The Association between Metabolic Syndrome, Frailty and Disability-Free Survival in Healthy Community-dwelling Older Adults

Abstract

OBJECTIVES:

To examine the association between metabolic syndrome (MetS) and frailty, and determine whether co-existent MetS and frailty affect disability-free survival (DFS), assessed through a composite of death, dementia or physical disability.

DESIGN:

Longitudinal study.

SETTING AND PARTICIPANTS:

Community-dwelling older adults from Australia and the United States (n=18,264) from “ASPirin in Reducing Events in the Elderly” (ASPREE) study.

MEASUREMENTS:

MetS was defined according to American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines (2018). A modified Fried phenotype and a deficit accumulation Frailty Index (FI) were used to assess frailty. Association between MetS and frailty was examined using multinomial logistic regression. Cox regression was used to analyze the association between MetS, frailty and DFS over a median follow-up of 4.7 years.

RESULTS:

Among 18,264 participants, 49.9% met the criteria for MetS at baseline. Participants with Mets were more likely to be pre-frail [Relative Risk Ratio (RRR): 1.22; 95%Confidence Interval (CI): 1.14, 1.30)] or frail (RRR: 1.66; 95%CI: 1.32, 2.08) than those without MetS. MetS alone did not shorten DFS while pre-frailty or frailty alone did [Hazard Ratio (HR): 1.68; 95%CI: 1.45, 1.94; HR: 2.65; 95%CI:1.92, 3.66, respectively]. Co-existent MetS with pre-frailty/frailty did not change the risk of shortened DFS.

CONCLUSIONS:

MetS was associated with pre-frailty or frailty in community-dwelling older individuals. Pre-frailty or frailty increased the risk of reduced DFS but presence of MetS did not change this risk. Assessment of frailty may be more important than MetS in predicting survival free of dementia or physical disability.

Introduction

Frailty is a geriatric syndrome associated with an increased risk of adverse health outcomes (1, 2). Although frailty has been conceptualized as a wasting syndrome akin to “failure to thrive,” in recent years, it is recognized that frailty is associated with several characteristics of metabolic syndrome (MetS), such as obesity and hyperglycemia (3, 4). MetS is a combination of atherogenic dyslipidemia, elevated blood pressure, insulin resistance and elevated blood glucose, driven by excess calories in relation to energy expenditure and concomitantly increased waist circumference, creating a pro-thrombotic and pro-inflammatory state (5). It is an important global health issue, as MetS increases the risk of developing cardiovascular and cerebrovascular disease (CVD) (6). Prior studies have described associations between MetS and frailty (4, 7, 8). In addition, MetS has been shown to increase risk of all-cause mortality in middle aged men (9) and in older adults (10), while frailty was shown to better predict mortality than did MetS, regardless of age (11). A systematic review and meta-analysis did not find a significant association between MetS and incident dementia, including Alzheimer’s disease (AD), but another report found that MetS increased the incidence of vascular dementia (12). In another study, MetS was associated with an increased risk of incident disability and deteriorated functional performance (13). On the other hand, frailty has been associated with incident AD and the rate of cognitive decline in older persons (14), and also with physical disability (15). Given the common pathophysiology potentially underpinning MetS and frailty (7, 8), it may be expected that in combination, MetS and frailty could have additive or synergistic effects on adverse health outcomes like dementia, physical disability or death. No previous studies have examined the role of MetS, in the presence of frailty, on the important geriatric outcome of disability-free survival (DFS) which is a composite of the first event of dementia, independence-limiting physical disability and mortality. DFS was the primary outcome of the ASPirin in Reducing Events in the Elderly (ASPREE) clinical trial (16), a large cohort of more than 19,000 community-dwelling older individuals from Australia and the United States (US) who were relatively healthy at enrolment, and notably free of established major CVD, dementia or physical disability (17). The comprehensive health status of ASPREE participants was available for analysis of MetS as well as for the characterization of frailty.

