodine-induced hypothyroidism.] Chinese Journal of Endocrinology and Metabolism. [Chinese] 2001;17(2):71–4.
26. Yang F, Li J, Shan ZY, Teng XC, Teng D, Guan HX et al. [Incidences of hyperthyroidism in communities with dierent
iodine intake levels: a ve-year prospective and comparative epidemiological survey.] Chinese Journal of
Endocrinology and Metabolism. [Chinese] 2006;22(6):523–7.
27. Sang DZ, Ruo Deng XR, Duo JI LS, Min G, Qiong P. [Analysis of IDD surveillance in Tibet autonomy region in 2002.]
Chinese Journal of Endocrinology. [Chinese] 2003;22(4):344–6.
28. Dong FZ, Zhan PL, He CX, Yu ZX. [Analysis for the 5 years results of 169 monitor places for endemic goitre of Shanxi
province.] Journal of Practical Endemic Diseases. [Chinese] 1986;1(3):65–8.
29. Fu LX, Zhong MY. [Primary exploration of children intelligence and special education in IDD areas of Changping in
Guandong province.] Zhong Guo Di Fang Bing Fang Zhi Za Zhi. [Chinese] 1991;6(Suppl.):73.
30. Fu ZL, Liu T, Song W, Wang JH, Liu WY, Liang HZ et at. [Survey of subclinical cretin among children in endemic goiter
areas after supplement iodine.] Zhong Guo Di Fang Bing Fang Zhi Za Zi. [Chinese] 1992;7(Suppl.):3–5.
31. Gao HK, Dong HC, Zhang GC, Song BX, Sun BY. [Observation for eect of preventing endemic goitre by 1/20000
concentration iodised salt.] Chinese Journal of Control of Endemic Diseases. [Chinese] 1982;4:33–6.
32. Guo YL, Zheng XP, Xin ZF. [Study of pathogenesis of increasing incidence of hyperthyroidism after administration
of iodised salt in iodine deciency disorders region.] Chinese Journal of Endocrinology. [Chinese] 2003;22(3):226–8.
33. Huang YF, Ji WP. [Observation for eect of iodised salt for preventing iodine deciency disorders in 15 years.]
Modern Preventive Medicine. [Chinese] 2003;30(5):712–13.
34. Liu J, Wang RH, Xiao BZ, Liao WF, Li XS, Chen J. Evaluation of the eect of universal salt iodization preventing iodine
deciency disorders over 8 years in Chongqing. J Trop Med. 2006;6(12):1245–8.
35. Ouyang Q. [Assessment of eect of iodised salt for preventing iodine deciency disorders goitre in 12 years in
Danba county Bawang Xiang.] Journal of Occupational Health and Damage. [Chinese] 2008;23(5):317–18.
36. Peng NC, Shi LX, Ma QL, Yang LF, Zhang TW, Gu ZD et al. [Assessment of overall monitoring iodine deciency
disorders over ten years in Guizhou province.] Chinese Journal of Endocrinology and Metabolism. [Chinese]
2001;17(6):329–33.
37. Su XH, Yu J, Li Y, Zhang SB, Wang WH, Su M et al. [Epidemiological investigation of endemic goitre and cretinism in
Jixian village. Chinese Journal of Endocrinology.] [Chinese] 2003;22(4):353–4.
38. Wang SK, Kou GH, Huo CJ, Liu JH. [An epidemiological investigation of preventing iodine deciency disorders
status by iodised salt after 15 years in Shangluo district]. China Public Health. [Chinese] 1992;8(3):102–4.
39. Wang SK, Kou GH, Huo CJ, Liu JH. [An epidemiological investigation of preventing iodine deciency disorders
status by iodised salt after 15 years in Shangluo district.] Endemic Diseases Bulletin. [Chinese] 1992;7(4):26–9.
40. Wang YH, Zhao XQ. Liu ZM, Yi NY, Zhang QY. [Study on newborn cord blood TSH levels in iodine deciency areas
after salt iodisation supplementation.] Shanxi Clinical Medicine Journal. [Chinese] 1999;8(4):262–4.
41. Xu LP, Yu HY. [Observation for eect of preventing endemic goitre in six years by 1/20000 iodised salt in Xiahe
village, Lushan county.] He-nan Health and Prevention. [Chinese] 1986;4:67–7.

26
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
42. Yang SH. [Observational study of the ecacy of 1/50000 iodised salt for preventing and treatment of endemic
goitre in Heba community of Majiang county, Guizhou province.] Journal of Guiyang Medical Collage. [Chinese]
1983;8(2):32–40.
43. Guizhou province Center of Disease Control and Prevention. [Observation of the eect of prevention and
treatment of 1/50000 iodised salt for endemic goitre in Heba community, Majiang county, Guizhou.] Guozhou
Medicine. [Chinese] 1984;8(2):9–12.
44. Tianjin Medical College. [Investigation for endemic goitre and cretin disease in Heba, Guizhou: changes of iodine
metabolism and pituitary-thyroid function before and after supplementation of iodised salt in three years.]
Guiyang Medical College Journal. [Chinese] 1983;8(2):46–55.
45. Yang SH. [Summary of observation for preventing eect of endemic goitre by 1/50000 concentration of iodised
salt in Majiang county, Heba village, Guizhou.] Chinese Journal of Endemicology. [Chinese] 1982;1(3):155–60, 171.
46. Yang ZM, Liu YX, Wang XW, Zhang YL, Lu YF, Cui XM et al. [Assessment for eect of KIO3 iodised salt with 1/30000
concentration for prevention of iodine deciency disorders.] Chinese Journal of Endemic Diseases Prevention.
[Chinese] 1993;8(2):107, 110.
47. Ye SY. [Observation of eect of iodised salt for prevention and treatment of endemic goitre for four years in Shilong
community, Duyun county, Guizhou province.] Guizhou Medicine. [Chinese] 1984;8(2):46–8.
48. Zhao ZR, Tan SZ, Saho JC, Tang DH. [Observation of eect of preventing endemic goitre by 1/20000 iodised salt.]
Endemic Diseases Information. [Chinese] 1982;1:68–75.
49. Zhao ZY, Lin BX, Xian S. [A report of investigating endemic goitre and cretinism in Napo county, Guangxi 17 years
after introduction of iodised salt.] Guangxi Medical Collage Journal. [Chinese] 1984;1(4):67–70.
50. Du FR, Li QH, Li DK, Lin HB. [An epidemiological study of endemic goitre and cretinism disease after
supplementation of iodised salt seventeen years in Napo county.] Chinese Journal of Endocrinology. [Chinese]
1985;4:338–41.
51. Zhong H. [Evaluation of eect of prevention of iodine deciency disorders in Alashan Zuoqi district.] Journal of
Preventive Medicine Information. [Chinese] 2010;26(1):80–1.
52. Zuo SM, Zhu CK, Wang Y Wang MX. [Surveillance on iodine deciency disorders in Qiannan state: a twenty-year
analysis.] Chinese Journal of Endocrinology. [Chinese] 2006;25(4):437–9.
53. Hintze G, Emrich D, Richter K, Thal H, Wasielewski T, Köbberling J. Eect of voluntary intake of iodinated salt on
prevalence of goitre in children. Acta Endocrinol (Copenh). 1988;117:333–8.
54. Romano R, Jannini EA, Pepe M, Grimaldi A, Olivieri M, Spennati P et al. The eects of iodoprophylaxis on thyroid
size during pregnancy. Am J Obstet Gynecol. 1991;164:482–5.
55. Gongora y Lopez J, Mejia CF. [Two years of treatment of goiter with iodized salt in the department of Caldas.] Med
Cir (Colombia). [Spanish]. 1952;16:357–71.
56. I’Ons A, Jooste PL, Weight MJ, Huskisson J. A eld clinical trial of the short-term eects of iodised salt on the iodine
status of rural primary school children. S Afr Med J. 2000;90:30–6.
57. Kimball OP. The prevention of goiter in Detroit and Cleveland. JAMA. 1931;97:1877–9.
58. Pongpaew P, Tungtrongchitr R, Phonrat B, Supawan V, Schelp FP, Intarakhao C et al. Nutritional status of school
children in an endemic area of iodine deciency disorders (IDD) after one year of iodine supplementation.
Southeast Asian J Trop Med Public Health. 1998;29(1):50–7.
59. Saowakhontha S, Sanchaisuriya P, Pongpaew P, Tungtrongchitr R, Supawan V, Intarakhao C et al. Compliance of
population groups of iodine fortication in an endemic area of goitre in Northeast Thailand. J Med Assoc Thai.
1994;77:449–54.
60. Sooch SS, Ramalingaswami V. Preliminary report of an experiment in the Kangra Valley for the prevention of
Himalayan endemic goitre with iodised salt. Bull World Health Organ. 1965;32:299–315.
61. Sooch SS, Deo MG, Karmarkar MG, Kochupillai N, Ramachandran K, Ramaligaswami V. Prevention of endemic goitre
with iodised salt. Natl Med J India. 2001;14(3):185–8.
62. Sooch SS, Deo MG, Karmarkar MG, Kochupillai N, Ramachandran K, Ramalingaswami V. Prevention of endemic
goitre with iodised salt. Bull World Health Organ. 1973;49:307–12.

27
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
63. Chen ZP, Liu XL, Lin XY, Wang D, Pei KF, Zou LY et al. [Evaluation of intelligence of children born after salt iodization
intervention in IDD areas.] Zhong Guo Di Fang Bing Xue Za Zhi. [Chinese] 1991;10(Suppl.):13–15.
64. Chen ZH, Chen JA ,Qiu YN ,Zhang WH. [Eect of universal salt iodization (USI) on children’s intelligence quotient.]
Hai Xia Yu Fang Yi Xue Za Zhi. [Chinese] 2005;11(13):1–2.
65. Dong HQ, Liu CS, Guo LB. [Comparison and analysis between IDD and ISD areas on bone-age and intelligence
among children aged 7–12 years.] Zhong Guo Di Fang Bing Fang Zhi Za Zhi. [Chinese] 1988;3(Suppl.):11–12.
66. Fu LX, Zeng QZ, Deng LQ, Chen QF. [The preventive eect of iodized salt on intelligence and physical development
in endemic cretinism of Changping of Guangdong.] Symposium of the Third Chinese Endemic Goiter and
Cretinism. [Chinese] 1987:153–5.
67. He SJ, Yang ZG, Zhang BC, Long SF, Zhou WX. [Comparison analysis of children’s intelligence among minority of
Miao, Dong, Han in Ziandongnan state.] Zhong Fuo Di Fang Bing Xue Za Zhi. [Chinese] 1993;12(1):38–40.
68. Yang Z, He S, Long H Zhang BC, Zhou WX . [The relationship between iodised salt and children’s intelligence
among minority of Miao, Dong, Han.] Zhong Fuo Di Fang Bing Xue Za Zhi [Chinese] 1991;10(Suppl.):266.
69. Li JQ, Yan YQ, Zhang ZJ. [Inuence of iodine deciency on physical and psychological development.] Zhong Guo Di
Fang Bing Fang Zhi Za Zhi. [Chinese] 1991;6(Suppl.):1–3.
70. Li JQ, Yan YQ, Zhang ZJ. [Inuence of iodine deciency inuence on physical and psychological development.]
Zhong Guo Di Fang Bing Xue Za Zhi. [Chinese] 1991;10(Suppl.):40–4.
71. Yan YQ, Liu JH, Wang X, Zhang ZJ, Li JQ, Zhao XQ. [A survey in Tongjin of severe IDD area after 10 years iodised salt
implementation.] Zhong Guo Di Fang Bing Za Zhi. [Chinese] 1991;10(4):40.
72. Zou JY, Wang DQ, Liu BG, Li JQ, Ding XG, Yuan XY, et.al. [Iodine inuence on intelligence development among
children.] Endemic Disease Bulletin. [Chinese] 1987;2(5q):293–5.
73. Pan LX, Wang SJ, Shu ER. [Application of Raven intelligence test for IDD area.] Zhong Guo Di Fang Bing Xue Za Zhi.
[Chinese] 1995;10(Suppl.):251.
74. Shen JZ. [Eect of iodine intervention on intelligence of rural children in iodine deciency areas.] Zhong Guo Di
Fang Bing Fang Xhi Za Zhi. [Chinese] 1991;6(Suppl.):19–21.
75. Shu YQ. [Analysis of children’s intelligence in IDD areas after correction of iodine deciency.] Symposium of the
Third Chinese Endemic Goitre and Cretinism. [Chinese] 1987;197–8.
76. Teng RT, Xiao CL, Wang JC, Chen ZD. [The analysis of children's intelligence level in dierent periods in iodine
deciency disorders light-ward in Liaoning Province.] Zhong Guo Di Fang Bing Fang Zhi Za Zhi. [Chinese]
2009;24(3):190–1.
77. Wang J, Rong G, Yu L, et. al. [Inuences of universal salt iodization on children's intelligence and psychomotor
development]. Zhong Guo Gong Gong Wei Sheng Za Zhi. [Chinese] 2005;21(9):1054–55.
78. Wang H, Shi FK, Li KD, Wang D. [A survey report of intelligence of children born after and before salt iodization IDD
areas.] Zhong Guo Di Fang Bing Xue Za Zhi. [Chinese] 1987;6(3):144–6.
79. Wang D, Chen ZP, Wang H, Fu HY, Wang HX, Liu DF. [Intelligence problems of "normal" children in iodine deciency
areas.] Zhong Guo Di Fang Bing Xue Za Zhi. [Chinese] 1987;6(3):137–40.
80. Wang FB, Cao ZM, Yao CX, Ha HX, Ma CY, Mao CT et al. [Survey of children's intelligence improved in IDD area after
intervention of salt iodization.] Nin Xia Yi Xue Za Zhi. [Chinese] 1992;14(3):168–70.
81. Han DQ, Ha HX, Mao CT, Ma CY, Gao HT, Wang FB et al. [The inuence of iodine deciency on intelligence
development of children.] Nin Xia Yi Xue Za Zhi. [Chinese] 1990;12(6):362–4.
82. Wang FB, Cao ZM, Ha HX, Ma CY, Mao CT, You WN et al. [Survey of children's intelligence after 20 years
implementation of salt iodization in Liupanshan.] Zhong Guo Di Fang Bing Za Zhi. [Chinese] 1991;6(Suppl.):36–7.
83. Wang C. [The damage of IDD on intelligence.] He Bei Yi Yao. [Chinese] 1994;16(5):288.
84. Zheng HM, Wang Y, Li XF. [A study of prevention of Iodine deciency on social and cultural development.] Henan
Yu Fang Yi Xue Za Zhi. [Chinese] 1995;6(2):63–6.
85. Wang Y, Liu ZT, Yu HY, Wang YC, Zheng HM, Zheng S et al. [A survey of endemic sub-clinical cretinism in Lushan
county of Henan Province.] Zhong Guo Di Fang Bing Fang Zhi Za Zhi. [Chinese] 1991;6(Suppl.):11–13.

28
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
86. Zhu CK, Zhang TW, Ke JJ, Ma QL, Zuo SM, Shi ZF. [Edible salt iodization eect on mental development among
children in IDD areas.] Di Fang Bing Tong Xun. [Chinese] 1993;8(3):35–8.
87. Zeng GH. [Preventive eect of salt iodization on mild damage of neural system among children in IDD areas.]
Zhong Guo Di Fang Bing Xue Za Zhi. [Chinese] 1991;10(Suppl.):129–32.
88. Zeng GH, Zeng XQ, Li SR, Ke YQ, Chen XG, Zhou CK et. al. [Mild mental retardation caused by iodine deciency: a
type of IDD.] Zhong Guo Di Fang Bing Za Zhi. [Chinese] 1995;14(4):205–8.
89. Zhang TW, Shi ZG, Zen GH, Lu L, Zhou CK, Lin YZ et al. [Trend observation and analysis of thyroid function and
intelligence among children in IDD areas after iodine intervention.] Zhong Guo Di Fang Bing Xue Za Zhi. [Chinese]
1991;10(Suppl.):70.
90. Zhang TW, Shi ZG, Zen GH, Lu L, Zhou CK, Lin YZ et al. [Trend observation and analysis of thyroid function and
intelligence among children in IDD area after iodine intervention.] Gui Zhou Yi Yao. [Chinese] 1992;16(6):354–7.
91. Zuo SM, Zhu CK, Ke JJ, Hou YF, Sun HQ. [Survey and analysis of intelligence development and bone age among
children after iodine implementation 13 years later in IDD areas.] Zhong Guo Di Fang Bing Xue Za Zhi. [Chinese]
1996;15(4):231–3.
92. Zhu CK, Zhang TW, Ke JJ, Shi ZF, Zhuo SM. [Impact on IQ development of children by iodised salt in iodine
deciency disorders region.] Chinese Journal of Endemic Diseases. [Chinese] 1995;14(2):88–90.
93. Chen CZ, You ZS. [Observation for the short term eect of preventing goitre by iodised salt in the west Fujian.]
Fujian Medical Journal. [Chinese] 1984;6:18–22.
94. Chen CZ, You ZS. [Short term eect of preventing goitre by iodised salt in the west Fujian.] Endemic Disease
Information. [Chinese] 1984;4:38–9.
95. Foo LC, Zainab T, Nakudin M, Letchuman GR. Salt: an ineective vehicle for iodine delivery to young children in
rural Sarawak. Ann Endocrinol (Paris). 1996;57:470–5.
96. Golkowski F, Buziak-Bereza M, Tromiuk M, Baldys-Waligorska A, Szybinsk Z, Huszno B. Increased prevalence of
hyperthyroidism as an early and transient side-eect of implementing iodine prophylaxis. Public Health Nutr.
2007;10(8):799–802.
97. Guo GR. [Observation for eect of preventing endemic goitre in three years by supplementation of 1/50000
iodised salt in Houchang, Ziyuan county, Guizhou.] Guiyang Medicine. [Chinese] 1984;8(2):29–34.
98. Ibanez Gonzalez RA, Guiraun F, Escobar del Rey F, Morata Garcia, Ortiz de Landázuri E. [Results obtained with
prophylaxis of fortied salt in endemic goiter.] Rev Clin Esp. [Spanish] 1956;61:285–90.
99. Kimiagar M, Azizi F, Navai L, Yassai M, Nafarabadi T. Survey of iodine deciency in a rural area near Tehran:
association of food intake and endemic goitre. Eur J Clin Nutr. 1990;44:17–22.
100. Pedersen IB, Knudsen N, Jorgensen T, Perrild H, Ovesen L, Laurberg P. Large dierences in incidences of overt
hyper- and hypothyroidism associated with a small dierence in iodine intake: a prospective comparative register-
based population survey. J Clin Endocrinol Metab. 2002;87(10):4462–9.
101. Pedersen IB, Laurberg P, Knudsen N, Jorgensen T, Perrild H, Ovesen L et al. Increase in incidence of hyperthyroidism
predominantly occurs in young people after iodine fortication of salt in Denmark. J Clin Endocrinol Metab.
2006;91(10):3830–4.
102. Pedersen IB, Laurberg P, Knudsen N, Jorgensen T, Perrild H, Ovesen L et al. An increased incidence of overt
hypothyroidism after iodine fortication of salt in Denmark; a prospective population study. J Clin Endocrinol
Metab. 2007;92(8):3122–7.
103. Tang BD, Yang FY, Zhao H, Xu SW, Zhou QH, Zhang YW. [Observation for eects of dierent concentration
iodised salt plus iodine oil capsule in dierent severity areas of iodine deciency disorders.] Chinese Journal of
Endocrinology. [Chinese] 1992;22(4):224–6.
104. Tazhibayev S, Dolmatova O, Ganiyeva G, Khairov K, Ospanova F, Oyunchimeg D et al. Evaluation of the potential
eectiveness of wheat our and salt fortication programs in ve Central Asian countries and Mongolia, 2002–
2007. Food Nutr Bull. 2008;29(4):55–265.
105. Wang YY, Yang S. Improvement in hearing among otherwise normal schoolchildren in iodine-decient areas of
Guizhou, China, following use of iodized salt. Lancet. 1985;2(8454):518–20.

29
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
106. Wang BZ. [Observation for iodised salt in the prevention of iodine deciency disorders.] Journal of Hebei Medical
College for Continuing Education. [Chinese] 2001;18(4):22–3.
107. Weber P, Manz F. Prophylaxis of endemic goitre with iodised salt in the Federal Republic of Germany. J Trace Elem
Electrolytes Health Dis. 1987;1(1):55–6.
108. Xu J, Song SS, Sun DC. [Observation for eect of preventing endemic goitre by 1/20000 iodised salt in three years.]
Endemic Diseases Information. [Chinese] 1984;3:56–7.
109. Yang SH, Zhang BC. [Report on the Eects of 1/50,000 Iodized Salt on Endemic Goiter(EG) and Cretinism(EC)
Prevention In 4 Years in Qiandongnan Autonomous Prefecture.] Guizhou Medicine. [Chinese] 1984;8(2):53–5.
110. Zhang SH, Wang XG, Wang RT. [Observation for the eect of by supplementation of 1/10000 KIO3 iodised salt three
years for prevention and treatment of endemic goitre.] Endemic Diseases Bulletin. [Chinese] 1988;3(4):54–6.
111. Zimmermann MB, Wegmuller R, Zeder C, Torresani T, Chaouki N. Rapid relapse of thyroid dysfunction and goiter in
school-age children after discontinuation of salt iodisation. Am J Clin Nutr. 2004;79:642–5.
112. Aghini-Lombardi F, Pinchera A, Antonangeli L, Rago T, Fenzi GF, Nanni P et al. Iodised salt prophylaxis of endemic
goiter: an experience in Toscana (Italy). Acta Endocrinol. 1993;129:497–500.
113. Azizi F, Navai L, Fattahi F. Goiter prevalence, urinary iodine excretion, thyroid function and anti-thyroid antibodies
after 12 years of salt iodization in Shahriar, Iran. Int J Vitam Nutr Res. 2002;72(5):291–5.
114. Baczyk M, Ruchala M, Pisarek M, Pietz L, Junik R, Sowinski J et al. Changes in thyroid morphology and function in
children in western Poland as a result of intensied iodine prophylaxis. J Pediatr Endocrinol Metab. 2007;20(4):511–
5.
115. Bauch K, Seitz W, Förster S, Keil U. [Dietary iodine deciency in East Germany following introduction of
interdisciplinary preventive use of iodine.] Z Gesamte Inn Med. [German] 1990;41(1):8–11.
116. Bimenya GS, Olico-Okui KD, Mbona N, Byarugaba W. Monitoring the severity of iodine deciency disorders in
Uganda. Afr Health Sci. 2002;2(2):63–8.
117. Cerqueira C, Knudsen N, Ovesen L, Perrild H, Rasmussen LB, Laurberg P et al. Association of iodine fortication with
incident use of antithyroid medication: a Danish nationwide study. J Clin Endocrinol Metab. 2009;94(7):2400–5.
doi:10.1210/jc.2009-0123.
118. Charania BA. A study of the results of ten years supply of iodized salt in Gigli and Hunsa. EMR Health Serv J.
1988;5:12–18.
119. Ali A, Khan MRK, Malik ZUR, Charania BA, Bhojani FA, Baig SM. Impact of the long term supply of iodised salt to the
endemic goitre area. J Pak Med Assoc. 1992;138–40.
120. Chen KP, Lee TY, Hsu PY, Sung CC, Chen CY, Chou HM et al. [Studies on the eect of salt iodization on endemic
goiter, Taiwan. I. Mass survey on goiter of school children.] Journal of Formaosan Medical Association. [Chinese]
1976;75:471–82.
121. Chen ZJ, Chen YC. [Assessment of the eect of iodised salt supplement for preventing iodine deciency disorders.]
Chinese Public Health. [Chinese] 1999;15(8):760.
122. Chen MZ. [Investigation of iodine deciency disorders incidence before and after iodised salt supplementation in
primary school children.] Literature and Information on Prevention Medicine. [Chinese] 2001;7(4):429.
123. Chen ZJ, Chen YC, He JK. [Assessment of eect of dierent ways of iodised salt supplementation for eliminating
iodine deciency disorders.] Chinese Prevention Medicine. [Chinese] 2002;3(2):142.
124. Dai L, Liu DF, Su HJ, Zhang YF. [Eectiveness evaluation on comprehensive intervention measures for eliminating
iodine deciency disorders in coastal areas of Xiamen.] Practical Preventive Medicine. [Chinese] 2008;15(2):417–9.
125. Dai L. [Survey of iodine nutrition in children before/after iodized salt intervention.] Practical Preventive Medicine.
[Chinese] 2000;17(1):30.
126. Fei LK, Wang YS, Qin XY, He FX. [Assessment of the eect of iodised salt for prevention of iodine deciency
disorders.] Modern Prevention Medicine. [Chinese] 1996;23(4):227–9.
127. Fu LX, Xu XP, Liu M, Wang CM. [The eect of iodized supplement on children's intelligence development in iodine
deciency.] Zhong Guo Di Fang Bing Za Zhi. [Chinese] 2001;16(3):140–2.

