Vitamin and Carotenoid Status in Older Women: Associations With the Frailty Syndrome

The Johns Hopkins Medical Institutions, Baltimore, Maryland, USA.
The Journals of Gerontology Series A Biological Sciences and Medical Sciences (Impact Factor: 5.42). 07/2006; 61(6):600-7. DOI: 10.1093/gerona/61.6.600
Source: PubMed
We investigated the relationship of micronutrient deficiencies with the frailty syndrome in older women living in the community.
Frailty status and serum micronutrients were assessed in a cross-sectional study of 754 women, 70-80 years old, from the Women's Health and Aging Studies I and II.
Among nonfrail, prefrail, and frail women, respectively, geometric mean serum concentrations were 1.842, 1.593, and 1.376 micromol/L for total carotenoids (p <.001); 2.66, 2.51, and 2.43 micromol/L for retinol (p =.04); 50.9, 47.4, and 43.8 nmol/L for 25-hydroxyvitamin D (p =.019); 43.0, 35.8, and 30.9 nmol/L for vitamin B(6) (p =.002); and 10.2, 9.3, and 8.7 ng/mL for folate (p =.03). Frail women were more likely to have at least two or more micronutrient deficiencies (p =.05). The age-adjusted odds ratios of being frail were significantly higher for those participants whose micronutrient concentrations were in the lowest quartile compared to the top three quartiles for total carotenoids, alpha-tocopherol, 25-hydroxyvitamin D, and vitamin B(6). The association between nutrients and frailty was strongest for beta-carotene, lutein/zeaxanthin, and total carotenoids (odds ratio ranging from 1.82 to 2.45, p =.05), after adjusting for age, sociodemographic status, smoking status, and body mass index.
Frail women are more likely to have relatively low serum carotenoid and micronutrient concentrations and are more likely to have multiple micronutrient deficiencies. Future longitudinal studies are needed to examine the relationships between micronutrient concentrations and frailty in older women.


Available from: Linda P Fried, Jan 23, 2016
Nutrition and Aging
Vitamin and Carotenoid Status in Older Women:
Associations With the Frailty Syndrome
Elisabete Michelon,
Caroline Blaum,
Richard D. Semba,
Qian-Li Xue,
Michelle O. Ricks,
and Linda P. Fried
The Johns Hopkins Medical Institutions, Baltimore, Maryland.
The Institute of Geriatrics and Gerontology, Pontifical Catholic
University of Rio Grande do Sul, Brazil.
The University of Michigan, Ann Arbor.
Objective. We investigated the relationship of micronutrient deficiencies with the frailty syndrome in older women
living in the community.
Methods. Frailty status and serum micronutrients were assessed in a cross-sectional study of 754 women, 70–80 years
old, from the Women’s Health and Aging Studies I and II.
Results. Among nonfrail, prefrail, and frail women, respectively, geometric mean serum concentrations were 1.842,
1.593, and 1.376 lmol/L for total carotenoids ( p , .001); 2.66, 2.51, and 2.43 lmol/L for retinol ( p ¼ .04); 50.9, 47.4,
and 43.8 nmol/L for 25-hydroxyvitamin D ( p ¼ .019); 43.0, 35.8, and 30.9 nmol/L for vitamin B
( p ¼ .002); and 10.2,
9.3, and 8.7 ng/mL for folate (p ¼ .03). Frail women were more likely to have at least two or more micronutrient
deficiencies (p ¼ .05). The age-adjusted odds ratios of being frail were significantly higher for those participants whose
micronutrient concentrations were in the lowest quartile compared to the top three quartiles for total carotenoids,
a-tocopherol, 25-hydroxyvitamin D, and vitamin B
. The association between nutrients and frailty was strongest for
b-carotene, lutein/zeaxanthin, and total carotenoids (odds ratio ranging from 1.82 to 2.45, p ¼ .05), after adjusting for age,
sociodemographic status, smoking status, and body mass index.
Conclusion. Frail women are more likely to have relatively low serum carotenoid and micronutrient concentrations and
are more likely to have multiple micronutrient deficiencies. Future longitudinal studies are needed to examine the
relationships between micronutrient concentrations and frailty in older women.
RAILTY is a geriatric syndrome characterized by
a multisystem reduction in physiological reserve and
vulnerability to stressors. It has been shown to predict
adverse health outcomes in older adults, including hospi-
talization, institutionalization, disability, falls, and death (1).
Phenotypic characteristics associated with a syndrome of
frailty have recently been described in a population-based
study of older adults (2,3). A validated definition describes
frailty as having at least three of five criteria that include
weight loss, weakness, low exercise tolerance, slow walking
speed, and low physical activity.
Although previous studies (4–8) have focused primarily
on the roles of weight loss and sarcopenia in frailty,
micronutrient malnutrition is also thought to play an
important role. Deficiencies of micronutrients are known
to be common among older adults (9), but the role that
micronutrient deficiencies may play in the frailty syndrome
has not been characterized. Micronutrient deficiencies have
also been associated with many conditions and factors
associated with the frailty syndrome, including an increased
risk of chronic diseases (10), impaired immune function
(11), decreased antioxidant activity (12), an increased risk of
osteoporosis (13), as well as peripheral vascular disease and
atherosclerosis (14), and a more rapid aging process (6,15).
The goal o f our study was to determine if decreased
micronutrient concentrations, measured either as defined
deficiencies or relatively low serum concentrations, are
associated with prevalent frailty. Because micronutrients are
generally enzyme cofactors responsible for critical steps in
metabolism, or have antioxidant activity, such research is
a first step in gaining insight into the potential contribution
of micronutrient malnutrition to the frailty syndrome and
potential pathophysiological mechanisms (3–7). Such hy-
pothesized pathways could include increased oxidative
stress from low antioxidant micronutrients and problems
such as osteoporosis, disability, and falls associated with
vitamin D deficiency. We hypothesized that micronutrient
malnutrition is associated with frailty and that multiple
deficiencies are more common in frail women.