The purpose of the present post hoc analysis of the ASPREE data was: firstly, to determine the cross-sectional association between MetS and frailty and, secondly, to assess whether the presence of MetS and frailty, independently or in combination, increased the risk of shortened disability-free survival.

Methods

Study design and participants

Details of ASPREE (17, 18) and its results have been published previously, including the finding that low-dose aspirin use did not reduce the risk of incident frailty or the trajectory of frailty (19, 20). A total of 19,114 participants (median age 74.0 years, interquartile range or IQR: 6.1 years) were recruited to the ASPREE trial, and 18,264 had all relevant data to analyze MetS and frailty at baseline. Exclusion criteria included a history of a cardiovascular event or established cardiovascular disease or atrial fibrillation, blood pressure ≥180/≥105 mm Hg, dementia or a score of less than 78 on Modified Mini-Mental State Examination (3MS), physical disability as defined by severe difficulty or inability to perform any one of six basic activities of daily living (BADLs), a condition with high current or recurrent risk of bleeding, anemia, current use of other antiplatelet or antithrombotic medication, current use of aspirin for secondary prevention or a condition likely to cause death within five years (17).

Instruments

Metabolic Syndrome (ACC/AHA 2018)

We used the diagnostic criteria for MetS from the 2018 Guideline on the Management of Blood Cholesterol: A Report of the ACC/AHA Task Force on Clinical Practice Guidelines of 2018 (21), which requires any three or more of the following five criteria: (1) elevated triglycerides (TG) ≥175 mg/dL (≥2.0 mmol/L) or lipid lowering treatment; (2) elevated fasting blood glucose ≥100 mg/dL (5.6 mmol/L) or drug treatment for elevated glucose; (3) reduced high-density lipoprotein cholesterol (HDL-C): in men, < 40 mg/dL (1.0 mmol/L) and in women, < 50 mg/dL (1.3 mmol/L) or lipid lowering treatment; (4) elevated blood pressure demonstrated by any of the following: systolic blood pressure ≥ 130 mm Hg and or diastolic blood pressure ≥ 85 mm Hg or drug treatment for hypertension; and (5) elevated waist circumference ≥102 cm for men and ≥88 cm for women.

Assessment of Frailty

Fried phenotype (proposed and validated by Fried and colleagues in the Cardiovascular Health Study) (1) and deficit accumulation Frailty Index (proposed and validated by Rockwood and colleagues in the Canadian Study of Health and Aging) (22) were used for frailty assessment. These two instruments are different but considered complementary (23).

Modified Fried Frailty Phenotype

A modified version of Fried phenotype (1, 18) was used and included low body mass index (BMI) of <20kg/m² substituting for unintentional weight loss, slowest 20% in gait speed adjusted for height and gender and lowest 20% in grip strength adjusted for BMI and gender, as in previous analyses (19, 24). The participants were defined as frail if they satisfied at least three of the following five criteria, and pre-frail if they met one or two of the criteria: (1) BMI < 20 kg/m², (shrinking); (2) lowest 20% of grip strength taking into account gender and weight (weakness); (3) for three days or more during the last week the participant reported that ‘I felt that everything I did was an effort’ and/or ‘I could not get going’ according to the Center for Epidemiological Studies-Depression 10 (CES-D10) scale (25) (exhaustion); (4) time to walk 3 meters (10 feet) was in the lowest 20% taking into account gender and height (slowness), and (5) no walking outside home in the last two weeks, or the longest amount of time walking outside without sitting down to rest was less than 10 minutes (low activity) according to LIFE disability questionnaire responses (26, 27).

Deficit Accumulation Frailty Index (FI)

The second measure of frailty was a deficit accumulation FI of 66 items (28) using ASPREE annual data collected at study visits across multiple domains, including cognition, activity, functional engagement, mental health, medical comorbidities, concomitant medications, laboratory values and self-rated health status; where this construct was based on methods from Rockwood K. et al. (22). Participants were classified as non-frail (<=0.10), pre-frail (>0.10 and <=0.21) or frail (>0.21), in line with cut-offs used previously (29). Full details of the ASPREE FI are published elsewhere (28, 30).

