Components of Metabolic Syndrome and the Risk of Disability among the Elderly Population

The direct relationship between metabolic syndrome (MetS) and function disability has not been established. The aim of the present study was to investigate the relationship between MetS and functional disability in the elderly. This retrospective observational study included 1,778 participants aged 60–84 years from the National Health and Nutrition Examination Survey (1999–2002). Impairments in activities of daily living (ADL), instrumental activities of daily living (IADL), leisure and social activities (LSA), lower extremity mobility (LEM), and general physical activities (GPA) were assessed. Additionally, the associations between the features of MetS and disability were evaluated. MetS was associated with a high prevalence of functional dependence in ADL, IADL, LSA, LEM, and GPA. After adjusting for potential confounders, a high number of MetS components was found to be associated with increased disability (P = 0.002). Additionally, associations were observed between MetS components, including abdominal obesity and high triglycerides levels, and functional dependence in ADL, IADL, LSA, LEM, and GPA (all, P < 0.05). A linear increase in disability might be associated with the number of MetS components in an elderly population. Additionally, MetS abnormalities, particularly abdominal obesity and high triglycerides levels, might be highly predictive of functional dependence in the elderly.

Metabolic components and functional dependence. The associations between the number of MetS components and the predicted total disability are presented in Table 2. In the adjusted analysis, the predicted total disability was higher in the MetS group than in the non-MetS group (P < 0.05). There was a strong linear increase in the predicted total disability with the increase in the number of MetS components. In the unadjusted analysis, the β coefficients of the predicted total disability considering 3, 4, and 5 MetS components were 0.446, 0.591, and 0.875 respectively (P-value for the trend < 0.001). After additional adjustment, the β coefficient of the predicted total disability considering 5 MetS components was 0.573 (P-value for the trend < 0.05). Moreover, hypertriglyceridemia and central obesity were significantly associated with higher predicted total disability in the fully adjusted models (P < 0.05).
The associations between the different metabolic components and each disability are presented in Table 3. After adjusting for covariates in Models 1-5, central obesity was found to be significantly associated with higher predicted ADL disability, IADL disability, LEM disability, and GPA disability (P < 0.05). High triglyceride levels were significantly associated with higher predicted ADL disability, IADL disability, LSA disability, and GPA disability (P < 0.05).
In table 4, there was a strong linear increase in predicted total disability with an increase in the number of MetS components in female elderly (p for trend < 0.05), but not in male. In both women and men, abdominal obesity and hypertriglyceridemia was significantly associated with higher predicted total disability in fully adjusted models (p < 0.05).

Discussion
The present study found that the risk of disability was significantly proportional to the number of MetS components. Additionally, abdominal obesity and high triglyceride levels were significantly associated with higher disability in the fully adjusted models. Very few studies have focused on the elderly population to investigate the association between MetS and disability [5][6][7] . To our knowledge, our study is the first to demonstrate that MetS is a potential risk factor for disability in the elderly US population.
Previous studies have shown that an increase in the rate of mobility dysfunction is associated with MetS 7-9 ; however, the prevalence of MetS and definition of disability differed among the studies. Additionally, the majority of studies only focused on mobility limitations and MetS 8,9 . There is limited information on the association between the number of MetS components and the different domains of functional disability. Our findings of the association between MetS and function disability are consistent with the findings of previous studies in elderly people [7][8][9] . We found an association between the number of MetS components and total functional disability, as well as between the different MetS components and each domain of functional disability (ADL, IADL, LSA, LEM, and GPA).
In a 3-year longitudinal study of 1,155 elderly participants (the CHIANTI study), MetS was found to be associated with ADL disability 10 . The Health, Aging and Body Composition Study included 2,920 participants aged 70-79 years who were followed up for 4.5 years 8 . The study reported that MetS was associated with an adjusted relative risk (RR) of 1.46 (95% confidence interval [CI]= 1.30-1.63) for progression of mobility limitations, and the risk increased as the number of MetS components increased (P-value for the trend < 0.001). In a cross-sectional study of 422 individuals living in 2 communities in southern Brazil 5 , MetS was found to significantly increase the risk of functional dependence, depression, and cognitive impairment. However, the number of participants was lower in that study than in the present study, and functional scales, such as LSA, LEM, and GPA, were not assessed. In a 3-year longitudinal study of 1,606 Mexican-Americans aged 60-98 years (the Sacramento Area Latino Study on Aging), MetS was found to be correlated with progressive limitations in mobility and high rates of impairment of ADL and IADL 6 . However, some data were missing owing to non-response and loss to follow-up, and this information was imputed through multiple imputation methodologies. In a 7-year longitudinal study of 6,141 French people aged > 65 years, the presence of MetS was found to raise the risk of restricted mobility and IADL limitations 7 . However, a large number of participants were excluded owing to high rates of existing MetS and disability, and the exclusion likely weakened the observed associations rather than negated the Scientific RepoRts | 6:22750 | DOI: 10.1038/srep22750 results. Our study included a large population-based representative sample of older adults and provided further evidence for the association of each metabolic component with motion disability and daily living dysfunction. An increase in the number of MetS components was significantly associated with functional disability in the US adult population. It is important to determine the association of the number of MetS components with overall mobility dysfunction and the limitation in complex ADL, as limited information is available on the association between MetS and ADL/IADL 6,7 . To our knowledge, our study is the first to identify an association of MetS with LSA and GPA disability.
