Articles Effectiveness of a 6-year multidomain vascular care intervention to prevent dementia (preDIVA): a cluster-randomised controlled trial

Abstract

Background:
Cardiovascular risk factors are associated with an increased risk of dementia. We assessed whether a multidomain intervention targeting these factors can prevent dementia in a population of community-dwelling older people.

Methods:
In this open-label, cluster-randomised controlled trial, we recruited individuals aged 70-78 years through participating general practices in the Netherlands. General practices within each health-care centre were randomly assigned (1:1), via a computer-generated randomisation sequence, to either a 6-year nurse-led, multidomain cardiovascular intervention or control (usual care). The primary outcomes were cumulative incidence of dementia and disability score (Academic Medical Center Linear Disability Score [ALDS]) at 6 years of follow-up. The main secondary outcomes were incident cardiovascular disease and mortality. Outcome assessors were masked to group assignment. Analyses included all participants with available outcome data. This trial is registered with ISRCTN, number ISRCTN29711771.

Findings:
Between June 7, 2006, and March 12, 2009, 116 general practices (3526 participants) within 26 health-care centres were recruited and randomly assigned: 63 (1890 participants) were assigned to the intervention group and 53 (1636 participants) to the control group. Primary outcome data were obtained for 3454 (98%) participants; median follow-up was 6·7 years (21 341 person-years). Dementia developed in 121 (7%) of 1853 participants in the intervention group and in 112 (7%) of 1601 participants in the control group (hazard ratio [HR] 0·92, 95% CI 0·71-1·19; p=0·54). Mean ALDS scores measured during follow-up did not differ between groups (85·7 [SD 6·8] in the intervention group and 85·7 [7·1] in the control group; adjusted mean difference -0·02, 95% CI -0·38 to 0·42; p=0·93). 309 (16%) of 1885 participants died in the intervention group, compared with 269 (16%) of 1634 participants in the control group (HR 0·98, 95% CI 0·80-1·18; p=0·81). Incident cardiovascular disease did not differ between groups (273 [19%] of 1469 participants in the intervention group and 228 [17%] of 1307 participants in the control group; HR 1·06, 95% CI 0·86-1·31; p=0·57).

Interpretation:
A nurse-led, multidomain intervention did not result in a reduced incidence of all-cause dementia in an unselected population of older people. This absence of effect might have been caused by modest baseline cardiovascular risks and high standards of usual care. Future studies should assess the efficacy of such interventions in selected populations.

Funding:
Dutch Ministry of Health, Welfare and Sport; Dutch Innovation Fund of Collaborative Health Insurances; and Netherlands Organisation for Health Research and Development.

Full text

Articles
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
1
Effectiveness of a 6-year multidomain vascular care
intervention to prevent dementia (preDIVA):
a cluster-randomised controlled trial
Eric P Moll van Charante*, Edo Richard*, Lisa S Eurelings, Jan-Willem van Dalen, Suzanne A Ligthart, Emma F van Bussel,
Marieke P Hoevenaar-Blom, Marinus Vermeulen, Willem A van Gool
Summary
Background Cardiovascular risk factors are associated with an increased risk of dementia. We assessed whether a
multidomain intervention targeting these factors can prevent dementia in a population of community-dwelling
older people.
Methods In this open-label, cluster-randomised controlled trial, we recruited individuals aged 70–78 years through
participating general practices in the Netherlands. General practices within each health-care centre were randomly
assigned (1:1), via a computer-generated randomisation sequence, to either a 6-year nurse-led, multidomain cardiovascular
intervention or control (usual care). The primary outcomes were cumulative incidence of dementia and disability score
(Academic Medical Center Linear Disability Score [ALDS]) at 6 years of follow-up. The main secondary outcomes were
incident cardiovascular disease and mortality. Outcome assessors were masked to group assignment. Analyses included
all participants with available outcome data. This trial is registered with ISRCTN, number ISRCTN29711771.
Findings Between June 7, 2006, and March 12, 2009, 116 general practices (3526 participants) within 26 health-care
centres were recruited and randomly assigned: 63 (1890 participants) were assigned to the intervention group and
53 (1636 participants) to the control group. Primary outcome data were obtained for 3454 (98%) participants; median
follow-up was 6∙7 years (21 341 person-years). Dementia developed in 121 (7%) of 1853 participants in the intervention
group and in 112 (7%) of 1601 participants in the control group (hazard ratio [HR] 0∙92, 95% CI 0∙71–1∙19; p=0·54).
Mean ALDS scores measured during follow-up did not dier between groups (85·7 [SD 6·8] in the intervention
group and 85·7 [7·1] in the control group; adjusted mean dierence –0∙02, 95% CI –0∙38 to 0∙42; p=0∙93). 309 (16%)
of 1885 participants died in the intervention group, compared with 269 (16%) of 1634 participants in the control group
(HR 0∙98, 95% CI 0∙80–1∙18; p=0∙81). Incident cardiovascular disease did not dier between groups (273 [19%] of
1469 participants in the intervention group and 228 [17%] of 1307 participants in the control group; HR 1∙06, 95% CI
0∙86–1∙31; p=0∙57).
Interpretation A nurse-led, multidomain intervention did not result in a reduced incidence of all-cause dementia in
an unselected population of older people. This absence of eect might have been caused by modest baseline
cardiovascular risks and high standards of usual care. Future studies should assess the ecacy of such interventions
in selected populations.
Funding Dutch Ministry of Health, Welfare and Sport; Dutch Innovation Fund of Collaborative Health Insurances;
and Netherlands Organisation for Health Research and Development.
Introduction
Dementia aects more than 36 million individuals
worldwide and its prevalence is expected to increase
substantially during the next few decades.1 WHO and the G8
acknowledge the major societal challenge this increase will
cause and call for strategies aiming to prevent dementia.2
Observational studies have repeatedly shown an
association of vascular and lifestyle-related risk factors with
incident dementia in people older than 65 years (>90% of all
patients with dementia).3 Furthermore, population-based
autopsy studies suggest that in addition to Alzheimer’s
disease pathology, vascular pathologies underlie a
substantial proportion of dementias.4 An estimated 30% of
cases of Alzheimer’s disease might be attributable to
potentially modifiable, mostly vascular risk factors.5 This
finding suggests the potential to prevent dementia if these
risk factors are addressed.6 Randomised controlled trials
targeting vascular and lifestyle-related risk factors with
cognitive decline or dementia as a (secondary) outcome
have mostly addressed single risk factors, including
hypertension, physical inactivity, unhealthy diet, and
smoking.7 Findings were mixed and meta-analyses
regarding antihypertensive treatment to prevent dementia
have reached divergent conclusions.8,9 This discrepancy
might be explained by dierences between study
populations, short follow-up periods (≤2 years), insucient
sample sizes, and attrition bias.10 Fears of side-eects from
antihypertensive treatment, including counterproductive
eects on cognition, further complicate treatment decisions
on blood pressure lowering in older people.11
Published Online
July 26, 2016
http://dx.doi.org/10.1016/
S0140-6736(16)30950-3
*Contributed equally
Department of General Practice
(E P Moll van Charante MD,
S A Ligthart MD,
E F van Bussel MD), and
Department of Neurology
(E Richard MD, L S Eurelings MD,
J-W van Dalen MSc,
M P Hoevenaar-Blom PhD,
Prof M Vermeulen MD,
Prof W A van Gool MD),
Academic Medical Centre,
Amsterdam, Netherlands; and
Department of Neurology,
Radboud University Medical
Centre, Nijmegen, Netherlands
(E Richard)
Correspondence to:
Prof Willem A van Gool,
Department of Neurology,
Academic Medical Centre/
University of Amsterdam,
Meibergdreef 9, 1100 DD
Amsterdam, Netherlands
w.a.vangool@amc.uva.nl