30
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
128. Gatti V, Moralejo AB, Simioni P, Garcia JRAR. [Intake of iodine-enriched salt. Excretion of urinary iodine in Rosario
and Bahia Blanca residents.] Medicina (B Aires). [Spanish] 1980;40(4):392–6.
129. Han SQ, Zeng Q, Li W, Li YM, Lan FR, Ma W et al. [Assessment of the eect of iodised salt intervention for preventing
iodine deciency disorders in Tianjin during 1995 to 2005.] Journal of Environment and Health. [Chinese]
2006;23(6):522–4.
130. Heydarian P, Ordookhani A, Azizi F. Goitre rate, serum thyrotropin, thyroid authoantibodies and urinary iodine
concentration in Tehranian adults before and after national salt iodisation. J Endocrinol Invest. 2007;30:404–10.
131. Hou X, Wang J, Qi Q, Zhu LX, Xue XF, Tang YP et al. [Epidemiologic comparative study of hyperthyroidism before
and after salt iodisation in dierent iodine intake regions of Jilin province.] Chinese Journal of Endocrinology and
Metabolism. [Chinese] 2003;19(3):104–5.
132. Hou WZ, Wang JH. [Assessment for eect of prevention and treatment of endemic goitre by iodised salt in Lantian
county.] Occupation and Health. [Chinese] 2009;25(4):393–4.
133. Hu YL, Liao FM, Zhang AQ, Liu CY. [Assessment for eect of preventing iodine deciency disorders by iodised salt in
32 years in Taiping village in Qingling mountain area.] Chinese Primary Health Care. [Chinese] 1998;12(6):24–5.
134. Jooste PL, Weight MJ, Lombard CJ. Short-term eectiveness of mandatory iodization of table salt, at an elevated
iodine concentration, on the iodine and goiter status of schoolchildren with endemic goiter. Am J Clin Nutr.
2000;71:75–8.
135. Kimball OP. Iodized salt for the prophylaxis of endemic goiter. JAMA. 1946;130:80–1.
136. Lv SM, Xie LJ, Zhou RH, Chong ZS, Jia LH, Ma J et al. [Control of iodine deciency disorders following 10 year
universal salt iodization in Hebei Province of China.] Biomedical and Environmental Sciences. [Chinese]
2009;22:472–9.
137. Mostafavi H. Eect of adequate salt iodization on the prevalence of goiter: a cross- sectional comparative study
among school children aged 6–18 years during 1989 and 1995. Pak J Med Sci. 2005;21(1):53–5.
138. Nicod JL. [Endemic goitre in Switzerland and its prophylaxis with iodized salt.] [French] Bull World Health Organ.
1953;9:259–73.
139. Regalbuto C, Scollo G, Pandini G, Ferrigno R, Pezzino V. Eects of prophylaxis with iodised salt in an area of
endemic goitre in north-eastern Sicily. J Endocrinol Invest. 2010;33:300–305. doi:10.3275/6715.
140. Rueda Williamson R, Pardo Tellez F. [Prevention of endemic goitre in Colombia.] Bol Ocina Sanit Panam. [Spanish]
1966;61(6):495–503.
141. Salvaneschi JP, Salvaneschi BG, Moralejo AB, Garcia JRAR. [Endemic goiter in Argentina epidemiologic studies
before (1968) and during (1986) prophylaxis with iodine salt in the city of Buenos Aires.] Medicina (B Aires).
[Spanish] 1991;52(2):99–105.
142. Scrimshaw N, Vega Franco L, Arellano R, Sagastume C, Méndez JI, De León R. [Eect of salt iodization on the
prevalence of endemic goiter in school children in Guatemala.] Bol Ocina Sanit Panam. [Spanish] 1966;60(3):222–8.
143. Szybinski Z, Delange F, Lewinsski A, Podoba J, Rybakowa M, Wasik R et al. A programme of iodine supplementation
using only iodised household salt is ecient – the case of Poland. Eur J Endocrinol. 2001;144:331–7.
144. Vejbjerg P, Knudsen N, Perrild H, Laurberg P, Carle A, Pedersen IB et al. Lower prevalence of mild hyperthyroidism
related to a higher iodine intake in the population: prospective study of a mandatory iodisation programme. Clin
Endocrinol. 2009;71:440–5. doi:10.1111/j.1365-2265.2008.03493.x.
145. Wang SL, Wang WD. [Observation for eect of preventing endemic goitre by 1/10000 iodised salt.] Endemic
Diseases Information. [Chinese] 1981;2:48–53.
146. Wang SY, Zhao Q, Yu ZH. Summary for preventing endemic goitre and endemic cretinism for 21 years in Chengde
suburbs. Endemic Diseases Information 1985;(4):61–5.
147. Wang YH, Jia JH, Cu XH, Zhao GB, Pan TR, Xu SQ et al. [Eect of universal iodine supplementation in Anhui province
for one year – studies on iodine status and thyroid function of children.] Chinese Journal of Endocrinology.
[Chinese] 2000;19(2):121–3.
148. Wang Y, Zhang Z, Ge P, Wang Y, Wang S. Iodine deciency disorders after a decade of universal salt iodization in a
severe iodine deciency region in China. Indian J Med Res. 2009;130:413–17.

31
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
149. Zhang YX, Yang HX, Zhu XN, Li L, Yang C, Cao YQ et al. [Exploring the impact of damage children intelligence
quotient by iodine deciency in Gansu province.] Chinese Journal of Epidemiology. [Chinese] 2003;24(1):76–7.
150. Wei JL. [Observation for eect of preventing endemic goitre by iodised salt in Daxin County.] Guangxi Medicine.
[Chinese] 1985;7(4):194–6.
151. Xue YX, Zhang JX. [Investigating and monitoring the prevention of endemic goitre by iodised salt in Hami District.]
Endemic Diseases Bulletin. [Chinese] 1993;8(4):118–19.
152. Yang LZ, Yang WY. [Evaluation of prevention and control of iodine deciency disorder in Jiaojiang District, Taizhou
city, Fujian province, 1995–2009.] Disease Surveillance. [Chinese] 2011;26(3):244–8.
153. Yuan BM, Zhou MT, Gen XX. Observation for eect of preventing endemic goitre by iodised salt in 7–14 years
children in endemic area. J Prev Med Info. 1993;(Suppl. 1):140.
154. Yusuf HK, Rahman AM, Chowdhury FP, Mohiduzzaman M, Banu CP, Sattar MA et al. Iodine deciency disorders in
Bangladesh, 2004–05: ten years of iodized salt intervention brings remarkable achievement in lowering goitre and
iodine deciency among children and women. Asia Pac J Clin Nutr. 2008;17(4):620–8.
155. Zhou ZZ, Ma XW, Meng JH, Chao ZP. [Eect analysis of iodised salt for preventing iodine deciency disorders in
eight years.] Chinese Journal of Endemic Diseases. [Chinese] 2004;23(1):90.
156. Zimmermann MD, Hess SY, Adou P, Toresanni T, Wegmuller R, Hurrell RF. Thyroid size and goitre prevalence after
introduction of iodised salt: a 5-y prospective study in schoolchildren in Cote d’lvoire. Am J Clin Nutr. 2003;77:663–7.
157. Jia ZJ, Zhang XY, Wang JL, Duan SZ, Lu YL. [Assessment of eect of iodised salt for preventing iodine deciency
disorders in Heshui County.] Health Occupation Education. [Chinese] 2004;22(17):117.
158. Yang SH. [Observation for eect of preventing endemic goitre and cretinism by 1/50000 iodised salt in Pingliang,
Kaili, Guizhou province.] Guizhou Medicine. [Chinese] 1984;8(2):49–52.
159. Huang WJ, Chen ZH, Lin SG, Lin ZH, Wang MH. [Children's intelligence quotient level before and after universal salt
iodisation (USI) in iodine deciency disorder regions in Fujian province.] China Journal of Endocrinology. [Chinese]
2004;23(1):64–5.
160. Liu P, Liu SJ. Research on prevention and control iodine deciency disorders of dierent districts in China. Chinese
Clinical Trial Registry and WHO ICTRP Search Portal: http://apps.who.int/trialsearch/Trial.aspx?TrialID=ChiCTR-
ECH-09000327 Date of registration: 08 January 2009.
161. Marquis GS, Aboud FS. A study to measure the eect of switching the salt supply from non-iodized to iodized
on cognitive development in Ethiopia. Registered 20 April 2011. WHO ICTRP Search Portal (http://apps.who.int/
trialsearch/Trial2.aspx?TrialID=ChiCTR-ECH-09000327, accessed 24 October 2014).
162. Schulz KF, Altman DG, Moher D, CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting
parallel group randomised trials. Int J Surg. 2011;9(8):672–7. doi:10.1016/j.ijsu.2011.09.004.
163. Angermayr L, Clar C. Iodine supplementation for preventing iodine deciency disorders in children. Cochrane
Database Syst Rev. 2004;(2):CD003819. doi:10.1002/14651858.CD003819.pub2.
164. Hetzel BS, Mano MT. A review of experimental studies of iodine deciency during fetal development. J Nutr.
1989;119:145–151.
165. Delange F. Administration of iodized oil during pregnancy: a summary of the published evidence. Bull World
Health Organ. 1996;74(1):101–8.
166. Zimmermann MB. The eects of iodine deciency in pregnancy and infancy. Paediatr Perinat Epidemiol.
2012;26(Suppl. 1):108–17. doi:10.1111/j.1365-3016.2012.01275.x.
167. UNICEF/WHO. World Summit for Children: mid-decade goal – iodine deciency disorders. In: UNICEF–WHO Joint
Committee on Health Policy Special Session, Geneva, 27–28 January 1994. Geneva: World Health Organization;
1994 (http://www.ceecis.org/iodine/01_global/01_pl/01_01_1994_summit.pdf, accessed 24 October 2014).
168. Prevention of cardiovascular disease: pocket guidelines for assessment and management of cardiovascular risk.
Geneva: World Health Organization; 2007 (http://www.who.int/cardiovascular_diseases/guidelines/PocketGL.
ENGLISH.AFR-D-E.rev1.pdf, accessed 24 October 2014).

32
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
169. WHO Secretariat, Andersson M, de Benoist B, Delange F, Zupan J. Prevention and control of iodine deciency in
pregnant and lactating women and in children less than 2-years-old: conclusions and recommendations of the
technical consultation. Public Health Nutr. 2007;10(12A):1606–11. doi:10.1017/S1368980007361004.
170. Bleichrodt N, Shrestha RM, West CE, Hautvast JG, van de Vijver FJR, Born MP. The benets of adequate iodine intake.
Nutr Rev. 1994;54(4 Pt 2):S72–S78.
171. Dunn JT, Delange F. Damaged reproduction: the most important consequence of iodine deciency. J Clin
Endocrinol Metab. 2001;86(6):2360–3.
172. Speeckaert MM, Speeckaert R, Wierckx K, Delanghe JR, Kaufman JM. Value and pitfalls in iodine fortication and
supplementation in the 21st century. Br J Nutr. 2011;106(7):964–73. doi:10.1017/S000711451100273X.
173. Galofre JC, Fernandez-Calvet L, Rios M, Garcia-Mayor RV. Increased incidence of thyrotoxicosis after iodine
supplementation in an iodine sucient area. J Endocrinol Invest. 1994;17(1):23–7.
174. Todd CH, Allain T, Gomo ZAR, Hasler JA, Ndiweni M, Oken E. Increase in thyrotoxicosis associated with iodine
supplements in Zimbabwe. Lancet. 1995;346:1653–64.
175. Bourdoux P, Ermans AM, Mukalay WA, Mukalay A, Filetti S, Vigneri R. Iodine-induced thyrotoxicosis in Kivu, Zaire.
Lancet. 1996;347:552–3.
176. Dunn JT, Semigran MJ, Delange F. The prevention and management of iodine-induced hyperthyroidism and its
cardiac features. Thyroid. 1998;8(1):101–6.

33
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Annex 1. Electronic search strategy
Unless otherwise stated, search terms are free text terms; MeSH = Medical Subject Heading
(Medline medical index term); exp =exploded MeSH; the dollar sign ($) stands for any character(s);
the question mark (?) = to substitute for one or no characters; tw = text word; pt = publication type;
sh = MeSH; adj = adjacent.
The Cochrane Library (Issue 5 2011) and EMBASE (1966 to 18 June, 2011), MEDLINE (PubMed 1966
to 31 May 2011); The Virtual Health Library of the Pan-American Health Organization (May 2011);
WHO International Clinical Trials Registry Platform (18 June 2011); and the WHO Global Health
Library regional databases (June 2011) used a similar search strategy:
#1 iodine deciency disorders/[MeSH term, all subheadings included]
#2 goitre
#3 endemic goitre
#4 cretinism
#5 hyperthyroidism
#6 hypothyroidism
#7 #1~#6/OR
#8 iodized salt [in all elds]
#9 salt
#10 salt iodization
#11 iodine fortication
#12 #8~#11/OR
#13 #7 AND #12
A similar search strategy was used for The China National Knowledge Infrastructure (May 2011);
VIP (the register of Chinese trials developed by the Chinese Cochrane Centre); and WANFANG,
using the following translated terms:
#1 碘缺乏病
#2 克汀病
#3 甲状腺肿
#4 地方性甲肿
#5 甲状腺机能亢进
#6 甲亢
#7 甲状腺机能低下
#8 甲低
#9 #1~#8/OR
#10 碘
#11 碘盐
#12 #10 OR #11
#13 #9 AND #12

34
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Annex 2. Characteristics of included studies
Aghini-Lombardi et al., 1993 (112)
Methods Multiple cross-sectional observational study conducted in Toscana, Italy
Participants All schoolchildren (6–14 years) from two villages
Interventions Iodized salt at 15 parts per million (ppm) became available in the two villages. Before the
intervention, no iodized salt was available for purchase.
Outcomes 1. Goitre rate measured by palpation
2. Thyroid volume determined by ultrasound examination
(only in the follow-up survey and therefore not included in this review)
3. Urinary iodine concentration
Publication details Published in a peer-reviewed journal
Stated aim of study To evaluate the ecacy of local legislative measures of iodized salt 10 years after introducing
the iodized salt
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 15 ppm
6. Duration: 10 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre prevalence)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding of participants or personnel
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with similar number of
inhabitants and school-aged children at both time
points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Same villages surveyed

35
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Azizi et al., 2002 (113)
Methods Multiple cross-sectional observational study conducted in Shahriar, Islamic Republic of Iran
The sampling technique was stratied random sampling, resulting in 48 urban and rural
centres as primary sampling units. The same 48 primary sampling units selected for the
rst survey were sampled during the follow-up surveys. The rst survey was conducted
before iodized salt was available and the second survey was conducted 12 years after the
introduction of iodized salt in the country.
Participants All adults and children in the randomly selected households
Interventions Iodized salt at a concentration of 40 ppm (parts per million) in the form of potassium iodide
was made available in the area.
Outcomes 1. Goitre
2. Urinary iodine excretion (UIE; µg/L)
3. Thyroid hormones – levels of tri-iodothyronine (T3), thyroxine (T4) and thyroid-stimulating
hormone (TSH)
4. Anti-thyroid antibodies (anti-M and anti-Tg)
5. Hypothyroidism (reported qualitatively)
6. Hyperthyroidism (reported qualitatively)
Publication details Published in a peer-reviewed journal
Stated aim of study To compare the prevalence of goitre, UIE, thyroid hormone levels, and thyroid antibodies
between 1983 and 1995
Notes 1. Design: multiple cross-sectional
2. Age: all
3. Group: non-pregnant
4. Salt consumption: unknown
5. Iodine concentration: 40 ppm
6. Duration: 12 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre prevalence)
The data on prevalence of goitre data were given by percentage only, it was not possible to
extract exact the number of goitres. The rate of consuming iodized salt was investigated and
found to be approximately 50% of households consuming iodized salt from 1991 to 1994, and
90% in 1995 at the time of the follow-up survey.
Authors reported qualitatively that there was no dierence in the prevalence of hypo- or
hyperthyroidism between the two surveys.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding of participants or personnel
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Multiple cross-sectional samples drawn in similar
fashion with more participants at follow-up than
baseline
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Multiple cross-sectional samples drawn in similar
fashion and same villages surveyed

36
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Baczyk et al., 2007 (114)
Methods Multiple cross-sectionalobservational study conducted in Wielkopolska Region, Poland
Cross-sectional surveys of goitre, urinary iodine excretion (UIE), and adverse eects were
conducted before and 9years after obligatory salt iodization in Poland.
Participants Children aged 8–12 years, living in urban and rural areas and of both sexes
Interventions Iodized salt at a concentration of 30 mg/kg (potassium iodide) was made available.
Outcomes 1. Goitre as measured by ultrasound
2. Hyperthyroidism
3. Hypothyroidism
4. UIE
5. Anti-thyroglobulin antibodies (TgAb) (data not used because only measured at baseline)
6. Anti-thyroid microsomal antibodies (data not used because only measured at follow-up)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To provisionally determine all changes of children's thyroid pathology occurring after the
introduction of salt iodization
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 30 ppm (parts per million)
6. Duration: 9 years
7. Baseline iodine deciency disorder (IDD) status: moderate IDD based on median UIE
• All data were presented as percentages without condence intervals; the number of
events was calculated by the percentages and the total number of participants.
• 1992 and 2005 data were compared as the endpoint but not 2000.
• For UIE only, the population median was reported without a range and therefore the
data could not be used in the quantitative analysis (baseline: 49 µg/L and follow-up:
107 µg/L).
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Participants and personnel not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Unclear for laboratory outcomes; not blinded for
goitre
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional surveys with similar sample
sizes
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Sampling methodology in two surveys not
described in sucient detail to discern potential bias

37
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Bauch et al., 1990 (115)
Methods Multiple cross-sectional observational study conducted in the former East Germany
Adults were selected from four geographic regions of East Germany to measure urinary
iodine concentration before the introduction of the blanket prophylaxis programme of salt
iodization at the national level in East Germany. Salt iodization was introduced in salt for
human consumption in 1985 and for animal consumption in 1986. The rst cross-sectional
survey was conducted in 1985. In 1986/1987, a second survey was conducted to measure
urinary iodine concentration from persons living in the same areas of East Germany after the
introduction of iodized salt.
Participants Sample of adult population of East Germany
Interventions Iodized salt was provided at a concentration of 25 ppm (parts per million) in a national
programme.
Outcomes Urinary iodine excretion (UIE)
Publication details Published in a peer-reviewed journal in German
Stated aim of study To evaluate the eect of the blanket prophylaxis programme of salt iodization at the national
level in East Germany
Notes 1. Design: multiple cross-sectional
2. Age: adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 25 ppm
6. Duration: 1–2 years
7. Baseline iodine deciency disorder (IDD) status: not reported
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Participants not blinded
Blinding of outcome assessment
(detection bias)
High risk Not blinded
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional surveys with larger sample
size at follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Similar sampling methodology in two surveys

38
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Bimenya et al., 2002 (116)
Methods Multiple cross-sectional observational study conducted in Uganda.
A sample of 2880 schoolchildren aged 6–12 years from 72 primary schools in six districts of
Uganda were selected for evaluation of goitre in 1999, 5 years after the provision of iodized
salt in the form of universal salt iodization in the country of Uganda. The results of goitre were
compared to the results of a similar study conducted in 1991, 3 years before universal salt
iodization in Uganda.
Study districts were randomly selected but geographic representation was ensured through
stratication of country. Probability proportionate to size (PPS) sampling was used to select
the sample of schools and random sampling was used to select the students from each
school; 40 students attending selected schools were included.
Participants Schoolchildren aged 6–12 years
Interventions Iodized salt was provided at a concentration of 50 ppm (parts per million) through universal
salt iodization in the country.
Outcomes 1. Goitre
2. Urinary iodine excretion (UIE; lack of baseline data precluded use of this outcome in this
review)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To monitor the universal salt iodization programme by determining the prevalence of goitre
in the country, establishing the proportion of household populations consuming adequately
iodized salt, and determining the levels of iodine intake using UIE
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm
6. Duration: 5 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre prevalence)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Participants not blinded
Blinding of outcome assessment
(detection bias)
High risk Assessors not blinded
Incomplete outcome data (attrition bias) Unclear risk Multiple cross-sectional surveys with unclear
number of participants at baseline
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Sampling methodology for baseline survey not
described

39
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Cerqueira et al., 2009 (117)
Methods Multiple cross-sectional observational study conducted in Denmark
This study was a nationwide registry study covering all inhabitants in Denmark during the period
1 January 1995 to 31 December 2007. Every person in Denmark is assigned a unique, permanent,
10-digit identication number at birth or immigration. The identication number facilitates the use
of registers and makes individual longitudinal follow-up possible. Pharmacies are required by law
to register all prescriptions dispensed at an individual level, which makes the registry highly valid.
The register includes information on the identication number of the patient, date of dispensing,
ATC (Anatomical, Therapeutic, Chemical) code, strength, and quantity dispensed (in dened daily
doses). Both anti-thyroid medication and thyroid hormone are sold solely as prescription drugs
in Denmark and were registered. The incidences of thyroid conditions were indicated through
monitoring the incident use of anti-thyroid medication.
Participants All inhabitants in Denmark
Interventions Iodized salt was provided at a concentration of 13 ppm (parts per million; potassium iodide)
through the national programme. It was estimated at 50 µg/day.
Outcomes The number of incident users of anti-thyroid medication
Publication details Published in a peer-reviewed journal in English
Stated aim of study To analyse the nationwide eect of the Danish iodine fortication programme on the
incidence of hyperthyroidism, by monitoring the incident use of anti-thyroid medication
Notes 1. Design: multiple cross-sectional
2. Age: children and adults
3. Group: not specically pregnant women
4. Salt consumption: 4.5 g/day
5. Iodine concentration: 13 ppm
6. Duration: 12 years
7. Baseline iodine deciency disorder (IDD) status: combination of a moderate IDD region and
a mild IDD region (based on median µg/L)
• Iodine intake was estimated at 50 µg/day.
• Data in this paper allow the estimate of salt consumption of 4.5 g/day; however, the study of
Pedersen et al., 2002 (100) also has data regarding salt iodization in Denmark, which suggests
a salt consumption of 5.5–6 g/day.
• Iodine fortication was at a concentration of 8 ppm potassium iodide to all salt from June 1998
on a voluntary basis. Iodization of salt became mandatory in July 2000, at a concentration of
13 ppm potassium iodide. All salt for household use and for commercial production of bread
was fortied.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Participants and personnel not blinded
Blinding of outcome assessment
(detection bias)
High risk Not blinded
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional surveys with large numbers
of participants at all time points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Methodology for selecting samples same at all time
points

40
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Charania et al., 1988 (118, 119)
Methods Multiple cross-sectional observational study conducted in Pakistan.
Villages in the Gilgit and Hunza regions of Pakistan were stratied by accessibility to a main
road and then randomly selected. Households within villages were then randomly selected.
Villagers in randomly selected households had goitre and urinary iodine concentration
measured. The rst survey was conducted in 1978 and is described as baseline (pre-
iodization) though the manuscript says salt fortied with iron became available in Pakistan
in 1977. Iodized salt then became available in this region of Pakistan. A second survey was
completed in 1987 (9 years later). Samples of participants were selected in the same way
and came from the same villages.Goitre was measured by palpation, using World Health
Organization (WHO) classication.
Participants Children (less than 5 years of age) and ALL (children 5 years of age and older, adolescents and
adults reported together)
Interventions Iodized salt was provided through a national programme, at a concentration of 30 ppm (parts
per million).
Outcomes 1. Goitre
2. Urinary iodine excretion (UIE; only measured at follow-up so not included in this review)
Publication details Two articles published in peer-reviewed journals in English, reporting the same results of this
study
Stated aim of study 1.To ascertain the magnitude, extent and pattern of iodized salt consumption
2. To ascertain the impact of the supply of iodized salt to the target area
3. To ascertain the eect of health education on the community and the extent of their
knowledge of the benets of iodized salt
4. To establish the dierence in impact of iodized salt on areas with easy access as compared
with remote areas, and on the roadside villages as compared with remote villages
Notes 1. Design: multiple cross-sectional
2. Age: children (and ALL ages but adults not reported separately)
3. Group: not specically pregnant women
4. Salt consumption: 8 g/day (estimated)
5. Iodine concentration: 70 ppm (however, the results show that the amount of salt in
households that was actually iodized varied by accessibility of village)
6. Duration: 9 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD based on goitre (not reported
using WHO criterion of school-aged children)
• Iodine consumption 300 µg/person/day
Iodine concentration in salt was found to dier by household. Accessibility to the road
was associated with the presence of iodized salt in the household.