Disability-Free Survival (DFS)

DFS was defined as the time to the first of any one of three events, including death, dementia (based on Diagnostic and Statistical Manual of Mental Disorders, fourth edition [DSM-IV] criteria) or persistent physical disability (31, 32). Mortality was confirmed from death information collected from at least two independent sources. Both immediate cause of death and major underlying illness causing the trajectory to death and the diagnosis of dementia were adjudicated by international expert panels. Persistent physical disability was considered to have occurred when a participant reported having an inability to perform, severe difficulty in performing, or requiring assistance to perform, any one of six basic activities of daily living that had persisted for at least six months. In addition, if BADL information was unavailable, the adjudicated eligibility for ‘admission to residential care’ was included in the definition of persistent physical disability endpoint (31, 33). The ASPREE study outcomes were evaluated by annual in-person visits, medical record reviews and by regular 6-month telephone calls (17).

Statistical analysis

The prevalences of MetS, non-frailty, pre-frailty and frailty were determined at baseline. Demographic data were described using means and standard deviations or percentages where appropriate and analyzed using analysis of variance (ANOVA) and chi-square, respectively. We assessed the collinearity between variables by running the variance inflation factor (VIF) analysis as well as examining the correlation matrix. There was no significant collinearity between variables. The association of factors with MetS and frailty categories was determined using multinomial logistic regression models, and relative risk ratios (RRRs) and 95% were reported. The potential confounders which were adjusted for included sociodemographic factors (age, gender, ethno-racial origin and education), lifestyle factors (smoking history and alcohol use) and chronic conditions/morbidities (chronic kidney disease, depression and previous cancer). Shortened DFS was examined according to both frailty scales, Fried phenotype and deficit accumulation FI. The association between concomitant MetS and frailty with DFS over the follow-up period was examined by Cox proportional-hazards regression model and reported as the hazard ratios (HRs) with 95% CIs. Proportional hazards assumptions were checked using Schoenfeld residuals. The final model was adjusted for socio-demographic factors, lifestyle factors and chronic conditions/morbidities. In the supplementary analyses of physical disability and dementia, all-cause mortality was treated as a competing risk as individual first events contributing to the composite, and Fine-Gray competing-risks regression analysis for sub-distribution hazard ratios (SHR) and 95% CIs was performed. Cumulative incidences were used to show event risks based on regression models. In sub-analyses, we examined the effect of sex, CVD and non-CVD mortality on these relationships. All p-values lower than 0.05 were considered statistically significant. Statistical analysis was performed using STATA, version 17 (34).

Ethics and trial registration

The ASPREE clinical trial is registered with the International Standard Randomized Controlled Trial Number Register (ISRCTN83772183) and ClinicalTrials.gov (NCT01038583). All participants signed informed consent. All federal and local regulations on human subjects’ research were followed with institutional approvals. The Intellectual Property and Ethics Committee of Monash University has approved this current project (Reference no. V6VVQTXZ; 29 November 2019) as well as by MUHREC (ethics #2021/30049).

Results

Among 18,264 participants, 49.9% met the criteria for MetS at baseline and of these, 41.6% were pre-frail and 2.6% were frail according to Fried phenotype compared with 36.3% pre-frail and 1.7% frail in those without MetS. According to FI, of those with MetS, 49.7% were pre-frail and 11.9% were frail while of those without MetS, 31.5% were pre-frail and 4.1% were frail. The distribution of correlates of MetS is shown in Table 1. Female sex, higher waist circumference, Hispanic/Latino ethno-racial origin, living alone, lower education (12 years or less), former smoking and current excess drinking were more common among participants with MetS.

Table 1.

