Abstract
Objectives
Insulin resistance determined by Homeostasis Model of Insulin Resistance (HOMA-IR) has been associated with functional decline in non-diabetic older subjects. However, insulin is not routinely assessed. The study evaluated the predictive value of non-insulin-dependent IR surrogates on functional decline in non-diabetic older men and women.
Design and participants
Prospective cohort study over 5 years. The study included 615 older participants from the Toledo Study of Healthy Aging.
Methods
Frailty was assessed by the Frailty Trait Scale-5 (FTS-5) at baseline and after 5 years follow-up. 193 subjects experienced functional decline (2.5-point reduction in the FTS-5 score). Multivariate regression models analysed the effect of five described IR surrogates on functional decline considering potential confounders.
Results
Among evaluated IR proxies, triglyceride glucose-body mass index (TyG-BMI) and HOMA-IR were significantly associated with an increased risk of functional decline (odd ratio (95% confidence interval) TyG-BMI: 1.16 (1.05, 1.28), p = 0.0035 and HOMA-IR: 1.59 (1.15, 2.21), p = 0.0056) among all participants. When stratified by gender, HOMA-IR was related to functional decline in men [2.02 (1.13, 3.59), p = 0.0173] and TyG-BMI in women [1.19 (1.05, 1.35), p = 0.0057].
Conclusions
Only TyG-BMI index mimics the predictive capacity of insulin-based IR marker. The predictive ability of IR indexes is gender-specific, being TyG-BMI the only index able to predict functional decline in women and HOMA-IR in men.
Keywords: Insulin resistance, Functional decline, Frailty, HOMA-IR, Triglyceride glucose-body mass index
1. Introduction
Frailty is a multisystemic syndrome that, since its inception, has provided significant insights into the common decline in function seen in aging individuals. One of its key benefits is the ability to identify those aged people at risk for severe adverse events, such as death, disability, hospitalization, or institutionalization. This identification offers a chance to halt and even reverse the progression of non-catastrophic disability [1]. Therefore, a better understanding of the factors related to frailty in an aging population could aid in preserving functional independence and preventing frailty, which is a critical public health priority.
Insulin resistance (IR) is a critical factor in the pathogenesis of various metabolic disorders and is increasingly recognized for its role in the aging process and age-related conditions including frailty [2,3]. Previous studies have shown a positive association between insulin resistance assessed by the Homeostasis Model of Insulin Resistance (HOMA-IR) and frailty both in cross-sectional and prospective studies [[4], [5], [6]], suggesting that IR may play a role in the pathological development of frailty. In fact, skeletal muscle is the main organ for insulin-induced glucose metabolism and the loss of skeletal muscle has been linked to both IR [7] and frailty [8]. Furthermore, diminished insulin sensitivity is closely associated with age-related changes in body composition, which primarily include an increase in fat mass and a decrease in lean mass [9,10]. Those latter changes seem to be tightly linked to frailty. In this sense, recent research identified trunk fat mass but not lean mass as the mediator of the association of insulin resistance with the risk of functional decline at 2.99 years of follow-up in non-diabetic community dwelling Spanish older subjects [11].
Of importance to note that chronic inflammation plays a crucial role in aging and frailty, acting as a key mediator [12,13]. Additionally, IR and IR-related conditions such as type 2 diabetes and prediabetes syndrome are tightly associated with inflammation [[14], [15], [16]].Therefore, evaluating the assocation between two conditons closely linked to inflammation in non-diabteic older subjects is well justified.
IR is commonly assessed by the HOMA-IR that depends on measuring serum insulin levels. However, despite its clinical relevance, insulin levels are not routinely assessed in everyday clinical practice. Alternatively, other non-insulin-based indices including the triglyceride glucose index (TyG), the triglyceride glucose-body mass index (TyG-BMI), the triglyceride to high-density lipoprotein- cholesterol ratio (TG/HDLc), and metabolic score for insulin resistance (METS-IR), have become recognized as reliable and sensitive indicators of IR in recent years [17,18]. In this sense, TyG index, which is calculated based on fasting glucose and triglyceride, is considered a reliable surrogate of IR for predicting diabetes mellitus, cardiovascular disease and mortality [19,20]. Moreover, Yin et al., have recently shown that increased TyG index and some of its related parameters (TyG-BMI) are correlated with prevalent frailty in a cross-sectional approach [21]. Albeit some studies have reported a positive association between some of IR surrogates and frailty presence, no available evidence is present regarding the predictive capacity of non-insulin-based IR indices on functional decline among non-diabetic older adults.
In addition, of importance to note that there is a sex-based difference in whole body insulin resistance [22]. For instance, women exhibit greater insulin sensitivity than men, but this metabolic advantage tends to diminish following menopause or when insulin resistance progresses to hyperglycemia and diabetes. These differences could be attributed to differences in body fat and muscle mass distribution and function, in inflammation and in sex hormones between males and females [22]. Furthermore, the prevalence of frailty is higher in women compared to men [23]. Therefore, it is important to evaluate if insulin resistance proxies are gender-specific in their predictive capacity for functional decline among older adults.
Taken as a whole, the main aim of this study is to evaluate the predictive value of various IR surrogates on functional decline in non-diabetic older men and women.
2. Materials and methods
2.1. Study participants
This longitudinal study analyzed data from participants in the Toledo Study of Healthy Aging (TSHA), a Spanish cohort focused on analyzing cognitive and functional determinants in community-dwelling individuals aged 65 years or older from the province of Toledo, Spain, with well-established methods that have been previously reported [24]. Ethical approval of the study protocol and informed consent were obtained from the Clinical Research Ethics Committee of the Toledo Hospital Complex, Spain. All participants provided written informed consent.
The study was conducted in accordance with the Declaration of Helsinki and approved by the Clinical Research Ethics Committee of the Complejo Hospitalario de Toledo (Toledo Study for Healthy Aging cohort; 22/May 30th, 2005).
For the present analysis, we used the baseline data from the first wave (2006–2009). The outcome (functional decline) was evaluated after a median follow-up period of 5 years (range 4.1–5.9 years) until the second wave analysis (2011–2013). A total of 615 non-diabetic older subjects with complete data at baseline were included in the study.