MetS has been shown to be a highly significant and independent predictor of the decline in mobility 7,9 . In an elderly population, MetS may be an obvious risk factor for functional disability. A previous study hypothesized that excessive free radicals and chronic inflammation are involved in the pathophysiological process of physical impairment 11 . High levels of C-reactive protein (CRP) and interleukin (IL)-6 have been shown to be associated with disability and low physical performance 12 , and the decline in physical function may be related to catabolic effects 13 . Kao et al. analyzed community-dwelling elderly people from the NHANES (1999-2002) and demonstrated that the serum total bilirubin levels within the normal range were inversely associated with the odds of functional dependence 14 . Serum bilirubin has been considered to have antioxidant and anti-inflammatory effects. Elevated homocysteine, which causes tissue injury by oxidative stress, has been shown to be associated with multiple domains of disability mediated by muscle strength and gait speed 15 . Previous studies have shown that patients with MetS have elevated levels of homocysteine 16 , CRP 17 , and proinflammatory mediators 18 . Total bilirubin levels have been shown to be inversely associated with MetS 19 . The cumulative effects of MetS components in the elderly population might result in chronic systemic inflammation that predisposes individuals to functional dependence and disability. The presence of MetS appears to worsen the age-related loss of serotoninergic innervation and responsiveness, leading to a high risk of depression 5 and promoting increased functional dependence. In a prospective cohort study of 1,122 participants aged ≥ 71 years who were living in the community, lower extremity function was found to be a predictor of subsequent disability 20 . Previous studies have shown that muscle strength was inversely associated with the incidence of MetS 21,22 , and MetS might result in slow gait speed and related functional dependence. The domains of IADL and LSA are associated with cognitive decline. MetS has been shown to be associated with cognitive decline and dementia 23 . Birdsill et al. found that low cerebral blood flow was associated with low memory function in patients with MetS 24 . Therefore, reducing the risk of MetS may be important for preserving cerebral blood flow and cognitive function. The presence of MetS has been shown to be associated with a low brain volume in patients with manifest arterial disease, even those without diabetes 25 . Cerebrovascular changes may weaken the frontal-subcortical circuit responsible for balance, gait, and movement.
In the present study, central obesity and high triglyceride levels were significantly associated with functional disability. These findings are consistent with previous findings and provide further information on the relationship between functional status and MetS. In a prospective study of 6,141 participants aged > 65 years (the Three-City cohort study), central obesity and high triglyceride levels were found to be associated with limitations in mobility and IADL 7 . In an English longitudinal study of 3,055 community-dwelling adults aged ≥ 65 years, frailty was found to be associated with high abdominal circumference, even among underweight older participants 26 . A previous study showed that WC, a marker for central obesity, could predict the risk for most disability outcomes across the preretirement age period 27 . Studies have shown that abdominal adiposity is associated with high levels of biomarkers for oxidative stress 28 and low-grade systemic inflammation 29 . Plasma leptin, secreted by adipocytes, has been shown to be associated with body fat mass in obese and lean participants 30 , and it may induce the production of IL-6, CRP, and other acute phase reactants. Plasma leptin has also been shown to exert an atherogenic effect by generating oxidative stress in endothelial cells 31 , and excessive oxidative stress has been shown to play an important role in the mechanisms of age-associated frailty 32 and skeletal muscle damage 33 , leading to functional disability. We found a positive independent relationship between disability and MetS, which   Table 3. Regression coefficients of the component of metabolic syndrome for elderly patients with eachdomain of disability. Model 1 = unadjusted Model 2 = Model 1 + age, gender, and race-ethnicity Model 3 = Model 2 + (serum AST, serum total cholesterol, and serum total bilirubin) Model 4 = Model 3 + (smoking, hypertension, type 2 diabetes mellitus, and stroke) Model 5 = Model 4 + (Cognition impairment*) Each disability was defined as any difficulty in performing one or more activities within a given domain and assessed as the sum of each major domain at baseline and analyzed as a continuous variable. *Cognition impairment was defined as the lowest quartile of DSST score.