Participants in this study were women, aged 70–80 years,
who participated in the Women’s Health and Aging Studies
(WHAS) I and II, two complementary, population-based
studies designed to evaluate the causes and course of
Journal of Gerontology: MEDICAL SCIENCES Copyright 2006 by The Gerontological Society of America
2006, Vol. 61A, No. 6, 600–607
by guest on November 6, 2015 from
Page 1
physical disability in older women living in the community.
WHAS I was recruited from an age-stratified random
sample of women aged 65 years or older selected from
Medicare enrollees residing in 12 contiguous ZIP code areas
in Baltimore (16). Women were screened to identify self-
reported physical disability that was categorized into four
domains by report difficulty with tasks in the following
areas: (a) mobility, (b) upper extremity function, (c) higher
functioning household management, and (d) self-care.
WHAS I e nrolled the one-third most disabled women age
65 or older, which were those women with disability in two
or more domains. Of the 1409 women who met study
eligibility criteria, 1002 agreed to participate in the study in
1992. There were no major differences in sociodemographic
or reported health characteristics between eligible partic-
ipants and those who declined. Standardized questionnaires
were administered in the participant’s home by trained
interviewers. Two weeks later, a trained registered nurse
conducted an examination of each study participant in her
home, using a standardized protocol that included physical
performance measures and a directed physical examination.
Approximately 75% of women also consented to phlebot-
omy performed during a separate visit by a trained
phlebotomist who followed a standardized protocol.
WHAS II was specifically designed to be a companion
study for WHAS I and includes a cohort of women, aged
70–79 years, selected to be representative of the two-thirds
least disabled women living in the community. Participants
were selected via age-stratified random samples from the
same sampling frame as in WHAS I, and were screened
using the same four domains of physical function. Eligible
women had either no disability or disability in only one
domain. In 1994, 880 women were found eligible for
WHAS II, and 436 consented to participate. Those agreeing
were more highly educated and reported more diseases that
those who refused, but did not differ signifi cantly in
disability characteristics. An interview stand ardized to that
performed in WHAS I was administered in The Johns
Hopkins Functional Status Laboratory. Trained technicians
then c onducted a standardized examination that included
a directed physical examination and physical performance
measures. Phlebotomy was performed in 93% of WHAS II
participants by a trained phlebotomist following the same
protocol as that used in WHAS I. Details on the study
methods and sampling design of the WHAS studies are
published elsewhere (16,17).
For our analyses, we used a combined sample linking the
two WHAS studies with a methodology that has been
developed by the WHAS research team (18). The sample
consists of women participating in WHAS I or WHAS II
who are 70–80 years of age (n ¼ 864). Appropriate weights
have been calculated to adjust for differential selection
probability with respect to age and disability status from the
sampling frame. Of the 864 women in the combined sample,
756 had blood samples available for analysis of micro-
nutrients. Two individuals did not have complete informa-
tion for health status classification and were excluded from
our analysis.
Definition and Classification of Frailty
Individuals were catego rized as nonfrail, prefrail, or frail
according to criteria recently developed by Fried and
colleagues (2). These criteria describe a frailty phenotype,
validated in a population-based sample of older adults and
based on the presence or absence of five measurable
characteristics: (a) shrinking (reported unintentional weight
loss of 10 pounds or more in the prior year), (b) weakness
(measured grip strength in the lowest quintile at baseline),
(c) poor endurance and energy (self-report of exhaustion),
(d) slowness (the slowest quintile on test of timed walking
speed), and (e) low physical activity level (lowest quintile of
physical activity scale). This frailty phenotype was origi-
nally developed in the Cardiovascular Health Study (2) and
was also assessed in WHAS I and II (Table 1). Individuals
were defined as frail by the presence of three or more of the
five components. Individuals with none of the components
were categorized as nonfrail, and those with one or two
components as prefrail, based on other work indicating that
this group is at high risk of progression to frailty (2). In the
Cardiovascular Health Study cohort, individuals categorized
as frail had a significantly increased 7-year hazard ratio for
incident falls, worsening mobility, hospitalization, and death
(2). The hazard ratios were lower for those individuals
categorized as nonfrail, and intermediate for those catego-
rized as prefrail.
Components of the frailty phenotype were measured
using standardized questions (weight loss and exhaustion)
and proto cols. Physical activity level was determined using
the modified Minnesota Leisure Time Activities Question-
naire (16). A weighted score of kilocalories expended per
week was calculated from questions on frequency and dura-
Table 1. Baseline Characteristics of Women Aged 70–80 Years
Who Participated and Did Not Participate in the Blood Draws for
WHAS I and WHAS II Studies
Blood Drawn
N ¼ 756
No Blood Drawn
N ¼ 108 p
Age, y* 74.4 74.8 .228
White race, % 78 71 .133
Marital status widowed, % 49 61 .143
Educational level, y* 11.7 10.3 .006
Income in US$/y* 21815 11604 .001
Never smokers, % 52 46 .268
Drink at least once a week, % 26 12 .012
Body mass index (weight/height
)* 28.4 28.5 .216
Self-reported weight loss past year, % 21 20 .985
Appetite fair/poor, % 15 29 .001
Wear dentures, % 61 70 .100
Problems chewing/swallowing, % 10 15 .090
Multivitamin users, % 20 18 .558
Self-perceived health fair/poor, % 25 61 .001
Domains of disability, % .001
1 or 2 40 39
3 or 4 20 54
Health status, % .001
Nonfrail 48 28
Prefrail 42 50
Frail 10 22
*Results presented as means.
by guest on November 6, 2015 from
Page 2
tion of walking for exercise, dancing, bowling, performing
moderately strenuous househo ld and outdoor chores, and
participating in any regular exercise program. Weaknes s
was measured by level of maximal grip strength. Data on
maximal grip strength were obtained using a JAMAR hand-
held dynamometer (model BK-7498; Fred Sammons, Inc.,
Burr Ridge, IL) and was measured three times with each
hand. The best measure in the dominant hand was used.