41
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
High risk Not blinded
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional surveys with more
participants at follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Methodology for selecting samples said to be the
same but sample size was twice as large at follow-
up, putting the methodology into question

42
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Chen et al., 1976 (120)
Methods Multiple cross-sectional observational study conducted during 1964–1965 and June
1971 in Taiwan, China
Participants Total of 155 primary schools, including 89 926 children at baseline and 77 605 children at
follow-up
Interventions Universal salt iodization was implemented in the Taiwan in 1969. There were ve areas of
investigation, mainly categorized by baseline iodine status:
1. hyperendemic (goitre >40% in boys: >50% girls) – iodized salt measured at 28 ppm (parts
per million);
2. mesoendemic (goitre 20–40% in boys: 25–50% girls) – iodized salt measured at 29 ppm;
3. hypoendemic (goitre 2–20% in boys: 5–25% girls) – iodized salt measured at 29 ppm;
4. nonendemic (goitre <5% in boys and girls) – iodized salt measured at 33 ppm;
5. aboriginal (not dened) – iodized salt measured at 28 ppm.
Outcomes Goitre prevalence
Publication details Published in English language in a peer-reviewed journal
Stated aim of study To examine the eectiveness of salt iodization on goitre prevalence in schoolchildren in
Taiwan
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 28–33 ppm
6. Duration: 6 years
7. Baseline iodine deciency disorder (IDD) status: moderate IDD (based on goitre
prevalence)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported but unknown whether
outcome assessors were blinded
Incomplete outcome data (attrition bias) Low risk Low attrition rate (high participant numbers at both
time points)
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Samples selected in a similar way at both time
points

43
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Chen et al., 1984 (93, 94)
Methods Cohort observational study conducted in China
Participants 5329 villagers were investigated
Interventions Iodized salt was provided. The iodine concentration reduced across the time:
• 18.948 mg/kg in the rst month;
• 16.543 mg/kg in the third month;
• 9.77 mg/kg in the sixth month.
No control intervention was used.
Outcomes 1. Goitre prevalence
2. Urinary iodine concentration (only reported qualitatively)
3. Test for tri-iodothyronine (T3), thyroxine (T4), thyroid-stimulating hormone (TSH)
concentration (only reported qualitatively)
Publication details Published in a Chinese language journal
Stated aim of study To estimate the eect of iodized salt on goitre prevalence
Notes 1. Design: cohort observational
2. Age: children and adults
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration:
• 18.948 mg/kg in the rst month; 20 ppm (parts per million)
• 16.543 mg/kg in the third month; 16 ppm
• 9.77 mg/kg in the sixth month; 10 ppm
6. Duration: 6 months
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre prevalence;
did not use World Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported/unknown
Incomplete outcome data (attrition bias) Low risk Low attrition
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Methods not described in detail

44
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Chen et al., 1991 (63)
Methods Quasi-experimental study conducted in Liaoyang, Fuxin, Fushun, Xinchengzi of Liaoning
province, China
Children of villages who were born before iodization of salt were compared to children in the
villages born after iodization of salt.
Participants Children aged 5–15 years at the time that intelligence was measured (1988)
Cretinism cases were excluded from the study.
Interventions The experiment group was made up of children whose mothers received iodine from iodized
salt during pregnancy and children were exposed to iodized salt in infancy; the comparison
group comprised children whose mothers did not receive iodine from salt during pregnancy
and children who were not exposed to iodized salt until an undetermined age.
There were three cohorts (not clear if this categorization was by village or by group of
children at baseline based on urinary iodine excretion [UIE]):
• mild iodine decient;
• moderate iodine decient;
• severe iodine decient.
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese language journal
Stated aim of study Not stated
Notes Intelligence measured with Combined Raven’s Test for Rural China
Comparisons made within iodine deciency category (not clear if the categories were
analogous to the village), to control for potential biases
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown (not reported)
6. Duration: unknown length of intervention group exposure to iodized salt other than
gestation
(9 months) compared to comparison group
7. Baseline iodine deciency disorder (IDD) status: three cohorts (not clear if this
categorization was by village or by group of children at baseline based on UIE):
• mild iodine decient
• moderate iodine decient
• severe iodine decient
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No mention of blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No mention of blinding
Incomplete outcome data (attrition bias) Unclear risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Groups comparable at baseline – same villages/
same IDD status

45
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Chen et al., 1999 (121)
Methods Multiple cross-sectionalstudy, before–after design carried out in the Hui-an county, Fujian
province, China
Participants Randomly sampled 8–10-year-old children from 30 primary schools, each school sampled 40
children
Total:
• 1343 children in 1995
• 1247 children in 1997
Interventions A survey was conducted in 1995 before the availability of iodized salt. Salt then became
iodized throughout the province at a concentration of 33 ppm (parts per million). There was
then another survey conducted with the same methods in 1997.
Outcomes 1. Goitre prevalence
2. Urinary iodine concentration µg/L (only median reported so included in summary table but
not meta-analysed)
Publication details Published in a Chinese language journal
Stated aim of study Report of monitoring of universal salt iodization programme
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 33 ppm (parts per million)
6. Duration: 2 years
7. Baseline iodine deciency disorder (IDD) status: moderate IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
High risk No blinding
Incomplete outcome data (attrition bias) Low risk Two cross-sectional surveys of similar sample size
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Same schools sampled at both time points to
decrease bias

46
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Chen et al., 2001 (122)
Methods Multiple cross-sectional study, before–after design, conducted in Nonghai city, Fuanjian
province, China
Participants 30 primary schools selected; 40 children randomly sampled from each school; a total of 1200
children examined in both before and after supplementation of the iodized salt
Interventions A survey in the schoolchildren was conducted in 1995 before the availability of iodized salt.
In 1996, the government started providing families iodized salt on a monthly basis. A second
survey was conducted in the same schools in 1997. The concentration of Iodine in the salt
was 33 ppm (parts per million; 1/30 000).
Outcomes 1. Goitre prevalence
2. Urinary iodine concentration (median reported; results found in summary table)
3. % of students with urinary iodine <100 µg/L
Publication details Published in a Chinese language journal
Stated aim of study To observe the eectiveness of using iodized salt for iodine deciency disorder (IDD)
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. concentration: 33 ppm
6. Duration: 2 years
7. Baseline IDD status: moderate IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported/unknown
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional design with same sample
size at both time points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Same methodology to choose participants from
same population groups at each time point

47
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Chen et al., 2002 (123)
Methods Multiple cross-sectional observational study conducted in Hui-an, Fujian, China
Primary schoolchildren aged 8–10 years were randomly selected for inclusion in the baseline
survey of goitre before the introduction of iodized salt. A follow-up survey was conducted
3years after the introduction of iodized salt and goitre was again measured in randomly
sampled schoolchildren.
Participants Schoolchildren aged 8–10 years
Interventions Two methods were used to introduce iodized salt in two separate provinces. In Hui An County
(HAC), the government distributed iodized salt on the basis of 6 kg/person/year and non-
iodized salt was taken completely o the market. In Quan Gang Region (QGR), iodized salt
was introduced in the market and non-iodized salt was also still allowed to be in the market.
There were two cohorts:
• HAC (only iodized salt available);
• QGR (iodized and non-iodized salt on the market).
Outcomes 1. Goitre prevalence
2. UIE (reported as median in manuscript and data found in summary table without meta-
analysis)
3. % of children with urinary iodine excretion (UIE) <100 µg/L (no baseline numbers provided
and therefore this outcome could not be included in the analyses or tables)
Publication details Published in a Chinese language journal
Stated aim of study To compare the eect of the two methods of availability of iodized salt on the prevention of
iodine deciency disorder (IDD)
(For the purpose of the systematic review, the data from the post-distribution surveys at both
localities were compared to the pre-distribution survey in the same location.)
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 3 years
7. Baseline IDD status: moderate IDD (based on goitre)/mild IDD (based on UIE)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation (selection
bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment
(selection bias)
High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional survey with similar number
of participants at each time point
Selective reporting (reporting bias) Unclear risk No baseline data provided for % of children with
UIE <100 µg/L
Other bias Unclear risk Methodology for follow-up survey was not
described though it was assumed to be the same as
baseline as the sample size was similar.

48
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Chen et al., 2005 (64)
Methods Quasi-experimental study conducted in Jiangshan township, Longyan city, Fujian province, China
Children of the severe iodine deciency disorder (IDD) area that were born before iodization
of salt (comparison group) were compared to children in the villages born after iodization
of salt (experimental group). Iodized salt was supplied to the village, beginning in 1982.
Children were sampled (sampling methodology not reported) in 1990 and intelligence tests
administered.
Participants All children aged 7–14 years living in the villages at time that intelligence was measured were
eligible and sampling was undertaken but methods not described (1985)
Interventions Universal salt iodization was carried out at unknown concentration.
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Combined Raven’s Test in China (CRT-C2)
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown (not reported)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) and
duration of exposure during infancy and early childhood varied from 1 to 6 years
7. Baseline IDD status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition was not reported
but sample sizes were fairly similar and samples
relatively large
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Children sampled from the same villages, to reduce
bias

49
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Dai et al., 2008 (124, 125)
Methods Multiple cross-sectional study conducted in Xiamen city, Fujian province, China
Schoolchildren aged 8–10 years were randomly sampled from 90 schools before
supplementation of iodized salt in 1995. Goitre and urinary iodine excretion (UIE) were
measured in these children. Another similar cross-sectional survey was conducted 12 years
after introduction of iodized salt.
Participants Schoolchildren aged 8–10 years
Interventions Comprehensive intervention was used in local salt mines and manufacturing facilities, to all
salt on the market. Coverage was reported at 90% at follow-up survey.
There were three cohorts; though the intervention was the same, the results were divided
into:
• urban;
• suburban;
• rural.
Outcomes 1. Goitre prevalence
2. UIE (reported as median in manuscript and results found in summary table but not used in
meta-analysis)
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To assess the eect of iodized salt on goitre and UIE in school-aged children
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 12 years
7. Baseline iodine deciency disorder (IDD) status: based on UIE/(based on goitre):
• urban: adequate iodine status/(mild IDD)
• suburban: adequate iodine status/(moderate IDD)
• rural: mild IDD/(moderate IDD)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Multiple cross-sectional survey with fewer
participants at follow-up than baseline
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Survey methodology similar at all time points and
probability proportionate to size (PPS) sampling
used

50
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Dong et al., 1988 (65)
Methods Quasi-experimental study conducted in Jiuling and Daxi villages in China
Children of each village who were born before iodization of salt were compared to children in
the village born after iodization of salt.
Participants Children aged 7–13 years at the time that intelligence was measured
Interventions Experiment group was made up of children whose mothers received iodine from iodized
salt during pregnancy and children were exposed to iodized salt in infancy; the comparison
group comprised children whose mothers did not receive iodine from salt during pregnancy
and children were not exposed to iodized salt until 1–3 years of age.
There were two cohorts:
• iodine-decient village (before and after salt iodization);
• iodine-sucient village (before and after salt iodization).
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Comparisons made within the village, to control for potential biases
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 33 ppm (parts per million) or 20 ppm (as stated in article)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus 1–3
years of life more than comparison
7. Baseline iodine deciency disorder (IDD) status: based on goitre:
• iodine-decient village: moderate IDD
• iodine sucient village: adequate iodine nutrition
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No mention of blinding
Blinding of outcome assessment
(detection bias)
High risk No mention of blinding
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Samples comparable at baseline (similar cultural and
educational conditions and same villages)

51
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Fei et al., 1996 (126)
Methods Multiple cross-sectional observational study conducted in Rizhao city, Shandong
province, China
A random sample of adults and children was selected for a survey of goitre before iodization
of salt. Another survey was conducted 10 years after iodization of salt, to measure goitre in
the population.
Participants Adults and children living in Rizhao city, Shandong province, China
Interventions Iodized salt was provided at a concentration of 33 ppm (parts per million).
Outcomes 1. Goitre prevalence
2. urinary iodine excretion (UIE)
3. % of children with UIE <100ug/L
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To assess the eect of supplementation of iodized salt on goitre and UIE
Notes 1. Design: multiple cross-sectional
2. Age: children and adults
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 33 ppm
6. Duration: 10 years
7. Baseline IDD status: severe IDD (based on goitre prevalence)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No mention of blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No mention of blinding
Incomplete outcome data (attrition bias) Unclear risk Multiple cross-sectional survey with >40 000
participants at baseline and only approximately
4000 at follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias High risk Methods were unclear but the very large dierence
in sample size makes the potential for bias due to
sampling dierences high

52
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Foo et al., 1996 (95)
Methods Cohort observational study in Malaysia
The original study design was a randomized controlled trial (RCT) conducted in Malaysia. The
groups were randomly allocated to receive iodized salt or iodized water. The randomization
was not described.
The data used in this analysis from this study were the baseline and follow-up data from the
group that received the iodized salt. The randomization to groups did not include a control
group and therefore the comparison between group data could not be used. A baseline
survey was conducted. The group received iodized salt for 1 year, at which time a follow-up
survey in the same participants was conducted.
Participants Women aged 15–40 years and children 6years and younger
Interventions Iodized salt was provided at 50 ppm (parts per million).
A second village received iodized water but the results from that village were not included in
the review.
There were two cohorts:
• women;
• children.
Outcomes 1. Goitre prevalence (data were only reported for women)
2. Urinary iodine excretion (UIE; median population urinary iodine concentration is reported
for before and after intervention; however, the range is not reported. The data from the
median could not be used in the quantitative summary analysis)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To compare the eect of iodized salt and iodized water on goitre and UIE in women and
children in rural villages in Malaysia
Notes 1. Design: cohort observational
2. Age: children and adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm
6. Duration: 1 year
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre; did not use
World Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not RCT
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk No loss to follow-up reported
Selective reporting (reporting bias) Unclear risk All outcomes reported for women but goitre not
reported for children
Other bias Low risk Groups chosen to be similar in living conditions,
access to road and city, etc.

53
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Fu et al., 1987 (66)
Methods Quasi-experimental study conducted in Qiaotou and Dunzhi of Dayuan township, China
Children of each village who were born before iodization of salt were compared to children in the
village born after iodization of salt.
Participants Children aged 4–15 years at the time that intelligence was measured (1987)
Cluster sampling method used for selection of participants
Interventions The experiment group was made up of children whose mothers received iodine from iodized salt
during pregnancy and children were exposed to iodized salt in infancy; the comparison group
comprised children whose mothers did not receive iodine from salt during pregnancy and children
were not exposed to iodized salt until 1–2 years of age.
• One cohort
• Children born before iodization of salt versus children born after iodization of salt
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication
details Published in a Chinese journal
Stated aim of
study Not stated
Notes Intelligence measured with Chinese Binet Scale
Comparisons made within village, to control for potential biases
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm (parts per million)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus 1 to 2 years
of life more than comparison
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre; did not use World
Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection
bias)
High risk Not RCT
Blinding of participants and
personnel (performance bias)
High risk No mention of blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No mention of blinding
Incomplete outcome data (attrition
bias)
Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Children in both groups were from the same village but
comparison was between 10–11 year olds and 4–5 year
olds

54
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Fu et al., 2001 (127)
Methods Multiple cross-sectional observational study conducted in Shangping township and
Fuchen township, China
The study involved sampling of children in each township before and after salt iodization and
intelligence was measured. Iodization began in 1984 after 9-year-old children were measured.
In 1999, 9-year-old children were tested, all of whom were born after salt iodization. The
sampling method was not reported.
Participants Children aged 9 years at the time that intelligence was measured
Interventions There were two cohorts (both populations sampled before and after salt iodization):
• Shangping (iodine-decient village);
• Fuchen (iodine-sucient village).
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Good-enough Test (draw-a-person test) at baseline and Chinese
Binet Scale at follow-up
Comparisons made within village, to control for potential biases
1. Design: multiple cross-sectional
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm (parts per million)
6. Duration: 9 years
7. Baseline iodine deciency disorder (IDD) status: as noted by authors but indicator used not
specied:
• Shangping: severe IDD
• Fuchen: adequate iodine status
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No mention of blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No mention of blinding
Incomplete outcome data (attrition bias) High risk As multiple cross-sectional, attrition not reported;
however, sampling methods were also not reported
and the n values for after salt iodization were much
lower than those before iodization
Selective reporting (reporting bias) Unclear risk % of participants with intelligence quotient (IQ) <70
not reported as expected
Other bias High risk Dierent measures of IQ used at baseline and follow-up

55
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Gatti et al., 1980 (128)
Methods Multiple cross-sectional observational study conducted in Argentina
Observational study of changes in urinary iodine excretion (UIE) from before to after the
provision of iodized salt in a region of Argentina. Data were collected from men and women
in the towns of Rosario and Bahia Blanca, Argentina. Only the sample in Rosario had complete
data to use in this review. The same sample of men and women had UIE measured at baseline
(1966/1967). Salt in the region (Santa Fe province) was iodized by law in the year 1970. A
follow-up measurement was made in 1977, 7 years after salt iodization in the region. Change
in UIE was measured.
Participants Adults (men and women) living in Rosario, Argentina, who met the inclusion criteria
Interventions Iodized salt was provided at the provincial level through mandatory fortication of salt at a
concentration of 33 ppm.
Outcomes UIE (µg iodine/g creatine)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To determine the contribution of iodine from iodized salt to the overall iodine intake in
persons living in an area of little naturally occurring iodine in the water
Notes 1. Design: multiple cross-sectional
2. Age: adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 33 ppm urinary iodine excretion (UIE)
6. Duration: 7 years
7. Baseline iodine deciency disorder (IDD) status: not reported
The study had four time points and it is not clear whether at each it was a cross-sectional
survey or whether it was repeated measures on the same individuals. The selection process
for the included participants was not explained.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants
and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Blinding not reported
Incomplete outcome data (attrition bias) High risk The study population was very small. It is not clear
that the same people were measured at each follow-
up. If it were the same people, there was greater
than 70% loss to follow-up at time points 2 and 3
relative to baseline and a 35% loss to follow-up at
time point 4 relative to baseline.
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Methodology for sampling not reported in detail

56
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Golkowski et al., 2007 (96)
Methods Cohort observational study conducted in the southern part of Poland
The authors compared thyroid dysfunction in participants prior to (1989–1990) and after
implementation of iodine prophylaxis at a national level (1997–1999).
A nutritional survey was conducted in randomly selected adults living in southern Poland in
1989–1990. A total of 4176 adults were surveyed. All those participants were contacted again in
1997 to take part in a similar survey in 1997–1999; 1648 participants answered the invitation to
participate in 1997. Of those, 1424 were selected, based on exclusion criteria (i.e. had not taken
thyroid medicine in their lifetime). Analysis was only of the 1424 who had complete data at baseline
and follow-up.
Participants Adults (age 16 years and older) who had not taken thyroid medicine in their lifetime, living in
the southern part of Poland
Interventions Iodized salt was provided at a concentration of 30 ppm (potassium iodide), through
mandatory fortication of household salt.
Outcomes 1.Thyroid-stimulating hormone (TSH) level (median reported in manuscript: results in
summary table but not meta-analysed)
2. Thyroid volume (only measured at follow-up)
3. Urinary iodine excretion (UIE) (only measured at follow-up)
4. Hyperthyroidism
5. Hypothyriodism
6. Positive anti-thyroid microsomal antibodies
Publication details Published in a peer-reviewed journal in English
Stated aim of study To assess the prevalence of hyperthyroidism just after implementation of iodine prophylaxis
among adults from an area with iodine deciency
Notes 1. Design: cohort observational
2. Age: adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 30 ppm
6. Duration: 2 years
7. Baseline iodine deciency disorder (IDD) status: not reported
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported/unknown
Incomplete outcome data (attrition bias) High risk Only 34% of persons measured at baseline were also
measured at follow-up (analysis included only those
with measures at both times)
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Subjects from same area of Poland and selected
randomly for invitation into study

57
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Gongora, 1952 (55)
Methods Non-randomized controlled trial (RCT) conducted in Colombia
Salt iodization was introduced into one community (Caldas, Colombia) and not into another
community (Bogotá, Colombia). The investigators selected participants randomly from
schools, to measure goitre before iodized salt was introduced. After 2 years of use of iodized
salt in the intervention community, another cross-sectional study, where the sample was
said to have been selected in the same fashion as the previous study, was conducted in both
communities. The community of Bogotá acted as a control because iodized salt was not used
in that community.
The methods of selection of participants were not described.
Methods of measurement of goitre were not described but it was noted that the same
methods were used in both surveys.
Participants School-aged children in the two communities
Interventions Iodized salt was introduced to the community at a concentration of 50 ppm (parts per
million).
Outcomes Goitre
Publication details Published in a peer-reviewed journal in Spanish
The complete original report could not be accessed. These results were abstracted from the
review article, de León Méndez R. [Ecacia del enriquecimiento de la sal con preparados de
yodo, como medio de prevencion del bocio endémico]/Eectiveness of the enrichment of
salt with iodine preparations as a means of preventing endemic goiter. Boletín de la Ocina
Sanitaria Panamericana 1966;61(1):1–26 (http://hist.library.paho.org/Spanish/BOL/v61n1p1.pdf,
accessed 27 October 2014).
Stated aim of study To evaluate the salt iodization programme in Colombia
Notes 1. Design: non-RCT
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm (parts per million)
6. Duration: 2 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors'
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not RCT
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
Unclear risk No blinding reported
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) High risk 20% in treatment/29% control – no explanation for
loss to follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias High risk Though both groups were considered severe IDD at
baseline, owing to high goitre prevalence, the control
group had almost half the prevalence of goitre of that
of the experimental group at baseline and the control
group was from the capital while the experiment was
from less of an urban centre.