Subjects with diabetes were excluded from the study. Participants were classified as diabetic if they have been previously diagnosed with diabetes. To ensure accurate classification, individuals who self-reported having diabetes (with clinical records checked when necessary) or were taking any medications listed under A10 (drugs used in diabetes) in the Anatomical Therapeutic Chemical (ATC) classification system were excluded.
2.2. Study main variables
2.2.1. Insulin resistance proxies
The following insulin resistance proxies were assessed:
Triglyceride-glucose index (TyG) was calculated as follow: natural logarithm (fasting triglyceride [mg/dL] * fasting glucose [mg/dL]/2) [25].
Triglyceride to high-density lipoprotein cholesterol (TG/HDLc) was measured by dividing fasting triglyceride (mg/dL) by the fasting HDLc (mg/dL) [26].
Triglyceride glucose-body mass index (TyG-BMI) as calculated as follow: TyG x body mass index [Kg/m2] [27].
METS-IR was calculated as natural logarithm ((2* fasting glucose [mg/dL])+ fasting triglycerides [mg/dL])* body mass index [Kg/m2])/(Ln(high-density lipoprotein cholesterol) [28].
Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) was calculated as follows: fasting insulin (mU/L) x fasting glucose (mg/dL)/405 [29]. HOMA-IR was natural log transformed given its skewed distribution. One unit increment in ln HOMA-IR was considered to compute odd ratios. Subjects with HOMA-IR values greater than 10 (suggesting a non-fasting state) were excluded.
2.2.2. Adjusting variables
The adjusting variables were selected according to previous research and biological plausibility for confounding effects. The adjusting variables included: age, sex, and education level. For education level participants were divided into three categories: no education, basic education (incomplete primary school), and middle or advanced education (completed primary school or higher). The reference category for adjusting by educational level was no education. Additional adjusting variables included cardiovascular disease (CVD) including myocardial infarction, congestive heart failure and angina pectoris, cerebrovascular disease (CeVD) including stroke and transient ischemic attack. Renal function was evaluated by estimated glomerular filtration rate (eGFR) calculated by using the creatinine-based Chronic Kidney Disease Epidemiology collaboration (CKD-EPI) formula [30]. The number of prescription and non-prescription drugs within the ATC classification system taken by the participant was calculated. Waist-to-hip ratio (WHR) was calculated by dividing waist circumference (cm) by hip circumference (cm). The functional status for basic activities of daily living was assessed using the Katz Index [31], which evaluates the dependence in bathing, dressing, toileting, transferences, continence and feeding. A patient was considered disabled if they were dependent for any of these activities.
All the above-mentioned data were assessed at wave 1 visit.
2.2.3. Functional decline
Functional decline was evaluated according to the Frailty Trait Scale-5 (FTS-5) as previously described [32]. Briefly, the FTS-5 considers the evaluation of five dimensions including energetic balance and nutrition (BMI criteria), physical activity (the Physical Activity Scale for the Elderly (PASE) score), nervous system (Romberg test score), strength (grip strength dynamometry), and walk speed (walk speed in 3 m). Each item ranges from 0 (the best) to 10 (the worst).
FTS-5 score was assessed at baseline (wave 1) and after a median follow-up period of five years (wave 2). Functional decline during follow-up for all subjects was determined as the worsening in 2.5 points (5% of maximum score) for the FTS score.
2.3. Statistical analysis
Continuous variables were presented as mean and standard deviation (SD) while number and percentage (N, %) was provided for categorical variables. Insulin resistance proxies were expressed as median ± interquartile range [IQR]. Differences between groups were tested using Mann-Whitney and Chi-square test, respectively.
To assess the associations between IR proxies and functional decline a multivariate logistic regression model was applied adjusted for the following variables: age, gender, education level, CVD, CeVD, eGFR, number of medications, WHR and its interaction with sex, disability and baseline FTS-5 scores. Spearman correlation was used to analyze the correlation between non-insulin dependent IR proxies and HOMA-IR.
Analyses were performed using the statistical package R for Windows (Vienna, Austria) version 3.6.1, and a p value <0.05 was considered statistically significant in all analyses.
3. Results
3.1. Baseline characteristics of study subjects
Table 1 presents the descriptive data of the 615 non-diabetic study participants (average age: 72.87 ± 4.48 years, 57.7% females) from TSHA included in the analysis. Of these, 193 participants experienced functional decline as determined by a 2.5-point increase in their FTS-5 score at follow-up. Compared to those without functional decline, this group had lower values of estimated glomerular filtration rate, took higher average number of medications, had lower educational level, and had a lower baseline FTS-5 score (p < 0.05). However, there were no significant differences between the two groups with respect to other parameters evaluated such as comorbidities (cardiovascular and cerebrovascular disease), body composition parameters (BMI and WHR), and disability for basic activities of daily living evaluated by the Katz Index.
Table 1.
Baseline characteristics of the study population. Data are compared depending on the worsening in 2.5 points in the FTS-5 score.
| No worsening | Worsening | P value | |
|---|---|---|---|
| Number | 422 | 193 | |
| Age, years | 72.65 (4.49) | 73.36 (4.41) | 0.0629 |
| Gender (men) | 187 (44.31) | 73 (37.82) | 0.1306 |
| CVD | 32 (7.58) | 20 (10.36) | 0.2502 |
| CeVD | 16 (3.80) | 4 (2.07) | 0.2628 |
| eGFR (mL/min/1.73 m2) | 81.47 (13.40) | 77.51 (14.39) | 0.0004 |
| Number of medications | 3.51 (2.52) | 4.12 (2.72) | 0.0116 |
| BMI (kg/m2) | 29.08 (4.27) | 29.19 (4.81) | 0.8129 |
| WHR | 0.93 (0.09) | 0.93 (0.10) | 0.9531 |
| Glucose (mg/dL) | 98.52 (14.84) | 98.91 (12.31) | 0.2033 |
| Education level | |||
| No school | 296 (70.48) | 122 (63.21) | 0.0009 |
| Primary education incomplete | 53 (12.62) | 47 (24.35) | |
| Middle/advanced education | 71 (16.90) | 24 (12.44) | |
| Basal FTS-5 score | 19.20 (6.47) | 15.15 (6.49) | 0.0000 |
| Katz Index score | 5.95 (0.29) | 5.98 (0.14) | 0.3204 |
| Disability (BADL) | 60 (14.39) | 23 (12.11) | 0.4477 |
| Insulin resistance proxies | |||
| TyG index | 4.57 [4.43, 4.74] | 4.62 [4.49, 4.78] | 0.0196 |
| TG/HDLc ratio | 1.83 [1.30, 2.71] | 2.02 [1.35, 3.02] | 0.1267 |
| TyG-BMI index | 131.47 [118.91, 147.14] | 133.51 [120.67, 149.08] | 0.2762 |
| METS-IR | 0.35 [0.32, 0.36] | 0.34 [0.32, 0.36] | 0.0535 |
| Ln HOMA-IR | 0.67 [0.31, 1.10] | 0.79 [0.32, 1.28] | 0.0536 |
Numerical variables are expressed as mean ± standard deviation (SD) while discrete variables are expressed as number and percentages (in parentheses). Insulin resistance proxies are expressed as median ± interquartile range [IQR].