Walking speed was measured using a standardized protocol
for timed walking. The participant could use a cane, walker,
or other walking aide, but not the aid of another person. The
length of walk in meters divided by the time in seconds was
used to calculate the walking speed (16). Cutoff points in
each measure that met the criteria for frailty were
standardized to those used for women in the Cardiovascular
Health Study (2).
Demographic characteristics, self-rated health, and in-
formation about appetite and eating were measured in the
WHAS questionnaires. Chronic diseases were adjudicated
by WHAS co-investigators on the basis of the questionnaire,
physical examination, and physician contact (16). Physical
function and disability status were assessed by the
questionnaires and have been described in detail elsewhere
(16). Participants were questioned about 15 distinct physical
tasks in four functional domains: mobility, upper extremity
function, self-care tasks, and higher functioning skills. For
each task the participant was asked, ‘By yourself, that is,
without help from another person or special equipment, do
you have any difficulty ,performing task.?’ The
responses were dichotomized into ‘yes’ difficulty versus
‘no’ difficulty.
Functional groups were derived from several standardized
scales (Activities of Daily Living [ADL], Instrumental
Activities of Daily Living [IADL], Rosow-Breslau) (16,17).
Individual tasks were grouped as follows: 1) mobility
(walking 2–3 blocks, climbi ng 10 steps, getting in or out
of bed or chairs, and heavy housework); 2) upper extrem-
ity function (raising arms, grasping, and lifting up to
10 pounds); 3) higher funct ioning tasks (IADL) (using
telephone, light housekeeping, meal preparation, and
personal shopping); and 4) self-care tasks (ADL) (bathing
and showering, getting in and out of bed or chair, dressing,
eating, and using the toilet).
A woman was defined as having a physical disability in
a particular group if she reported difficulty in any of the
tasks specific to that category. However, individuals often
reported difficulty for tasks in several groups. Therefor e,
study participants were further differentiated into six
mutually exclusive domains of disability derived from the
original four groups: upper extremity only, mobility only,
both mobility and upper extremity, higher functioning, self-
care, and higher functioning and self-care. The latter three
generally included people who had mobility and/or upper
extremity difficulty in addition to their higher functioning
and/or self-care difficulty (16,17).
Laboratory Analysis
Nonfasting blood samples were obtained by venipuncture
between 9:00
AM and 2:00 PM. Processing, aliquoting, and
freezing were carried out at the Core Genetics Laboratory of
The Johns Hopkins Univers ity School of Medicine
following a standardized protocol. Blood samples were
delivered to Quest Diagnostics Laboratories (Teterboro, NJ)
on the day of blood drawing for assays conduct ed by this
commercial laboratory. Plasma carotenoids, retinol, and a-
tocopherol were determined by high performance liquid
chromatography in the laboratory of one of the investigators
(R.D.S.) (19). Total carotenoids were calculated as the sum
of a-carot ene, b-carotene, b-cryptoxanthin, lutein/zeaxan-
thin, and lycopene (in micromoles per liter). The other
serum biochemical measurements were performed by Quest
Diagnostics Laboratories. Serum concentration of 25-
hydroxyvitamin D [25(OH)D] was measured using a radio-
receptor assay (20). Vitamin B
status was assessed by
pyridoxal 5-phosphate measurements using high perfor-
mance liquid chromatography (21). Serum vitamin B
folate were measured by radioimmunoassay (RIA) (22).
Within-run and between-run coefficients of variation were
10.7% and 23.9% for a-carotene, 7.0% and 19.1% for b-
carotene, 4.7% and 8.5% for b-cryptoxanthin, 4.1% and
4.6% for lutein/zeaxanthin, 10.0% and 14.0% for lycopene
(in micromoles per liter), 4.1% and 9.7% for a-tocopherol,
and 7.5% and 9.6% for 25-hydr oxyvitamin D.
Micronutrient deficiencies were defined using standard
cutoffs: for serum retinol , 1.05 lmol/L; for vitamin E, a-
tocopherol , 11.6 lmol/L; for vitamin D, 25(OH)D , 30
nmol/L; for vitamin B
(plasma pyridoxal 59-phosphate) ,
30 nmol/L; for vitamin B
, 300 pg/mL; and for folate ,
5.0 ng/mL (23). Cutoff s for carotenoids have not been
established and therefore were not included in the analysis
of prevalence of deficiencies. A serum creatinine concen-
tration .1.4 mg/dL was considered consistent with renal
disease. Elevated liver enzymes (alanine aminotransferase
.74 U/L, alkaline phosphatase .154.5 U/L, and/or
aspartate aminotransferase .68 U/L) were considered
consistent with significant hepatic disease. In seasonal
analyses, seasons were defined as summer (June–August),
autumn (September–November), winter (December–
February), and spring (March–May), with summer as the
reference category.
Statistical Analysis
Probability weights specific to WHAS I and II were used
in the analysis to make inferences about community-
dwelling women aged 70–80 years based on ou r study
sample. The procedures used to calculate the weights for the
WHAS samples are described elsewhere (16,18). Descrip-
tive statistics were used to characterize the study population,
give the distribution of biochemical measurements of
micronutrients, and calculate the prevalence of deficiencies.