Articles
2
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
We assessed the eects of a 6-year nurse-led multi-
domain intervention targeting vascular and lifestyle-
related risk factors on the prevention of dementia in
people aged 70–78 years from the general population.
Methods
Study design and participants
The Prevention of Dementia by Intensive Vascular care
(preDIVA) trial was a pragmatic, multisite, cluster-
randomised, open-label trial done in 116 general practices
(family practices) in 26 health-care centre buildings in the
Netherlands. The study protocol has been published
previously.12 We used a population-based approach, inviting
all community-dwelling older people (aged 70–78 years)
registered with a participating general practice (>98% of
the Dutch population is registered) to participate in the
trial. The only exclusion criteria were dementia and other
disorders likely to hinder successful long-term follow-up
according to the general practitioner (family doctor), such
as terminal illness and alcoholism. Recruitment was from
June 7, 2006, until March 12, 2009 (see appendix for full
details of the enrolment procedure). The study was
approved by the medical ethics committee of the Academic
Medical Center, Amsterdam, Netherlands. Participants
gave written informed consent before their baseline visit.
Randomisation and masking
After completion of all baseline visits at a health-care
centre (to avoid non-random dierences in consent rates
between general practices in the same building), cluster
randomisation took place with general practice as the
unit of randomisation, to minimise contamination at the
level of general practice. A centralised computer
algorithm was used by the Clinical Research Unit, not
involved in the study in any other way, with health-care
centres as blocks, and general practices as clusters, in
equal proportions for both conditions, allowing a
maximum dierence of 250 participants between groups,
to accommodate dierences in cluster size and number
of clusters per health-care centre (median 4 [IQR 3–6]).
Within each health-care centre at least one general
practice was randomised to the intervention group. All
outcome assessors were masked to group allocation and
were not involved in intervention activities. The final
clinical assessment was done by an independent
investigator who was masked to group allocation.
Procedures
The intervention consisted of visits to a practice nurse in
the general practice every 4 months, for a period of
6 years (18 visits). During these visits, the nurse assessed
the following cardiovascular risk factors: smoking habits,
diet, physical activity, weight, and blood pressure. Blood
glucose and lipid concentrations were assessed every
2 years and when indicated otherwise. On the basis of
these assessments, individually tailored lifestyle advice
was given according to a detailed protocol conforming
with prevailing Dutch general practitioner guidelines on
cardiovascular risk management13 and supported by
motivational interviewing techniques. If indicated, drug
treatment for hypertension, dyslipidaemia, and type 2
diabetes mellitus was initiated or optimised and
Research in context
Evidence before this study
We searched PubMed, the ISRCTN Registry, the ClinicalTrials.gov
database, and the WHO International Clinical Trials Registry
Platform for articles published between the date of database
inception and Feb 19, 2016, to identify randomised controlled
trials of multidomain interventions. Search terms were
“prevention” and “dementia”, “cognitive impairment”,
or “Alzheimer’s disease”. We included trials of lifestyle
interventions in combination with drug treatment in individuals
aged 60 years or older with a duration of at least 2 years, and
dementia as a primary outcome. These criteria were based on the
2010 National Institutes of Health evidence report on preventing
Alzheimer’s disease and cognitive decline. We identified two
randomised controlled trials. The Multidomain Alzheimer
Preventive Trial (MAPT; NCT00672685) has been completed but
not yet published. The Finnish Geriatric Intervention Study to
Prevent Cognitive Impairment and Disability (FINGER), which
assessed a multidomain intervention consisting of lifestyle
interventions, drug treatment, and cognitive training in adults
aged 60–77 years, showed a small excess in improvement in a
composite score of tests for cognitive functioning after 2 years in
the intervention group compared with control.
Added value of this study
To our knowledge, preDIVA is the first large, long-term trial to
assess the effectiveness of a multidomain cardiovascular
intervention in an unselected population of older people with
all-cause dementia as a primary outcome. Although overall
findings were neutral, subgroup analyses suggested potential
beneficial effects on non-Alzheimer’s disease and on all-cause
dementia in participants who were adherent to the
intervention, especially those with untreated hypertension
at baseline.
Implications of all the available evidence
The strong association between cardiovascular risk and
all-cause dementia suggests an opportunity for dementia
prevention. Our study was done in a public health context, in
which small, sustained changes can have substantial long-term
effects in an unselected population of older people. In
health-care systems with high standards of usual care such as in
the Netherlands, the potential for preventing dementia by
improving cardiovascular risk factor management might have
been too small to see significant beneficial effects.
See Online for appendix