High serum triglyceride levels were significantly associated with functional disability in the present study. Serum triglyceride is a clinically valuable marker of MetS. Elevated serum triglyceride levels were shown to be associated with insulin resistance 34 , and the presence of the insulin resistance-hyperinsulinemic-metabolic complex obstructs successful healthy ageing. A previous study showed that insulin resistance may be considered a biomarker of age-related diseases and reduced lifespan 35 . Moreover, the presence of insulin resistance may be considered a biomarker of susceptibility to mobility limitations among the elderly 36 . Furthermore, Cholerton et al. 37 found that insulin resistance is associated with impaired cognitive function in the elderly. High triglyceride levels were associated with insulin resistance, which was related with cognition impairment and mobility limitation. Additionally, functional disability was associated with elevated serum triglyceride levels.
Our findings indicate that MetS and its individual components can be considered a cluster of risk factors for functional dependence, highlighting the increasing number of MetS components that are related to the worsening of physical disabilities. The present study has several strengths. The study included a community-based, racially mixed sample and discussed the effects of the number of MetS components and each MetS component on disabilities. In addition, multiple different instruments were used to evaluate physical function, leisure activity, and social activity. Furthermore, each MetS component and functional dependence domain was included for evaluation.
The present study has some limitations. First, this study was an observational, retrospective analysis of an existing database from a single period, rather than an analysis of long-term repeated observations with limited causal inferences. Second, the disability outcomes were self-reported, which may have led to over-reporting in participants who had metabolic disorders, such as obesity, or vascular disorders, such as hypertension. Third, data on potential confounding factors, such as IL-6, plasma leptin, and homocysteine, were not available in the NHANES datasets from 1999 to 2002. Forth, despite adjusting for a large number of potential confounding factors, residual confounding effects from unmeasured confounders of the association between MetS and disability might have been present. Fifth, we could not truly control for the diagnosed medical illness, as most information on medical history was self-reported. Finally, our sample was collected from 1999 to 2002. Therefore, the sample may not accurately reflect the current US population. However, based on a recent study on the prevalence of MetS in the US 38 , MetS continues to be common in older, non-Hispanic whites, which is consistent with our sample population.

Conclusion
A linear increase in disability might be associated with the number of MetS components in an elderly population. It is important to evaluate the limitations in complex ADL, as well as mobility difficulties in the elderly. Additionally, MetS abnormalities, particularly abdominal obesity and high triglycerides levels, might be highly predictive of functional dependence in the elderly. Therefore, the prevention and treatment of MetS would be useful for promoting normal aging and preventing disability. A further prospective study is needed to establish the causality of MetS components in functional dependence. Our study provides epidemiologic evidence to support further studies on interventions for preventing disability in the elderly.

Methods
Study design and participants. The NHANES is a population-based health survey of non-institutionalized US citizens. It was performed by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC). This was a continuous annual survey rather than a periodic survey. The   A total of 21,004 participants were included in the NHANES dataset from 1999 to 2002. We excluded participants aged < 60 years (N = 17,298) and aged ≥ 85 years (N = 409). A total of 3,297 participants were aged between 60 and 84 years. We further excluded 1,519 persons with missing data, including disability questionnaire screening (N = 308), laboratory measurements (N = 251), missing data on metabolic syndrome (N = 960) were also excluded. Therefore, 1,778 participants were finally included in the present study.
The NHANES data can be downloaded from the NHANES website and analyzed without any permission. The Institutional Review Board of NCHS approved the NHANES, and our study did not require approval as it used de-identified data. Functional dependence. Nineteen questions on physical function were asked to evaluate the functional status of the participants. The questionnaire was designed to determine the condition of dependence, and an individual's level of difficulty can be evaluated without using any special equipment. The questions were based on functional mobility and transfers, household productivity, manipulation of surroundings, and social integration. The 19 questions were divided into the following 5 domains: (1) ADL (eating, dressing, walking, and getting out of bed); (2) IADL (managing money, doing house chores, and preparing meals); (3) LSA (attending social events, going out to watch movies, and in-home leisure activities); (4)GPA (standing, bending, stooping, sitting, lifting, grasping, and reaching); and (5) LEM (walking one-quarter of a mile and walking up 10 steps). The following 4 levels of difficulty were used to evaluate each domain: no difficulty, some difficulty, much difficulty, and unable to perform. Total disability was assessed as the sum of the 5 major domainsat baseline and analyzed as a continuous variable, with a range from 0 to 19. Furthermore, we defined each disability as difficulty performing one or more activities within the domainand analyzed the domains as continuous variables, with ranges from 0 to 4 (ADL), 0 to 3 (IADL), 0 to 3 (LSA), 0 to 2 (LEM), and 0 to 7 (GPA).