Baseline characteristics of the ASPREE participants by MetS as defined according to ACC/AHA 2018 (N=18,264)

Baseline characteristics	Total	No MetS	MetS	P value
Participants, n (%)	18,264	9,149 (50.1)	9,115 (49.9)	<0.001
Fried frailty, n (%)				<0.001
Non-frail	10,760	5,675 (52.7)	5,085 (47.3)
Pre-frail	7,115	3,322 (46.7)	3,793 (53.3)
Frail	389	152 (39.1)	237 (60.9)
Frailty index n (%)				<0.001
Non-frail	9,387	5,886 (62.7)	3,501 (37.3)
Pre-frail	7,416	2,885 (38.9)	4,531 (61.1)
Frail	1,459	377 (25.8)	1,082 (74.2)
Age, year, mean, (SD)	75.1 (4.5)	75.2 (4.6)	75.0 (4.4)	<0.001
Age group, n (%)				<0.001
65-74 y	10,688	5,349 (50.1)	5,339 (49.9)
75-84 y	6,893	3,408 (49.4)	3,485 (50.6)
≥85 y	683	392 (57.4)	291 (42.6)
Gender, n (%)				0.001
Men	7,983	4,251 (53.3)	3,732 (46.8)
Women	10,281	4,898 (47.6)	5,383 (52.4)
Body mass index, kg/m2	28.1 (4.7)	26.5 (4.1)	29.7 (4.7)	<0.001
Waist circumference, cm, mean (SD)	97.1 (12.8)	92.8 (11.8)	101.4 (12.4)	<0.001
Living status, n (%)				0.138
Living alone	5,967	2,942 (49.3)	3,025 (50.7)
Living with others	12,297	6,207 (50.5)	6,090 (49.5)
Ethno-racial, n (%)				<0.001
Australian white	15,590	7,920 (50.8)	7670 (49.2)
US white	1,065	564 (53.0)	501 (47.0)
African-American	878	371 (42.3)	507 (57.7)
Hispanic/Latino	469	180 (38.4)	289 (61.6)
Others2	261	114 (43.7)	147 (56.3)
Education (y), n (%)				<0.001
12 years or less	10,464	4,922 (47.0)	5,542 (53.0)
More than 12 years	7,800	4,227 (54.2)	3,573 (45.8)
Smoking status, n (%)				<0.001
Current	701	353 (48.8)	348 (51.2)
Former	7,428	3,554 (47.9)	3,874 (52.1)
Never	10,135	5,242 (51.7)	4,893 (48.3)
Alcohol consumption, n (%)				<0.001
Never drinker	3,186	1,404 (44.1)	1,782 (55.9)
Former drinker	1,086	500 (46.0)	586 (54.0)
Current moderate drinker3	13,533	7,050 (52.1)	6,483 (47.9)
Current higher drinker4	459	195 (42.5)	264 (57.5)
Components of MetS5
Increased waist circumference, n (%)				<0.001
No	7,610	5,583 (73.4)	2,027 (26.6)
Yes	10,654	3,566 (33.5)	7,088 (66.5)
Elevated TG, n (%)				<0.001
No	10,665	8,663 (81.2)	2,002 (18.8)
Yes	7,599	486 (6.4)	7,113 (93.6)
Low HDL cholesterol, n (%)				<0.001
No	10,474	8,583 (82.0)	1,891 (18.1)
Yes	7,790	566 (7.3)	7,224 (92.7)
Hypertension, n (%)				<0.001
No	2,713	2,195 (80.9)	518 (19.1)
Yes	15,551	6,954 (44.7)	8,597 (55.3)
Hyperglycemia, n (%)				<0.001
No	12,628	8,044 (63.7)	4,584 (36.3)
Yes	5,636	1,105 (19.6)	4,531 (80.4)

Abbreviations: MetS: Metabolic syndrome; SD: Standard deviation; TG: Triglycerides; HDL-C: High density lipoprotein-cholesterol; Explanatory notes:

^1.Two participants were excluded because they had recorded a response to less than the minimum of 50 items for deficit accumulation FI

^2.Other ethno-racial groups include Asian, Native Hawaiian, other pacific islanders, Maori, American Indian, aboriginal or Torres Strait Islanders (TSI)

^3.Moderate alcohol drinker: no more than eight standard drinks a week and four drinks on any one day

^4.Current higher drinker: more than eight standard drinks a week and more than four drinks on any day

^5.The diagnostic criteria for metabolic syndrome are from 2018 AHA/ACC Guideline on the Management of Blood Cholesterol: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines; and require any three or more of the following five criteria: 1) triglycerides ≥175 mg/dL (≥2.0 mmol/L) or lipid-lowering treatment, 2) fasting glucose ≥ 100 mg//dL (5.6 mmol/L) or drug treatment for elevated glucose, 3) reduced high-density lipoprotein cholesterol: a) in men, < 40 mg/dL (1.0 mmol/L), b) in women, < 50 mg/dL (1.3 mmol/L), or lipid-lowering treatment 4) elevated blood pressure demonstrated by any of the following: a) systolic blood pressure ≥ 130 mm Hg or b) diastolic blood pressure ≥ 85 mm Hg, c) antihypertensive drug treatment in a patient with a history of hypertension, 5) increased waist circumference ≥102 cm for men and ≥88 cm for women.

The associations between MetS and frailty, using both frailty scales, are shown in Table 2. MetS was associated with a 22% more likelihood of being pre-frail (RRR: 1.22; 95% CI: 1.14, 1.30) and a 66% more likelihood of being frail (RRR: 1.66; 95% CI: 1.32, 2.08), according to Fried phenotype after adjustment (Model 2). This association was stronger using the deficit accumulation FI in the adjusted Model (RRR for pre-frailty: 2.69; 95% CI: 2.51, 2.89 and frailty: 5.12; 95% CI: 4.45, 5.90). All five components of MetS showed a statistically significant association with pre-frailty and frailty according to both Fried phenotype and FI except hypertension and hyperglycemia for Fried pre-frail and frail groups, respectively (Table 2).

Table 2.

Multinomial logistic regression showing the relationship between MetS, the components of MetS (ACC/AHA 2018) and pre-frailty/frailty

Characteristics	Non-frail	Model 1: Unadjusted analysis		Model 2: Adjusted analysis
		Pre-frail (RRR; 95% CI)	Frail (RRR; 95% CI)	Pre-frail (RRR; 95% CI)	Frail (RRR; 95% CI)
According to Fried Phenotype
MetS	Reference	1.27 (1.20, 1.35)	1.74 (1.41, 2.14)	1.22 (1.14, 1.30)	1.66 (1.32, 2.08)
Components of MetS
Increased waist circumference	Reference	1.34 (1.26, 1.43)	1.61 (1.30, 1.99)	1.34 (1.25, 1.44)	1.55 (1.23, 1.96)
Elevated TG	Reference	1.14 (1.08, 1.22)	1.48 (1.21, 1.82)	1.10 (1.03, 1.17)	1.42 (1.14, 1.77)
Low HDL-C	Reference	1.14 (1.07, 1.21)	1.51 (1.23, 1.85)	1.07 (1.00, 1.14)	1.38 (1.11, 1.72)
Hypertension	Reference	1.08 (1.00, 1.17)	2.26 (1.54, 3.31)	1.00 (0.91, 1.09)	2.03 (1.35, 3.05)
Hyperglycemia	Reference	1.26 (1.19, 1.35)	1.11 (0.90, 1.39)	1.25 (1.17, 1.34)	1.08 (0.86, 1.37)
According to Frailty Index
MetS	Reference	2.64 (2.48, 2.81)	4.83 (4.26, 5.46)	2.69 (2.51, 2.89)	5.12 (4.45, 5.90)
Components of MetS
Increased waist circumference	Reference	3.63 (3.40, 3.88)	8.67 (7.42, 10.13)	4.01 (3.72, 4.32)	10.94 (9.18, 13.03)
Elevated TG	Reference	2.15 (2.02, 2.29)	3.18 (2.84, 3.57)	2.18 (2.03, 2.34)	3.34 (2.94, 3.80)
Low HDL-C	Reference	2.00 (1.88, 2.13)	2.97 (2.65, 3.32)	1.91 (1.79, 2.05)	2.76 (2.42, 3.13)
Hypertension	Reference	1.83 (1.68, 2.01)	2.71 (2.22, 3.30)	1.78 (1.61, 1.96)	2.65 (2.13, 3.29)
Hyperglycemia	Reference	1.61 (1.50, 1.72)	2.37 (2.11, 2.65)	1.74 (1.62, 1.86)	2.78 (2.44, 3.16)