58
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Guo, 1984 (97)
Methods Cohort observational study conducted in Houchang community, Ziyun county, Guizhou
province, China
The study report provided data of before–after supplementation of iodized salt.
Participants All inhabitants of Houchang community
Rates of examination for before and after, respectively: 96.19% (4039/4199) and 96.76%
(4269/4412)
Interventions Iodized salt was provided through universal salt iodization at 20 ppm (parts per million).
Outcomes 1. Goitre prevalence
2. Physiologically enlarged thyroid gland (not included in this review)
3. Tri-iodothyronine (T3; not reported/used to determine prevalence of hypothyroidism)
4. Thyroxine (T4; added to summary table)
5. Thyroid-stimulating hormone (TSH; median value reported in manuscript: results in
summary table but not meta-analysed)
6. Urinary iodine concentration (UIC; median value reported in manuscript: results in
summary table but not meta-analysed)
7. Hypothyroidism
8. Rate of absorption of I-131 by thyroid gland (only done in persons described as having
iodine deciency and therefore not included in this review)
Publication details Published in a Chinese language journal
Stated aim of study Not stated
Notes 1. Design: cohort observational
2. Age: children and adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20 ppm
6. Duration: 3 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported/unknown
Incomplete outcome data (attrition bias) Low risk High response rates
Selective reporting (reporting bias) Low risk Most outcomes reported fully
Other bias Low risk Entire community measured at each time point
and number of deaths/births and emigration/
immigration between surveys accounted for

59
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Han et al., 2006 (129)
Methods Multiple cross-sectional observational study conducted in Tianjin, China
The participants were sampled by probability proportionate to size sampling (PPS), according
to the population. Samples were taken from 30 schools. One hundred and fty housewives
were also sampled, though the selection process was not described.
Participants Included pupils aged 8–10 years sampled from 30 schools, 600 class ve pupils and 994 pupils
aged 8–10 years whose drinking water iodine concentration was (median) >150 (300) µg/L
150 housewives sampled, though selection process not described
Interventions A survey was conducted in 1995, then in 1996 universal salt iodization was implemented.
A follow-up study was conducted in 2005. The concentration of iodine in the salt was not
reported.
There was only one cohort, as results were reported for women and children combined.
Outcomes 1. Urinary iodine excretion (UIE; median reported and results included in a summary table but
not meta-analysed)
2. Goitre prevalence
Publication details Published in a Chinese language journal
Stated aim of study To evaluate the eciency over 10 years of iodized salt to prevent and control iodine
deciency disorder (IDD)
Notes 1. Design: multiple cross-sectional
2. Age: children and adults
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 10 years
7. Baseline IDD status: severe based on goitre/mild based on UIE
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported/unknown
Incomplete outcome data (attrition bias) High risk Multiple cross-sectional study but sample sizes
much smaller at follow-up than at baseline
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Probability proportionate to size sampling (PPS)
described and reported that it was same at all time
points but no reason given for very dierent sample
sizes

60
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
He et al., 1993 (67, 68)
Methods Quasi-experimental study conducted in Qiandongnan State of Guizhou Province, China
Children of villages who were born before iodization of salt were compared to children in the
villages born after iodization of salt.
Participants Children aged 7–14 years at time that intelligence was measured (1990)
Interventions The experiment group was made up of children whose mothers received iodine from iodized
salt during pregnancy and children were exposed to iodized salt in infancy; the comparison
group comprised children whose mothers did not receive iodine from salt during pregnancy
and children were not exposed to iodized salt until 1–3 years of age. The entire study area was
considered an iodine-decient area at baseline.
There were three cohorts:
• Miao ethnic group;
• Dong ethnic group;
• Han ethnic group.
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Comparisons made within ethnic groups, to control for potential biases
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20 ppm (parts per million)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus 1–3
years of life more than comparison
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on goitre; did not use
World Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment
(selection bias)
High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No mention of blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No mention of blinding
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Cluster sampling method used to form
representative groups – each of three racial groups
compared to same group, to reduce bias

61
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Heydarian et al., 2007 (130)
Methods Multiple cross-sectional observational study conducted in Tehran province,
Islamic Republic of Iran
Cluster random sampling was used to select adults for each survey. The rst survey was
conducted in 1983–1984 and the second survey was conducted in 1999–2000. National
salt iodization became law in 1994. Goitre was assessed by palpation and graded using
World Health Organization (WHO) classication. Serum samples were collected and thyroid
stimulating hormone (TSH) and anti-thyroid peroxidase (TPOAb) and anti-thyroglobulin
(TgAb) antibodies were measured. Free T4 was also measured in those with abnormal TSH
levels. Casual morning urine samples were taken from a subsample of participants, to
measure urinary iodine excretion (UIE).
Participants Adults aged 20 years or older living in Tehran province, Islamic Republic of Iran
Interventions Iodized salt was provided at 40 ppm (parts per million; potassium iodide).
Outcomes 1. Goitre
2. TSH
3. Positive anti-thyroid peroxidase antibodies (TPOAb+)
4. Positive anti-thyroglobulin antibodies (TgAb+)
5. Hypothyroidism
6. Hyperthyroidism
Publication details Published in peer-reviewed journal in English
Stated aim of study To compare goitre rate, serum TSH, thyroperoxidase antibodies, thyroglobulin antibodies,
thyroid function, and UIE in the adult population of Tehran before and after national salt
iodization
Notes 1. Design: multiple cross-sectional
2. Age: adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 40 ppm
6. Duration: 5 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre; did not use
WHO criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
High risk Unclear but not likely
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with larger sample size at
follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias High risk Sampling method described and not the same for
the two surveys

62
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Hintze et al., 1988 (53)
Methods Randomized controlled trial (RCT) conducted in Göttingen, Germany and a small rural area
in the vicinity
All schoolchildren were invited to participate. Of the 1160 parents contacted, nearly 50% did
not respond, 227 did not wish to participate and the parents of a total of 334 boys and girls
agreed to the study and allowed their children to participate. Measurements were taken in
the schools of body weight and height, goitre classied by World Health Organization (WHO)
standards, and neck circumference. After baseline measurements, students were randomly
assigned to group A or group B. The randomization method was not explained. Group A was
assigned to buy from the market and consume only iodized salt, while group B was assigned
to buy from the market and consume only non-iodized salt. Groups were not blinded to
treatment but assessors were blinded. Follow-up measurements of goitre were undertaken at
2 and 4 years and urinary iodine excretion (UIE) was measured yearly. The nal 4-year follow-
up measure is used in this review.
Participants Schoolchildren averaging 10 years of age at baseline and with a goitre prevalence of 30.5%
Interventions 1. Iodized salt (iodate, 20 ppm iodine) (n = 146)
2. Control (plain salt) (n = 188)
Outcomes 1. Goitre
2. Neck circumference
3. UIE (µg iodine/g creatinine; results in summary table because cannot be combined with the
other RCT which reports UIE as µg iodine/24 h in pregnant women)
Publication details Published in a peer-reviewed journal in English
Stated aim of study Quote “The availability of iodinated salt containing 20 mg of iodine as iodate/kg salt
consumed on a voluntary basis enabled us to investigate its eect on goitre prevalence and
iodine excretion in urine in a longitudinal, prospective, randomised study over 4 years.”
Notes 1. Design: RCT
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: 5 g salt/day (estimated)
5. Iodine concentration: 20 ppm (parts per million)
6. Duration: 4 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Iodine intake estimated at 100 µg/day
If at any time over the course of the study a child presented with grade II goitre, a thyroid
nodule, or a UIE of 200 µg/g creatinine, he/she was excluded from the study. There was a total
of ve participants in each group lost to -follow-up for these reasons.

63
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
Unclear risk Randomization methods not explained
Allocation concealment (selection bias) Unclear risk Allocation methods not explained
Blinding of participants and personnel
(performance bias)
High risk Participants and providers not blinded to treatment
Blinding of outcome assessment
(detection bias)
Low risk Assessors were blinded
Incomplete outcome data (attrition bias) Low risk Details on attrition were unclear; however, loss to
follow-up appeared to be similar between groups
and ve participants in each group were lost to
follow-up because of presenting with grade II goitre,
a thyroid nodule or high UIE
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk No other factors to potentially increase bias
identied

64
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Hou et al., 2003 (131)
Methods Multiple cross-sectional observational study conducted in three districts of Jilin
province, China
There was only a baseline measure before salt iodization in one village (Helongcheng) and
that information is used in this review.
Participants Probability proportionate to size (PPS) sampling used to get a representative sample of
inhabitants in Helongcheng village of Long county
The rst survey was conducted in 1986, iodized salt was introduced in 1993 and universal salt
iodization achieved in 1996. The follow-up survey was conducted in 1999.
Interventions Monitoring of salt iodization showed concentration levels varied from 21 ppm (parts per
million) to 47 ppm, with most measures above 40 ppm.
Outcomes 1. Goitre (only provides prevalence with no denominator value and therefore data cannot be
used in meta-analysis)
2. urinary iodine excretion (UIE; measures done dierently at baseline and follow-up and
therefore cannot be used in meta-analysis)
3. Incidence of hyperthyroidism (measured only in adults)
Publication details Published in a Chinese language journal
Stated aim of study To compare hyperthyroidism before and after universal salt iodization
Notes 1. Design: multiple cross-sectional
2. Age: all (adults for hyperthyroidism)
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20–50 ppm
6. Duration: 3 years (of universal salt iodization and 6 years of some exposure to iodized salt)
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on goitre; did not use
World Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Two cross-sectional studies but number of
individuals measured was not reported
Selective reporting (reporting bias) Unclear risk Prevalence reported but number of individuals not
reported
Other bias Low risk PPS sampling using same methodology said to have
been used in both surveys

65
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Hou and Wang, 2009 (132)
Methods Multiple cross-sectional observational study conducted in Lantian county, Shanxi
province, China
A survey to measure goitre was conducted in 1970, whereby all residents of Lantian county
were participants. The same county was surveyed again in 2006 for goitre.
Participants All residents of Liantian county
Interventions Iodized salt was provided at an unreported concentration.
Outcomes Goitre
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To assess the eect of iodized salt on the prevention of goitre
Notes 1. Design: multiple cross-sectional
2. Age: children and adults
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 36 years
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on goitre; did not use
World Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Unclear but not likely
Incomplete outcome data (attrition bias) Low risk Two multiple cross-sectional surveys with all
inhabitants surveyed
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk All inhabitants surveyed using same census
methodology

66
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Hu et al., 1998 (133)
Methods Multiple cross-sectional observational study conducted in Taiping village, Hanyin
county, Shanxi province
All villagers of Taiping village were surveyed for goitre and cretinism in the village before the
introduction of iodized salt. Thirty-two years after the introduction of iodized salt, all villagers
were again surveyed for goitre and cretinism.
Participants All villagers
Interventions Iodized salt was provided at less than 20 ppm (parts per million) for 25 years, then 33 ppm for
5 years and nally 50 ppm for 1year.
Outcomes 1. Goitre
2. Cretinism
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To assess the eect of supplementation of iodized salt for 32 years
Notes 1. Design: multiple cross-sectional
2. Age: adults and children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm
6. Duration: 32 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre; did not use
World Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Unclear but not likely
Incomplete outcome data (attrition bias) Low risk Two multiple cross-sectional surveys with entire
village surveyed and high participation rate
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk All inhabitants surveyed using same census
methodology

67
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Huang et al., 2004 (159)
Methods Multiple cross-sectional study and quasi-experimental study conducted in Fujian, China
Multiple cross-sectional: in 1994, investigators selected primary schoolchildren in grades
2–5, to test their intelligence quotient (IQ) level. Salt iodization began in 1994, then in 2002
another survey was conducted in children of the same age from the same schools.
Quasi-experimental: in 2002, children were divided into those born before 1994 and those
born after 1994. Those born before were the comparison group because they did not have
exposure to iodized salt during gestation or the rst 2years of life and those born after 1994
did have exposure to iodized salt during rst 2years of life.
Participants Multiple cross-sectional study: baseline 257 and follow-up 252 children, all grades 2–5
Quasi-experimental study: experiment 52 schoolchildren and comparison 200 schoolchildren
Interventions Iodized salt was provided to inhabitants through universal salt iodization.
Outcomes Mental intelligence:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To understand the eect of universal salt iodization on intelligence quotient level of children.
Notes 1. Design: multiple cross-sectional and quasi-experimental
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 8 years
7. Baseline iodine deciency disorder (IDD) status: severe (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment
(selection bias)
High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional study had similar numbers
at baseline and follow-up and in quasi-experiment
attrition was not applicable
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk The groups were comparable at baseline and
students from the same school measured in the
multiple cross-sectional study, to reduce bias

68
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Ibanez Gonzalez et al., 1956 (98)
Methods Cohort observational study conducted in Guejar Sierra, Granada, Spain
A random sample of the population living in the town of Guejar Sierra, Granada, Spain was
evaluated for goitre before and after the introduction of iodized salt in the community. The sample
was described as random but methods of sampling were not described. The methods for measuring
goitre were not described; 16 months after the introduction of the iodized salt in the community,
the same participants were evaluated again for goitre. The loss to follow-up was not reported.
Participants Adult and child participants randomly selected from the community; no exclusion criteria described
Interventions Iodized salt was provided at a concentration of 16.7 ppm (parts per million).
Outcomes Goitre (only reported prevalence of goitre and thus could not be used in the quantitative analysis)
Publication
details
Published in a peer-reviewed journal in Spanish
The complete original report could not be accessed. These results were abstracted from the review
article, de León Méndez R. [Ecacia del enriquecimiento de la sal con preparados de yodo, como medio
de prevencion del bocio endémico]/Eectiveness of the enrichment of salt with iodine preparations
as a means of preventing endemic goiter. Boletín de la Ocina Sanitaria Panamericana 1966;61(1):1–26
(http://hist.library.paho.org/Spanish/BOL/v61n1p1.pdf, accessed 27 October 2014).
Stated aim of
study Not stated
Notes 1. Design: cohort observational
2. Age: adults and children
3. Group: not specically pregnant women
4. Salt consumption: 10 g/day (estimation reported in study)
5. Iodine concentration: 16.7 ppm
6. Duration: 16 months
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre):
• intake of 167 µg iodine/day
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Unclear if there was blinding of the outcome
assessor
Incomplete outcome data (attrition bias) Unclear risk Sample sizes not reported so attrition unknown
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Random selection of individuals in community,
taking into account various age groups

69
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
I'Ons et al., 2000 (56)
Methods Non-randomized controlled trial (RCT) conducted in South Africa
Non-randomized controlled trial where two schools were selected for convenience and
because there was no iodized salt available in the region where the schools were located.
All children in both schools participated. Goitre, measured by palpation and dened using
World Health Organization criteria, and urinary iodine excretion (UIE), were measured at
baseline and after a 4-month follow-up period. Children in the experiment school were all
given iodized salt for use in the household and children in the control school were given non-
iodized salt for use in the household.
Participants All schoolchildren in two communities, aged 7–16 years
Interventions 1. Iodized salt
2. Plain salt
Outcomes 1. Goitre
2. UIE (reported as median; results in summary table)
3. % of population with UIE <100 µg/L
Publication details Published in a peer-reviewed journal in English
Stated aim of study To evaluate the short-term eect of salt iodized according to the recently revised South
African salt legislation on the iodine status of rural primary schoolchildren
Notes 1. Design: non-RCT
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 53 ppm (parts per million)a
6. Duration: 4 months
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on UIE)
a rst 2 batches of salt had 18 and 12 ppm iodine, and the last batch (last 4–5 weeks) had 53
ppm unclear if iodide or iodate; n = 39)
The authors conclusions state “The sample size and duration of intervention were inadequate
to estimate the eect of iodized salt on the prevalence of goitre accurately and to detect
changes in goitre rates, but is eective in eradicating mild and borderline iodine deciency
within a period of 4 months in primary schoolchildren. We compared the change from
baseline of two groups before–after salt iodization.”
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized
Allocation concealment (selection bias) Unclear risk Unclear
Blinding of participants and personnel
(performance bias)
Low risk Participants were blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Not reported
Incomplete outcome data (attrition bias) High risk Control group had more than double the drop-out
rate relative to the intervention group
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Schools were chosen because they were very
comparable and the students and families very
comparable

70
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Jia et al., 2004 (157)
Methods Cohort observational study, quasi-experimental and multiple cross-sectional
observational study conducted in Heshui county, Gansu province, China
A survey was conducted ofallresidents of Heshui county in 1990. In that year, universal
salt iodization was implemented. Another study ofallresidents was conducted in 2000.
The results of all persons aged over 10years in 2000 are presented as cohort study results
because they were measured at both time points. Individuals aged 0–10 years in 1990 are a
comparison group in the quasi-experiment, as they had no exposure to iodized salt and the
intervention group is those children aged 0–10 years at follow-up because they did have
exposure to iodized salt. The measure of UIE was only reported for 7–14 year olds and is a
multiple cross-sectional study.
Participants All residents of the county were measured for goitre (cohort) and children aged 7–14 years
were measured for urinary iodine concentration (multiple cross-sectional).
Interventions Iodized salt was provided through universal salt iodization at an unknown concentration.
Outcomes 1. Goitre (all)
2. Urinary iodine concentration (children 7–14 years)
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To assess the eect of iodized salt on goitre and urinary iodine concentration
Notes 1. Design: multiple cross-sectional, cohort observational and quasi-experimental
2. Age: all for goitre (cohort), children for goitre (quasi-experimental), children for urinary
iodine excretion (UIE; multiple cross-sectional)
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 10 years
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on UIE)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Low attrition reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk All inhabitants measured at both time points and in
quasi-experiment it was children from same villages

71
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Jooste et al., 2000 (134)
Methods Multiple cross-sectional observational study conducted in South Africa
Schools in four communities in South Africa were surveyed for goitre and urinary iodine
excretion (UIE). The schools were not representative of any larger population but were
selected to have participants of a variety of socioeconomic statuses. In large schools, every
other child in grades 4–7 was selected to participate and in smaller schools all children in
grades 4–7 were selected to participate. The baseline survey measured goitre and UIE during
the month before universal salt iodization was begun in South Africa. The follow-up survey
occurred 1year later, using the same sampling technique.
Participants Children in grades 4–7 in selected communities in South Africa
Interventions Iodized salt (potassium iodate) provided at a concentration of 40–60 µg/g through
mandatory fortication
Outcomes 1. Goitre
2. UIE (median included in summary table)
3. % population with UIE <100ug/L
4. % of population with UIE >200 µg/L (data not used in the current review as this is not a cut-
o for excessive UIE and not an indicator of status
Publication details Published in a peer-reviewed journal in English
Stated aim of study To evaluate, after 1 year, the eectiveness of mandatory iodization of salt at an iodine
concentration higher than that occurring under optional iodization, on the goitre rates and
iodine status of schoolchildren living in a goitre-endemic area in South Africa
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 40–60 ppm (parts per million)
6. Duration: 1 year
7. Baseline iodine deciency disorder (IDD) status: moderate IDD (based on goitre)/moderate
IDD (based on UIE)
Limited data from quantitative analysis of iodine concentration in salt suggest that the
fortication target of 40–60 ppm was not being reached 1year after universal salt iodization
in South Africa. The mean of the 18 samples of salt was only 25 ppm.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple-cross-sectional with similar number of
participants
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Participants of similar age and chosen using same
methodology

72
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Kimball, 1931 (57)
Methods Non-randomized controlled trial (RCT) conducted in the United States of America (USA)
Two cross-sectional surveys were conducted in a community that started a salt iodization
programme (Detroit, Michigan, USA) and in a community without salt iodization, which acted
as a control (Cleveland, Ohio, USA).The rst survey was before iodized salt was introduced
into the Detroit community and the second survey was 7 years after the introduction of
iodized salt in the Detroit community. No iodized salt was introduced into the Cleveland
community during this time. The samples were reported to be selected in a similar fashion
but details were not reported. The samples were from schoolchildren; however, selection and
sample sizes were not reported. Similarity of samples at baseline is unknown; however, the
prevalence of goitre at baseline is similar in the two communities and they are in the same
region of the country.
Participants School-aged children attending schools in either Detroit Michigan or Cleveland Ohio, USA,
715 years of age
Interventions Iodized salt was introduced in the community.
Outcomes 1. Goitre (only goitre prevalence is reported and thus the data could not be used in the
quantitative analysis in this report; only added to summary table)
Publication details Published in English in a peer-reviewed journal
The complete original report could not be accessed. These results were abstracted from the
review article de León Méndez R. [Ecacia del enriquecimiento de la sal con preparados de
yodo, como medio de prevención del bocio endémico]/Eectiveness of the enrichment of
salt with iodine preparations as a means of preventing endemic goiter. Boletín de la Ocina
Sanitaria Panamericana 1966;61(1):1–26 (http://hist.library.paho.org/Spanish/BOL/v61n1p1.pdf,
accessed 27 October 2014).
Stated aim of study To evaluate the salt iodization programme in Detroit Michigan under natural conditions and
compare that to a control area without salt iodization.
Notes 1. Design: non-RCT
2. Age: child
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 7 years
7. Baseline iodine deciency disorder (IDD) status: severe (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk No randomization
Allocation concealment (selection bias) High risk No allocation
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Only prevalence reported, so number of participants
and attrition unknown
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Not clear if groups were comparable at baseline,
though goitre prevalence was similar

73
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Kimball, 1946 (135)
Methods Multiple cross-sectional study conducted in Michigan, United States of America (USA)
All schoolchildren in three counties in Michigan, USA were measured for goitre in 1924 before
the introduction of iodized salt to that state. In 1924, iodized salt was introduced to the state
of Michigan at a concentration of 20 ppm (parts per million). In 1928 and in 1936, all the
schoolchildren in the same schools in the same four counties were again measured for goitre.
Participants All schoolchildren in the four counties selected for measurement in the state of Michigan,
USA
Interventions Iodized salt was provided at state level at a concentration of 20 ppm.
Outcomes 1. Goitre
2. Adverse eects (qualitatively reported that NO adverse eects were identied or reported)
Publication details Published in English in a peer-reviewed journal
Stated aim of study To learn about the eciency of iodized salt and to determine any harmful eects from its
continued use
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20 ppm
6. Duration: 4 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Unclear but not likely
Incomplete outcome data (attrition bias) Low risk Two cross-sectional studies with similar sample size
Selective reporting (reporting bias) Low risk In terms of adverse eects, the authors note that
none were reported
Other bias Low risk All children in the same schools measured, to reduce
potential bias

74
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Kimiagar et al., 1990 (99)
Methods Cohort observational study conducted in the Islamic Republic of Iran
Participants 2034 participants from 368 families were surveyed; of these, a total of 1323 persons were
clinically examined.
Interventions The study randomly selected 60 villages from total of 180 villages in the endemic area, and
selected 368 families for inclusion for a total of 1323 participants. A baseline survey was
conducted and then all households were provided iodized salt directly for 3 years. A follow-
up survey was undertaken at 2 years and at 3 years post introduction of iodized salt.
Outcomes 1. Goitre
2. Tri-iodothyronine (T3)
3. Thyroxine (T4)
4. Thyroid-stimulating hormone (TSH)
5. Urinary iodine excretion (UIE; this outcome was the only outcome measured at follow-up
and therefore the only data included in meta-analysis)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To collect up-to-date baseline data on all aspects of the problem of iodine deciency
disorder (IDD) in this area of the Islamic Republic of Iran and to arrive at a suitable prophylaxis
programme
Notes 1. Design: cohort observational
2. Age: adults and children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 40 ppm (parts per million; 20 µg/kg [1/50 000] for 2years’ duration;
40 µg/kg [1/25 000] for 1 year’s duration)
6. Duration: 3 years
7. Baseline IDD status: severe IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
Unclear risk No blinding reported
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) High risk Loss to follow-up almost 20%
Selective reporting (reporting bias) High risk Only UIE measured/reported at follow-up
Other bias Low risk Included individuals randomly selected from
villages that had been randomly selected

75
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Li et al., 1991 (69–72)
Methods Quasi-experimental study conducted in four villages of Heilongjiang Province in China
Children of the village that were born before iodization of salt were compared to children in
the village born after iodization of salt.
Participants Children aged 5–14 years at time that intelligence was measured (1985)
Interventions The experiment group was made up of children whose mothers received iodine from iodized salt
during pregnancy and children were exposed to iodized salt in infancy; the comparison group
comprised children whose mothers did not receive iodine from salt during pregnancy and children
were not exposed to iodized salt until 1–3 years of age. All villages were areas of severe iodine
deciency at baseline.
There were two cohorts:
• Jixian village (severe iodine-deciency area);
• Taihui village (mild iodine-deciency area).
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Though the study was conducted in ve villages, there is only sucient report of results to
include two village in the quasi-experimental study. (The standard deviation of the second
village is not included, so only the % of children with IQ <70 points can be included for that
village.)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Sampling methodology not reported
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus rst
years of life more than comparison (exact time of exposure dierence between groups not
reported)
7. Baseline iodine deciency disorder (IDD) status:
• Jixian: severe IDD (based on goitre)
• Taihui: mild IDD (as stated by authors but values and indicators not provided)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding reported
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Sampling methods not reported and, as quasi-
experimental, attrition not reported
Selective reporting (reporting bias) High risk Only the results of two villages are reported
Other bias Low risk Results of experiment and comparison within
village, to reduce potential bias