Articles
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
3
antithrombotic drugs were started. Medication adherence
was improved where appropriate. Five educational
sessions for all nurses were organised during the course
of the study to strengthen the consistency of the
intervention. Participants in the control group received
usual care, according to the prevailing standards for
cardiovascular risk management.
Baseline data for demographic characteristics,
cardiovascular and family history, medication use, and
self-reported diet and smoking habits were collected and
cross-checked with the participants’ electronic health
records. Physical activity was assessed with the LASA
Physical Activity Questionnaire (LAPAQ), disability with
the Academic Medical Center Linear Disability Score
(ALDS), cognitive function with the Mini-Mental State
Examination (MMSE) and Visual Association Test (VAT),
and depressive symptoms with the 15-item Geriatric
Depression Scale (GDS-15; appendix).
Anthropometrics and blood pressure were measured
by a standardised protocol and blood samples were
obtained for lipid spectrum and blood glucose
measurements. Genomic DNA was stored and used for
apolipoprotein-E (APOE) genotyping.
All measurements were repeated during 2-yearly
follow-up assessments. To allow participants recruited
early into the trial to continue follow-up until all
assessments were completed, the study was extended up
to 8 years for participants randomised in 2006–07.
The diagnosis of dementia was made according to the
Diagnostic and Statistical Manual of Mental Disorders,
fourth edition (DSM-IV)14 and classified into Alzheimer’s
disease, vascular dementia, dementia with Lewy bodies,
and other dementia types according to current guidelines
(appendix).
We also measured adherence to the intervention,
defined as the proportion of visits attended out of all
planned visits.
Outcomes
The primary outcomes were cumulative incidence of
dementia and disability score (ALDS) at 6 years of follow-
up. Disability was chosen because any eect of our
intervention on either cardiovascular disease or dementia
would ultimately translate into disability. The main
secondary outcomes were incident cardiovascular disease
(myocardial infarction, stroke, and peripheral arterial
disease) and cardiovascular and all-cause mortality. Other
secondary outcomes were cognitive decline as measured
by MMSE and VAT, symptoms of depression as measured
by GDS-15, blood pressure, body-mass index (BMI),
blood lipid concentrations, and glucose concentration.
Dementia subtype was not prespecified in the original
protocol but was added as an endpoint before the analysis
of data had begun.
Outcome data were collected during follow-up visits,
supplemented by information from general practitioners’
electronic health records and the National Death Registry.
An independent outcome adjudication committee
consisting of neurologists, old age psychiatrists,
geriatricians, cardiologists, and general practitioners
analysed all clinical outcomes masked to group allocation.
As a quality check and to minimise the risk of false-
positive diagnoses, dementia diagnoses were re-evaluated
after 1 year (appendix).
Serious adverse events were defined as events that
were fatal or life-threatening, or resulted in significant or
persistent disability, and needed admission to hospital.
Events were included if the condition was stated as the
reason for admission or if the diagnosis was listed in the
hospital discharge letter to the general practitioner.
Statistical analysis
We based our sample size calculation on the age-specific
cumulative incidence of dementia, as available in 2004.15
Enrolment of 3700 participants would provide 80%
power to detect a 33% between-group dierence in the
cumulative incidence of dementia, with a two-sided
α level of 0∙05, and compensating for an estimated 33%
dropout rate and unknown intracluster coecient.16 The
33% between-group dierence was deemed realistic on
the basis of published data.17
A planned interim analysis by an independent
committee after the 4-year follow-up assessments on
dementia, disability, and mortality, resulted in the
recommendation to continue the trial with no change to
the protocol (appendix). The final analyses were completed
by the study group and verified by an independent
biostatistician. Analyses included all participants with
available outcome data. Person-years were calculated from
the date of randomisation to the date of dementia
diagnosis, death, or the last visit. For binary time-to-event
outcomes, a random eects Cox proportional hazards
model was used, accounting for clustering of participants
within practices and health-care centres. A similar random
eects linear multiple measurements model was used for
continuous measures, including disability, blood pressure,
BMI, laboratory values, cognition, and depressive
symptoms. Each continuous factor was adjusted for
baseline imbalance and treatment-by-time interaction.
Details of the analyses are provided in the appendix.
Sensitivity analyses for the primary outcome included a
per-protocol analysis (see appendix for details), best-case
and worst-case scenario, an analysis including all cases
of possible dementia, and models adjusting for additional
variables including cardiovascular risk factors. The eect
of values not missing at random on repeated
measurements outcomes was assessed in a sensitivity
analysis using a joint model. Subgroup analyses were
done for sex, age (split at the median), hypertension
severity (according to WHO grades), cardiovascular
history, APOE genotype (any vs no ε4 allele), and
antihypertensive treatment at baseline (appendix).
All analyses were done with SPSS version 22 and
R version 3.2 (see appendix for a more detailed description

Articles
4
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
of analyses and R-codes employed). This trial is registered
with ISRCTN, number ISRCTN29711771.
Role of the funding source
The funders of the study had no role in study design,
data collection, data analysis, data interpretation, or
writing of the report. All authors had full access to all the
data in the study and the corresponding author had final
responsibility for the decision to submit for publication.
Results
In 26 health-care centres representing 116 general
practices, 7772 people were potentially eligible (figure 1).
After 1010 (13%) people were excluded, 6762 were invited
to participate in the trial by letter. Of these, 3526 (52%)
provided informed consent. Mean cluster size was 30
(SD 21). 1890 participants from 63 practices were
randomly assigned to the intervention group and
1636 participants from 53 practices to the control group.
Figure 1: Trial profile
*Number comprises participants assessed at ≥6 years, those who discontinued but who were retrieved for primary outcome, those who were excluded from further
assessments because they developed dementia during follow-up, and those who died during follow-up; the total of these numbers presented in the figure exceeds
that given here because of deaths occurring after the final assessment or after the diagnosis of dementia. †Number comprises participants assessed at ≥6 years, those
who discontinued but who were retrieved for survival analysis, those who were excluded from further assessments because they developed dementia during follow-
up, and those who died during follow-up; the total of these numbers presented in the figure exceeds that given here because of deaths occurring after the final
assessment or after the diagnosis of dementia.
7772 individuals in source population
(116 general practices in
26 health-care centres)
6762 eligible individuals invited to
participate
3600 individuals assessed for eligibility
3526 participants (in 116 general practices)
randomly assigned
1010 individuals excluded by general practitioner
316 not eligible
107 probable dementia
92 limited life expectancy
158 did not speak Dutch
337 unknown
3162 (47%) no informed consent
74 (2%) excluded
38 withdrew
21 probable dementia
7 deceased before randomisation
3 complication
2 language problem
2 unknown
1 dementia
1636 participants (in 53 general practices)
assigned to control
1312 assessed at 2-year follow-up
1053 assessed at 4-year follow-up
892 assessed at ≥6-year follow-up
69 dementia
178 deceased
607 discontinued
474 withdrew
133 relocated
572 retrieved for
primary
outcome
43 dementia
605 retrieved for
survival
analysis
91 deceased
1601 included in analysis of primary
dementia outcome*
112 dementia
1634 in analysis of survival
269 deceased†
1890 participants (in 63 general practices)
assigned to intervention
1477 assessed at 2-year follow-up
1217 assessed at 4-year follow-up
1035 assessed at ≥6-year follow-up
79 dementia
179 deceased
748 discontinued
604 withdrew
144 relocated
711 retrieved for
primary
outcome
42 dementia
743 retrieved for
survival
analysis
130 deceased
1853 included in analysis of primary
dementia outcome*
121 dementia
1885 in analysis of survival
309 deceased†