In addition to described deficiencies, for all the micro-
nutrients, the population was categorized into quartiles
based on micronutrient concentration. This was done
because deficiencies are described for the general population
and may not necessarily be meaningful in older women, and
we hypothesized that low levels may have physiological
significance, even if they did not reach the level of
‘deficiency.’’ For the carotenoids, which do not have
described deficiencies, quartiles were considered the
by guest on November 6, 2015 from
Page 3
appropriate method of analysis. All the measurements of
micronutrients had a skewed distribution and were analyzed
using logarithmic transformation and presented as geometric
means and 95% confidence intervals.
We first compared individuals who had blood tests and
those who did not, using the Pearson chi-square test. Chi-
square and trend tests were used to examine the associations
between micronutrients and dichotomized covariates, i.e.,
frailty status, and means. Values of p .05 were considered
statistically significant. Binomial logistic regression models
were used to examine associations between low micro-
nutrient concentrations and frailty. Low micronutrient
concentrations were defined as the lowest quartile of the
distribution of each micronutrient. Women were categorized
in two groups as frail versus not (prefrail and nonfrail
groups combined). The odds ratio (OR) for being frail was
then calculated for the lowest quartile of each micro-
nutrient’s distribution, using all other quartiles combined as
the reference. Initial models were adjusted for age.
Sequential adjustment was done adding variables known
to be associated with both micronutrient deficiencies and
frailty status. Models were adjusted for age and sociodemo-
graphic characteristics (race [black, white], income [in
tertiles], education [completed high school vs not]); then
smoking status (current smoker vs not) was added; and
finally, body mass index (BMI, calculated from measured
height and weight).
Trends across frailty subgroups were examined using
generalized linear models for continuous variables and the
chi-square test for catego rical variables. Trends across frailty
subgroups were analyzed for differences in the baseline
characteristics of the participants, mean concentrations of
micronutrients, and prevalence of micronutrient deficiencies.
The statistical programs used were SAS (SAS Inst itute, Cary,
NC) (24) for data management and Stata (Stata Corporation,
College Station, TX) (25) for weighted analyses.
Participants with (n ¼ 756) and without (n ¼ 108) blood
drawing in the study are compared in Table 1. Those
without blood drawing had lower educational and income
levels than did those who participated in the blood drawing.
Those without blood drawing were less likely to report
drinking at least once a week, but more likely to report fair
or poor self-rated health and appetite. Disability and frailty
were more prevalent among those participants who did not
participate in blood drawing.
Social and demographic characteristics were compared by
frailty status (Table 2). Women classified as frail were older,
had a lower income, and had fewer years of education. The
frail group had a higher proportion of African Americans,
widows, current or former smokers, and a lower proportion
of weekly alcohol drinkers than did the other groups. It also
had a higher proportion with self-reported weight loss, fair
or poor appetite, denture use, and problems chewing or
swallowing. The mean BMI was lower in the nonfrail than
in the prefrail or frail subgroups, although the prevalence of
obesity (BMI 30) was higher among the nonfrail
participants. A higher proportion of frail individuals
reported their self-perceived health a s fair or poor and had
three or four domains of disability. All of these trends were
significant across the frailty groups from nonfrail to frail.
However, the proportion of women who reported use of
multivitamins did not differ among the three groups.
The unadjusted means of serum micronutrient concen-
trations are presented in Table 3. Mean serum concen-
trations of all micronutrients were lower in the frail than in
Table 2. Baseline Characteristics of Study Participants by Frailty Status
Health Status
Nonfrail Prefrail Frail
N ¼ 331 (%) N ¼ 337 (%) N ¼ 86 (%)
Mean age, y (95% CI) 74.1 (73.7–74.4) 74.4 (74.1–74.7) 75.8 (75.1–76.5) .001
White race, % 81 76 68 .025
Marital status widowed, % 45 50 61 .027
Educational level, y (95% CI) 12.5 (12.1–12.9) 11.2 (10.5–12.0) 9.5 (8.7–10.3) .014
Income , US $9,000, % 20 39 48 .001
Never smokers, % 56 51 39 .006
Drink at least once a week, % 35 20 9 .001
Drink .8 drinks per week, % 9 5 1 .016
Mean body mass index (95% CI) 26.2 (25.5–26.8) 28.5 (27.8–29.2) 29.2 (27.6–30.8) .001
Self-reported weight loss past year, % 17 23 29 .013
Appetite fair/poor, % 8 18 31 .001
Wear dentures, % 57 63 70 .036
Problems chewing/swallowing, % 6 11 27 .001
Multivitamin users, % 21 21 14 .422
Creatinine .1.4 mg/dL, % 3 5 7 .247
Elevated liver enzymes, % 3 3 9 .015
Self-perceived health fair/poor, % 12 32 59 .001
Domains of disability, % .001
1 or 2 30 53 37
3or4 6 26 60
Note:CI¼ confidence interval.
by guest on November 6, 2015 from
Page 4
the other groups. The trend toward lower serum micro-
nutrient concentrations in the frail group was statistically
significant for each of the carotenoids, retinol, 25(OH)D,
vitamin B
, and folate. The trend was not statistically
significant for a -tocopherol and vitamin B
. In ana-
lyses of micronutrient concentrations by season, significant
differences by season were found only for retinol, 25-
hydroxyvitamin D, and vitamin B
, with the generally the
highest concentrations noted in the summer season.