Metabolic syndrome criteria.
Covariates. The self-reported data included age, sex, race/ethnicity, smoking history, and medical history. Participants were asked the question "Do you now smoke cigarettes?" to determine the smoking status. Participants were considered to have heart disease if they had been diagnosed with the condition or if they experienced myocardial infarction, coronary artery disease, angina, or congestive heart failure. Stroke was self-reported. Participants were considered to have diabetes if they had been diagnosed with the condition by a doctor, if they were using anti-diabetic medication, or if the fasting glucose level was ≥126 mg/dL or the random glucose level was ≥ 200 mg/dL. Blood pressure was measured 3 to 4 times using a mercury sphygmomanometer in the right arm unless the participant has a specific condition, and the mean systolic and diastolic blood pressure values were recorded. Participants were considered to have hypertension if the mean blood pressure was ≥140/90 mmHg, if they were using anti-hypertensive medication, or if they had been diagnosed with the condition by a doctor. WC was measured to the nearest 0.1 cm at the high point of the iliac crest during minimal respiration. Digit Symbol Substitution Test (DSST), a subtest of the Wechsler Adult Intelligence Scale, can examine the visuospatial and motor speed of processing. DSST is a nonverbal test that evaluates psychomotor speed and represents a sensitivemeasure of frontal lobe executive functions. In DSST coding exercise, participants transcribe symbols matched to numbers using a legend. The score represented the number of correct items completed within period of 120 seconds and the maximum score is 133. Because of the absence of agree-upon threshold for normal DSST score 40 , we categorized the DSST score into four groups. The quartiles of DSST score were as follows: Q1 ≦ 28, 28 < Q2 ≦ 40, 40 < Q3 ≦ 53, and Q4 > 53. The lowest quartile (Q1) of DSST score is considered as cognition impairment.
The levels of total cholesterol, triglycerides, HDL-C, AST, and alanine aminotransferase were measured enzymatically using the Hitachi-704 analyzer (Boehringer Mannheim Diagnostics, Indianapolis, IN). Low-density lipoprotein cholesterol levels were calculated using the Friedewald formula 41 . Total bilirubin levels were assessed using automated biochemical profiling (Beckman Synchron LX20; Beckman Coulter Inc., Fullerton, CA), and serum glucose levels were measured using an enzymatic assay (Cobas Mira assay; Roche, Basel, Switzerland). Additionally, CRP levels were determined using a highly sensitive assay technique and were quantified with latex-enhanced nephelometry using a Behring Nephelometer System (Dade Behring, Deerfield, IL).
All the protocols used standardized methods with documented accuracy with respect to CDC reference methods. All detailed specimen collection information is available on the NHANES website.
Statistical analysis. The NHANES data sets from 1999 to 2000 and 2001 to 2002 were combined according to the NHANES analytic guidelines. The predicted values of self-reported dependence were divided into MetS and non-MetS groups. The Chi-square test was used to analyze categorical data, and the Mann-Whitney U test was used to analyze continuous data. The associations between self-report dependence and the number of MetS components or each individual MetS component were determined. From a clinical standpoint, demographic factors can influence the results; therefore, they were used in covariate adjustment. Five multiple linear regression models were constructed, and of these, 1 was unadjusted and the other 4 were adjusted. Model 1 was unadjusted. Model 2 included Model 1+ age, sex, and race/ethnicity adjustment. Model 3 included Model 2+ serum AST, serum total cholesterol, and serum total bilirubin adjustment. Model 4 included Model 3+ smoking, hypertension, type 2 diabetes mellitus, and stroke adjustment. Model 5 included Model 4 + cognition impairment adjustment. Moreover, we also evaluated effect on the association with MetS and disability in gender categorization. The P-values for the trend tests were determined by treating the number of MetS components as a continuous variable (1)(2)(3)(4)(5) in order to observe the associations between an increase in the number of MetS components and self-reported dependence. All statistical analyses were performed using SPSS (version 18.0 for Windows; IBM, Armonk, NY). Two-sided P-values < 0.05 were considered to indicate significant differences.