Articles
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
5
The intervention and control groups were well balanced
at baseline, apart from a 2 mm Hg dierence in systolic
blood pressure (table 1).
After a median follow-up of 2442 days (6∙7 years),
complete follow-up data for the primary outcome of
dementia were obtained for 3454 (98∙0%) participants,
yielding 21 341 person-years. Information about survival
was available for 3519 (99∙8%) participants.
Dementia developed in 121 (7%) of 1853 participants in
the intervention group and in 112 (7%) of 1601 participants
in the control group (hazard ratio [HR] 0∙92, 95% CI
0∙71–1∙19; p=0∙54; table 2; figure 2). No participants
diagnosed with dementia reverted to normal cognition
during the 1-year follow-up after diagnosis. Occurrence of
Alzheimer’s disease did not dier between groups
(table 2). Dementia other than Alzheimer’s disease
occurred less frequently in the intervention group than in
the control group (11 [1%] of 1743 participants vs 23 [2%] of
1512 participants, respectively; HR 0∙37, 95% CI
0∙18–0∙76; p=0∙007; appendix). Vascular dementia
occurred in seven (<1%) of 1739 participants in the
intervention group compared with 12 (1%) of
1501 participants in the control group (HR 0∙43, 95% CI
0∙17–1∙12; p=0∙09). Sensitivity analyses yielded similar
results (appendix).
In the per-protocol analysis, dementia occurred in
85 (6%) of 1403 participants in the intervention group
and in 107 (7%) of 1479 participants in the control group
(HR 0∙78, 95% CI 0∙58–1∙04; p=0∙09; appendix). In
participants with untreated hypertension at baseline,
dementia occurred in 31 (5%) of 646 in the intervention
group and in 36 (7%) of 522 in the control group
(HR 0∙69, 95% CI 0∙43–1∙11; p=0∙13; appendix). In
Intervention
(n=1890)
Control (n=1636)
Demographics
Age (years) 74·5 (2·5) 74·5 (2·5)
Sex
Male 850 (45%) 757 (46%)
Female 1040 (55%) 879 (54%)
Educational level
<7 years 455 (24%) 381 (23%)
7–12 years 1168 (62%) 1014 (62%)
>12 years 255 (13%) 218 (13%)
Missing data 12 (1%) 23 (1%)
Race
White 1817 (96%) 1578 (96%)
Other 40 (2%) 32 (2%)
Missing data 33 (2%) 26 (2%)
Medical history
Cardiovascular disease (excluding
stroke and TIA)
568 (30%) 476 (29%)
Missing data 4 (<1%) 2 (<1%)
Stroke or TIA 175 (9%) 172 (11%)
Missing data 7 (<1%) 5 (<1%)
Cardiovascular risk factors
Systolic blood pressure (mm Hg) 156·3 (22·0) 154·2 (20·5)
Missing data 1 (<1%) 2 (<1%)
Diastolic blood pressure (mm Hg) 81·4 (11·2) 81·5 (10·8)
Missing data 02 (<1%)
Total cholesterol (mmol/L) 5·2 (1·1) 5·3 (1·1)
Missing data 45 (2%) 28 (2%)
LDL cholesterol (mmol/L) 3·1 (1·0) 3·2 (1·0)
Missing data 62 (3%) 36 (2%)
Body-mass index (kg/m²) 27·6 (4·2) 27·3 (4·1)
Missing data 2 (<1%) 0
Waist circumference, female (cm) 102·3 (9·9) 101·6 (9·9)
Missing data 3 (<1%) 1 (<1%)
Waist circumference, male (cm) 97·5 (12·4) 97·4 (12·0)
Missing data 3 (<1%) 5 (<1%)
Type 2 diabetes 357 (19%) 289 (18%)
Missing data 0 0
Blood glucose (mmol/L) 5·8 (1·2) 5·9 (1·2)
Missing data 45 (2%) 28 (2%)
Current smoking 252 (13%) 216 (13%)
Missing data 4 (<1%) 3 (<1%)
Physical activity (WHO)17 1594 (86%) 1398 (85%)
Missing data 38 (2%) 33 (2%)
Genetic factors
APOE ε4, negative 1155 (61%) 996 (61%)
APOE ε4, heterozygous 412 (22%) 332 (20%)
APOE ε4, homozygous 36 (2%) 35 (2%)
Missing data 287 (15%) 273 (17%)
Medication use
Antihypertensive drug(s) 1028 (55%) 923 (57%)
Missing data 3 (<1%) 2 (<1%)
(Table 1 continues in next column)
Intervention
(n=1890)
Control (n=1636)
(Continued from previous column)
Cholesterol-lowering drug(s) 650 (34%) 550 (34%)
Missing data 3 (<1%) 3 (<1%)
Antiplatelet or anticoagulant
drug(s)
616 (33%) 550 (34%)
Missing data 3 (<1%) 2 (<1%)
Disability and neuropsychiatric assessment
ALDS 89 (86–89) 89 (86–89)
Missing data 5 (<1%) 9 (1%)
MMSE 28 (27–29) 28 (27–29)
Missing data 4 (<1%) 2 (<1%)
GDS-15 1 (0–2) 1 (0–2)
Missing data 3 (<1%) 2 (<1%)
VAT A 6 (5–6) 6 (5–6)
Missing data 9 (<1%) 9 (1%)
Data are mean (SD), n (%), or median (IQR). TIA=transient ischaemic attack.
ALDS=Academic Medical Center Linear Disability Score. MMSE=Mini-Mental State
Examination. GDS-15=15-item Geriatric Depression Scale. VAT A=Visual
Association Test A.
Table 1: Baseline characteristics