BADL: Basic Activities for Daily Living; BMI: body mass index; CeVD: cerebrovascular disease; CVD: cardiovascular disease; eGFR: estimated glomerular filtration rate; FTS-5: frailty trait scale-5; HDLc: High density lipoprotein cholesterol; HOMA-IR: homeostasis model assessment of insulin resistance; METS-IR: Metabolic score for IR; TyG: triglyceride-glucose; WHR: waist-to-hip ratio. Significant associations are highlighted in bold.
Regarding IR surrogates measured, older participants who suffered functional decline at follow-up presented higher values of TyG index than those of the no worsening group (p < 0.05). By contrast, albeit METS-IR and HOMA-IR tended to display lower and higher values respectively in the worsening group compared to participants who did not experience functional deterioration at follow-up, these differences did not reach statistical significance. Furthermore, no significant differences were detected for TyG-BMI and the TG/HDLc ratio between subjects who experienced functional decline and those who did not (p > 0.05) (Table 1).
3.2. Association of IR proxies with functional decline among non-diabetic older subjects
Among the IR- surrogates analyzed, the TyG-BMI index showed a significant positive association with an increased risk of functional decline in non-diabetic older adults from TSHA over a median follow-up period of 5 years after adjusting for age, gender, CVD, CeVD, eGFR, WHR, the number of medications, educational level, disability and basal FTS-5 (odd ratio (95% confidence interval), OR 1.16 (1.05–1.28), p = 0.0035) (Table 2). Additionally, one-logarithmic unit increase in HOMA-IR was linked to a higher risk of functional deterioration after adjusting for the same confounders [1.59 (1.15–2.21), p = 0.0056]. Although the TyG index was related to an increased risk of functional impairment among older participants, it did not reach statistical significance (p = 0.0626). Furthermore, no significant correlation was observed between TG/HDLc ratio or METS-IR index and incident worsening in FTS-5 score (Table 2).
Table 2.
Association of different insulin resistance proxies with functional decline over 5 years follow-up in non-diabetic older subjects form the Toledo Study of Healthy Aging.
| Insulin resitance proxies | OR (95%, CI) for functional decline | p value |
|---|---|---|
| TyG index | 2.37 (0.96, 5.87) | 0.0626 |
| TG/HDLc ratio | 1.09 (0.94, 1.25) | 0.2602 |
| TyG-BMI index | 1.16 (1.05, 1.28) | 0.0035 |
| METS-IR | 1.30 (0.44, 3.88) | 0.6345 |
| Ln HOMA-IR | 1.59 (1.15, 2.21) | 0.0056 |
BMI: body mass index; HDLc: high density lipoprotein cholesterol; HOMA-IR: homeostasis model assessment of insulin resistance. METS-IR: metabolic score for IR; TyG: triglyceride-glucose.
Data are expressed as odds ratio (95%, confidence interval) per 1-unit increment for IR surrogates except for the TyG-BMI index, which is expressed per 10-unit increment, and HOMA-IR which is expressed per 1-logarithmic unit increment. Significant associations are highlighted in bold.
These observations were evidenced regardless of the fact that all non-insulin dependent IR indices significantly correlated to IR determination by HOMA-IR (Supplementary Table S1).
3.3. TyG-BMI index stands out as the sole IR surrogate related to incident functional decline in non-diabetic older women
Baseline characteristics of male and female subjects from TSHA are presented in Supplementary Table S2. Briefly, women took a higher number of medications, exhibited higher levels of obesity and lower WHR values compared to men. In addition, women had a lower educational level, higher basal FTS-5 scores and greater disability for BADL. Regarding IR proxies, women had significant higher TyG-BMI index and lower METS-IR index values compared to men.
Stratified analysis by gender were used to further explore the correlation of IR proxies with functional decline among non-diabetic older women and men from the TSHA. Descriptive data of the 355 women included in the functional evaluation analysis are depicted in Table 3. A total of 120 females experienced an increase in 2.5-point of the FTS-5 score after a median follow-up period of 5 years. Similar to that observed in the whole sample, when compared to the no worsening group, women with functional decline had lower eGFR values, were on higher medication, had less education level and lower initial FTS-5 scores (p < 0.05).
Table 3.
Baseline characteristics of the female and male participants experiencing or not a worsening in 2.5 points in the FTS-5 score.