76
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Lv et al., 2009 (136)
Methods Multiple cross-sectional observational study conducted in Hebei province, China
Probability proportionate to size sampling (PPS) technique was used to select the participants for a
survey on urinary iodine excretion (UIE) and goitre.
1200 schoolchildren aged 8–10 years were selected from 30 schools (40 children in each school).
At the same time, salt samples were collected from those children’s households; 360 urinary
specimens were randomly selected from the 1200 children. The same sampling occurred at the
baseline survey, conducted before the provision of iodized salt and at the follow-up survey 10 years
after the provision of iodized salt through mandatory means in the province. The baseline survey
was conducted in 1995 then universal salt iodization was implemented. A follow-up survey was
conducted in 2005.
Participants Schoolchildren aged 8–10 years
Interventions Iodized salt was provided at a concentration of 35 ppm (parts per million), through
mandatory fortication.
Outcomes 1. UIE (median reported and results added to summary table)
2. % population UIE <100 µg/L
3. Adverse eects (% population UIE >300 µg/L)
4. Goitre
5. Iodine concentration in the salt (data not used in this review)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To assess the eect of the mandatory iodized salt programme in Hebei province, China
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 35 ppm (parts per million)
6. Duration: 10 years
7. Baseline iodine deciency disorder (IDD) status: mild IDD (goitre)/adequate iodine status (based
on UIE)
Proportion of inadequately iodized salt:
• 50.3% in 1995;
• 13.6% in 1997;
• 2% in 1999;
• 3% in 2001;
• 9.6% in 2005.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Unclear but not likely
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with similar sample sizes
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Probability proportionate to size (PPS) sampling
used for sampling methodology for all surveys

77
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Mostafavi et al., 2005 (137)
Methods Multiple cross-sectional observational study conducted in Shiraz, Islamic Republic of
Iran
Schools in Shiraz, Iran were randomly selected and subsequently students within schools
randomly selected for evaluation of goitre. A survey was conducted in 1989 prior to the
initiation of universal salt iodization in the Islamic Republic of Iran. Goitre was evaluated in
randomly selected students aged 6–18 years, using World Health Organization classication.
In 1995, 5years after the initiation of the universal salt iodization programme in the Islamic
Republic of Iran, a second survey using the same methods was conducted to measure the
impact of the salt iodization programme on goitre in schoolchildren.
Participants Children, 6–18 years of age,attending school in Shiraz, Islamic Republic of Iran
Interventions Iodized salt was provided at an unknown concentration through universal salt iodization in
the country.
Outcomes Goitre
Publication details Published in a peer-reviewed journal in English
Stated aim of study To evaluate the impact of the salt iodization programme on the prevalence of goitre in
schoolchildren in the city of Shiraz
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 5 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
High risk Not blinded
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with similar sample sizes at
both time points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Same sampling methodology used, to reduce
potential bias

78
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Nicod 1953 (138)
Methods Multiple cross-sectional observational studyconducted in Switzerland
The article provides an overview of the prophylaxis of goitre through the use of iodized salt
in various cantons of Switzerland. The study includes data for discharged military adults,
national prevalence of goitre, and level of fortication in each canton and when iodization
programmes were initiated. The data included in the review come from surveys of school-
aged children in two cantons where complete data from pre-iodization and post-iodization
of salt is available. No data are available on how students were selected to participate in the
evaluations. Numerous indicators of the thyroid size were reported, based on palpation of the
neck. The indicators were normal thyroid, large neck and advanced goitre.
Participants School-aged children in Vaud and Valais cantons, Switzerland
Interventions Iodized salt was provided at the canton level at 10 ppm (parts per million).
Outcomes 1. Goitre (only prevalence reported, with no sample size, so data included in summary table
only)
2. Other indicators of thyroid size (not dened but included in summary table as reported)
Publication details Published in French in a peer-reviewed journal
Stated aim of study To evaluate the success of the salt iodization programme in Switzerland
Notes 1. Design: multiple cross-sectional
2. Age: children (two separate cohorts by canton)
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 10 ppm
6. Duration: canton of Valais 6 years/canton of Vaud 30 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
• Most of the data provided in this manuscript are not specic enough to be extracted.
• Each canton decided when to introduce iodized salt and the level of fortication.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
High risk Not blinded
Incomplete outcome data (attrition bias) Unclear risk Sample sizes not provided
Selective reporting (reporting bias) High risk Only limited information on outcomes available for
extraction from article
Other bias Unclear risk The limited description of the manuscript leaves the
potential for additional bias unclear

79
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Pan et al., 1995 (73)
Methods Quasi-experimental study conducted in two townships of Jiling city, China
One township was classied as “mild IDD [iodine deciency disorder]” area and the other as
“severe IDD” area, based on historic prevalence of goitre.
Children aged 5–14.5 years of the townships were eligible and a sample was drawn for
intelligence testing. The sampling method was not described. Universal salt iodization
began; however, the date was not stated. Children were divided into two groups: children
born before the universal salt iodization (comparison) and children born after universal salt
iodization (experiment). The intelligence of children was measured using the Chinese Binet
Scale.
Participants Children aged 5–14.5 years at time that intelligence was measured (date not stated)
Interventions Universal salt iodization was provided at an unknown concentration.
Outcomes Intelligence of children:
• mean intelligence quotient (IQ; no standard deviation reported and no sample size by
group reported; data found in summary table and no meta-analyses)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 9 months’ gestation to 4 years (comparison group was born before universal salt
iodization and some were born up to 4 years before)
7. Baseline IDD status: mild IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment
(selection bias)
High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data
(attrition bias)
Unclear risk Quasi-experimental, so no attrition but sample sizes
in each group not reported
Selective reporting (reporting bias) Unclear risk % of population with low IQ (mild mental
retardation – MMR) not reported as expected
Other bias Unclear risk Children selected from same communities, to
reduce bias, but sampling method not mentioned
and sample sizes not provided

80
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Pedersen et al., 2002 (100–102)
Methods Cohort observational study conducted in Copenhagen and Aalberg, Denmark.
The entire population of two towns in Denmark with varying degrees of iodine suciency/
deciency at baseline were the study populations. Copenhagen was a town with mild iodine
deciency, while Aalberg was a town with moderate iodine deciency. The overall population
in the catchment areas of selected health centres in the towns were the populations studied
over time. The same denominator was used at baseline and at each follow-up. The results
were normalized to an incident rate of hypothyroidism or hyperthyroidism per 100 000
inhabitants per year. The numbers of incident cases of hypothyroidism or hyperthyroidism
were taken from the health laboratory registries and standardized to an incident rate of
cases/100 000/year (12 months).
The investigation took place from 1997 to 2005 and comprised four periods:
1. 1997–1998, baseline, was before iodization of salt was initiated and lasted
16 months in Aalborg and 14 months in Copenhagen;
2. 1999–2000, with iodization of salt on voluntary basis;
3. 2001–2002, with early mandatory iodization of salt;
4. 2003–2005, with late mandatory iodization of salt.
In the period with voluntary iodization of salt, the iodized salt was not used by the food
industry and it covered only around 50% of household salt. Accordingly, the average increase
in iodine intake was low (on average, 10 µg/day).
Participants Two open population cohorts representing all inhabitants of Aalborg and Copenhagen,
Denmark
Interventions Iodized salt was provided at a concentration of 8 ppm (parts per million) during a voluntary
phase and then at 13 ppm during mandatory salt iodization in Denmark.
Outcomes 1. Incident rate of hypothyroidism
2. Incident rate of hyperthyroidism
Publication details Three articles from this study published in peer-reviewed journals in English
Stated aim of study To compare the long-term eects of dierent grades of low iodine intake on the incidence of
hypothyroidism and hyperthyroidism in Denmark
Notes 1. Design: cohort observational
2. Age: all
3. Group: not specically pregnant women
4. Salt consumption: 5.5–6.0 g/day estimated
5. Iodine concentration: 13 ppm
6. Duration: 7 years (hypothyroidism), 5 years (hyperthyroidism)
7. Baseline iodine deciency disorder (IDD) status:
• Copenhagen: mild iodine deciency
• Aalberg: moderate iodine deciency
Intake of approximately 72 µg iodine/day
Data from adults and children collected but results presented as a total number for the entire
population

81
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Low attrition
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Methodology to collect data from all members of
population was consistent throughout study

82
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Pongpaew et al., 1998 (58)
Methods Non-randomized controlled trial (RCT) conducted in Thailand
Concurrent controlled study conducted in four schools in Thailand. Four schools were
randomly selected to participate in a baseline study of iodine status and the status of
other nutritional indicators. After the baseline study, each of the schools was assigned to
one of three interventions (iodinated salt, iodized water, iodized sh sauce) or control (no
intervention). In the iodized salt intervention arm, iodized salt was given to the students
on a regular basis, for use at all times. The participants were measured again after 1year of
intervention. The data from the iodized salt intervention school and the control school were
compared in this review.
Participants All schoolchildren; 68 intervention/63 control
Interventions The intervention results used in this analysis were of an intervention of iodized salt. The salt
was iodized at 50 g potassium iodate per 1000 kg salt (50 ppm [parts per million]).
Outcomes 1. Urinary iodine concentration
2. % of population with urinary iodine excretion (UIE) <100 µg/L
Publication details Published in a peer-reviewed journal in English
Stated aim of study To assess the best method of providing iodine to schoolchildren among the interventions of
salt, sh sauce and drinking water
Notes 1. Design: non-RCT
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: estimation reported at 5–10 g/day
5. Iodine concentration: 50 ppm
6. Duration: 1 year
7. Baseline iodine deciency disorder (IDD) status: mild IDD in intervention and adequate
iodine status in control (based on UIE)
• Intake of 250 µg iodine per day
• Ages and numbers of students included in the study not reported
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not RCT
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Reported there was no loss-to-follow up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias High risk Baseline characteristics of schools were
evaluated before it was decided if school would
be intervention or control and how each was
determined, not described

83
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Regalbuto et al., 2010 (139)
Methods Multiple cross-sectional observational study conducted in Sicily, Italy
This study compared the goitre prevalence in 1977, in 1994 during which the use of iodized
salt was less than 1% of the total salt consumed, and in 2007, 10 years after the promotion
of iodized salt, which began in 1996. The use of iodized salt reached 44% (28% to 55%).
Schoolchildren from dierent towns were surveyed for goitre, thyroid volume and urinary
iodine excretion (UIE). The selection of subjects was not described in detail.
Participants Schoolchildren areas dened as endemic and not endemic for iodine deciency, based on
baseline UIE
Two cohorts each from a dierent village
All results in manuscript reported by village:
• Bronte
• Maniace
Interventions 30 mg/kg iodized salt (potassium iodate) was used.
Outcomes 1. Goitre
2. Thyroid volume (mL; only assessed at the follow-up and therefore these data are not part of
the systematic review)
3. UIE (data in summary table because sample sizes not reported)
Publication details Published in peer-reviewed journal in English
Stated aim of study To assess the increase in iodized salt consumption and to re-evaluate goitre prevalence and
iodine intake in the population of the endemic area previously studied.
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 30 ppm (parts per million)
6. Duration: 10 years
7. Baseline iodine deciency disorder (IDD) status: based on goitre:
• Bronte: mild IDD
• Maniace: moderate IDD
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Not clear but not likely
Incomplete outcome data (attrition bias) High risk Multiple cross-sectional but much lower sample size
at follow-up
Selective reporting (reporting bias) High risk Sample sizes for UIE were not reported because of
low numbers of participants for this measurement
Other bias Unclear risk Methodology for sample selection not described

84
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Romano et al., 1991 (54)
Methods Randomized controlled trial(RCT) conducted in Italy
Pregnant women were randomly selected from a health-care facility. Women were rst
stratied by smoking status and then randomly assigned to either group A (iodized salt) or
group B (non-iodized salt). Each trimester, measurements were taken at the clinic.
Units of comparison: individuals
Participants Pregnant women with no personal history of thyroid diseases, not having been pregnant
before in the previous year, and not taking medications that could have aected thyroid
hormone metabolism
Interventions 1. Iodized salt provided at 20 ppm (parts per million; n = 17)
2. Control: plain salt, but not specically distributed (n = 18)
Outcomes 1. Thyroid volume (data not used in this review because not reported at follow-up for control
group)
2. UIE (µg iodine/24 h; added to summary table because cannot be combined with other RCT
because of units of reporting and measurement in children)
3. Serum thyroid-stimulating hormone (TSH) levels (data not used in this review because not
reported for follow-up)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To investigate the thyroid size during pregnancy by use of an accurate ultrasonographic
technique and to show the role of iodoprophylaxis in the prevention of an increase in thyroid
volume in a group of pregnant women who live in an area with moderate iodine deciency
Notes 1. Design: RCT
2. Age: adult
3. Group: pregnant women
4. Salt consumption: 7–10.6 g/day estimated by authors
5. Iodine concentration: 20 ppm (parts per million)
6. Duration: <1 year (9 months of pregnancy)
7. Baseline iodine deciency disorder (IDD) status: moderate IDD (based on author's
description but no data provided)
• Iodine intake 120–180 µg/day
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
Unclear risk Detail on randomization not published
Allocation concealment (selection bias) Unclear risk Detail on allocation concealment not published
Blinding of participants and personnel
(performance bias)
High risk Subjects not blinded – control group not given salt
while experiment group was provided with iodized
salt
Blinding of outcome assessment
(detection bias)
Low risk Assessor blinded
Incomplete outcome data (attrition bias) Low risk 0% attrition
Selective reporting (reporting bias) High risk TSH and thyroid volume not reported for follow-up
for both groups
Other bias Low risk Women had to meet many inclusion criteria, so the
groups were comparable at baseline

85
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Rueda Williamson and Pardo Tellez, 1966 (140)
Methods Multiple cross-sectional observational study conducted in Colombia
A representative sample of school-aged children from one department in Colombia was
selected for measurement of goitre status before iodization of salt in Colombia. The rst
survey was conducted in 1945. Salt was iodized at 50 ppm (parts per million), starting as pilot
in 1950, then a law was implemented in 1955 for the country. Large-scale iodization began in
1959 and in 1964 salt in the country varied from 60 to 75 ppm iodine, based on monitoring
activities. Another survey sample was measured for goitre 10 years after the implementation
of a law requiring iodization of salt. Goitre was dened using World Health Organization
standards.
Participants A representative sample of school-aged children (8700 at baseline and 2166 at follow-up)
Interventions Iodized salt was provided at a concentration of 60–75 ppm, through a mandatory
programme.
Outcomes Goitre
Publication details Published in a peer-reviewed journal in Spanish
Stated aim of study To measure the real eectiveness of salt iodization at the national scale for the prevention of
endemic goitre
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 60–75 ppm
6. Duration: 10 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
The coverage of salt iodization at 60 ppm or higher in the country was found to be 76% in
1964, which was 1year before the post-iodization survey was conducted. The coverage was
found to be 85% in 1966, 1year after the post survey was conducted. The coverage in 1965,
the year of the survey, could be estimated to be between these values.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Not clear but not likely
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with similar sample sizes at
both time points
Selective reporting (reporting bias) Low risk Only outcome measured was goitre and it was fully
reported
Other bias Low risk Both samples representative of the department
of Caldas, Colombia and selected using similar
methodology

86
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Salvaneschi et al., 1991 (141)
Methods Multiple cross-sectional observational study conducted in Argentina.
Before/after study design with cross-sectional representative surveys of goitre in city of
Buenos Aires, Argentina. Goitre was measured using World Health Organization methods and
denition. The rst survey occurred before the mandatory fortication of salt with iodine and
the second survey occurred 18 years after the implementation of mandatory fortication of
salt with iodine in the city.
Participants School-aged children (3882 at baseline and 3365 at follow-up)
Interventions Iodized salt was provided at a concentration of 33 ppm (parts per million).
Outcomes Goitre
Publication details Published in a peer-reviewed journal in Spanish
Stated aim of study To describe the prevalence of goitre in 1986 after more than 15 years of salt iodization in the
city of Buenos Aires and compare that value to those found in 1968 before the iodization of
salt in the city
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 33 ppm
6. Duration: 18 years
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
High risk Not blinded
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with similar samples sizes at
both time points
Selective reporting (reporting bias) High risk Goitre prevalence reported on fewer participants
than measured but reason not given
Other bias Low risk Both samples representative of the city of Buenos
Aires and selected using similar methodology –
probability proportionate to size (PPS) sampling

87
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Saowakhontha et al., 1994 (59)
Methods Non-randomized controlled trial (RCT) conducted in four villages of Thailand
The four villages sampled in the study were randomly selected from 12 villages. The
participants were randomly selected from the inhabitants of the villages. One of three
interventions (sh sauce, iodized salt, iodized water) or control were applied to each village.
All participants in each village received the intervention assigned to his/her village. The
iodized salt group received the salt at the household on a regular basis. The control group
received no intervention. The data used in this analysis were from the iodized salt and the
control groups.
Participants Women randomly selected in each of the villages from the age of 1 to 45years
Interventions • Group A: sh sauce
• Group B: iodized salt provided at 50 ppm (parts per million) to the households of the
randomly selected participants
• Group C: iodized water
• Group D: no intervention (conventional iodine intake, which might have include iodized
salt, sh sauce, iodine water, etc.)
Outcomes Urinary iodine excretion (UIE; the values were reported as µmol iodine/g creatinine; however,
the numeric values were 10–1000 times greater than one would expect and therefore were
not considered reliable and not included in the summary table or meta-analysis)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To evaluate the best method of providing iodine to this population
Notes 1. Design: non-randomized controlled study
2. Age: adults
3. Group: not specically pregnant women
4. Salt consumption: 5–10 g/day (estimated)
5. Iodine concentration: 50 ppm
6. Duration: 6 months
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
• Iodine intake estimated at 250–500 µg/day
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not RCT
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Low attrition
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias High risk Though intervention and control were severe IDD
areas, goitre prevalence at baseline in intervention
group was 68.9% and in control was 36%

88
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Scrimshaw et al., 1966 (142)
Methods Multiple cross-sectional observational study conducted in Guatemala
Before and after design to monitor the eect of fortication of salt with iodine on goitre rates
in schoolchildren in Guatemala. Two cross-sectional studies were conducted, one before
and one after the implementation of mandatory salt iodization in the country. The surveys
were representative of 19 of the 22 departments of the country. Coverage of salt iodization
at sucient concentration in the country was estimated at 80% at the year of the follow-up
survey.
Participants Schoolchildren aged 6–18 years, representative of 19 of 22 departments in the country
(28 787 at baseline and 19 442 at follow-up)
Interventions Iodized salt was provided at a concentration of 100 ppm (parts per million) via a national
programme.
Outcomes Goitre
Publication details Published in a peer-reviewed journal in Spanish
Stated aim of study To evaluate the ecacy of the salt iodization programme
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 100 ppm
6. Duration: 1.5 years
7. Baseline iodine deciency disorder (IDD) status: severe (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk Unclear but not likely
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional design with similar sample
sizes at both time points
Selective reporting (reporting bias) Low risk Only outcome measured was goitre and it was fully
reported
Other bias Low risk Both samples representative and selected using
similar methodology

89
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Shen, 1991 (74)
Methods Quasi-experimental study conducted in villages of Xinzhou county, Hubei province, China
Children of three villages in the county were eligible. All three villages began receiving
universal salt iodization in 1977. Children were randomly sampled in 1987 (10 years after
universal salt iodization) and divided into two groups: children born before (comparison) and
children born after universal salt iodization (experiment). The intelligence of children was
measured using the Chinese Binet Scale.
Participants Children aged 6–14 years at the time that intelligence was measured (1987)
Interventions Universal salt iodization was carried out at a concentration of 20 ppm (parts per million).
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <0 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20 ppm
6. Duration: 9 months’ gestation to 4 years (comparison group was born before universal salt
iodization and some were born up to 4 years before)
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on goitre; did not use
World Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding reported but not likely
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Comparison from same villages, to reduce bias
but sampling methods not reported and n value
in experiment group was lower (n = 48) than
comparison (n = 60)

90
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Shu, 1987 (75)
Methods Quasi-experimental study conducted in a suburb of Shenyang, Liaoning province, China
Children of three villages in the area were eligible. All three villages received universal
salt iodization for 9 years before the measurements were taken. The children born before
(comparison) were compared to children born after universal salt iodization (experiment)
Participants Children aged 5–15 years at time that intelligence was measured (date not reported)
Interventions Universal salt iodization was provided in unknown concentration.
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 9 months’ gestation to 6 years (comparison group was born before universal salt
iodization and some were born up to 6 years before)
7. Baseline iodine deciency disorder (IDD) status: severe IDD (as reported by authors but no
data provided)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding reported but not likely
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Comparison from same area for children born before
universal salt iodization but sample methodology
not reported and sample sizes dierent and
therefore comparability unknown; also, sampling
methods not reported and n value in intervention
group was much lower (n = 61) than in the
comparison group (n = 170)

91
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Sooch and Ramalingaswami, 1965 (60–62)
Methods Non-randomized controlled trial (RCT) conducted in Punjab, India
This study was set up as a prospective controlled study of the eect of iodization of salt on
goitre rates. The study was undertaken in the Himalayan endemic goitre belt of northern
India. Over 20 000 villagers were surveyed and the analysis of results was divided between
children aged 5–16 years attending school, children not attending school and the general
population of adults and children combined. The area of the study was divided into three
zones. One zone received salt with potassium iodide, one received salt with potassium
iodate, and the third was a control zone that did not receive fortied salt. Randomization and
blinding were not mentioned in the study reports.
All schoolchildren in each of the zones were surveyed at the beginning, in the middle and
at the end of the study period. In each zone, two thirds of the villages were sampled and all
of the villagers present on the day of survey were included in the surveys. The rst, baseline,
survey was in 1956, another survey was conducted 5years later (1962), and a nal survey was
conducted in 1968. From 1956 to 1962, two zones received the intervention and one zone
was the control. In 1956, the control zone was also provided salt fortied with salt and all
three villages were followed over time until the nal follow-up.
Additional data were reported in reference (62), which was ofmultiple cross-sectional design
describing the results after 12 years of salt iodization.
Participants All villagers in two thirds of the villages in the three zones and all school-aged children
Two cohorts reported:
• schoolchildren
• all other children and adults in the villages (results not presented separated by age)
Interventions 1. Iodized salt was provided to one zone in the concentration of 20 ppm (parts per million)
potassium iodide (i.e. 15.27 ppm iodine) and as 25 ppm potassium iodate (i.e. 14.83 ppm
iodine) to another zone (the results from the two intervention groups were combined and
used as intervention and compared to control).
2. Plain salt was provided.
Outcomes 1. Goitre
2. I-131 absorption rate of thyroid and subsequent excretion (there were no data for this that
were measured at baseline and follow-up and these data could not be used in the meta-
analysis)
3. Urinary iodine excretion (UIE) measured at follow-up of multiple cross-sections, but no
baseline measure for comparison
Publication details Published in peer-reviewed journals in English
Stated aim of study To demonstrate the eectiveness of iodine in salt in the prevention of goitre and to test the
relative eectiveness of potassium iodide and potassium iodate
Notes 1. Design: non-RCT (and multiple cross-sectional in a subsequent analysis)
2. Age: all and children (non-RCT and children in multiple cross-sectional)
3. Group: not specically pregnant women
4. Salt consumption: 15 g (estimated)
5. Iodine concentration: 20 ppm potassium iodide, and 25 ppm potassium iodate
6. Duration: 5 years (12 years in a subsequent multiple cross-sectional analysis)
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)