Articles
6
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
participants with untreated hypertension who were
adherent to the intervention, dementia occurred in
22 (4%) of 512 in the intervention group compared with
35 (7%) of 471 in the control group (HR 0∙54, 95% CI
0∙32–0∙92; p=0∙02; appendix). In participants with a
history free from cardiovascular disease who were
adherent to the intervention, dementia occurred in 45
(5%) of participants in the intervention group and in 69
(7%) of 934 participants in the control group (HR 0·64,
95% CI 0·44–0·94; p=0·02; appendix).
Disability did not dier between groups, with ALDS
scores at 6 years of 85∙7 in both study groups (adjusted
mean dierence –0∙02, 95% CI –0∙38 to 0∙42; p=0∙93;
table 3). 309 (16%) of 1885 participants died in the
intervention group, compared with 269 (16%) of 1634
participants in the control group (HR 0∙98, 95% CI
0∙80–1∙18; p=0∙81; table 2). Cardiovascular disease events
occurred in 273 (19%) of 1469 participants in the
intervention group and 228 (17%) of 1307 participants in
the control group (HR 1∙06, 95% CI 0∙86–1∙31; p=0∙57;
table 2).
Cognition and the number of depressive symptoms did
not dier between both groups (table 3). Systolic blood
pressure decreased more in the intervention group than
in the control group (adjusted mean dierence
–2∙06 mm Hg, 95% CI –3∙21 to –0∙90; p=0∙0005). BMI,
total cholesterol, and LDL cholesterol decreased in both
groups, but dierences were not significant (appendix).
Antihypertensive medication was started in 329 (58%)
of 568 participants in the intervention group and 231
(48%) of 479 in the control group, irrespective of baseline
blood pressure (odds ratio [OR] 1∙48, 95% CI 1∙16–1∙89;
p=0∙002). In participants with hypertension who were
not using antihypertensive drugs at baseline, these
proportions were 295 (67%) of 439 in the intervention
group compared with 203 (56%) of 364 in the control
group (OR 1∙62, 95% CI 1∙22–2∙17; p=0∙001). These
medication changes were associated with substantial
eects on blood pressure in both the intervention and
control groups (appendix). Changes in other medication
and lifestyle variables are given in the appendix.
There was no excess mortality in either group.
The median number of hospital admissions per
1000 participants per year was 117 in the intervention
group and 108 in the control group (adjusted mean
dierence –3, 95% CI –24 to 18; p=0∙78). There were no
significant dierences between groups in frequencies of
serious adverse events for hypotension, syncope,
electrolyte abnormalities, injurious falls, or acute kidney
injury or failure (appendix).
Discussion
In this randomised controlled trial, we recorded no eect
of 6∙7 years of nurse-led intensive vascular care on
incident all-cause dementia. We also observed no eect
on mortality, cardiovascular disease, or disability, despite
a greater improvement in systolic blood pressure in the
intervention group compared with the control group.
There are several possible explanations for this finding.
First, the contrast between study groups in cardiovascular
risk reduction was relatively small. As a result of the
pragmatic nature of the study, the intensity of the
vascular care that was delivered might have been
insucient to induce relevant eects on lifestyle change.
Moreover, in primary care settings already providing
high standards of cardiovascular risk management, it
might be dicult to improve overall ecacy, especially
for secondary cardiovascular prevention. This theory is
supported by subgroup analyses showing the strongest
eects of the intervention in participants with
Figure 2: Kaplan-Meier plot of cumulative incidence of dementia
To allow participants recruited early into the trial to continue follow-up until the
6-year assessment of the last participant was completed, the study was extended
for participants randomised early (ie, in 2006–07). The hazard ratio (HR) refers to
an analysis including all participants, up to 8 years of follow-up. The period
beyond the planned 6-year follow-up, concerning few participants, is shaded.
Number at risk
Control group
Intervention group
0
1601
1853
2
1540
1785
4
1437
1674
6
1130
1290
8
28
32
Follow-up (years)
HR 0·92 (95% CI 0·71–1·19); p=0·54
0
5
10
15
20
Cumulative incidence of dementia (%)
Control group
Intervention group
Intervention Control Hazard ratio (95% CI) p value
All-cause dementia 121/1853 (7%) 112/1601 (7%) 0·92 (0·71–1·19) 0·54
Alzheimer’s disease* 99/1831 (5%) 81/1570 (5%) 1·05 (0·78–1·41) 0·74
Non-Alzheimer’s dementia*† 11/1743 (1%) 23/1512 (2%) 0·37 (0·18–0·76) 0·007
Unspecified types of dementia* 11/1743 (1%) 8/1497 (1%) 1·24 (0·46–3·41) 0·67
Cardiovascular events‡ 273/1469 (19%) 228/1307 (17%) 1·06 (0·86–1·31) 0·57
Myocardial infarction 68/1503 (5%) 57/1339 (4%) 1·03 (0·71–1·49) 0·87
Stroke including TIA 120/1503 (8%) 102/1341 (8%) 1·05 (0·80–1·38) 0·74
Other§ 103/1495 (7%) 83/1333 (6%) 1·08 (0·78–1·50) 0·65
Death 309/1885 (16%) 269/1634 (16%) 0·98 (0·80–1·18) 0·81
Cardiovascular death¶ 63/1639 (4%) 60/1425 (4%) 0·91 (0·63–1·32) 0·63
Other¶ 126/1702 (7%) 125/1490 (8%) 0·87 (0·68–1·12) 0·28
Data are n (%), unless otherwise indicated. Further details are given in the appendix. TIA=transient ischaemic attack.
*Participants with a dementia subtype other than the one being analysed were left out of the denominator.
†Non-Alzheimer’s dementia includes vascular dementia (seven in the intervention group vs 12 in the control group);
Lewy body dementia (two vs six); Parkinson’s dementia (two vs two); frontotemporal dementia (none vs one); primary
progressive aphasia (none vs one); and other (none vs one). ‡Including fatal and non-fatal myocardial infarction and
stroke, and angina pectoris, TIA, and peripheral arterial disease. §Angina pectoris and peripheral arterial disease.
¶The cause of death was unknown for 120 participants in the intervention group and 84 in the control group;
therefore, numbers of cardiovascular and other causes of death do not add up to the total number of deaths.
Table 2: Clinical outcomes