| Women |
Men |
|||
|---|---|---|---|---|
| No worsening | Worsening | No worsening | Worsening | |
| Number | 235 | 120 | 187 | 73 |
| Age, years | 72.52 (4.13) | 73.17 (4.42) | 72.81 (4.92) | 73.68 (4.41) |
| CVD | 13 (5.53) | 12 (10.00) | 19 (10.16) | 8 (10.96) |
| CeVD | 8 (3.42) | 2 (1.67) | 8 (4.28) | 2 (2.74) |
| eGFR (mL/min/1.73 m2) | 82.62 (13.62) | 76.06 (15.34)*** | 80.02 (13.02) | 79.89 (12.41) |
| Number of medications | 3.94 (2.48) | 4.60 (2.66)* | 2.97 (2.47) | 3.33 (2.64) |
| BMI (kg/m2) | 29.81 (4.58) | 29.98 (5.07) | 28.16 (3.66) | 27.90 (4.07) |
| WHR | 0.89 (0.08) | 0.90 (0.07) | 0.97 (0.08) | 0.98 (0.12) |
| Glucose (mg/dL) | 97.11 (13.73) | 98.18 (13.53) | 100.12 (9.98) | 100.30 (15.99) |
| Education level | ||||
| No school | 164 (70.09) | 74 (61.67)** | 132 (70.97) | 48 (65.75) |
| Primary education incomplete | 36 (15.38) | 36 (30.00) | 17 (9.14) | 11 (15.07) |
| Middle/advanced education | 34 (14.53) | 10 (8.33) | 37 (19.89) | 14 (19.18) |
| Basal FTS-5 score | 21.06 (6.38) | 16.43 (6.43)*** | 16.86 (5.79) | 13.03 (6.04)*** |
| Katz Index score | 5.95 (0.27) | 5.97 (0.16) | 5.95 (0.30) | 5.99 (0.12) |
| Disability (BADL) | 42 (18.18) | 19 (16.10) | 18 (9.68) | 4 (5.56) |
| Insulin resistance proxies | ||||
| TyG index | 4.57 [4.45, 4.75] | 4.65 [4.50, 4.79]* | 4.56 [4.41, 4.74] | 4.58 [4.48, 4.75] |
| TG/HDLc ratio | 1.77 [1.26, 2.67] | 2.17 [1.40, 3.17]* | 2.00 [1.36, 2.82] | 1.86 [1.33, 2.75] |
| TyG-BMI index | 134.32[122.48, 151.24] | 138.92[124.70, 152.32] | 126.11 [116.32, 139.37] | 125.94 [116.05, 141.26] |
| METS-IR | 2.26 [2.15, 2.38] | 2.31 [2.19, 2.42] | 2.35 [2.22, 2.45] | 2.29 [2.19, 2.45] |
| Ln HOMA-IR | 1.98 [1.37, 3.10] | 2.40 [1.35, 3.62] | 0.67 [0.31, 1.06] | 0.75 [0.35, 1.24] |
Numerical variables are expressed as mean ± standard deviation (SD) while discrete variables are expressed as number and percentages (in parentheses). Insulin resistance proxies are expressed as median ± Interquartile range [IQR].
BADL: Basic Activities for Daily Living; BMI: body mass index; CeVD: cerebrovascular disease; CVD: cardiovascular disease; eGFR: estimated glomerular filtration rate; FTS-5: frailty trait scale-5; HDLc: High density lipoprotein cholesterol; HOMA-IR: homeostasis model assessment of insulin resistance; METS-IR: Metabolic score for IR; TyG: triglyceride-glucose; WHR: waist-to-hip ratio. *p < 0.05, **p < 0.01, ***p < 0.0001 versus non-worsening group.
With respect to IR surrogates, older women with functional decline at follow-up presented higher values of TyG index and TG/HDLc ratio compared to the no worsening group. By contrast, TyG-BMI index, METS-IR, and Ln HOMA-IR were not significantly different among the two groups of participants evaluated (Table 3).
The analysis of the relationship between IR proxies and incident FTS worsening in women identified a significant positive association between the TyG-BMI index and incident functional deterioration after adjusting for confounders [OR 1.19 (1.05–1.35), p = 0.0057] (Table 4). Notably, female participants in the highest quartile of the TyG-BMI index were at significant greater risk of functional decline compared to those in the lowest quartile [TyG-BMI index Q4 vs Q1, OR 2.76 (1.21, 6.30), p = 0.0158] (Table 4). In contrast, although the TyG index and TG/HDLc ratio appeared to be associated with a higher risk of functional impairment [OR 3.14 for TyG and 1.21 for TG/HDLc], these associations did not reach statistical significance. None of the other evaluated IR surrogates was significantly related to incident functional decline in non-diabetic older women (Table 4).
Table 4.
Association of different insulin resistance proxies with functional decline over 5 years follow-up in non-diabetic older women from the Toledo Study of Healthy Aging.
| IR- proxies | Number events | OR (95%, CI) for functional decline | p value |
|---|---|---|---|
| Q1 (lowest) | 24 | ref | |
| Q2 | 23 | 0.70 (0.32, 1.52) | 0.3680 |
| Q3 | 37 | 1.30 (0.62, 2.73) | 0.4800 |
| Q4 (highest) | 36 | 1.51 (0.73, 3.16) | 0.2695 |
| TyG index | 120 | 3.14 (0.91, 10.80) | 0.0690 |
| Q1 (lowest) | 25 | ref | |
| Q2 | 19 | 0.53 (0.24, 1.19) | 0.1221 |
| Q3 | 39 | 1.29 (0.62, 2.69) | 0.4905 |
| Q4 (highest) | 37 | 1.35 (0.64, 2.84) | 0.4318 |
| TG/HDLc ratio | 120 | 1.21 (1.00, 1.48) | 0.0534 |
| Q1 (lowest) | 26 | ref | |
| Q2 | 26 | 0.74 (0.34, 1.61) | 0.4506 |
| Q3 | 39 | 1.70 (0.79, 3.65) | 0.1729 |
| Q4 (highest) | 29 | 2.76 (1.21, 6.30) | 0.0158 |
| TyG-BMI index | 120 | 1.19 (1.05, 1.35) | 0.0057 |
| Q1 (lowest) | 23 | ref | |
| Q2 | 28 | 0.76 (0.35, 1.66) | 0.4945 |
| Q3 | 33 | 1.09 (0.50, 2.35) | 0.8316 |
| Q4 (highest) | 36 | 1.71 (0.79, 3.69) | 0.1722 |
| METS-IR | 120 | 3.82 (0.85, 17.16) | 0.0799 |
| Q1 (lowest) | 30 | ref | |
| Q2 | 23 | 0.66 (0.31, 1.42) | 0.2898 |
| Q3 | 29 | 0.99 (0.47, 2.07) | 0.9788 |
| Q4 (highest) | 38 | 1.43 (0.69, 2.97) | 0.3328 |
| Ln HOMA-IR | 120 | 1.40 (0.93, 2.13) | 0.1100 |
BMI: Body mass index; HDLc: High density lipoprotein cholesterol; HOMA-IR: homeostasis model assessment of insulin resistance. METS-IR: Metabolic score for IR; TyG: triglyceride-glucose.