92
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk No randomization mentioned in the published
manuscripts
Allocation concealment (selection bias) High risk No concealment mentioned in the published
manuscripts
Blinding of participants and personnel
(performance bias)
High risk No blinding mentioned in the published manuscripts
but unlikely
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding mentioned in the published manuscripts
Incomplete outcome data (attrition bias) Low risk All people in villages were measured and attrition
not reported but actual numbers were higher at
follow-up than at baseline, suggesting low attrition,
and measurement of individuals not captured at
baseline
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Population in experiment and control similar in
habits, customs, dietary practices, socioeconomic
status, ethnic origin and severity of goitre at baseline

93
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Szybinski et al., 2001 (143)
Methods Multiple cross-sectional observational study conducted in Poland
Children attending selected schools were surveyed before and 5 years after the initiation of a
national programme for the iodization of salt.
Participants 952 schoolchildren in 1994, and 884 in 1999
Interventions Iodized salt was provided with potassium iodide concentration of 25 ± 10 mg/kg from 1986,
and 30 ± 10 mg/kg from 1997; there was 5years’ follow-up.
Outcomes 1. Prevalence of goitre
2. Urinary iodine excretion (UIE; µg/L)
3. % population with UIE <100 µg/L
Publication details Published in a peer-reviewed journal in English
Stated aim of study To compare UIE and prevalence of goitre in schoolchildren before and after the initiation of
salt iodization in Poland and to analyse regional dierences in the eectiveness of iodine
prophylaxis
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 30 ppm (parts per million)
6. Duration: 5 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with similar sample sizes at
both time points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Schools were representative of dierent areas of
country but appears that only some schools from
baseline survey were selected for follow-up survey
(“randomly”)

94
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Tang et al., 1992 (103)
Methods Cohort observational study conducted in villages near Kunmin, China
Participants All inhabitants of four villages
Interventions A survey of goitre and urinary iodine excretion (UIE) was conducted in four villages. Three
villages then provided iodized salt to the villagers and after 3years the villagers were again
surveyed. The fourth village received iodized salt and iodized oil and was not included in this
analysis.
The three cohorts were:
• Houcun;
• Guanbo;
• Xiaocun.
The concentration of iodine varied by village:
• Houcun <10 ppm (parts per million);
• Guanbo <10 ppm;
• Xiaocun <20 ppm.
Outcomes 1. Goitre prevalence
2. Goitre cure rate in patients with goitre at baseline (not included in this review)
3. UIE (µg/L); there was probably a mistake in the UIE gures, in that the standard deviation
was too small; the meta-analysis was conducted with and without data from this study
4. Tri-iodothyronine (T3), thyroxine (T4), thyroid-stimulating hormone (TSH; not used in this
review because not considered practical indicators)
Publication details Published in a Chinese language journal
Stated aim of study To nd the appropriate level of iodine concentration in salt needed for eective goitre
prevention and control
Notes 1. Design: cohort observational
2. Age: all
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: <20 ppm
6. Duration: 3 years
7. Baseline iodine deciency disorder (IDD) status: based on goitre:
• Houcun moderate IDD
• Guanbo severe IDD
• Xiaocun severe IDD

95
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Though the methods claim a cohort study of
measurement of all villagers, the sample sizes at
follow-up were much higher in Guanbo than at
baseline (more than twice as high)
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Participants from same villages compared to one
another, to reduce bias

96
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Tazhibayev et al., 2008 (104)
Methods Cohort observational studyinvecountriesinEasternEuropeandCentralAsia
Urinary iodine levels were measured in children at sentinel sites in six countries before
salt iodization in those countries. The ve countries then began salt iodization at the
national level at 40 ppm. Four years after the initiation of the salt iodization programmes
in Kazakhstan, Mongolia, Tajikistan and Uzbekistan, and 3years in Azerbaijan, a follow-up
survey in the same children measured urinary iodine concentration again. Women were also
measured at the follow-up studies but their data are not included. Additionally, baseline data
were collected in Kyrgyzstan but, according to the authors, the salt iodization programme in
that country was not initiated and therefore there are no follow-up data.
Participants Children aged 2–15 years in households randomly selected from health clinic registries in the
sentinel sites in six countries; review includes the data from ve countries with baseline and
follow-up data
Cohorts:
• Azerbaijan
• Kazakhstan
• Mongolia
• Tajikistan
• Uzbekistan
Interventions Iodized salt was provided at 40 ppm (parts per million).
Outcomes 1. Urinary iodine excretion (UIE; µg/L)
2. % of population with UIE <100 µg/L
3. Haemoglobin concentration (not utilized in this review)
4. Ferritin (not utilized in this review)
5. Folic acid (not utilized in this review)
Publication details Published in peer-reviewed journal in English
Stated aim of study To determine the potential eectiveness of the food fortication programmes in improving
micronutrient status in a sentinel population in each country
Notes 1. Design: cohort observational
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 40 ppm (parts per million)
6. Duration: 4 years in Kazakhstan, Mongolia, Tajikistan and Uzbekistan, 3 years in Azerbaijan
7. Baseline IDD status: based on UIE:
• Kazakhstan, adequate iodine status
• Mongolia, mild iodine deciency
• Tajikistan, moderate iodine deciency
• Uzbekistan, adequate iodine status
• Azerbaijan, adequate iodine status

97
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Sample size at follow-up not reported so size
assumed to be same as baseline
Selective reporting (reporting bias) Low risk Only data from Kyrgyzstan not reported but stated
it was due to the iodization programme not being
initiated
Other bias Low risk Participants randomly selected from registries at
sentinel site locations in each country

98
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Teng et al., 2009 (76)
Methods Quasi-experimental study conducted in Tuanshan township, Huangdi township, Taihe
township and Jiangtun township, China
Children of each township that were born before iodization of salt were compared to children
in the townships born after iodization of salt.
Participants School-aged children at time that intelligence was measured
Interventions The experiment group was made up of children whose mothers received iodine from iodized
salt during pregnancy and children were exposed to iodized salt in infancy; the comparison
group comprised children whose mothers did not receive iodine from salt during pregnancy
and children were not exposed to iodized salt until up to 12 years of age.
There was one cohort.
Results from all townships were combined; the experiment group was students born after salt
iodization and the comparison group was children born before salt iodization.
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Combined Raven’s Test in China (CRT-C2)
Comparisons were made within village to control for potential biases.
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown (not reported)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus up
to 12 years of life more than comparison
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No mention of blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No mention of blinding
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk No sampling methodology was reported but it
is thought that all children were included and all
children from same villages, to reduce potential bias

99
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Vejbjerg et al., 2009 (144)
Methods Multiple cross-sectional observational study conducted in Denmark
Observational, cross-sectional study conducted in Denmark to evaluate the eect of
iodization of salt on thyroid function by comparing thyroid-stimulating hormone (TSH)
and hypo- and hyperthyroidism pre/post mandatory salt iodization; 4649 participants were
examined in 1997–1998 before salt iodization and 3570 were involved after the introduction
of a mandatory iodization programme in 2000. Participants in the rst survey were excluded
from the second survey and the second survey was age and sex matched to the rst survey.
Participants Adults living in Copenhagen or Aalborg, Denmark who were not taking iodine
supplementation and did not have thyroid disease
Interventions Iodized salt was provided at 13 ppm (parts per million), through mandatory legislation at the
national level.
Outcomes 1. Serum TSH (data not used in this review as crude values not sensitive indicator of status).
2. Hypothyroidism (dened using TSH and free thyroxine [fT4])
3. Hyperthyroidism (dened using TSH and fT4)
4. Thyroid volume by ultrasound (values not reported in results of manuscript and only used
to help diagnose hypo- and hyperthyroidism)
5. Urinary iodine excretion (UIE; only measured at follow-up and not included in this review)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To evaluate the impact of the introduction of a mandatory iodization programme on thyroid
hormone levels, and the prevalence of thyroid dysfunction in the population
Notes 1. Design: multiple cross-sectional
2. Age: adults
3. Group: not specically pregnant women
4. Salt consumption: 5.5–6.0 g/day estimated (from Pedersen 2002 study (100))
5. Iodine concentration: 13 ppm
6. Duration: 4 years
7. Baseline iodine deciency disorder (IDD) status: combination of data from a mild and a
moderate IDD area based on UIE – iodine intake estimated at 66 µg/person/day
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with similar sample sizes
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Samples were selected to be representative of
general population and sampling techniques were
the same for both surveys

100
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang and Wang, 1981 (145)
Methods Multiple cross-sectional observational study conducted in Anqiu county, Shandong
province, China
Participants 84% of inhabitants in 1966 and 83.9% in 1980 of nine villages
Results from nine groups are reported separately but the selection of those groups is not
described and therefore the total values are used in this review.
Interventions Salt was iodized at 100 ppm (parts per million) for 13 years.
Outcomes Goitre (as in many Chinese languages articles, “endemic goitre” and “physical enlargement of
the thyroid” are both reported – the combined value is used as the denition of goitre for use
in this review)
Publication details Published in Chinese in a peer-reviewed journal
Stated aim of study To assess the eect of iodized salt in the prevention of iodine deciency disorder (IDD)
Notes 1. Design: multiple cross-sectional
2. Age: all
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 100 ppm
6. Duration: 13 years
7. Baseline IDD status: severe IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional study with similar sample
sizes at both time points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Entire population included through census
methodology used in both surveys

101
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang et al., 1985 (146)
Methods Multiple cross-sectional observational study, before–after data, conducted in Chengde
city, China
A survey was conducted in 1961 and universal salt iodization began in 1962. A follow-up study
using the same methods was conducted in 1983.
Participants All inhabitants of four villages aged from 0 year to 60+ years
Interventions 50 ppm (parts per million) iodized salt was supplied from 1962 to 1982; no control was used.
Outcomes 1. Goitre prevalence (as in many Chinese languages articles, “goitre” and “physical enlargement
of the thyroid” or “thyroid enlargement” are both reported – the combined value is used as the
denition of goitre for use in this review)
2. Urinary iodine concentration (these data not included in this review because no baseline)
3. I-131 absorption rate of thyroid in a small sample of people with iodine deciency disorder
(IDD; these data not included because no baseline value)
4. Cretinism (reported in summary table)
Publication details Published in a Chinese language journal
Stated aim of study To test the eectiveness of salt iodization to prevent goitre and cretinism
Notes 1. Design: multiple cross-sectional
2. Age: children and adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm
6. Duration: 20 years
7. Baseline IDD status: severe IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional study with similar sample
sizes at both time points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Same survey methods used in both surveys

102
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang and Yang, 1985 (105)
Methods Cohort observational study conducted in China
120 schoolchildren living in an endemic area of iodine deciency in China were selected to
participate in an evaluation of hearing and markers of iodine status; 30 children were selected from
each of four communities, which were selected based on the endemic status of the community for
iodine deciency. Children were excluded if they had mental deciency as reported by the teacher,
had goitre, had a known cause of hearing impairment, or if hearing disorder due to improper
vestibular function as assessed by electronystagmography was detected. Three communities were
described as endemic for iodine deciency and one community was described as non-endemic.
The only comparisons that had outcomes of interest and thus were included in this review were
those from one endemic community measured before the introduction of iodized salt in the
community, and the same children measured 3 years after the introduction of iodized salt. No
information is given as to how the iodized salt was given to the children or if it was merely provided
to the community.
Participants Children 7–11 years of age at baseline attending school in randomly selected communities in
an endemic goitre area of China
Interventions Iodized salt was provided. There are no details regarding the provision of the salt.
Outcomes 1. Thyroxine (T4; not used in this review)
2. Thyroid-stimulating hormone (TSH)
3. Urinary iodine excretion (UIE; µg iodine/g creatinine)
4. I-131 uptake (not used in this review)
5. Hearing level (not used in this review)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To search for marginal cases of endemic cretinism and to assess the therapeutic and
prophylactic eect of iodized salt administration on the hearing of these inhabitants of
endemic areas
(Note: no mention was made of inclusion criteria related to cretinism, and children with
mental deciency were excluded from the study.)
Notes 1. Design: cohort observational
2. Age: children
3. Group: not specically pregnant
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 3 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (as reported by authors but no
data provided)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Reported no loss to follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Selection criteria for children pre and post were the same
and from same community, to reduce potential bias

103
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang et al., 1987 (78)
Methods Quasi-experimental study conducted in Huabeitun and Chengguanzhen villages in China
Children of the villages that were born before iodization of salt were compared to children in
the villages born after iodization of salt.
Participants All children aged 8, 9, 14 and 15 years living in the villages at time that intelligence was
measured (1985)
Interventions The experiment group was made up of children whose mothers received iodine from iodized
salt during pregnancy and children were exposed to iodized salt in infancy; the comparison
group comprised children whose mothers did not receive iodine from salt during pregnancy
and children were not exposed to iodized salt until 1–2 years of age.
There were two cohorts:
• Huabeitun (severe iodine deciency area at baseline);
• Chengguanzhen (mild iodine deciency area at baseline).
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
Though the text says % of children with IQ <70 points (mild mental retardation – MMR) is
reported, results are only presented for both villages combined.
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Sampling methodology not reported
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown (not reported)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus
1–2 years of life more than comparison
7. Baseline iodine deciency disorder (IDD) status:
• Huabeitun: severe IDD (based on goitre; did not use World Health Organization criterion
of school-aged children)
• Chengguanzhen: mild IDD (as reported by authors but no data provided)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding reported
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported (stated
drop-out was less than 15% though statement is
dicult to interpret)
Selective reporting (reporting bias) Unclear risk Results for % of children with intelligence quotient
(IQ) <70 points (mild mental retardation – MMR)
only reported for both villages combined
Other bias Low risk Children from same villages compared to one
another, to reduce potential bias

104
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang et al., 1987 (79)
Methods Quasi-experimental studyconducted in Xinzhou city, Shanghi province, China
Children of Jin village (severe iodine deciency disorder [IDD] area) where there was no salt
iodization were compared to children in Siping village (severe IDD area), which did have salt
iodization. Salt iodization was begun in Siping in 1974. The villages were surveyed in 1974
and again in 1985 after 11 years of universal salt iodization in Siping. Measurements were also
taken in an area without IDD but those data were not included in this review.
Participants Children aged 7–13 years at time that intelligence was measured (1985)
Interventions Universal salt iodization was provided in Siping village at unknown concentration.
Outcomes Intelligence of children and psychomotor function:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
• psychomotor function (selective response time, knock test, reversing hand, ratio of grasp
power)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 7–11 years (11 years intervention but some participants <11 years of age)
7. Baseline IDD status: moderate IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Villages comparable at baseline

105
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang et al., 1992 (80–82)
Methods Quasi-experimental studyconducted in Ninxia Hui autonomous region, China
Children of Shicha village (severe iodine deciency disorder [IDD] area) where there was no
salt iodization were compared to children in Shenzi village (severe IDD area), which did have
salt iodization. Salt iodization was begun in Siping in 1968. The villages were surveyed in 1968
and again in 1983 after 15 years of universal salt iodization in Shenzi. Measurements were also
taken in an area without iodine deciency disorder (IDD) but those data were not included in
this review. Stratied sampling was used to control for age, with 20 children of each year of
age selected randomly for survey and measurements. Children with “brain impairment” were
excluded.
Participants Children aged 7–14 years at time that intelligence was measured (1983)
Interventions Universal salt iodization was provided in Shenzi village at 50 ppm (parts per million).
Outcomes Intelligence of children
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Stated that villages were very similar in sociodemographics and environment
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm
6. Duration: 7–14 years (15 years intervention but some participants < 5 years of age)
7. Baseline IDD status: severe IDD (based on goitre; reported for entire population in Shenzi
and just for children in Shicha)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported but
sampling methodology consistent and sample sizes
similar
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Villages comparable at baseline

106
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang, 1994 (83)
Methods Quasi-experimental study conducted in Hebei province, China
Children living in a “severe IDD [iodine deciency disorder]” area were sampled
(methodology not mentioned) and divided into those born before and those born after
universal salt iodization. The intelligence of those children was measured.
Participants Children aged 7–14 years at time that intelligence was measured (date not stated)
Interventions Universal salt iodization was carried out at unknown concentration.
Outcomes Intelligence of children
• mean intelligence quotient (IQ; standard deviation not provided so data only included
in summary table and not meta-analysed)
• % of children with IQ <70 points (mild mental retardation – MMR; table does not include
the results for this variable though the variable is mentioned)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: exposure duration of experimental group unknown but comparison did not
have exposure during pregnancy
7. Baseline IDD status: combined data from an area of severe IDD area and an area of
adequate iodine status (as stated by authors but no data provided)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment
(selection bias)
High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data
(attrition bias)
Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Unclear risk % of population with intelligence quotient (IQ) <70
not reported as expected
Other bias Unclear risk Children within same villages compared, to reduce
bias, but sample size of comparison group much
larger than experimental group

107
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang et al., 2000 (147)
Methods Multiple cross-sectionalobservational study, before–after design, carried out in Hefei and
Caohu area, China
In two localities, a survey was conducted in 1995 before implementation of universal salt iodization
and it was then implemented in 1996, with a follow-up survey in 1997. In one locality, the universal
salt iodization began in 1995 and the follow-up survey was in 1996. In all three localities, the follow-
up was one year after introduction of universal salt iodization.
Participants All 7–14-year-old students in selected primary schools, a total of 2644 children, included
before supplementation of iodized salt in 1995, and 2395 after supplementation of iodized
salt for 1year in 1996
Interventions Iodized salt supplementation was carried out for one year, the concentration was not
mentioned and no control was used.
Outcomes 1. Urinary iodine concentration (median reported and therefore results presented in summary
tables)
2. Goitre prevalence (goitre is reported as prevalence in each locality, whereas the sample size is
given as a total for all localities, so the number of aected individuals cannot be calculated – results
found in summary table)
3. Perchlorate discharge test, thyroid gland antibody for students with goitre (data not used
because only reported for individuals with goitre)
4. Thyroid-stimulating hormone (TSH), tri-iodothyronine (T3) and thyroxine (T4; not used in this
review)
5. % urinary iodine excretion (UIE) <100 µg/L (only reported for follow-up)
6. % UIE >300 µg/L (only reported for follow-up)
Publication details Published in a Chinese language journal
Stated aim of study To understand the eects of universal salt iodization on IDD control and prevention and the
iodine nutrition and thyroid function of children
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 1 year
7. Baseline iodine deciency disorder (IDD) status: based on goitre/(based on UIE):
• Chaohu: severe IDD/(moderate IDD)
• Dabieshan: moderate IDD/(above requirements)
• Hefei: adequate iodine status/(mild IDD)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation (selection bias) High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding reported
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional study with similar sample
sizes at both time points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Children of same age and from same communities
compared to one another, to reduce potential bias

108
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang, 2001 (106)
Methods Multiple cross-sectional observational study, before–after design, conducted in Qingyuan
county, Hebei province
The study randomly sampled 10 primary and middle schools and examined all of the
students from the 10 sampled schools every year from 1995 to 2000.
Participants 7–14-year-old students from 10 primary and middle schools; 4867 students included before
supplementation of iodized salt in 1995, and 2000 in 2000 after supplementation of iodized
salt
Universal salt iodization began in 1995.
Interventions Universal salt iodization was carried out at a concentration of 20 ppm (parts per million).
Outcomes Goitre prevalence
Publication details Published in a Chinese language journal
Stated aim of study None stated
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20 ppm
6. Duration: 5 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) High risk Cohort observational study with much smaller sample
sizes at follow-up, with no explanation of attrition rate
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Methodology for sample selection not described

109
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang 2005 (77)
Methods Quasi-experimental study conducted in Lianoning province, China
Children without Down syndrome or other brain impairment were sampled (methodology
not mentioned) from Rongxing, Kuanbang, and Pingfangzi townships in June 2002 to
December 2003. Rongxing was an iodine-sucient area and the data were not used in this
review.
Kuanbang did not have iodized salt from 1985 to 1995 (comparison group). Pingfangzi did
have salt iodization. The ages of the children were not mentioned in the manuscript.
Participants Children of unknown age at time that intelligence was measured (2002–2003
Interventions Universal salt iodization was provided at unknown concentration (experiment).
There was no salt iodization from 1985 to 1995 when tested children were born (comparison
group).
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Combined Raven’s Test for Rural China
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: exposure duration of experimental group unknown but comparison did not have
exposure during pregnancy
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported but
sample size of experiment and control similar
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias High risk Exact ages of children not reported/comparability
between villages not clear

110
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wang et al., 2009 (148, 149)
Methods Multiple cross-sectional observational survey conducted in Gansu province, China
Surveys were conducted to assess the eectiveness of salt iodization for the reduction
of goitre in school-aged children. A baseline survey was conducted before universal
salt iodization was implemented and a follow-up survey conducted 10 years after
implementation. Children who were sampled for the measurement of goitre and urinary
iodine were randomly selected through multi-stage cluster sampling. In the rst stage,
schools were randomly selected and in the second stage, students within schools (40 students
per school) were randomly selected. The measurement included thyroid examination. In
a subsample, urinary iodine was measured (12 students per school). In 1997, 2 years after
universal salt iodization was implemented in the province, another survey was conducted to
measure the intelligence quotient (IQ) of students. After the baseline survey, universal salt
iodization was implemented in the country; 10 years later, a follow-up survey with the same
measures was completed in the same schools, using the sample methodology.
Participants School-age children in Gansu province China (8–10 years of age)
Interventions Iodized salt was provided at 50 ppm (parts per million) for 5 years and then 35 ppm, through
universal salt iodization.
Outcomes 1. Goitre
2. Urinary iodine excretion (UIE; median reported and included in summary table)
3. % of population with UIE <100 µg/L
Publication details Published in a peer-reviewed journal in English
Stated aim of study To assess the eects of universal salt iodization on the status of iodine deciency disorder
(IDD) in a region of traditionally severe iodine deciency
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 35 ppm (parts per million)a
6. Duration: 10 years
7. Baseline IDD status: severe IDD (based on goitre)/adequate iodine status (based on UIE)
a Note the concentration was 50 ppm for rst 5 years and then 35 ppm for second 5 years of
follow-up.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
High risk Not blinded
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional with similar sample sizes at
both time points
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Same methods used to select school-aged children
in both surveys, to reduce potential bias

111
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Weber and Manz, 1987 (107)
Methods Cohort observational study conducted in Germany
42 families were recruited for the study if the household had adults and children. A baseline
measure of urinary iodine excretion (UIE) was taken, then all households were provided with
iodized salt for use in the home during a time when use of iodized salt in processed foods was
illegal. The families were follow-up after 6months to measure UIE again.
Participants 78 adults and 73 children from 42 families
Four cohorts:
• children >6 years
• children 6–12 years
• adult men
• adult women
Interventions Iodized salt was provided at 20 ppm (parts per million) given to households for use.
Outcomes UIE (median reported as µg/day and included in summary table)
Publication details Published as a short communication in English
Stated aim of study Not specied
Notes 1. Design: cohort observational
2. Age: adults and children
3. Group: not specically pregnant women
4. Salt consumption: 12.2 g/day adult men; 9.5 g/day adult women; unknown in children
5. Iodine concentration: 20 ppm
6. Duration: 6 months
7. Baseline iodine deciency disorder (IDD) status: not reported
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk No loss to follow up
Selective reporting (reporting bias) Low risk All prespecied outcomes reported
Other bias Unclear risk Methods of sample selection were not reported

112
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Wei, 1985 (150)
Methods Multiple cross-sectional observationalstudy conducted in Daxin county, Guangxi, China
In 1966, 1623 villagers were selected in severe villages of endemic iodine deciency disorders
before supply of iodized salt. Universal salt iodization began in 1967 and in 1981 a follow-up
survey was conducted in the same villages.
Participants Villagers in selected villages in Guangxi, China (1623 at baseline and 543 at follow-up)
Six areas were surveyed at baseline’; however, only three were surveyed at follow-up. The data
from baseline and follow-up from three villages (combined) were used in meta-analysis.
Interventions Iodized salt was provided at unknown concentration.
Outcomes Goitre
Publication details Published in a Chinese provincial journal.
Stated aim of study To assess the eect of supplementation of iodized salt on goitre
Notes 1. Design: multiple cross-sectional
2. Age: children and adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: 15 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Multiple cross-sectional survey but sample size at
follow-up was much less than at baseline
Selective reporting (reporting bias) Low risk All prespecied outcomes reported
Other bias High risk Methods used for sampling not described but very
dierent sample size numbers suggest high risk of
bias