The prevalence of specific vitamin deficiencies by frailty
status is shown in Figure 1. The overall prevalence of
vitamin deficiencies was: vitamin B
(41.6%), vitamin B
(19.4%), vitamin D (14.2%), folate (12.1%), vitamin E
(3.4%), and vitamin A (1.1%). For all vitamins, there was
a trend toward higher prevalence of deficiencies among the
frail and a lower prevalence of deficiencies among the
nonfrail group; this trend was statistically significant for a-
tocopherol, vita min D, and vitamin B
. Using the criteria
for micronutrient deficiencies (23), we found that 43% of
women had no micronut rient deficiencies, 33% had one
deficiency only, and 24% had two or more deficiencies. In
the frail group there was a lower proportion of individuals
with no deficiencies and a higher proportion with two or
more deficiencies than other groups; this trend was
statistically significant (Figure 2).
Table 4 shows the results of logistic regression models of
the associations between frailty and low serum micro-
nutrient concentrations, including low carotenoid concen-
trations. In these models, nonfrail and prefrail data were
combined to focus on frailty and to create conservative
models. In all models specified, whether adjusted only for
age or for multiple variables, the likelihood of being frail (vs
nonfrail or prefr ail), as described by the OR, was always
higher for those with low micronutrient concentrations (the
lowest quartile of the distribution of serum micronutrients
compared to the upper quartiles combined). For example, in
Table 3. Mean Serum Concentration of Micronutrients by Frailty Status
Health Status
Nonfrail Prefrail Frail
N ¼ 331 N ¼ 337 N ¼ 86
Carotenoids (lmol/L)
a-carotene 0.097 (0.088–0.107) 0.075 (0.068–0.083) 0.058 (0.048–0.070)
b-carotene 0.440 (0.401–0.485) 0.363 (0.329–0.400) 0.296 (0.249–0.352)
b-cryptoxanthin 0.136 (0.126–0.147) 0.111 (0.101–0.121) 0.090 (0.077–0.106)
Lutein/zeaxanthin 0.410 (0.388–0.434) 0.345 (0.326–0.365) 0.323 (0.288–0.363)
Lycopene 0.589 (0.549–0.633) 0.545 (0.509–0.583) 0.460 (0.397–0.532)
Total carotenoids 1.842 (1.741–1.949) 1.593 (1.505–1.687) 1.376 (1.249–1.515)
Retinol (lmol/L) 2.66 (2.53–2.80) 2.51 (2.40–2.62 2.43 (2.25–2.63)
a-tocopherol (lmol/L) 22.27 (21.37–23.20) 21.88 (20.84–22.98) 19.61 (18.11–21.23)
25(OH) D (nmol/L) 50.90 (48.32–53.62) 47.43 (44.28–50.81) 43.85 (38.11–50.46)
Vitamin B
(nmol/L) 43.09 (38.79–47.86) 35.86 (32.03–40.16) 30.99 (26.47–36.29)
Vitamin B
(pg/mL) 452.94 (426.79–480.69) 428.24 (404.40–453.49) 417.20 (373.60–465.8)
Folate (ng/mL) 10.20 (9.47–10.99) 9.29 (8.64–9.98) 8.68 (7.44–10.13)
Note: Results presented as geometric means and 95% confidence intervals.
Figure 1. Prevalence of specific vitamin deficiencies by frailty status. *p , .05 by Mantel–Haenszel chi-square.
by guest on November 6, 2015 from
Page 5
the age-adjusted model for a-carotene, the OR for being frail
was 2.24 for the group in the lowest quartile of a-carotene
distribution compared to the group in the upper quartiles.
The strongest associations of low micronutrient concentra-
tion with frailty were noted for the carotenoids (a-carotene,
b-carotene, b-cryptoxanthin, lutein/zeaxanthin, and total
carotenoids), and these associations wer e statistically signif-
icant in age-adjusted models, with a-tocopherol, 25(OH)D,
vitamin B
, and folate reaching borderline significance.
In general, after adjustment for socioeconomic variables,
the OR describing the association between frailty and low
micronutrient concentrations became attenuated. However,
associations for several carotenoid measures were persis-
tently significant, including total carotenoids, lutein/
zeaxanthin, b-carotene, and b-cryptoxanthin. No major dif-
ferences in the associations were observed after adjustment
for smoking status and BMI, except for b-cryptoxanthin,
which lost statistical significance. Additional multivariate
models were run which adjusted for both seasonality and
multivitamin use; these did not change the significance of
the results shown the main multivariate models in Table 4.
Our study has shown a stepwise association of decreasing
serum micronutrient concentration with frai lty status in
older women, i.e., nonfrail, prefrail, and frail. Although this
trend was noted for every micronutrient tested, the strongest
trends were found for serum carotenoids. In addition, low
concentrations of serum carot enoids showed the strongest
associations with frailty status, compared to the other
micronutrients studied, and this association persisted even
after socioeconomic characteristics, smoking status, and
BMI were taken into account. Carotenoids have strong
antioxidant activity and are thought to protect against free
radical damage to tissues (19). A higher dietary intake or
serum concentrations of carotenoids are associated with
decreased risk of cardiovascular disease (26), cancer (27),
Figure 2. Number of vitamin deficiencies by frailty status. *p , .05 by Mantel–Haenszel chi-square.