Data are expressed as odds ratio (95%, confidence interval) per 1-unit increment for IR surrogates except for the TyG-BMI index, which is expressed per 10-unit increment, and HOMA-IR which is expressed per 1-logarithmic unit increment. Significant associations are highlighted in bold.
To further explore the potential role of triglyceride and BMI independently of glucose levels in the pathophysiology of functional decline, additional analysis was conducted considering only these two parameters (Triglyceride x BMI). After adjusting for confounders, Triglyceride-BMI (T-BMI) index showed a significant positive association with functional decline in women [OR 1.22 (1.03–1.44), p = 0.018]. Female participants in the third and fourth quartile were at greater risk of functional worsening compared to those in the lowest quartile [T-BMI index Q3 vs Q1, OR 2.29 (1.05, 5.00), p = 0.0378; Q4 vs Q1, OR 2.71 (1.22, 6.03), p = 0.0143].
3.4. Only HOMA-IR is prospectively associated with incident functional decline in non-diabetic older men
The analysis of the association between proxies of IR with functional decline was carried out in a total of 260 men from the TSHA. The men’s baseline characteristics are shown in Table 3. Briefly, 73 subjects (28%) experienced a 2.5-point worsening in the FTS-5 score over a period of 5 years. Those subjects had lower FTS-5 scores at baseline compared with the no worsening group. None of the other parameters depicted in Table 3 showed statistical differences between the male participants who experienced functional decline compared with the no worsening group.
Furthermore, among the different indexes analyzed, only HOMA-IR showed a significant association with an increased risk of frailty worsening, as indicated by an increase in 2.5 points in the FTS-5 score at five years follow-up [OR 2.02 (1.13, 3.59), p = 0.0173] (Table 5). Albeit male participants in the fourth quartile exhibited higher odds ratio for functional decline vs the first quartile [Ln HOMA-IR Q4 vs Q1, OR 2.09 (0.89, 4.91)], this increase did not reach statistical significance, p = 0.09. In contrast, none of the other evaluated non-insulin dependent surrogates was significantly related to incident functional decline in non-diabetic older men over five-years follow-up (Table 5).
Table 5.
Association of different insulin resistance proxies with functional decline over 5 years follow-up in non-diabetic older men from the Toledo Study of Healthy Aging.
| IR- proxies | Number events | OR (95%, CI) for functional decline | p value |
|---|---|---|---|
| Q1 (lowest) | 15 | ref | |
| Q2 | 21 | 1.34 (0.57, 3.17) | 0.5017 |
| Q3 | 18 | 1.91 (0.78, 4.65) | 0.1545 |
| Q4 (highest) | 19 | 1.17 (0.48, 2.82) | 0.7321 |
| TyG index | 73 | 2.40 (0.54, 10.70) | 0.2496 |
| Q1 (lowest) | 20 | ref | |
| Q2 | 21 | 1.37 (0.59, 3.17) | 0.4652 |
| Q3 | 12 | 0.92 (0.37, 2.28) | 0.8605 |
| Q4 (highest) | 20 | 1.15 (0.48, 2.78) | 0.7525 |
| TG/HDLc ratio | 73 | 1.03 (0.79, 1.33) | 0.8477 |
| Q1 (lowest) | 18 | ref | |
| Q2 | 20 | 0.62 (0.26, 1.49) | 0.2899 |
| Q3 | 16 | 0.96 (0.39, 2.36) | 0.9284 |
| Q4 (highest) | 19 | 1.69 (0.67, 4.28) | 0.2696 |
| TyG-BMI index | 73 | 1.12 (0.93, 1.34) | 0.2341 |
| Q1 (lowest) | 23 | ref | |
| Q2 | 17 | 0.62 (0.26, 1.44) | 0.2652 |
| Q3 | 15 | 0.60 (0.25, 1.45) | 0.2597 |
| Q4 (highest) | 18 | 0.91 (0.39, 2.13) | 0.8238 |
| METS-IR | 73 | 0.51 (0.08, 3.01) | 0.4539 |
| Q1 (lowest) | 17 | ref | |
| Q2 | 19 | 0.83 (0.35, 1.97) | 0.6800 |
| Q3 | 14 | 1.01 (0.40, 2.57) | 0.9783 |
| Q4 (highest) | 23 | 2.09 (0.89, 4.91) | 0.0914 |
| Ln HOMA-IR | 73 | 2.02 (1.13, 3.59) | 0.0173 |
BMI: Body mass index; HDLc: High density lipoprotein cholesterol; HOMA-IR: homeostasis model assessment of insulin resistance. METS-IR: Metabolic score for IR; TyG: triglyceride-glucose.
Data are expressed as odds ratio (95%, confidence interval) per 1-unit increment for IR surrogates except for the TyG-BMI index, which is expressed per 10-unit increment, and HOMA-IR which is expressed per 1-logarithmic unit increment. Significant associations are highlighted in bold.
4. Discussion
In this study, we have evaluated the capacity of different IR proxies in predicting functional decline in community dwelling non-diabetic older adults from the Toledo Study of Healthy Aging. We have reported that only TyG-BMI index mimics the predictive capacity of HOMA-IR in older subjects. Furthermore, the current findings provided first insights into gender-specific capacity of insulin resistance proxies to predict functional decline in older adults. Interestingly TyG-BMI was the only index able to predict functional decline in women, whereas in men, only HOMA-IR was associated with incident functional deterioration over a median follow-up period of five years.
IR is a critical factor in the pathogenesis of frailty [3] with skeletal muscle tissue being one of the primarily drivers of this metabolic condition [33]. Previous research has reported a positive correlation between IR, assessed by the HOMA-IR, and frailty in both cross-sectional and prospective studies [[4], [5], [6]]. Consistent with this, we have reported that higher levels of HOMA-IR are linked to increased risk of functional deterioration assessed as an increase in 2.5 points in the FTS-5 score in non-diabetic subjects over 65 years old.