113
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Xu et al., 1984 (108)
Methods Cohort observational study conducted in Yushugou village, Dajiagou community, China
All villagers were surveyed in 1980 before universal salt iodization began. The entire
population of the villages was again surveyed each year until 1984. The data from 1980 and
1984 were used in this meta-analysis as a pre-iodization and post-iodization measure.
Participants All villagers in Yushuge of Dajiage commune (an area dened as severely iodine decient at
baseline)
Interventions Universal salt iodization was provided at 50 ppm (parts per million) iodized salt.
Outcomes 1. Goitre incidence
2. Goitre prevalence
3. Goitre cure rate (data not included in this review)
Publication details Published in a Chinese journal
Stated aim of study To investigate the eect of 50 ppm concentration of iodized salt on goitre prevention and
control
Notes 1. Design: cohort observational
2. Age: children and adults
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 50 ppm
6. Duration: 3 years
7. Baseline iodine deciency disorder (IDD) status: mild IDD (based on goitre; did not use
World Health Organization criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk All villagers surveyed and the sample size was
actually larger at follow-up
Selective reporting (reporting bias) Low risk All prespecied outcomes reported
Other bias Low risk Same census methodology used to measure all
villages with high response rate, so the samples were
comparable

114
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Xue and Zhang, 1993 (151)
Methods Multiple cross-sectional observational study conducted in Hami, Xinjiang, China
Before–after data of supplementation of iodized salt were provided. A survey was conducted
in 1975 in areas dened as mild iodine deciency disorder (IDD), serious IDD, and non-disease
(manner of classication of area not explained). Universal salt iodization began in 1975 and a
follow-up survey was conducted in the same areas in 1992.
Participants In 1975, as the baseline, 15 365 people in mild endemic area, 112 074 in the severe endemic
area, and 3893 in the non-endemic area
In 1992, after supplementation of iodized salt, 2687 in mild endemic area, 48 416 in severe
area and 5454 in non-endemic area
Interventions Iodized salt supplementation was carried out for 18 years (1975 to 1992), the concentration
was 35 ppm (parts per million) and there was no control intervention.
Outcomes 1. Prevalence of goitre
2. Concentration of thyroid hormones (tri-iodothyronine [T3], thyroxine [T4], thyroid-
stimulating hormone [TSH]; only measured at follow-up)
3. Urinary iodine concentration (only measured at follow-up)
4. Prevalence of hyperthyroidism (only measured at follow-up)
Publication details Published in a Chinese journal
Stated aim of study To better understand the eects of iodized salt distribution
Notes Data on thyroid hormones and urinary iodine not included for analysis, owing to lack of data
before supplementation of iodized salt
1. Design: multiple cross-sectional
2. Age: children and adults
3. Group: any group other than specically pregnant women
4. Salt consumption: 10 g/day (not clear how this value was determined)
5. Iodine concentration: 35 ppm
6. Duration: 18 years
7. Baseline IDD status: mild IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Multiple cross-sectional survey but sample size at
follow-up only half the baseline sample size
Selective reporting (reporting bias) Low risk All prespecied outcomes reported
Other bias High risk Sampling methodology not described but very
dierent sample sizes suggest high risk of bias

115
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Yang, 1984 (158)
Methods Multiple cross-sectional observational study (outcome of cretinism) and cohort
observational study (all other outcomes)conducted in Pingliang community, Guizhou,
China
Surveys compared before versus after introduction of iodized salt in Pingliang community,
Kaili county, Guizhou province, China, an endemic iodine deciency disorder (IDD) area.
Because all inhabitants were surveyed, goitre, urinary iodine concentration and thyroid
function were measured following a cohort design. The measure of cretinism was multiple
cross-sectional (incident cretinism).
Participants The entire population of adults and children of surveyed community – 3117 people in 1979
(95% inhabitants), 3067 people in 1982 (92% inhabitants)
Interventions Iodized salt was provided at a concentration of <20 ppm (parts per million).
Outcomes 1. Goitre
2. Tri-iodothyronine (T3), thyroxine (T4), thyroid-stimulating hormone (TSH; not used in this
review)
3. Urinary iodine concentration, (µg iodine/g creatinine – only reported in children and no
standard deviation; results provided in summary table)
4. Rate of absorption of I-131 of thyroid in individuals with iodine deciency (because these
outcomes were limited to those with iodine deciencies, the data are not included in this
review)
5. Incident cretinism (multi-cross-sectional design)
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To assess the eects of prevention and treatment of IDD through iodized salt
Notes 1. Design: multiple cross-sectional and cohort observational
2. Age:adults and children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: <20 ppm
6. Duration: 3 years
7. Baseline IDD status: severe IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional survey but sample size
similar/cohort showed low attrition
Selective reporting (reporting bias) Low risk All prespecied outcomes reported
Other bias Low risk Census methodology used at both times with high
response rate for cohort and multiple cross-sectional
survey conducted in same areas, to reduce risk of
bias

116
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Yang and Zhang, 1984 (109)
Methods Cohort observational study conducted in Qiandongnan autonomous prefecture, China
A survey of goitre and cretinism of the entire population of the prefecture was conducted
in 1978–1979. Universal salt iodization was then implemented. A follow-up survey was
conducted in 1982.
Participants The entire population of adults and children of surveyed communities in prefecture (227 001
people measured at both time points and those data reported)
Interventions Iodized salt was provided at a concentration of <20 ppm (parts per million).
Outcomes 1. Goitre
2. Tri-iodothyronine (T3), thyroxine (T4), thyroid-stimulating hormone (TSH; not used in this
review)
3. Urinary iodine concentration (µg iodine/g creatinine – reported as median and range (only
provided for 3 of 40 communes surveyed); results provided in summary table
4. Rate of absorption of I-131 of thyroid in individuals of “target group” – not clear what this
group was (data not included in this review)
5. Incident cretinism (only reported for follow-up and therefore could not be included in this
review)
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To assess the eects of prevention and treatment of iodine deciency disorder (IDD) through
iodized salt
Notes 1. Design: cohort observational
2. Age:adults and children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: <20 ppm
6. Duration: 3 years
7. Baseline IDD status: mild IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Low attrition (entire prefecture)
Selective reporting (reporting bias) High risk Some outcomes only reported for specic
communes of prefecture
Other bias Low risk Same census methodology used at both times, with
high participation

117
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Yang and Yang, 2011 (152)
Methods Multiple cross-sectional observational study conducted in China
The study randomly selected a primary school, with 40–100 schoolchildren aged 8–10 years.
The rst survey was conducted before the provision of iodized salt in the community and the
follow-up surveys, conducted yearly, were conducted up to 15 years after the initiation of
iodization of salt in the community. The data include every year from 1995 to 2009, 1995 and
2009 data only were compared in this review.
Participants 8–10-year-old schoolchildren
Interventions Iodized salt was provided at a concentration of 35 ppm (parts per million), through provision
at the population level.
Outcomes 1. Goitre
2. Urinary iodine excretion (UIE; median reported and results in summary table)
3. % population <100 µg/L UIE
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To monitor the eect of preventing iodine deciency disorder (IDD) by the provision of
iodized salt
Notes 1. Design: multiple cross-sectional
2. Age:children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 35 ppm
6. Duration: 15 years
7. Baseline IDD status: mild IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Unclear risk Multiple cross-sectional survey with much lower
sample size at follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias High risk Sampling methodology was not detailed and very
dierent sample sizes suggest high risk of bias

118
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Yuan et al., 1993 (153)
Methods Multiple cross-sectional observational study conducted in Nanhai and Beihe
communities of Qianjiang county, Sichuan, China
The rst survey was conducted in 1978 before provision of iodized salt, and the second
survey was conducted in 1992 after provision of iodized salt for 15 years.
Participants Schoolchildren of primary and middle school
Interventions Iodized salt was provided at unknown concentration.
Outcomes Goitre
Publication details Published as a short report in Chinese
Stated aim of study To investigate the eect of iodized salt provision on goitre
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: <20 ppm (parts per million)
6. Duration: 15 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation (selection bias) High risk Not randomized controlled trial (RCT )
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
Unclear risk Not reported
Blinding of outcome assessment (detection bias) Unclear risk Not reported
Incomplete outcome data (attrition bias) Unclear risk Multiple cross-sectional survey but sample
size at follow-up was half of that of baseline
Selective reporting (reporting bias) Low risk All prespecied outcomes reported
Other bias High risk Sampling methodology was not detailed and
very dierent sample sizes suggest high risk
of bias

119
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Yusuf et al., 2008 (154)
Methods Multiple cross-sectional observational study conducted in Bangladesh
Nationally representative surveys on nutritional status in children and women of reproductive
age in Bangladesh were conducted in numerous years and the data provide national
estimates of IDD before and after the introduction on iodized salt in that country.Probability
proportionate to size (PPS) sampling was used to select households for inclusion, to select a
nationally representative sample. The rst sample used in this review was from the year 1993.
The Salt Byelaw was passed in Bangladesh in 1994 and at that time substantial investments
were made in the infrastructure for salt iodization. By 1995, all salt factories were equipped
for iodization (the original law was passed in 1989 but no eort was made to enforce it
and there was little iodization in the country until 1994). The data from a second nationally
representative survey undertaken 10 years later (2005) was used to compare the iodine status
of children and women pre and post fortication of salt with iodine.
Participants Children aged 6–12 years or women aged 15–44 years living in Bangladesh
Interventions Iodized salt was provided at the national level, at a concentration of 45–50 ppm (parts per
million).
Outcomes 1. Goitre (only percentages without sample size reported at baseline; data found in summary
table)
2. UIE (values only given for follow-up)
3. % of population with urinary iodine excretion (UIE) <100 µg/L (only percentages without
sample size reported at baseline; data found in summary table)
Publication details Published in English in a peer-reviewed journal
Stated aim of study To monitor the situation towards the elimination of iodine deciency disorder (IDD) in
Bangladesh
Notes 1. Design: multiple cross-sectional (nationally representative)
2. Age: adults and children
3. Group: not specically pregnant
4. Salt consumption: unknown
5. Iodine concentration: 45–50 ppm
6. Duration: 10 years
7. Baseline IDD status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
High risk Not blinded
Incomplete outcome data (attrition bias) Unclear risk Sample sizes not provided for baseline data
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Sampling procedures and weighing reported to have
been conducted, to ensure surveys were comparable
and nationally representative

120
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zeng, 1991 (87–90)
Methods Quasi-experimental study conducted in the townships of Heba, Shilong and Qianling in
Guizhou province China
Participants Children 7–13 years of age during time of survey of 1989–1990
Interventions The experiment group was made up of children whose mothers received iodine from iodized
salt during pregnancy and children were exposed to iodized salt in infancy; the comparison
group comprised children whose mothers did not receive iodine from salt during pregnancy
and children were not exposed to iodized salt until 1–3 years of age.
There were three cohorts:
• Heba (iodine-decient village at baseline) – salt iodized at 20 ppm (parts per million);
• Shilong (iodine-decient village at baseline) – salt iodized at 20 ppm;
• Qianling (iodine-sucient village at baseline) – salt iodized at unknown concentration
(not reported).
Outcomes Intelligence:
• Mean intelligence quotient (IQ)
• % of children with IQ <70 (mild mental retardation – MMR)
Publication details Published in a Chinese language journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Comparisons made within village to control for potential biases
1. Design: quasi-experimental
2. Age: children (7–13 years)
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20 ppm (Heba and Shilong) or unknown (Qianling)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus
1 to 3 years of life more than comparison
7. Baseline iodine deciency disorder (IDD) status (based on goitre; did not use World Health
Organization criterion of school-aged children):
• Heba: severe IDD
• Shilong: moderate IDD
• Qianling: adequate iodine status
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Comparable comparison groups that were of similar size
and selected in similar manner

121
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zhang et al., 1988 (110)
Methods Cohort observational study conducted in Xiguanying community, Beipiao city, Liaoning
province, China
15 villages’ residents of an endemic iodine deciency disorder (IDD) area; established records
for each villager – 18 252 villagers in 1974 before supplementation of iodized salt and 13 248
in 1977 after 3years’ supplementation of iodized salt were included
Participants All villagers in selected villages
Interventions Iodized salt was provided via universal salt iodization, at a concentration of 100 ppm
(parts per million).
Outcomes 1. Goitre
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To investigate the eect of prevention of goitre through provision of iodized salt
Notes 1. Design: cohort observational
2. Age: children and adults
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 100 ppm
6. Duration: 3 years
7. Baseline IDD status: mild IDD (based on goitre; did not use World Health Organization
criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Not blinded
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) High risk >20% attrition
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Sampling methodology not explained

122
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zheng et al., 1995 (84, 85)
Methods Quasi-experimental study conducted in Yue village of Lushan county and Lian village of
Ye county in Henan Province China
Participants Children 9–10 years of age during time of survey (date not specied)
Interventions Children were divided into those born before the implementation of universal salt iodization
in 1979. The experiment group was made up of children whose mothers received iodine
from iodized salt during pregnancy and children were exposed to iodized salt in infancy; the
comparison group comprised children whose mothers did not receive iodine from salt during
pregnancy and children were not exposed to iodized salt as infants and young children.
Outcomes • Intelligence quotient (IQ; table is incomplete and no data can be used for this review)
• % of children with IQ <70 (mild mental retardation – MMR; table is incomplete and no
data can be used for this review)
Publication details Published in a Chinese language journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Comparisons made within village to control for potential biases
1. Design: quasi-experimental
2. Age: children (9–10 years)
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown
6. Duration: intervention group exposed to iodized salt during gestation (9 months) and the
comparison was not (the duration of exposure as infants and young children was not stated)
7. Baseline iodine deciency disorder (IDD) status: based on goitre:
• Yue: severe IDD (did not use World Health Organization (WHO) criterion of school-aged
children)
• Lian: mild IDD (did use WHO criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Unclear risk Table in article incomplete
Other bias Low risk Children of same village compared to one another
and only 1 year apart in age and therefore
comparable

123
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zhou et al., 2004 (155)
Methods Multiple cross-sectional observational study conducted in Zhaotong city, Yuannan
province, China
Randomly sampled students from one primary school
A survey of students aged 8–10 years was conducted before iodization of salt and then
another sample of students were surveyed 8 years after introduction of iodized salt.
Participants Schoolchildren, 8–10 years of age
Interventions Iodized salt was provided at a concentration of 15 ppm (parts per million).
Outcomes 1. Goitre
2. urinary iodine excretion (UIE; median reported; results in summary table)
Publication details Published in a peer-reviewed journal in Chinese
Stated aim of study To assess the eect of provision of iodized salt on goitre and UIE
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 15 ppm
6. Duration: 8 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
High risk No blinding
Incomplete outcome data (attrition bias) High risk Only 40 subjects measured at follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk Sampling methodology not detailed

124
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zhu et al., 1993 (86)
Methods Quasi-experimental study conducted in Qingshui, Sizhai and Shilong townships, China
Children in each village born before iodization of salt were compared to children of the same
village born after iodization of salt.
Participants Children aged 7–14 years at time that intelligence was measured (1991)
Interventions Experiment group was made up of children whose mothers received iodine from iodized salt
during pregnancy and children were exposed to iodized salt in infancy; the comparison group
comprised children whose mothers did not receive iodine from salt during pregnancy and
children were not exposed to iodized salt until 1 year of age.
There were three cohorts:
• Qingshui (severe iodine-decient area);
• Sizhai (mild iodine-decient area);
• Shilong (severe iodine-decient area).
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Comparisons made within villages, to control for potential biases
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20 ppm (parts per million)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus 1
year of life more than comparison
7. Baseline iodine deciency disorder (IDD) : based on goitre:
• Qingshui: severe IDD (did not use World Health Organization (WHO) criterion of school-
aged children)
• Sizhai: mild IDD (did not use WHO criterion of school-aged children)
• Shilong: severe IDD (did not use WHO criterion of school-aged children)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk Blinding not reported
Blinding of outcome assessment
(detection bias)
Unclear risk Blinding not reported
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Unclear risk % of children with intelligence quotient (IQ) <70
points not reported
Other bias Low risk Children within same communities compared to one
another, to reduce potential bias

125
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zhu et al., 1995 (92)
Methods Quasi-experimental study conducted in Niuchang, Qingshui and Yiangfeng townships in
China
Children of each village that were born before iodization of salt were compared to children in
the village born after iodization of salt.
Participants Children aged 7–14 years at time that intelligence was measured (1993); children “with brain
impairment from non-iodine factors” were excluded
Interventions The experiment group was made up of children whose mothers received iodine from iodized
salt during pregnancy and children were exposed to iodized salt in infancy; the comparison
group comprised children whose mothers did not receive iodine from salt during pregnancy
and children were not exposed to iodized salt until 1–3 years of age.
There were three cohorts:
• Niuchang (severe iodine deciency area at baseline);
• Qingshui (severe iodine deciency area at baseline);
• Yiangfeng (severe iodine deciency area at baseline).
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Comparisons made within village, to control for potential biases
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 20 ppm (parts per million)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus 1
year of life more than comparison
7. Baseline iodine deciency disorder (IDD) status: based on goitre (did not use World Health
Organization criterion of school-aged children)
• Niuchang: severe IDD
• Qingshui: severe IDD
• Yiangfeng: severe IDD
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No mention of blinding
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Unclear risk % of children with intelligence quotient (IQ) <70 not
reported as expected
Other bias Low risk Children within same communities compared to
one another, to reduce potential bias (communities
were of dierent ethnic groups)

126
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zimmermann et al., 2003 (156)
Methods Multiple cross-sectionalobservational study, conducted in six remote villages in the
Danane health district of western Côte d'Ivoire.
Surveys of goitre and other indicators of iodine status were conducted in schools of the
villages. All children, aged 5–14 years of age, who were present at school on the day of
measurement were included in the sample. The rst survey was conducted in 1997 before
iodized salt was available in the villages, and children from two schools were included.
In 1998, universal salt iodization was initiated. Other surveys using similar methods were
conducted in schoolchildren in the same villages, yearly until 2001. Students from two
schools were surveyed in 1997 and 1998 and six schools were included in the surveys in
1999, 2000 and 2001.
Participants All children aged 5–14 years, present at school when surveys were conducted
Interventions Iodized salt was provided at a concentration of 30–50 ppm (parts per million).
Outcomes 1. Goitre
2. Urinary iodine excretion (UIE; median reported and values used in summary table)
3. % of population with UIE <100 µg/L
4. Serum thyroid-stimulating hormone (TSH; not used in this review)
5. Hypothyroidism
6. Serum thyroxine (not measured at follow-up and therefore not included in this review)
7. Thyroid volume (not used in this review)
Publication details Published in a peer-reviewed journal in English
Stated aim of study To measure the thyroid size, urinary iodine and thyroid hormones before and after the
introduction of universal iodized salt
Notes 1. Design: multiple cross-sectional
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: 30–50 ppm (parts per million)
6. Duration: 3 years
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)/moderate
IDD (based on UIE)
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk Multiple cross-sectional study with similar sample
sizes at each time point
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Sampling methodology similar at all time points and
children same age, to reduce risk of bias for secular
trends with age

127
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zimmermann et al., 2004 (111)
Methods Cohort observational study conducted in northern Morocco
A cohort of children were followed from pre-iodization to 1year post-iodization and
then 14 months after cessation of iodization of salt. Children were recruited and baseline
measurement performed. Households were then provided with iodized salt. After 1year, the
same measurements were conducted on the same children. Owing to nancial constraints,
iodized salt could no longer be provided to the households of the children. Fourteen months
after the cessation of provision of iodized salt to the households, the same measurements
were again conducted on the selected households.
Participants Schoolchildren aged 6–16 years
Interventions Iodized salt was provided at 25 ppm (parts per million), through provision to households.
Outcomes 1. Goitre
2. Urinary iodine excretion (UIE; median reported; results in summary table)
3. % of population with UIE <100 µg/L
3. Whole-blood thyrotropin (not used in this review)
4. Serum total thyroxine (not used in this review)
5. Serum thyroglobulin (not used in this review)
6. Thyroid volume (not used in this review)
7. Hypothyroidism
Publication details Published in a peer-reviewed journal in English.
Stated aim of study To describe the evolution of thyroid dysfunction after the discontinuation of salt iodization in
a cohort of children in an area of severe endemic goitre
Notes 1. Design: cohort observational
2. Age: children
3. Group: not specically pregnant women
4. Salt consumption: 7.3–11.6 g/day
5. Iodine concentration: 25 ppm
6. Duration: 1 year
7. Baseline iodine deciency disorder (IDD) status: severe IDD (based on goitre)/severe IDD
(based on UIE)
• Estimated intake of iodine 183–290 µg/person/day
Data used in this review are those of baseline and follow-up after introduction of iodized salt;
however, the authors also followed up children 14 months after the cessation of salt iodization.
The authors’ conclusions were that there was a rapid regression of iodine status back to that of
insuciency in this sample, within 14 months of cessation of iodization of salt.
A number of variables were reported that were not synthesized in the analyses in this review, all of
which indicated an improvement in thyroid function with iodization of salt and a return to poorer
thyroid function with cessation of iodization of salt. Adverse eects were not detected.
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation (selection bias) High risk Not randomized controlled trial (RCT )
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment (detection bias) Unclear risk No blinding reported
Incomplete outcome data (attrition bias) Low risk No loss to follow-up
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Unclear risk All children from two schools included in survey
– selection criteria for schools not reported

128
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Zuo et al., 1996 (91)
Methods Quasi-experimental study conducted in Wenan county and Shadu county, China
Children at two primary schools that were born before iodization of salt were compared to
children in the schools born after iodization of salt.
Participants Children aged 7–13 years at time that intelligence was measured
Interventions The experiment group was made up of children whose mothers received iodine from iodized
salt during pregnancy and children were exposed to iodized salt in infancy; the comparison
group comprised children whose mothers did not receive iodine from salt during pregnancy
and children were not exposed to iodized salt until 1–3 years of age.
There were two cohorts:
• students at Qixing primary school;
• students at Wagua primary school.
Outcomes Intelligence of children:
• mean intelligence quotient (IQ)
• % of children with IQ <70 points (mild mental retardation – MMR)
Publication details Published in a Chinese journal
Stated aim of study Not stated
Notes Intelligence measured with Chinese Binet Scale
Students in school were randomly selected for measurement of IQ
Comparisons made within schools to control for potential biases
1. Design: quasi-experimental
2. Age: children
3. Group: any group other than specically pregnant women
4. Salt consumption: unknown
5. Iodine concentration: unknown (not stated in article)
6. Duration: intervention group exposed to iodized salt during gestation (9 months) plus 1–3
years of life more than comparison
7. Baseline iodine deciency disorder (IDD) status:
• Qixing: adequate iodine status (based on goitre; did not use World Health Organization
criterion of school-aged children)
• Wagua: severe as reported by authors, but data not reported
Risk of bias table
Bias
Authors’
judgement Support for judgement
Random sequence generation
(selection bias)
High risk Not randomized controlled trial (RCT)
Allocation concealment (selection bias) High risk Not RCT
Blinding of participants and personnel
(performance bias)
High risk No blinding
Blinding of outcome assessment
(detection bias)
Unclear risk No mention of blinding
Incomplete outcome data (attrition bias) Low risk As quasi-experimental, attrition not reported
Selective reporting (reporting bias) Low risk All outcomes reported
Other bias Low risk Students from same schools in same villages compared
to one another, to reduce potential bias

129
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Annex 3. Summary table: goitre prevalence with and without iodization of salt
Study ID Country Study design Age group
Concentration of iodine in
salt, ppm
Goitre prevalence
without salt iodization
Goitre prevalence with salt
iodization
Bimenya et al.,
2002 (116)
Uganda Multiple cross-sectional Children 50 74.3 60
Hou et al., 2003
(131)
China Multiple cross-sectional Children and adults 20–50 18.2 5.5
Ibanez-Gonzalez
et al., 1956 (98)
Spain Cohort observational Children (<10years) <20 40.0 8.0
Children (10–14.9years) <20 61.3 27.9
Adolescents
(15–19.9years)
<20 61.7 33.4
Adults (≥20years) <20 80.0 64.7
Jia et al., 2004
(157)
China Cohort Children Unknown 1.9 0.00
Kimball et al., 1931
(56)
United States
of America
Non-RCT Children Unknown (iodized salt) 36 12
Unknown (control) 34 30
Nicod, 1953 (138) Switzerland Multiple cross-sectional Children (6-year follow-up Valais
canton)
10 71.2 29.5
Children (30-year follow-up
Vaud canton)
10 Avenches: 78.9
Moudon: 73.5
Payerne: 78.0
7.1
5.0
4.0
Wang et al., 2000
(147)
China Multiple cross-sectional School-age children Unknown Chaohu: 59.6
Dabeisan: 28.5
Hefei: 4.6
22.0
11.9
9.5
Yusuf et al., 2008
(154)
Bangladesh Multiple cross-sectional Children 45–50 49.9 6.2
Women 45–50 55.6 11.7
ppm: parts per million; RCT: randomized controlled trial.