Abbreviations: MetS: Metabolic syndrome; RRR: Relative Risk Ratio; CI: Confidence Intervals; SD: Standard deviation; TG: Triglycerides; HDL-C: High density lipoprotein-cholesterol; Model 1: Unadjusted analysis; Model 2: Association between MetS and frailty adjusted for age, gender, ethno-racial origin, education, smoking history, alcohol intake and chronic conditions/morbidities (chronic kidney disease, depression, previous cancer history)

Table 3 shows Cox proportional hazards model for MetS according to MetS-specific frailty categories and shortened DFS over 4.7 years of follow-up. Fried pre-frailty (HR: 1.68; 95% CI:1.45, 1.94) and frailty (HR: 2.65; 95% CI: 1.92, 3.66) without co-existent MetS were associated with 1.7-fold and 2.7-fold risk of shortened DFS, respectively. The concomitant presence of MetS to Fried pre-frailty (HR: 1.70; 95% CI: 1.47, 1.96) or frailty (HR: 3.02; 95% CI: 2.29, 3.98) did not accentuate the risk of shortening of DFS before or after covariate adjustments; with similar results using FI (Table 3).

Table 3.

Cox proportional hazards regression for disability-free survival of participants with and without metabolic syndrome (ACC/AHA 2018) according to Fried phenotype and Frailty Index

MetS by Fried Phenotype	Model 1	Model 2
	HR (95% CI)	HR (95% CI)
	(n =18,264; DFS loss=1,736)	(n=17,142; DFS loss=1,656)
No MetS, Non-frail (n= 5,675)	Reference	Reference
No MetS, Pre-frail (n= 3,322)	2.28 (1.98, 2.61)	1.68 (1.45, 1.94)
No MetS, Frail (n= 152)	5.45 (4.00, 7.41)	2.65 (1.92, 3.66)
MetS, Non-frail (n= 5,085)	0.98 (0.84, 1.15)	0.96 (0.82, 1.13)
MetS, Pre-frail (n= 3,793)	2.19 (1.91, 2.51)	1.70 (1.47, 1.96)
MetS, Frail (n= 237)	5.49 (4.23, 7.11)	3.02 (2.29, 3.98)
MetS by Frailty Index	(n =18,262; DFS loss=1,735)	(n=17,140; DFS loss=1,655)
No MetS, Non-frail (n= 5,886)	Reference	Reference
No MetS, Pre-frail (n= 2,885)	2.34 (2.02, 2.69)	1.90 (1.63, 2.21)
No MetS, Frail (n= 377)	5.77 (4.66, 7.14)	4.05 (3.22, 5.11)
MetS, Non-frail (n= 3,501)	0.94 (0.79, 1.12)	1.00 (0.84, 1.19)
MetS, Pre-frail (n= 4,531)	1.63 (1.42, 1.87)	1.50 (1.29, 1.74)
MetS, Frail (n= 1,082)	4.19 (3.55, 4.93)	3.17 (2.64, 3.81)

Abbreviations: HR: Hazard Ratios; CI: Confidence Intervals; DFS: disability-free survival; Model 1: Unadjusted analysis; Model 2: Adjusted for age, gender, ethno-racial origin, education, smoking history, alcohol intake and chronic conditions/morbidities (chronic kidney disease, depression, previous cancer history).