IR is often evaluated by using the HOMA-IR, which relies on measuring serum insulin levels but despite its clinical relevance, insulin levels are not commonly measured in the routine of clinical practice. Recent studies have suggested the use of non-insulin based indices relied on anthropometric data and/or biochemical parameters such as the TyG, the TyG-BMI, the ratio TG/HDLc, and METS-IR as reliable and sensitive indicators of IR [17,18] in predicting different outcomes including diabetes mellitus, cardiovascular disease and mortality [19,20]. However, there are very scarce evidence analyzing the relationship of the above-mentioned indices with the risk of functional decline among older adults. In the present study, we have reported that among all the indices assessed, only the TyG-BMI index was able to mimic the predictive capacity of HOMA-IR in older subjects. Noteworthy, TyG-BMI index was positively and prospectively related to an increased risk of functional decline in non-diabetic older subjects over five-years. Meanwhile, although the TyG index was related to an increase in the odds ratio for functional deterioration, it did not reach statistical significance (p = 0.06). In line with this Yin et al., have recently shown that increased TyG index and some of its related parameters including TyG-BMI were correlated with prevalent frailty in a cross-sectional study [21]. Furthermore, the results obtained from a prospective cohort study suggested that elevated TyG index and a high-stable trajectory of the TyG index were significantly associated with an increased risk of frailty, even after adjusting for potential confounders, especially in older urban residents with high BMI [34], emphasizing the role of obesity in this association. For instance, BMI is the most common measurement to assess obesity and has been shown to be related to frailty [35,36]. In line with this, it is well accepted that reduced insulin sensitivity is closely related to age-associated changes in body composition, primarily characterized by an increase in fat mass and a decrease in lean mass [9,10]. In this sense, IR causes an increase in the production of free fatty acids that are converted to hepatic triglyceride [37]. Body composition changes appear to be closely related to frailty. Supporting this, recent research identified trunk fat mass but not lean mass as the mediator of the association of IR with the risk of functional decline over a 2.99 follow-up period in non-diabetic community dwelling Spanish older subjects [11]. The requirement of including BMI and triglycerides in any calculation to obtain a significant association of IR proxies with FTS-5 worsening, reinforces the idea of the outstanding role of body structure in the impact of IR on functional outcomes in older people. This latter observation is further supported by the fact that albeit all non-insulin resistance indices were significantly associated with HOMA-IR, only the TyG-BMI index was able to predict functional decline among older subjects.
Research has reported a sex-based differences in whole body insulin resistance [22]. Specifically, women exhibit greater insulin sensitivity compared to men. However, this metabolic advantage typically decreases after menopause or as insulin resistance advances to hyperglycemia and diabetes. These differences may be due to differences in body fat and muscle mass distribution and function, in inflammation and in sex hormones between males and females [22]. Additionally, it is well known that the prevalence of frailty is higher in women than in men [23]. Taking into account these considerations, we evaluated whether IR proxies are gender-specific in their predictive capacity of functional decline among older adults. Present results showed that women had worse functional status as evidenced by higher basal FTS-5 scores and were more disabled for basic activities of daily living compared to men. Additionally, women tended to higher obesity rates, reflected in higher BMI values. Interestingly, the TyG-BMI index showed statistical differences between men and women, with women presenting higher values. Among IR surrogates evaluated, the TyG-BMI index stood out as the sole predictor of functional decline determined as an increase in 2.5 points the in FTS-5 scores among women after a median follow-up period of 5 years and after adjusting for different confounders. Additionally, female participants in the highest quartile of the TyG-BMI index were at significant greater risk of functional decline compared to those in the lowest quartile (p = 0.0158). To further support the role of BMI and triglyceride, two parameters tightly linked to fat mas, in their association with functional decline in female participants, a sub analysis was carried out considering only these two parameters. Triglyceride multiplied by BMI was positively related to a higher risk of functional decline in women. These suggests that both body fat mass parameters play a role in the pathogenesis of frailty independently of glucose levels.
By contrast, in men only HOMA-IR was associated with incident functional deterioration. Previous studies have reported an inverse correlation between HOMA-IR and physical impairment, specifically gait speed in non-diabetic older men but not in women [38]. Our findings are the first in providing insights into gender-specific capacity of insulin resistance proxies to predict functional decline in non-diabetic older adults. In fact, most of the available evidence related to the association of IR surrogates with frailty and functional deterioration are assessed in older subjects without stratifying by sex. To our knowledge the only study that have stratified by sex is the reported by Yin and colleagues. In this cross-sectional research, the association of TyG and its related indexes including TyG-BMI with frailty was explored after stratification analyses and interaction tests without significant interactions observed [21].
4.1. Strengths and limitations
The present study has several strengths, including the detailed characterization of a large cohort of non-diabetic, community-dwelling older adults and the utilization of the FTS-5 to address, for the first time, the association of different IR surrogates with incident functional decline. Additionally, the study's prospective design, with assessments of functional status at median follow-up periods of 5 years, provides valuable insights into the potential factors linked to IR-related events. Another potential strength of this study includes the stratification by sex, which have permitted the evaluation of the gender-specific predictive capacity of IR surrogates. Moreover, the present data suggest that routine blood test results could help warn about a potential risk of functional decline in older people, especially in women.
However, one limitation of the study is its inability to establish the causal relationship between TyG-BMI index and HOMA-IR with functional status deterioration in older adults. Further research is definitely needed to clarify this aspect.
5. Conclusion
Most of IR surrogates exhibit limited efficacy in predicting functional decline in older subjects. Only TyG-BMI index mimics the predictive capacity of HOMA-IR which requires insulin determination in older subjects. The clinical implications of these findings suggest that the TyG-BMI index may serve as a practical alternative to insulin-based markers for predicting functional decline, particularly in settings where insulin measurements are not readily available. Furthermore, the ability of IR indexes to indicate the risk of functional deterioration in the FTS-5 score is gender-specific, being TyG-BMI the only index able to predict functional decline in women, whereas in men, only HOMA-IR were associated with functional decline. Thus, a routine blood test could help to calibrate the risk for functional decline in older people, especially in women.Those findings highlight the need for tailored approaches and indicate that gender should be considered when selecting insulin resistance indexes for assessing the risk of functional decline in clinical practice.