130
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Annex 4. Summary tables: urinary iodine excretion with and without iodized salt
Randomized controlled trials and non-randomized controlled trials
Without iodized salt With iodized salt
Study ID
Study
design Country
Age
group
Concentration
of iodine in salt,
ppm nMedian Units nMedian Units
Hintze et al., 1988 (53) RCT Germany Children 20 160 45.1 µg iodine/g creatinine 126 60.1 µg iodine/g creatinine
I'Ons et al., 2000 (57) Non-RCT South
Africa
Children >40 69 94 µg iodine per liter of
urine
39 216 µg iodine per liter of
urine
Pongpaew et al., 1998 (58) Non-RCT Thailand Children >40 93 Baseline 129
Follow-up 302
µg iodine per liter of
urine
98 96
195
µg iodine per liter of
urine
Romano et al., 1991 (54) RCT Italy Adults 20 18 50.0 µg iodine/24 h 17 100.0 µg iodine/24 h
Saowakhontha et al., 1994 (59) Non-RCT Thailand Adults 50 Values reported were a factor of 10–1000× greater than expected and therefore considered not
reliable and not included here
ppm: parts per million; RCT: randomized controlled trial.

131
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Urinary iodine excretion in cohort observational studies
Without iodized salt With iodized salt
Study ID Country
Age
group
Concentration
iodine in salt,
ppm nMedian Units nMedian Units
Foo et al., 1996 (95) Malaysia All >40 33 ( aged ≤6 years)
48 (women)
28.1
36.8
µg iodine per liter
of urine
33
48
33.4
168.2
µg iodine per
liter of urine
Guo, 1984 (97) China All 20 31 (goitre adults)
30 (normal adults)
33 (children with cretinism)
13 (normal children)
16.5
24.5
30.0
34.1
µg iodine/g
creatinine
31
32
33
13
99.7
99.2
131.2
170.6
µg iodine/g
creatinine
Weber and Manz, 1987 (107) Germany All 20 NR Men 55
Women 39
6–12 years 21
<6 years 21
µg/day 63
49
33
28
µg/day
Yang and Zhang, 1984 (109) China All <20 NR Cretinism
patients 3.9
Non-cretinism
patients 19.9
µg iodine/g
creatinine
NR 139.9
134.3
µg iodine/g
creatinine
Yang et al., 1984 (158) China All <20 158 Hebei
150 Pingliang
160 Guading
23.3
29.1
20.7
µg iodine/g
creatinine
104
105
62
128.4
153.0
178.0
µg iodine/g
creatinine
Zimmermann et al., 2004
(111)
Morocco Children 25 159 18 µg iodine per liter
of urine
159 181 µg iodine per
liter of urine
NR: not reported; ppm: parts per million.

132
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Urinary iodine excretion in multiple cross-sectional observational studies
Without iodized salt With iodized salt
Study ID Country
Age
group
Concentration
iodine in salt, ppm nMedian Units nMedian Units
Baczyk et al., 2007 (114) Poland Children 30 NR 49 µg iodine per liter of urine NR 107 µg iodine per liter of urine
Chen et al., 1999 (121) China Children 33 214 71.82 µg iodine per liter of urine 210 105.11 µg iodine per liter of urine
Chen et al., 2001 (122) China Children 33 210 60.10 µg iodine per liter of urine 210 103.4 µg iodine per liter of urine
Chen et al., 2002 (123) China Children Unknown 147 (HAC)
64 (QGR)
68.53
61.25
µg iodine per liter of urine 147
63
140.11
99.69
µg iodine per liter of urine
Dai et al., 2008 (124, 125) China Children Unknown 364 (city)
338 (suburb)
367 (rural)
137.5
102.12
94.66
µg iodine/g creatinine 257
210
210
271.12
240.42
198.14
µg iodine/g creatinine
Han et al., 2006 (129) China All Unknown 1829 96.1 µg iodine per liter of urine 358 228.1 µg iodine per liter of urine
Hou et al., 2003 (131) China All 20–50 NR 187.6 µg iodine per liter of urine NR 247.5 µg iodine per liter of urine
Jooste et al., 2000 (134) South
Africa
Children >40 565 21.59 µg iodine per liter of urine 536 186.7 µg iodine per liter of urine
Lv et al., 2009 (136) China Children 35 360 160.1 µg iodine per liter of urine 362 212.3 µg iodine per liter of urine
Regalbuto et al., 2010 (139) Italy Children 30 NR mean (SD)
Bronte: 129.3 (16.6.)
Maniace: 68.2 (12.6.)
µg iodine per liter of urine NR 192.0
(117.2.)
160.8
(79.5.)
µg iodine per liter of urine

133
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Without iodized salt With iodized salt
Study ID Country
Age
group
Concentration
iodine in salt, ppm nMedian Units nMedian Units
Wang et al., 2009 (148, 149) China Children 35 362 119.9 µg iodine per liter of urine 356 191.8 µg iodine per liter of urine
Yang and Zhang, 1984
(109)
China All <20 NR cretinism patients 3.9
non- patients 19.9
µg iodine/g creatinine NR 139.9
134.3
µg iodine/g creatinine
Yang and Yang, 2011 (152) China Children 35 100 109.4 µg iodine per liter of urine 100 216.9 µg iodine per liter of urine
Zhou et al., 2004 (155) China Children <20 262 52.5 µg iodine per liter of urine 12 342.1 µg iodine per liter of urine
Zimmermann et al., 2003
(156)
Côte
d'Ivoire
Children 30–50 419 28 µg iodine per liter of urine 526 104 µg iodine per liter of urine
NR: not reported; ppm: parts per million.
Note: In the studyconducted by Yusuf et al. in 2008 (154), in children and adults in Bangladesh, the percentage of the population with urinary iodine excretion <10µg/L decreased from 71.0% to 33.8% (children)
and 70.2% to 38.7% (adults) from before the introduction of iodized to after the introduction of salt iodized at a concentration of >40 ppm.

134
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Annex 5. Summaries of the risk of bias of included studies
Risk of bias graph: review of authors’ judgements about each risk of bias item, presented as
percentages across all included studies
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
0% 25% 50% 75% 100%
Low risk of bias Unclear risk of bias High risk of bias

135
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
Risk of bias summary
Aghini-Lombardi et al., 1993 (112)
Azizi et al., 2002 (113)
Baczyk et al., 2007 (114)
Bauch et al., 1990 (115)
Bimenya et al., 2002 (116)
Cerqueira et al., 2009 (117)
Charania et al., 1988 (118, 119)
Chen et al., 1976 (120)
Chen et al., 1984 (93, 94)
Chen et al., 1991 (63)
Chen et al., 1999 (121)
Chen et al., 2001 (122)
Chen et al., 2002 (123)

136
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
Risk of bias summary
Chen et al., 2005 (64)
Dai et al., 2008 (124, 125)
Dong et al., 1988 (65)
Fei et al., 1996 (126)
Foo et al., 1996 (95)
Fu et al., 1987 (66)
Fu et al., 2001 (127)
Gatti et al., 1980 (128)
Golkowski et al., 2007 (96)
Gongora, 1952 (55)
Guo, 1984 (97)
Han et al., 2006 (129)
He et al., 1993 (67, 68)

137
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
Risk of bias summary
Heydarian et al., 2007 (130)
Hintze et al., 1988 (53)
Hou et al., 2003 (131)
Hou and Wang, 2009 (132)
Hu et al., 1998 (133)
Huang et al., 2004 (159)
Ibanez Gonzalez et al., 1956 (98)
I’Ons et al., 2000 (56)
Jia et al., 2004 (157)
Jooste et al., 2000 (134)
Kimball, 1931 (57)
Kimball, 1946 (135)
Kimiagar et al., 1990 (99)

138
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
Risk of bias summary
Li et al., 1991 (69–72)
Lv et al., 2009 (136)
Mostafavi et al., 2005 (137)
Nicod 1953 (138)
Pan et al., 1995 (73)
Pedersen et al., 2002 (100–102)
Pongpaew et al., 1998 (58)
Regalbuto et al., 2010 (139)
Romano et al., 1991 (54)
Rueda Williamson and Pardo Tellez, 1966 (140)
Salvaneschi et al., 1991 (141)
Saowakhontha et al., 1994 (59)
Scrimshaw et al., 1966 (142)

139
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
Risk of bias summary
Shen, 1991 (74)
Shu, 1987 (75)
Sooch and Ramalingaswami, 1965 (60–62)
Szybinski et al., 2001 (143)
Tang et al., 1992 (103)
Tazhibayev et al., 2008 (104)
Teng et al., 2009 (76)
Vejbjerg et al., 2009 (144)
Wang and Wang, 1981 (145)
Wang et al., 1985 (146)
Wang and Yang, 1985 (105)
Wang et al., 1987 (78)
Wang et al., 1987 (79)

140
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
Risk of bias summary
Wang et al., 1992 (80–82)
Wang et al., 1994 (83)
Wang et al., 2000 (147)
Wang, 2001 (106)
Wang 2005 (77)
Wang et al., 2009 (148, 149)
Weber and Manz, 1987 (107)
Wei, 1985 (150)
Xu et al., 1984 (108)
Xue and Zhang, 1993 (151)
Yang, 1984 (158)
Yang and Zhang, 1984 (109)
Yang and Yang, 2011 (152)

141
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
Risk of bias summary
Yuan et al., 1993 (153)
Yusuf et al., 2008 (154)
Zeng, 1991 (87–90)
Zhang et al., 1988 (110)
Zheng et al., 1995 (84, 85)
Zhou et al., 2004 (155)
Zhu et al., 1993 (86)
Zhu et al., 1995 (92)
Zimmermann et al., 2003 (156)
Zimmermann et al., 2004 (111)
Zuo et al., 1996 (91)

142
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Annex 6. GRADE summary of evidence
Iodized salt and cognitive function
Patient or population: children
Settings: all settings
Intervention: iodized salt
Comparative values* and risks** (95% CI)
Outcomes (study design)
Assumed risk
Comparison
Corresponding risk
Iodized salt
Relative eect***
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE) Comments
Cognitive function – (RCT) 0 (0)
Cognitive function – (non-RCT) 0 (0)
Cognitive function – (cohort ) 0 (0)
Cognitive function – (quasi-experimental)
mean IQ
Mean IQ – in comparison group
80.39 points
Mean IQ – in intervention group
8.18 higher
(6.71 to 9.65 higher)
12995
(18 studies)
⊕⊕⊝⊝
low1,2,3
Cognitive function – (multiple cross-sectional)
mean IQ
Mean IQ – in comparison group
82.31 points
Mean IQ – in intervention group
10.45 higher
(4.79 to 16.11 higher)
2262
(2 studies)
⊕⊕⊝⊝
low1,3
Cognitive function – (quasi-experimental)
IQ <70
107 per 1000 30 per 1000
(23 to 38)
RR 0.28
(0.22 to 0.36)
12761
(16 studies)
⊕⊕⊝⊝
low1,2,3
Cognitive function – (multiple cross-sectional)
IQ <70
51 per 1000 12 per 1000
(4 to 41)
RR 0.24
(0.07 to 0.82)
509
(1 study)
⊕⊕⊝⊝
low1,2,3
Only 1
study
*Mean value for the comparison group and the group exposed to iodized salt.
**The assumed risk for categorical variables is the risk in the group unexposed to iodized salt. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative eect
of the intervention (and its 95% CI).
*** Calculated as risk ratio.
IQ: intelligence quotient; CI: condence interval; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 Many studies did not provide sucient detail to assess risk of bias.
2 Quasi-experimental studies.
3 Large protective eect with iodized salt.

143
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Iodized salt for iodine deciency
Patient or population: children, adults, general population
Settings: all settings
Intervention: iodized salt
Comparative values* and risks** (95% CI)
Outcomes (study design)
Assumed risk
Comparison
Corresponding risk
Iodized salt
Relative eect**
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE) Comments
Iodine deciency UIE <100µg/L (non-RCT) Study population RR 0.42
(0.14 to 1.29)
233
(2 studies)
⊕⊕⊝⊝
low1,2
349 per 1000 147 per 1000
(49 to 450)
Iodine deciency UIE <100µg/L (cohort) Study population RR 0.40
(0.26 to 0.60)
1118
(6 studies)
⊕⊕⊝⊝
low2
640 per 1000 256 per 1000
(167 to 384)
Iodine deciency UIE <100µg/L (multiple cross-sectional) Study population RR 0.45
(0.34 to 0.59)
7252
(10 studies)
⊕⊕⊝⊝
low2,3,4
707 per 1000 318 per 1000
(240 to 417)
* The assumed risk is the risk in the comparison group. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative eect of the intervention (and its 95% CI).
**Calculated as risk ratio.
CI: condence interval; RCT: randomized controlled trial; RR: risk ratio; UIE: urinary iodine excretion.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 Non-RCTs.
2 95% CI crossed null.
3 Many studies lack reporting to assess true risk of bias.
4 Very large protective eect of iodized salt.

144
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Iodized salt for urinary iodine excretion
Patient or population: general population
Settings: all settings
Intervention: iodized salt
Comparative values* (95% CI)
Outcomes (study design)
Assumed risk
Comparison
Corresponding risk
Iodized salt
No of participants
(studies)
Quality of the evidence
(GRADE) Comments
UIE – (cohort)
µg/L iodine in urine
Mean UIE in comparison group
97.11 µg/L
Mean UIE in intervention group
59.22 higher
(50.40 to 68.04 higher)
1307
(3 studies)
⊕⊕⊝⊝
low1,2
UIE – (cohort)3
µg iodine/g creatinine in urine
Mean UIE in comparison group
42.90 µg iodine/g creatinine
Mean UIE in intervention group
87.35 higher
(49.74 to 124.97 higher)3
244
(2 studies)
⊕⊕⊕⊝
moderate2
UIE – (multiple cross-sectional)
µg iodine/L urine
Mean UIE in comparison group
68.62 µg/L
Mean UIE in intervention group
72.35 higher
(44.54 to 100.17 higher)
6760
(5 studies)
⊕⊕⊕⊝
moderate2
UIE – (multiple cross-sectional)
µg iodine/g creatinine
Mean UIE in comparison group
42.78 µg iodine/g creatinine
Mean UIE in intervention group
104.11 higher
(55.28 to 152.94 higher)
1773
(4 studies)
⊕⊕⊕⊝
moderate2
*Mean value for the comparison group and the group exposed to iodized salt.
CI: condence interval; MD: mean dierence; RR: risk ratio; UIE: urinary iodine excretion.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 Many studies lacked detail needed to assess true risk of bias.
2 Very large positive eect of iodized salt on UIE.
3 The standard deviation for one of the two studies was 10-fold less than the other study and these values could not be veried; therefore, the data were not pooled. Both studies showed a signicant increase in
UIE with iodized salt (MD=66 (45, 87); MD = 105 (104, 106).

145
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Iodized salt for goitre
Patient or population: general population
Settings: all settings
Intervention: iodized salt
Comparative risks* (95% CI)
Outcomes
Assumed risk
Comparison
Corresponding risk
Iodized salt
Relative eect**
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE) Comments
Goitre – (RCT) 225 per 1000 238 per 1000
(155 to 365)
RR 1.06
(0.69 to 1.62)
286
(1 study)
⊕⊕⊝⊝
low1,2
Only 1 study
Goitre – (non RCT) 422 per 1000 249 per 1000
(152 to 401)
RR 0.59
(0.36 to 0.95)
32219
(3 studies)
⊕⊕⊕⊝
moderate3,4
Goitre – (quasi-experimental) 19 per 1000 2 per 1000
(1 to 2)
OR 0.1
(0.08 to 0.13)
42367
(1 study)
⊕⊕⊝⊝
low3,5
Only 1 study
Goitre – (cohort) 156 per 1000 47 per 1000
(36 to 64)
RR 0.3
(0.23 to 0.41)
754387
(11 studies)
⊕⊕⊕⊝
moderate5
Goitre – (multiple cross-sectional) 154 per 1000 28 per 1000
(22 to 34)
RR 0.18
(0.14 to 0.22)
1963247
(34 studies)
⊕⊕⊝⊝
low5,6
*The assumed risk is the risk in the comparison group. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative eect of the intervention (and its 95% CI).
**Calculated as Peto odds ratio or risk ratio.
CI: condence interval; OR: odds ratio; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 Unclear randomization and allocation of concealment.
2 95% CI crosses null.
3 Non-randomized.
4 Though the results of one study suggested an opposite result of the others, there were too few events to generate a meaningful result (5 in intervention and 4 in control).
5 Very large protective eect of iodized salt.
6 Many studies lack sucient detail to understand true risk of bias.

146
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Iodized salt for cretinism
Patient or population: children
Settings: all settings
Intervention: iodized salt
Comparative risks* (95% CI)
Outcomes
Assumed risk
Comparison
Corresponding risk
Iodized salt
Relative eect**
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE) Comments
Cretinism – (RCT) 0 (0)
Cretinism – (non-RCT)
0 (0)
Cretinism – (quasi-experimental) 0 (0)
Cretinism – (cohort) 0 (0)
Cretinism – (multiple cross-sectional) 30 per 1000 4 per 1000
(2 to 6)
OR 0.13
(0.08 to 0.2)
8273
(2 studies)
⊕⊕⊕⊝
moderate1
* The assumed risk is the risk in the comparison group. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative eect of the intervention
(and its 95% CI).
**Calculated as Peto odds ratio.
CI: condence interval; OR: odds ratio; RCT: randomized controlled trial.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 The reduction in risk of cretinism was very large

147
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Iodized salt for adverse eects
Patient or population: general population
Settings: all settings
Intervention: iodized salt
Comparative risks* (95% CI)
Outcomes
Assumed risk
Comparison
Corresponding risk
Iodized salt
Relative eect**
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE) Comments
Positive ATMA – (cohort) 49 per 1000 123 per 1000
(95 to 161)
OR 2.51
(1.93 to 3.27)
2848
(1 study)
⊕⊕⊝⊝
low
Only 1 study
Positive ATMA – (multiple cross-sectional) 31 per 1000 40 per 1000
(29 to 53)
OR 1.27
(0.94 to 1.71)
6172
(2 studies)
⊕⊝⊝⊝
very low1,2
Elevated TgAb – (multiple cross-sectional) 33 per 1000 47 per 1000
(36 to 63)
OR 1.43
(1.08 to 1.89)
6172
(2 studies)
⊕⊝⊝⊝
very low2
Elevated UIE – (multiple cross-sectional) 222 per 1000 300 per 1000
(236 to 387)
RR 1.35
(1.06 to 1.74)
722
(1 study)
⊕⊕⊝⊝
low1,2
Only 1 study
* The assumed risk is the risk in the comparison group. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative eect of the intervention (and its 95% CI).
**Calculated as Peto odds ratio or risk ratio.
ATMA: anti-thyroid microsomal antibody; CI: condence interval; OR: odds ratio; RR: risk ratio; TgAb anti-thyroglobin antibody; UIE: urinary iodine excretion.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 Point estimates fall on both sides of null.
2 Downgraded due to inconsistency, which caused imprecision and therefore not downgraded again for imprecision.

148
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Iodized salt for thyroid hormones
Patient or population: general population
Settings: all settings (low-, middle-, high-income countries)
Intervention: iodized salt
Comparative risks* (95% CI)
Outcomes (study design)
Assumed risk
Comparison
Corresponding risk
Iodized salt
No of participants
(studies)
Quality of the
evidence
(GRADE) Comments
Thyroid-stimulating hormone – (cohort) Mean thyroid-stimulating hormone in
comparison group
11.7
Mean thyroid-stimulating hormone in
intervention group
9.80 lower
(10.85 to 8.75 lower)
60
(1 study)
⊕⊕⊝⊝
low
Only 1 study
Thyroid-stimulating hormone – (multiple
cross-sectional)
Mean thyroid-stimulating hormone in
comparison group
2.5
Mean thyroid-stimulating hormone in
intervention group
0.18 higher
(0.28 lower to 0.64 higher)
5645
(2 studies)
⊕⊝⊝⊝
very low1,2
Thyroglobulin – (multiple cross-sectional) Mean thyroglobulin in comparison group
10.9
Mean thyroglobulin in intervention group
8.73 lower
(9.31 to 8.14 lower)
624
(1 study)
⊕⊕⊝⊝
low
Only 1 study
*Mean value for the comparison group and the group exposed to iodized salt.
CI: condence interval.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 Point estimates fall on both sides of the null.
2 Downgraded for inconsistency, which caused imprecision and therefore not downgraded again for imprecision.

149
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Iodized salt for hyperthyroidism
Patient or population: general population
Settings: all settings
Intervention: iodized salt
Comparative risks* (95% CI)
Outcomes (study design)
Assumed risk
Comparison
Corresponding risk
Iodized salt
Relative eect**
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE) Comments
Hyperthyroidism – (cohort) 2 per 1000 2 per 1000
(2 to 3)
OR 1.36
(1.12 to 1.66.
202848
(2 studies)
⊕⊕⊝⊝1
low
Hyperthyroidism – (multiple cross-sectional 5 per 1000 4 per 1000
(4 to 5)
OR 0.96
(0.92 to 1.00)
1999998
(5 studies)
⊕⊝⊝⊝1,2
very low
*The assumed risk is the risk in the comparison group. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative eect of the intervention (and its 95% CI).
**Relative eect measured using Peto odds ratio.
CI: condence interval; OR: odds ratio.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 Body of evidence not downgraded for any reason.
2 Downgraded due to imprecision.

150
WHO • Eect and safety of salt iodization to prevent iodine deciency disorders: a systematic review with meta-analyses
Iodized salt for hypothyroidism
Patient or population: general population
Settings: all settings (low-, middle-, high-income countries)
Intervention: iodized salt
Comparative risks* (95% CI)
Outcomes (study design)
Assumed risk
Comparison
Corresponding risk
Iodized salt
Relative eect**
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE) Comments
Hypothyroidism (cohort) 1 per 1000 1 per 1000
(1 to 1)
OR 1.14
(0.84 to 1.53)
203425
(4 studies)
⊕⊝⊝⊝
very low1,2
Hypothyroidism (multiple cross-sectional) 43 per 1000 48 per 1000
(40 to 58)
OR 1.13
(0.94 to 1.36)
11375
(4 studies)
⊕⊝⊝⊝
very low1,2
* The assumed risk is the risk in the comparison group. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative eect of the intervention (and its 95% CI).
**Relative eect measured using Peto odds ratio.
CI: condence interval; OR: odds ratio.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our condence in the estimate of eect.
Moderate quality: Further research is likely to have an important impact on our condence in the estimate of eect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our condence in the estimate of eect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1 Point estimates fall on both sides of the null.
2 Downgraded due to inconsistency, which also caused imprecision and therefore not downgraded for imprecision.

For more information, please contact:
Department of Nutrition for Health and Development
World Health Organization
Avenue Appia 20, CH-1211 Geneva 27, Switzerland
Fax: +41 22 791 4156
Email: nutrition@who.int
www.who.int/nutrition
ISBN 978 92 4 150828 5