Table 4. Odds Ratio for Being Frail vs. Not Frail (Pre-Frail and Non-Frail Combined) at the Lower Quartile of the Distribution of
Serum Nutrients Compared to the Upper Quartiles
Micronutrients Age-adjusted*
Adjusted for
Adjusted for
Adjusted for
a-carotene 2.24 (1.34–3.74) 1.31 (0.70–2.47) 1.26 (0.67–2.36) 1.28 (0.67–2.45)
b-carotene 2.38 (1.41–4.01) 1.81 (1.03–3.17) 1.70 (0.97–2.96) 1.82 (1.03–3.21)
b-cryptoxanthin 2.34 (1.38–3.99) 1.77 (0.98–3.19) 1.60 (0.87–2.92) 1.63 (0.87–3.08)
Lutein/zeaxanthin 2.92 (1.75–4.88) 2.62 (1.49–4.61) 2.48 (1.38–4.46) 2.45 (1.32–4.53)
Lycopene 1.39 (0.82–2.37) 1.21 (0.63–2.30) 1.20 (0.62–2.32) 1.20 (0.61–2.36)
Total Carotenoids 2.50 (1.51–4.14) 1.87 (1.06–2.32) 1.76 (0.99–3.14) 1.91 (1.06–3.46)
Retinol 1.31 (0.77–2.20) 1.03 (0.58–1.80) 1.07 (0.61–1.88) 1.16 (0.65–2.07)
a-tocopherol 1.64 (0.95–2.84) 1.23 (0.64–2.36) 1.26 (0.65–2.44) 1.27 (0.64–2.52)
25 (OH) D 1.71 (1.00–2.94) 1.53 (0.83–2.81) 1.54 (0.84–2.82) 1.57 (0.86–2.86)
Vitamin B
1.79 (0.99–3.24) 1.45 (0.74–2.81) 1.45 (0.75–2.83) 1.69 (0.85–3.36)
Vitamin B
1.28 (0.74–2.23) 1.34 (0.75–2.39) 1.46 (0.81–2.61) 1.35 (0.74–2.45)
Folate 1.62 (0.92–2.83) 1.44 (0.76–2.73) 1.41 (0.76–2.63) 1.45 (0.76–2.76)
*Odds ratio (95% confidence interval).
by guest on November 6, 2015 from
Page 6
macular degener ation (28), and sarcopenia (19). These
known links may possibly explain the strength and
persistence of the associations between low carotenoid
levels and frailty status noted in the present study. Among
the other micronutrients, all had sequentially decreasing
levels in nonfrail, prefrail, and frail women, with only two
(vitamin B
and a-tocopherol) not reaching statistical
significance for the trend.
The prevalence of vitamin deficiencies was higher among
the frai l compared to the nonfrail and prefrail women. The
trends across frailty status were significant for vitamin B
a-tocopherol, and 25(OH)D deficiencies, and there was
a similar but not statistically significant trend for folate.
Measurement of folate status may provide limited insight
because low serum folate concentrations are consi dered to
indicate n egative folate balance but not necessarily depleted
body stores associated with functional changes (29). The
trend of increasing deficiency prevalence with frailty is
biologically plausible because deficiencies of B vitamins
, and folate) are associated with homocysteine
elevation and subsequent increased risk of cardiovascular
disease (30) and decreased cognitive function (31). Cardio-
vascular disease has been linked to the frailty syndrome
(32). Such a biological mechanism, while plausible, needs
further investigation.
The incre asing prevalence of 25(OH)D deficiency in
a dose-response type of relationship from nonfrail to frail is
also biologically plausible. A high prevalence of vitamin D
deficiency among older adults has been observed in other
epidemiological studies (33). Vitamin D deficiency is
associated with osteoporosis, myopathy (34), disability
(35), and falls (36), all of which are associated with the
frailty syndrome.
Vitamin B
showed the highest prevalence of deficiency
in our total sample. Lower vitamin B
concentrations were
observed among the prefrail and frail, but no significant
trend in the prevalence of deficiency across the frailty
subgroups was observed. This high prevalence of vitamin
deficiency has also been observed in other epidemi-
ological studies (9,37), although the significance of these
findings for older populations is not well established. Rich
dietary sources of vitamin B
include fortified cereals, white
potatoes and other starchy vegeta bles, noncitrus fruits,
poultry, and beef. Vitamin B
deficiency is known to affect
humoral and cellular immune responses in older adults (38).
In summary, the association between low micronutrient
levels and frailty was increased for every micronutrient
studied, and was particularly strong and persistent for the
carotenoids. When sociodemographic variables were in-
cluded in multivariate models, the associations of frailty (vs
nonfrail and prefrail) with decreased a-carotene, b-crypto-
xanthin, and lycopene were no longer statistically significant
although the associations with b-carotene, lutein/zeaxanthin,
and total carotenoids, although attenuated, persisted. Further
adjustment for smoking status and BMI did not substantially
change our findings. The carotenoids comprise a major
portion of circulating antioxidants, and oxidative stress is
associated with aging, multiple comorbiditie s, and disr uption
in the inflammatory system. This study raises the potential of
an independent association of carotenoids with frailty.
Our study has some limitations. First, 13% of the women
in our sample refused blood draws. Tho se who refused had
lower educational and income levels, and were more likely to
be disabled and to be classified as frail. However, because the
most impaired women in the sample may have not
participated in the blood studies, the associations that we
have found are likely to underestimate associations that
would have been found if all the women had been available
for study. Second, serum measurements of micronutrients
tend to have high within-person variability, particularly when
single measures and nonfasting samples are used, as was the
case in our study (39). Single measurements can lead to
misclassification, and the cutoff points routinely used to
define deficiency may identify an acute rather than a chronic
deficiency syndrome (39). Repeated measures could improve
the precision of the estimates and strengthen the observed
associations. Third, this is a cross-sectional study, and the
direction of the associations cannot be deduced from the
study design. Women who are frail may have more difficulty
shopping and preparing vitamin- and carot enoid-rich foods at
home, or the lack of carotenoids and vitamins could
contribute to the pathogenesis of frailty as discussed above.
Disability has also been associated with financial difficulty in
acquiring food among older women (40). Fourth, the study
did not include dietary assessment, thus, it relies solely on
biochemical measurements of nutritional status.
Further investigations are needed to elucidate the
potential biological pathways by which micronutrient
deficiencies could increase the risk of frailty. Longitudinal
studies may help determine whether preexisting micro-
nutrient deficiencies in nonfrail individuals increase the risk
of subsequently develo ping frailty. Such future studies may
provide insight into the role of micronutrients in the frailty
syndrome and identify potential strategies to prevent frailty.