Articles
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
7
hypertension who were not on antihypertensive
medication at baseline and for participants with no
history of cardiovascular disease. Additionally, a
substantial Hawthorne eect could have occurred, with
the 2-yearly screenings prompting interventions in high-
risk cases in both the intervention and control groups.
The decrease in blood pressure in the control group,
particularly during the first 2 years (appendix), could
reflect this eect. This finding might have been further
enhanced by the 2011 update of the cardiovascular risk
management guideline, which recommended a more
proactive primary prevention in people older than
70 years.18 Second, in view of the pragmatic, public health
approach of the trial, we did not specifically select a
population with increased cardiovascular risk, potentially
limiting the overall eect of the intervention. Third, our
population was aged 70–78 years, whereas most
observational data show an association between midlife
(age 40–60 years) vascular risk factors and dementia.
Although blood pressure reduction in patients aged
80 years or older was associated with a trend for reduced
incidence of dementia in the HYVET trial,19 an inverse
association of blood pressure with dementia and survival
has been suggested in older age groups.20,21 Our results
mitigate fears that antihypertensive treatment in older
age groups evokes cognitive decline. Moreover, they
show that such an intervention is safe, which is in
accordance with findings from the recent SPRINT trial.22
Major strengths of our study are the long intervention
period, the blinded adjudication of outcomes including a
1-year follow-up after the diagnosis of dementia, and
completeness of follow-up on all-cause dementia (98∙0%)
and mortality (99∙8%). The pragmatic design and
population-based sample result in a high external validity
of our findings, further strengthened by the fact that our
population is comparable to a population from national
(cohort) data.23
At the time of the preDIVA study design, the available
data from the Syst-Eur study suggested a 55% risk
reduction of dementia through modest blood pressure
reduction.17 This finding led to our estimated—and at that
time seemingly conservative—relative risk reduction of
33% for the multidomain intervention in our power
analysis. Recently, it was estimated that up to 30% of all
cases of dementia might be attributable to seven modifiable
risk factors, with a population attributable risk of 6∙8% for
hypertension only, in European populations.5 This finding
is in line with meta-analyses of antihypertensive treatment
eects reporting all-cause dementia risk reductions of only
2–9%.8,9 The HR of 0∙92 reported in our study is consistent
with these findings, although our study was underpowered
to detect such an eect size. On the basis of the fully
adjusted per-protocol analysis (appendix), a 24% lower
dementia hazard in participants who were adherent to the
intervention would translate into an absolute risk reduction
of 1∙7% (from 7∙2% to 5∙5%).
Intervention Control Adjusted mean
difference (95% CI)‡
p value Time-by-treatment
interaction (years,
95% CI)
p value
n* Mean (SD)† n* Mean (SD)†
ALDS score 1484 85·7 (6·8) 1326 85·7 (7·1) 0·02 (–0·38 to 0·42) 0·93 –0·02 (–0·17 to 0·12) 0·74
Systolic blood pressure (mm Hg) 1494 148·0 (19·4) 1334 149·6 (20·7) –2·06 (–3·21 to –0·90) 0·0005 0·69 (0·29 to 1·08) 0·0006
WHO normotension (mm Hg) 344 136·9 (17·9) 307 135·9 (18·2) 0·74 (–1·34 to 2·81) 0·49 0·80 (0·06 to 1·54) 0·03
WHO hypertension (mm Hg)§ 1150 151·3 (18·6) 1027 153·7 (19·7) –2·93 (–4·29 to –1·57) <0·0001 0·65 (0·19 to 1·10) 0·006
Diastolic blood pressure (mm Hg) 1495 77·4 (10·5) 1334 78·8 (10·9) –1·15 (–1·84 to –0·46) 0·001 0·57 (0·37 to 0·78) <0·0001
WHO normotension (mm Hg) 344 74·7 (10·0) 307 75·0 (10·3) 0·16 (–1·15 to 1·47) 0·81 0·60 (0·22 to 0·99) 0·002
WHO hypertension (mm Hg)§ 1151 78·2 (10·5) 1027 79·9 (10·9) –1·71 (–2·41 to –1·02) <0·0001 0·55 (0·31 to 0·78) <0·0001
Waist circumference, female (cm) 818 96·7 (12·4) 716 96·7 (12·3) –0·20 (–1·02 to 0·62) 0·63 0·08 (–0·12 to 0·28) 0·43
Waist circumference, male (cm) 665 102·2 (10·2) 604 101·8 (10·1) –0·20 (–0·76 to 0·37) 0·50 0·18 (0·00 to 0·35) 0·0049
Body-mass index (kg/m²) 1492 27·4 (4·8) 1334 27·1 (4·7) 0·06 (–0·10 to 0·23) 0·45 –0·03 (–0·10 to 0·03) 0·32
Total cholesterol (mmol/L) 1310 5·0 (1·1) 1172 5·1 (1·1) –0·02 (–0·09 to 0·04) 0·49 0·02 (0·00 to 0·04) 0·038
LDL cholesterol (mmol/L) 1309 2·8 (1·0) 1167 3·0 (1·0) –0·03 (–0·09 to 0·03) 0·30 0·01 (0·00 to 0·03) 0·10
Glucose (mmol/L) 1307 6·1 (1·6) 1168 6·1 (1·6 0·02 (–0·06 to 0·10) 0·56 0·00 (–0·03 to 0·03) 0·97
VAT A 1484 5·3 (1·1) 1325 5·3 (1·1) –0·02 (–0·09 to 0·04) 0·48 0·01 (–0·02 to 0·03) 0·52
MMSE score 1494 28·2 (2·1) 1330 28·3 (2·0) –0·02 (–0·14 to 0·10) 0·73 0·01 (–0·03 to 0·05) 0·59
GDS-15 score 1490 1·8 (2·2) 1333 1·7 (2·2) 0·01 (–0·09 to 0·12) 0·79 0·00 (–0·04 to 0·04) 0·93
Number of observations in analysis (intervention vs control): ALDS (Academic Medical Center Linear Disability Score): 4184 vs 3701; systolic blood pressure: 4198 vs 3633;
diastolic blood pressure: 4202 vs 3633; body-mass index: 4199 vs 3645; waist circumference, female: 2262 vs 1962; waist circumference, male: 1794 vs 1579; total
cholesterol: 3079 vs 2634; LDL cholesterol: 3064 vs 2610; glucose: 3032 vs 2586; VAT A (Visual Association Test A): 4129 vs 3586; MMSE (Mini-Mental State Examination):
4323 vs 3724; GDS-15 (15-item Geriatric Depression Scale): 4177 vs 3640. *Number of participants available for analysis. †Mean (SD) of repeated measurements after
baseline. ‡Adjusted for baseline and clustering within centres and individuals, taking all measurements at all timepoints into account. §WHO hypertension: systolic blood
pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg.
Table 3: Continuous outcomes