The cumulative hazard estimates of shortened DFS are shown in Figures 1 and 2. Supplementary tables 1-3 and Supplementary figures 1-6 show the associations between concomitant MetS and pre-frailty/frailty with the components of DFS individually (i.e., dementia, physical disability and all-cause mortality) according to both frailty scales. The results indicate that the co-existence of MetS with pre-frailty/frailty did not increase the risk of loss of DFS than that of the association between pre-frailty/frailty alone; and all three components of DFS, with physical disability having the highest risk association. In Fine-Gray competing-risks regression analysis, the addition of MetS resulted in a significant increase in risk of dementia in those who were frail according to Fried phenotype (but not for FI). For physical disability, the addition of MetS did not affect the strong association between pre-frailty/frailty and physical disability, regardless of the frailty scale. (Supplementary tables 1-3).

Figure 1.

Nelson-Aalen cumulative incidence of loss of disability-free survival (DFS) +/− metabolic syndrome according to Fried Phenotype

Figure 2.

Nelson-Aalen cumulative incidence of loss of disability-free survival (DFS) +/− metabolic syndrome according to Frailty Index

Stratified Cox regression by sex for DFS in those with MetS and frailty is shown in Supplementary Table 4. Effect sizes for the shortening of DFS among males and females did not show significant differences. Stratified Cox proportional hazards regression for CVD- and non-CVD mortality demonstrated that MetS alone did not increase CVD- or non-CVD mortality risk in this cohort of older adults. Pre-frailty/frailty alone increased both CVD- and non-CVD mortality (except Fried pre-frail group which did not show increased risk for CVD- mortality). The concomitant presence of MetS and pre-frailty/frailty did not further increase the mortality risks. (Supplementary Table 5).

Discussion

In this study, 50% of the participants met the criteria for MetS; and MetS was associated with 22% and 66% more likelihood of being pre-frail and frail, respectively, at baseline. Moreover, pre-frailty and frailty without concomitant MetS shortened DFS as well as increased the risk of each component of DFS i.e., incident dementia, persistent physical disability and mortality. Co-existent MetS with pre-frailty/frailty did not change the risks of shortened DFS.

Prevalence of MetS

The prevalence estimates of MetS in the literature across populations vary, largely based on the criteria used to define MetS (35). In our study, according to ACC/AHA 2018 definition of MetS, the prevalence was higher, even among these relatively healthy community-dwelling older participants, than estimated Australian, US or global prevalences (36-38).

Association between MetS, its components and frailty

Earlier studies found that persons with MetS have an increased risk of early physical limitations and higher vulnerability to adverse health outcomes (39, 40). More recently, MetS was associated with frailty in a cohort study (8) and two cross-sectional analyses (4, 7). However, these studies used only Fried phenotype to assess frailty. By considering frailty defined with two scales, our study adds further robust evidence for the relationship between MetS and frailty in a large cohort of older individuals free of established major CVD, cognitive impairment or physical disability.

In our study, all five components of MetS demonstrated a positive association with pre-frailty and frailty, according to both frailty scales (except hypertension and hyperglycemia for Fried pre-frail and frail groups, respectively). In previous studies, the association between components of MetS and frailty varied from study to study. Our finding that increased waist circumference, reflecting abdominal obesity, had the largest relative risk ratio of all components is in accordance with the findings of previous studies (3, 4, 41-43). For hypertension, few studies found that persons with frailty had higher odds (42, 43), while others did not find any association (4). Our analysis had the benefit of large numbers in the cohort and a high proportion of participants with hypertension, strengthening our findings. Regarding hyperglycemia, there is evidence in the literature of a significant association between hyperglycemia and frailty (8, 44), though not all studies reported this relationship (4). All those studies used Fried phenotype only for frailty assessment whereas we confirmed the associations with the FI scale. Additionally, our study results showed an association between elevated TG and low HDL-C with frailty according to Fried phenotype and FI. In contrast, previous studies have had mixed findings according to Fried phenotype (4, 42, 43).