Such strategies may include micronutrient supplementation
and improvement in quality of dietary intake in older adults,
but further evidence may be needed from large controlled
clinical tri als.
This work was supported by National Institute on Aging (NIA) Grant
RO1 AG027012, National Institutes of Health-National Center for Research
Resources, Outpatient Unit-General Clinical Research Center grant
RR00722, RO1 AI41956, R37 AG19905, and NIA Contract NO1-
AG12112. Dr. Michelon was supported by a post-doctoral scholarship
from Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico
(CNPq)-Brazil. Dr. Blaum was supported by NIA-K08 AG00749-02.
We thank Dana Totin Moncrief and Barbara Dancheck for carotenoid
and vitamin E analyses.
Address correspondence to Richard D. Semba, MD, MPH, Associate
Professor, Johns Hopkins School of Medicine, 550 N. Broadway, Suite
700, Baltimore, MD 21205. E-mail:
1. Fried LP, Walston J. Frailty and failure to thrive. In: Hazzard WR,
Blass JP, Ettinger WH Jr, Halter JB, Ouslander J, eds. Principles of
Geriatric Medicine and Gerontology. 4
ed. New York: McGraw Hill;
2. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J.
Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci
Med Sci. 2002;56A:M146–M156.
by guest on November 6, 2015 from
Page 7
3. Walston J, McBurnie MA, Newman A, et al. Frailty and activation of
the inflammation and coagulation systems with and without clinical
comorbidities: results from the Cardiovascular Health Study. Arch
Intern Med. 2002;162:2333–2341.
4. Bales CW, Ritchie CS. Sarcopenia, weight loss, and nutritional frailty
in the elderly. Annu Rev Nutr. 2002;22:309–323.
5. Campbell AJ, Buchner DM. Unstable disability and the fluctuations of
frailty. Age Ageing. 1997;26:315–318.
6. Meydani M. Nutrition interventions in aging and age-associated
disease. Ann N Y Acad Sci. 2001;928:226–235.
7. Strawbridge WJ, Shema SJ, Balfour JL, Higby HR, Kaplan GA.
Antecedents of frailty over three decades in an older cohort. J Gerontol
B Psychol Sci Soc Sci. 1998;53:S9–S16.
8. Morley JE, Baumgartner RN, Roubenoff R, Mayer J, Nair KS.
Sarcopenia. J Lab Clin Med. 2001;137:231–243.
9. de Groot LC, Hautvast JG, van Staveren WA. Nutrition and health of
elderly people in Europe: the EURONUT-SENECA Study. Nutr Rev.
10. Fairfield KM, Fletcher RH. Vitamins for chronic disease prevention in
adults: scientific review. JAMA. 2002;287:3116–3126.
11. High KP. Micronutrient supplementation and immune function in the
elderly. Clin Infect Dis. 1999;28:717–722.
12. Ames BN, Shigenaga MK, Hagen RM. Oxidants, antioxidants, and the
degenerative diseases of aging. Proc Natl Acad Sci U S A. 1993;90:
13. Holick MF. Vitamin D: a millennium perspective. J Cell Biochem.
14. Homocysteine Studies Collaboration. Homocysteine and risk of is-
chemic heart disease and stroke. A meta-analysis. JAMA. 2002;288:
15. Ames BN. Micronutrients prevent cancer and delay aging. Toxicol Lett.
16. Guralnik JM, Fried LP, Simonsick EM, Kasper D, Lafferty ME.
The Women’s Health and Aging Study: Health and Social Character-
istics of Older Women with Disability. Bethesda, MD: National Insti-
tute on Aging; NIH Publication No. 95-4009, 1995.
17. Fried LP, Bandeen-Roche K, Kasper JD, Guralnik JM. Association of
comorbidity with disability in older women: the Women’s Health and
Aging Study. J Clin Epidemiol. 1999;52:27–37.
18. Cappola AR, Bandeen-Roche K, Wand GS, Volpato S, Fried LP.
Association of IGF-1 levels with muscle strength and mobility in older
women. J Clin Endocrinol Metab. 2001;86:4139–4146.
19. Semba RD, Blaum C, Guralnik JM, Totin D, Ricks MO, Fried LP. Low
carotenoid and vitamin E status are associated with indicators of
sarcopenia among older women living in the community. Aging Clin
Exp Res. 2003;15:482–487.
20. Hadda J, Chyu K. Competitive protein-binding radioassay for 25-
hydroxycholecalciferol. J Clin Endocrinol Metab. 1971;33:40:
21. Kimura M, Kanehira K, Yokoi K. Highly sensitive and simple liquid
chromatographic determination in plasma of B-6 vitamers, especially
pyridoxal 59-phosphate. J Chromatogr. 1996;A722:296–301.
22. Stabler SP, Allen RH, Fried LP, et al. Racial differences in prevalence
of cobalamin and folate deficiencies in disabled elderly women. Am J
Clin Nutr. 1999;70:911–919.
23. Sauberlich HE. Laboratory Tests for the Assessment of Nutritional
Status. 2nd Ed. Boca Raton, FL: CRC Press; 1999.
24. SAS Institute. SAS Languages and Procedures: Introduction, Version
6. Cary, NC: SAS Institute; 1990.
25. Stata Corporation. Stata User’s Guide. College Station, TX: Stata
Press; 1996.
26. Gaziano JM. Antioxidants in cardiovascular disease: randomized trials.
Nutrition. 1996;12:583–
27. Flagg EW, Coates RJ, Greenberg RS. Epidemiologic studies of
antioxidants and cancer in humans. J Am Coll Nutr. 1995;14:419–427.