Articles
8
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
No other trials of multidomain interventions for
prevention of dementia of similar size and follow-up
duration have been reported, impeding direct comparison
of our trial with previous research. In the LIFE study, a
2-year moderate-intensity physical activity intervention
did not improve cognition or reduce incidence of dementia
in sedentary adults aged 70–89 years.24 By contrast, in the
FINGER trial, participants aged 60–77 years who received
a multidomain intervention for 2 years showed a small
excess improvement on a composite score based on tests
for cognitive functioning.25 The clinical relevance of this
eect is uncertain; whether this finding will translate into
the prevention of cognitive decline or dementia over time
is to be explored after a planned extended follow-up.
The suggested eect of the intervention on the
subgroup of participants with non-Alzheimer’s disease
dementia, most of whom had vascular dementia, should
be interpreted with caution, because of the small number
of cases. The nature of our intervention renders a
preventive eect on cerebrovascular damage more
plausible than an eect on the occurrence or progression
of Alzheimer’s disease. The mechanisms through which
vascular risk factors contribute to the development of
Alzheimer’s disease are unknown. Interaction between
small vessel disease and neurodegenerative changes, in
particular at the neurovascular unit, might partly explain
this association. Nevertheless, we did not find an eect of
vascular risk management on the development of clinical
symptoms of Alzheimer’s disease, as a result of
insucient contrast between study groups, or perhaps
by lack of causal interaction with the neurodegenerative
changes that underlie Alzheimer’s disease.
Our study has several limitations. First, we did not
perform detailed neuropsychological testing. In theory,
this drawback could have led to a type II error (ie, missing
a small treatment eect). However, rather than exploring
eects on surrogate endpoints, we chose a clinical
diagnosis of dementia as the outcome to draw conclusions
on prevention of dementia with unequivocal clinical
relevance. Additionally, the long period of follow-up further
ensured reliable detection of dementia and avoidance of
false-positive diagnoses of dementia. Second, not all
eligible individuals in the participating practices consented
to participation, potentially introducing recruitment bias,
although dierences in sex and age between participants
and non-participants seemed small and such selections
are inherent to preventive initiatives. Third, our
intervention was of modest intensity and resulted in
limited contrast. Our per-protocol analysis suggests better
response in participants who were adherent to the
intervention, and therefore we cannot exclude the
possibility that a more intensive intervention would have
yielded a larger eect, although such an intervention
might also be associated with more side-eects.
On the basis of our findings, future trials of multidomain
interventions could benefit from tighter controlled
intervention delivery and selection of individuals without
appropriate hypertension treatment. Sample-size cal-
culations for future studies will have to account for levels
of usual vascular care in the target population and the
recently reported declining age-adjusted risk of dementia
in some developed countries.26 Intervention at earlier ages
(eg, <60 years) will require a longer follow-up because of
the low incidence of dementia in midlife, for which a
classic randomised controlled trial design might fall
short.27 Multidomain interventions to prevent dementia
might have a greater eect in low-income and middle-
income countries in view of the projected increase in
hypertension, incident cardiovascular disease, and
dementia and lower levels of cardiovascular risk
management in these settings.28 Because the projected
worldwide increase in the prevalence of dementia during
the next decades will also be largely attributable to an
increased prevalence in low-income and middle-income
countries, new interventions must be low cost, safe, and
easy to implement in a wide range of settings.7
In conclusion, our study shows that long-term nurse-
led vascular care in an unselected population of
community-dwelling older people is safe but does not
result in a reduction in incidence of all-cause dementia,
disability, or mortality. However, our results do not rule
out clinically meaningful eects in people with untreated
hypertension who are adherent to the intervention.
Contributors
MV, ER, EPMvC, and WAvG conceived and designed the trial. WAvG
coordinated the trial. WAvG, EPMvC, and SAL designed and supervised
the intervention components (cardiovascular risk management, based
on national guidelines for primary care). LSE, SAL, and J-WvD
coordinated database management and outcome adjudication. J-WvD
and MPH-B performed the data analysis. All authors interpreted the
results; EPMvC, ER, and WAvG drafted the report. MV, WAvG, ER, and
EPMvC obtained funding. All authors revised the Article for important
intellectual content. WAvG is the principal investigator.
Declaration of interests
We declare no competing interests.
Acknowledgments
We sincerely thank all participants of the preDIVA study. The preDIVA
trial was supported by the Dutch Ministry of Health, Welfare and Sport
(grant number 50-50110-98-020), the Dutch Innovation Fund of
Collaborative Health Insurances (grant number 05-234), and the
Netherlands Organisation for Health Research and Development (grant
number 62000015). We are indebted to all practice nurses who delivered
the intervention and all general practitioners involved in the care for the
participants, including the Zorggroep Almere. We particularly thank our
project manager, Carin E Miedema for her outstanding role in
coordinating the trial. We acknowledge the eorts of the interim
committee members (Anton de Craen [deceased], Niek de Wit, and
Jan Stam). We thank the members of the independent outcome
adjudication committee. We thank Ronald Geskus for his statistical
advice and Wim Busschers for his independent, critical revision of the
statistical methods and analyses. We thank Rupert McShane for his
critical revision of the manuscript.
References
1 Prince M, Bryce R, Albanese E, Wimo A, Ribeiro W, Ferri CP.
The global prevalence of dementia: a systematic review and
metaanalysis. Alzheimers Dement 2013; 9: 63–75.
2 Prince M, Guerchet M, Prina M. Alzheimer’s Disease International.
Policy brief for Heads of Government: the global impact of
dementia 2013-2050. December, 2013. http://www.alz.co.uk/
research/G8-policy-brief (accessed June 13, 2016).