Author contribution
Author Contributions: Conceptualization, M.E.A., J.A. and L.R.-M.; methodology, M.E.A., J.A., J.A.C., B.M.-B., F.J.G.-G., P.S. and L.R.-M.; formal analysis, M.E.A., J.A., J.A.C., B.M.-B., F.J.G-G., P.S. and L.R.-M.; investigation, M.E.A., J.A., J.A.C., B.M.-B., F.J.G-G., P.S. and L.R.-M.; resources, F.J.G.-G., and L.R.-M.; data curation, M.E.A., J.A., J.A.C., and L.R.-M; writing—original draft preparation, M.E.A., J.A., J.A.C., B.M.-B., F.J.G-G., P.S. and L.R.-M.; writing—review and editing, M.E.A., J.A., J.A.C., B.M.-B., F.J.G-G., P.S. and L.R.-M.; funding acquisition, L.R.-M. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by grants from the Ministry of Economy and Competitiveness and co-financed by FEDER funds (Instituto de Salud Carlos III, PI23/00877, CIBERFES CB16/10/00464), Spanish Government.
Conflict of interest
The authors declare no competing interests.
Acknowledgments
We would like to thank the participants, cohort members and team researcher members.
Footnotes
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jnha.2024.100376.
Appendix A. Supplementary data
The following is Supplementary data to this article:
References
- 1.El Assar M., Rodríguez-Sánchez I., Álvarez-Bustos A., Rodríguez-Mañas L. Biomarkers of frailty. Mol Aspects Med. 2024;97 doi: 10.1016/j.mam.2024.101271. [DOI] [PubMed] [Google Scholar]
- 2.Krentz A.J., Viljoen A., Sinclair A. Insulin resistance: a risk marker for disease and disability in the older person. Diabet Med. 2013;30:535–548. doi: 10.1111/DME.12063. [DOI] [PubMed] [Google Scholar]
- 3.Perazza L.R., Brown-Borg H.M., Thompson L.V. Physiological systems in promoting frailty. Compr Physiol. 2022;12:3575–3620. doi: 10.1002/cphy.c210034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Pérez-Tasigchana R.F., León-Muñoz L.M., Lopez-Garcia E., Gutierrez-Fisac J.L., Laclaustra M., Rodríguez-Artalejo F., et al. Metabolic syndrome and insulin resistance are associated with frailty in older adults: a prospective cohort study. Age Ageing. 2017;46(5):807–812. doi: 10.1093/ageing/afx023. [DOI] [PubMed] [Google Scholar]
- 5.Sen Peng P., Kao T.W., Chang P.K., Chen W.L., Peng P.J., Wu L.W. Association between HOMA-IR and frailty among U.S. middle-aged and elderly population. Sci Rep. 2019;9 doi: 10.1038/s41598-019-40902-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rodríguez-Mañas L., Angulo J., Carnicero J.A., El Assar M., García-García F.J., Sinclair A.J. Dual effects of insulin resistance on mortality and function in non-diabetic older adults: findings from the Toledo Study of Healthy Aging. Geroscience. 2022;44:1095–1108. doi: 10.1007/s11357-021-00384-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nishikawa H., Asai A., Fukunishi S., Nishiguchi S., Higuchi K. Metabolic syndrome and sarcopenia. Nutrients. 2021;13 doi: 10.3390/NU13103519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Nascimento C.M., Ingles M., Salvador-Pascual A., Cominetti M.R., Gomez-Cabrera M.C., Viña J. Sarcopenia, frailty and their prevention by exercise. Free Radic Biol Med. 2019;132:42–49. doi: 10.1016/J.FREERADBIOMED.2018.08.035. [DOI] [PubMed] [Google Scholar]
- 9.Goodpaster B.H., Krishnaswami S., Resnick H., Kelley D.E., Haggerty C., Harris T.B., et al. Association between regional adipose tissue distribution and both type 2 diabetes and impaired glucose tolerance in elderly men and women. Diabetes Care. 2003;26:372–379. doi: 10.2337/diacare.26.2.372. [DOI] [PubMed] [Google Scholar]
- 10.St-Onge M.P., Gallagher D. Body composition changes with aging: the cause or the result of alterations in metabolic rate and macronutrient oxidation? Nutrition. 2010;26:152–155. doi: 10.1016/J.NUT.2009.07.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.El Assar M., Carnicero J.A., Angulo J., Cámara-Hernández V., García-García F.J., Rodríguez-Mañas L. Fat mass accounts for insulin resistance impact on functional decline and mortality in nondiabetic older adults. J Am Med Dir Assoc. 2024;25:448–453. doi: 10.1016/j.jamda.2023.09.015. [DOI] [PubMed] [Google Scholar]
- 12.Bilgin S., Aktas G., Kahveci G., Tel A.B., Kurtkulagi O., Duman T.T. Does mean platelet volume/lymphocyte count ratio associate with frailty in type 2 diabetes mellitus? Bratisl Lek Listy. 2021;122:116–119. doi: 10.4149/BLL_2021_017. [DOI] [PubMed] [Google Scholar]
- 13.Xu Y.S., Wang M.M., Chen D., Jiang X., Xiong Z.F. Inflammatory biomarkers in older adults with frailty: a systematic review and meta-analysis of cross-sectional studies. Aging Clin Exp Res. 2022;34:971–987. doi: 10.1007/S40520-021-02022-7. [DOI] [PubMed] [Google Scholar]
- 14.Aktas G. Association between the Prognostic Nutritional Index and Chronic Microvascular Complications in Patients with Type 2 Diabetes Mellitus. J Clin Med. 2023;12 doi: 10.3390/JCM12185952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Yaribeygi H., Farrokhi F.R., Butler A.E., Sahebkar A. Insulin resistance: review of the underlying molecular mechanisms. J Cell Physiol. 2019;234:8152–8161. doi: 10.1002/JCP.27603. [DOI] [PubMed] [Google Scholar]
- 16.Balci S.B., Tel B.M.A., Duman T., Ozkul F.N., Aktas G. A novel marker for prediabetic conditions: uric acid-to-HDL cholesterol ratio. Bratisl Lek Listy. 2024;125:145–148. doi: 10.4149/BLL_2023_130. [DOI] [PubMed] [Google Scholar]