28. Mares-Perlman JA, Millen AE, Ficek TL, Hankinson SE. The body of
evidence to support a protective role for lutein and zeaxanthin in
delaying chronic disease. Overview. J Nutr. 2002;132:518S–524S.
29. Bailey LB. Folate status assessment. J Nutr. 1990;120(Suppl 11):
30. Herrmann W. The importance of hyperhomocysteinemia as a risk factor
for diseases: an overview. Clin Chem Lab Med. 2001;39:666–674.
31. Duthie SJ, Whalley LJ, Collins AR, Leaper S, Berger K, Dearly IJ.
Homocysteine, B vitamin status, and cognitive function in the elderly.
Am J Clin Nutr. 2002;75:908–913.
32. Newman AB, Gottdiener JS, McBurnie MA, et al. Associations of
subclinical cardiovascular disease with frailty. J Gerontol A Biol Sci
Med Sci. 2001;56:M158–M166.
33. Jacques PF, Felson DT, Tucker KL, et al. Plasma 25-hydroxyvitamin D
and its determinants in an elderly population sample. Am J Clin Nutr.
34. Reichel H, Koeffler HP, Norman AW. The role of the vitamin D endo-
crine system in health and disease. N Engl J Med. 1989;320:981–991.
35. Semba RD, Garrett E, Johnson BA, Guralnik JM, Fried LP. Vitamin D
deficiency among older women with and without disability. Am J Clin
Nutr. 2000;72:1529–1534.
36. Janssen HC, Samson MM, Verhaar HJ. Vitamin D deficiency, muscle
function, and falls in elderly people. Am J Clin Nutr. 2002;75:611–615.
37. Bates CJ, Pentieva KD, Prentice A, Mansoor MA, Finch S. Plasma
pyridoxal phosphate and pyridoxic acid and their relationship to plasma
homocysteine in a representative sample of British men and women
aged 65 years and over. Brit J Nutr. 1999;81:191–201.
38. Rall LC, Meydani SN. Vitamin B6 and immune competence. Nutr Rev.
39. Gibson RS. Principles of Nutritional Assessment. New York: Oxford
University Press; 1990.
40. Klesges LM, Pahor M, Shorr RI, Wan JY, Williamson JD, Guralnik
JM. Financial difficulty in acquiring food among elderly disabled
women: results from the Women’s Health and Aging Study. Am J Publ
Health. 2001;91:68–75.
Received February 11, 2004
Accepted June 21, 2004
Decision Editor: John E. Morley, MB, BCh
by guest on November 6, 2015 from
Page 8
  • Source
    • "As energy and nutrient deficiency may affect mitochondrial function leading to muscle-related symptoms, including fatigue and weakness [26], protein intake is the major factor responsible for muscle protein anabolism in older persons [13]. Unopposed oxidative stress may be detrimental for skeletal muscle [29] and antioxidant vitamins may play a preventive role in reducing oxidative injury [11]. Overall, the present study findings suggested that poor nutritional intake is an important factor associated with the frailty syndrome where low intake of nutrients was associated with frailty even independently of energy intake. "
    Full-text · Dataset · Mar 2016
  • Source
    • "As energy and nutrient deficiency may affect mitochondrial function leading to muscle-related symptoms, including fatigue and weakness [26], protein intake is the major factor responsible for muscle protein anabolism in older persons [13]. Unopposed oxidative stress may be detrimental for skeletal muscle [29] and antioxidant vitamins may play a preventive role in reducing oxidative injury [11]. Overall, the present study findings suggested that poor nutritional intake is an important factor associated with the frailty syndrome where low intake of nutrients was associated with frailty even independently of energy intake. "
    Full-text · Article · Feb 2016
  • Source
    • "Therefore, studies using more detailed items about food intake aiming to understand how dietary habits affect elderly people's muscle strength seem necessary to make suggestions regarding the diet of elderly people for maintaining muscle strength. The smaller decline of KES in women who ate green or yellow vegetables almost every day may be explained by the antioxidant effects of carotenoids and vitamin C. Recent epidemiological studies in community-dwelling elderly show that low serum concentrations of carotenoids or vitamin C are associated with low muscle strength [4,6,293031. A cross-sectional study revealed that daily dietary intake of vitamin C and β-carotene is significantly associated with KES [6]. "
    [Show abstract] [Hide abstract] ABSTRACT: This cross-sectional and 4-year longitudinal cohort study aimed to clarify how various lifestyle-related variables affect knee extension strength in elderly Japanese women. The participants were community-dwelling women (n = 575) living in the Itabashi Ward of Tokyo, Japan aged 75-85 years at baseline (in 2008) who returned for a follow-up examination 4 years later (in 2012). Maximum isometric knee extension strength in the dominant leg was measured during comprehensive medical check-ups at baseline and follow-up. Interviews with participants included questions on their history of 11 diseases and lifestyle-related factors such as physical activity as well as dietary, smoking, and drinking habits. Cross-sectional and longitudinal analyses yielded inconsistent results regarding the associations between lifestyle-related factors and knee extension strength. While going out more frequently and regular physical exercise positively affected baseline knee extension strength, they did not affect knee extension strength in the longitudinal analysis. The longitudinal analysis revealed that more frequent intake of soy products or green and yellow vegetables at baseline decreased age-related knee extension strength decline. The inconsistent results from the cross-sectional and longitudinal analyses indicate that conducting both types of analyses is crucial for researching this type of subject. The present study demonstrates that the age-related decline in muscle strength is lower in those who frequently eat soy products or green and yellow vegetables. Thus, recommending higher intake of soy products, and green and yellow vegetables for the elderly might help maintain their muscle health.
    Full-text · Article · Jul 2015 · PLoS ONE
Show more