Articles
www.thelancet.com Published online July 26, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30950-3
9
3 Daviglus ML, Bell CC, Berrettini W, et al. National Institutes of
Health State-of-the-Science Conference statement: preventing
alzheimer disease and cognitive decline. Ann Intern Med 2010;
153: 176–81.
4 Matthews FE, Brayne C, Lowe J, McKeith I, Wharton SB, Ince P.
Epidemiological pathology of dementia: attributable-risks at death
in the Medical Research Council Cognitive Function and Ageing
Study. PLoS Med 2009; 6: e1000180.
5 Norton S, Matthews FE, Barnes DE, Yae K, Brayne C. Potential for
primary prevention of Alzheimer’s disease: an analysis of
population-based data. Lancet Neurol 2014; 13: 788–94.
6 Lincoln P, Fenton K, Alessi C, et al. The Blackfriars Consensus on
brain health and dementia. Lancet 2014; 383: 1805–06.
7 Andrieu S, Coley N, Lovestone S, Aisen PS, Vellas B. Prevention of
sporadic Alzheimer’s disease: lessons learned from clinical trials
and future directions. Lancet Neurol 2015; 14: 926–44.
8 Anderson C, Teo K, Gao P, et al. Renin-angiotensin system blockade
and cognitive function in patients at high risk of cardiovascular
disease: analysis of data from the ONTARGET and TRANSCEND
studies. Lancet Neurol 2011; 10: 43–53.
9 Peters R, Beckett N, Forette F, et al. Incident dementia and blood
pressure lowering in the Hypertension in the Very Elderly Trial
cognitive function assessment (HYVET-COG): a double-blind,
placebo controlled trial. Lancet Neurol 2008; 7: 683–89.
10 Gorelick PB. Blood pressure and the prevention of cognitive
impairment. JAMA Neurol 2014; 71: 1211–13.
11 Moonen JE, Foster-Dingley JC, de Ruijter W, et al. Eect of
discontinuation of antihypertensive treatment in elderly people on
cognitive functioning–the DANTE Study Leiden: a randomized
clinical trial. JAMA Intern Med 2015; 175: 1622–30.
12 Richard E, Van den Heuvel E, Moll van Charante EP, et al.
Prevention of dementia by intensive vascular care (PreDIVA):
a cluster-randomized trial in progress. Alzheimer Dis Assoc Disord
2009; 23: 198–204.
13 Stalman WAB, Scheltens T, Burgers JS, et al. Dutch guideline on
cardiovascular riskmanagement. Houten: Bohn Stafleu
van Loghum, 2006 (in Dutch).
14 American Psychiatric Association. Diagnostic and statistical manual
of mental disorders, fourth edn. Washington DC: American
Psychiatric Association, 1994.
15 Schrijvers EM, Verhaaren BF, Koudstaal PJ, Hofman A, Ikram MA,
Breteler MM. Is dementia incidence declining?: Trends in dementia
incidence since 1990 in the Rotterdam Study. Neurology 2012;
78: 1456–63.
16 Richard E, Ligthart SA, Moll van Charante EP, van Gool WA.
Methodological issues in a cluster-randomized trial to prevent
dementia by intensive vascular care. J Nutr Health Aging 2010;
14: 315–17.
17 Forette F, Seux ML, Staessen JA, et al. Prevention of dementia in
randomised double-blind placebo-controlled Systolic Hypertension
in Europe (Syst-Eur) trial. Lancet 1998; 352: 1347–51.
18 Dutch College of General Practitioners. Guideline on cardiovascular
risk management (first revision). Huisarts Wet 2012; 55: 14–28
(in Dutch).
19 Beckett NS, Peters R, Fletcher AE, et al. Treatment of hypertension
in patients 80 years of age or older. N Engl J Med 2008; 358: 1887–98.
20 Bejan-Angoulvant T, Saadatian-Elahi M, Wright JM, et al.
Treatment of hypertension in patients 80 years and older: the lower
the better? A meta-analysis of randomized controlled trials.
J Hypertens 2010; 28: 1366–72.
21 Euser SM, van BT, Schram MT, et al. The eect of age on the
association between blood pressure and cognitive function later in
life. J Am Geriatr Soc 2009; 57: 1232–37.
22 Wright JT Jr, Williamson JD, Whelton PK, et al. A randomized trial
of intensive versus standard blood-pressure control. N Engl J Med
2015; 373: 2103–16.
23 Ligthart SA, Richard E, van Gool WA, Moll van Charante EP.
Cardiovascular risk management in community-dwelling elderly:
opportunities for prevention. Eur J Prev Cardiol 2012; 19: 1365–72.
24 Sink KM, Espeland MA, Castro CM, et al. Eect of a 24-month
physical activity intervention vs health education on cognitive
outcomes in sedentary older adults: The LIFE Randomized Trial.
JAMA 2015; 314: 781–90.
25 Ngandu T, Lehtisalo J, Solomon A, et al. A 2 year multidomain
intervention of diet, exercise, cognitive training, and vascular risk
monitoring versus control to prevent cognitive decline in at-risk
elderly people (FINGER): a randomised controlled trial. Lancet 2015;
385: 2255–63.
26 Satizabal CL, Beiser AS, Chouraki V, Chene G, Dufouil C,
Seshadri S. Incidence of dementia over three decades in the
Framingham Heart Study. N Engl J Med 2016; 374: 523–32.
27 Richard E, Andrieu S, Solomon A, et al. Methodological challenges
in designing dementia prevention trials–the European Dementia
Prevention Initiative (EDPI). J Neurol Sci 2012; 322: 64–70.
28 WHO. Global recommendations on physical activity for health.
Chapter 4.4: Age group: 65 years old and above. Geneva:
World Health Organization, 2015: 29–32.