MetS, frailty and DFS

Previously no study examined the relationship between concomitant presence of MetS and pre-frailty/frailty with a composite outcome like DFS though some studies examined the associations with MetS or frailty with different components of DFS individually. For example, a prior study showed that frailty was associated with incident Alzheimer’s disease (AD) and the rate of cognitive decline in older persons (14) and frailty also impacted physical disability (15). However, no significant pooled association was found between MetS and incident dementia and AD though MetS increased the incidence of vascular dementia (12). Further, MetS was associated with an increased risk of incident disability and deteriorated functional performance (13). Regarding mortality, earlier studies also yielded mixed findings due to MetS in frail individuals (10, 45, 46). In an analysis of the US National Health and Nutrition Examination Survey (NHANES), MetS was associated with increased mortality risk in younger subjects (20 years to 65 years) but not among older adults (above 65 years), whereas a 41-item FI better predicted mortality than did the MetS, regardless of age (11). This age discrepancy could be due to a healthy survivor effect, and the study’s authors concluded that frailty was more suitable to predict mortality than MetS among older people. On this background, our study adds robust evidence that MetS alone did not increase the risk of shortened DFS or any of its components i.e., incident dementia, persistent physical disability or all-cause mortality; whereas pre-frailty/frailty according to either Fried phenotype or FI did increase the risks. Furthermore, co-existence of MetS with pre-frailty/frailty did not boost the risk of reduced DFS, irrespective of sex. Moreover, the additional burden of MetS did not increase the risk of CVD-related mortality in this cohort of well-functioning older adults possibly because of high rates of treatment of hypertension and diabetes, and statin use.

Strengths and limitations

Our study’s strengths include a large community-based cohort of older adults from two countries, with a substantial number of African-Americans and Hispanics/Latinos, and detailed in-person measurements allowing rigorous characterization of frailty by two scales and of MetS. Study participants were followed longitudinally for about five years and were from geographically and culturally diverse backgrounds. Nevertheless, the generalizability may not extend to other countries with different ethnicities. Furthermore, we used modified criteria for the Fried phenotype definition though many studies have assessed the shrinking or weight loss criterion by BMI (24) and our modified Fried criteria were previously published in different articles (18, 19, 28, 30).

Conclusions

In conclusion, this study showed that in a cohort of community-dwelling older adults, free from major CVD, dementia or physical disability at baseline, about half of the participants had MetS. The presence of MetS was significantly associated with pre-frailty and frailty according to two frailty scales. MetS alone did not increase the risk of shortened DFS or any of its components i.e., incident dementia, persistent physical disability or all-cause mortality according to either frailty scale over a median 4.7 years of follow-up. Furthermore, the presence of MetS in those who were pre-frail or frail did not increase the frailty-related risk of shortened DFS, or its components i.e., dementia, physical disability or mortality. Frailty might be a better predictor of DFS than MetS alone or in combination, suggesting that it may be beneficial to integrate frailty status in well-being assessment of community-dwelling older adults for a healthy life span.

Key Points.

• Half of the ASPREE participants had MetS, and the presence of MetS was associated with pre-frailty and frailty in healthy community-dwelling older people

• Pre-frailty or frailty without MetS shortened DFS but co-existent MetS with pre-frailty or frailty did not increase the risk.

• Frailty might be a better predictor of disability-free survival than MetS, suggesting that it could be beneficial to incorporate frailty status in health assessment of community-dwelling older adults for a healthy life span.