- 17.Zhang Y., Wang R., Fu X., Song H. Non-insulin-based insulin resistance indexes in predicting severity for coronary artery disease. Diabetol Metab Syndr. 2022;14:191. doi: 10.1186/s13098-022-00967-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Son D.-H., Lee H.S., Lee Y.-J., Lee J.-H., Han J.-H. Comparison of triglyceride-glucose index and HOMA-IR for predicting prevalence and incidence of metabolic syndrome. Nutr Metab Cardiovasc Dis. 2022;32:596–604. doi: 10.1016/j.numecd.2021.11.017. [DOI] [PubMed] [Google Scholar]
- 19.Lopez-Jaramillo P., Gomez-Arbelaez D., Martinez-Bello D., Abat M.E.M., Alhabib K.F., Avezum Á, et al. Association of the triglyceride glucose index as a measure of insulin resistance with mortality and cardiovascular disease in populations from five continents (PURE study): a prospective cohort study. Lancet Healthy Longev. 2023;4:e23–33. doi: 10.1016/S2666-7568(22)00247-1. [DOI] [PubMed] [Google Scholar]
- 20.Che B., Zhong C., Zhang R., Pu L., Zhao T., Zhang Y., et al. Triglyceride-glucose index and triglyceride to high-density lipoprotein cholesterol ratio as potential cardiovascular disease risk factors: an analysis of UK biobank data. Cardiovasc Diabetol. 2023;22 doi: 10.1186/S12933-023-01762-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Yin H., Guo L., Zhu W., Li W., Zhou Y., Wei W., et al. Association of the triglyceride-glucose index and its related parameters with frailty. Lipids Health Dis. 2024;23:150. doi: 10.1186/s12944-024-02147-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Gado M., Tsaousidou E., Bornstein S.R., Perakakis N. Sex-based differences in insulin resistance. J Endocrinol. 2024:261. doi: 10.1530/JOE-23-0245. [DOI] [PubMed] [Google Scholar]
- 23.Rohrmann S. Epidemiology of frailty in older people. Adv Exp Med Biol. 2020;1216:21–27. doi: 10.1007/978-3-030-33330-0_3. [DOI] [PubMed] [Google Scholar]
- 24.Garcia-Garcia F.J., Gutierrez Avila G., Alfaro-Acha A., Amor Andres M.S., De Los Angeles De La Torre Lanza M., Escribano Aparicio M.V., et al. The prevalence of frailty syndrome in an older population from Spain. The Toledo Study for Healthy Aging. J Nutr Health Aging. 2011;15:852–856. doi: 10.1007/s12603-011-0075-8. [DOI] [PubMed] [Google Scholar]
- 25.Zhao J., Fan H., Wang T., Yu B., Mao S., Wang X., et al. TyG index is positively associated with risk of CHD and coronary atherosclerosis severity among NAFLD patients. Cardiovasc Diabetol. 2022;21:123. doi: 10.1186/s12933-022-01548-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wang H., Wang C., Xuan X., Xie Z., Qiu Y., Qin H., et al. Association between triglyceride to high-density lipoprotein cholesterol ratio and type 2 diabetes risk in Japanese. Sci Rep. 2023;13:3719. doi: 10.1038/s41598-022-25585-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Jiang C., Yang R., Kuang M., Yu M., Zhong M., Zou Y. Triglyceride glucose-body mass index in identifying high-risk groups of pre-diabetes. Lipids Health Dis. 2021;20:161. doi: 10.1186/s12944-021-01594-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Bello-Chavolla O.Y., Almeda-Valdes P., Gomez-Velasco D., Viveros-Ruiz T., Cruz-Bautista I., Romo-Romo A., et al. METS-IR, a novel score to evaluate insulin sensitivity, is predictive of visceral adiposity and incident type 2 diabetes. Eur J Endocrinol. 2018;178:533–544. doi: 10.1530/EJE-17-0883. [DOI] [PubMed] [Google Scholar]
- 29.Matthews D.R., Hosker J.P., Rudenski A.S., Naylor B.A., Treacher D.F., Turner R.C. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412–419. doi: 10.1007/BF00280883. [DOI] [PubMed] [Google Scholar]
- 30.Chapter 1: definition and classification of CKD. Kidney Int Suppl (2011) 2013;3:19–62. doi: 10.1038/kisup.2012.64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Katz S., Ford A.B., Moskowitz R.W., Jackson B.A., Jaffe M.W. Studies of illness in the aged. The index of ADL: a standardized measure of biological and psychosocial function. JAMA. 1963;185:914–919. doi: 10.1001/JAMA.1963.03060120024016. [DOI] [PubMed] [Google Scholar]
- 32.García-García F.J., Carnicero J.A., Losa-Reyna J., Alfaro-Acha A., Castillo-Gallego C., Rosado-Artalejo C., et al. Frailty trait scale–short form: a frailty instrument for clinical practice. J Am Med Dir Assoc. 2020;21:1260–1266.e2. doi: 10.1016/j.jamda.2019.12.008. [DOI] [PubMed] [Google Scholar]
- 33.Freeman A.M., Acevedo L.A., Pennings N. StatPearls [Internet] StatPearls Publishing; Treasure Island (FL): 2023. Insulin Resistance. 2024 Jan–. [Google Scholar]
- 34.Yuan Y., Chen S., Lin C., Huang X., Lin S., Huang F., et al. Association of triglyceride-glucose index trajectory and frailty in urban older residents: evidence from the 10-year follow-up in a cohort study. Cardiovasc Diabetol. 2023;22 doi: 10.1186/s12933-023-02002-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Nishida Y., Yamada Y., Sasaki S., Kanda E., Kanno Y., Anzai T., et al. Effect of overweight/obesity and metabolic syndrome on frailty in middle-aged and older Japanese adults. Obes Sci Pract. 2023;10 doi: 10.1002/OSP4.714. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Yuan L., Chang M., Wang J. Abdominal obesity, body mass index and the risk of frailty in community-dwelling older adults: a systematic review and meta-analysis. Age Ageing. 2021;50:1118–1128. doi: 10.1093/AGEING/AFAB039. [DOI] [PubMed] [Google Scholar]
- 37.Kahn B.B., Flier J.S. Obesity and insulin resistance. J Clin Invest. 2000;106:473–481. doi: 10.1172/JCI10842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kuo C.K., Lin L.Y., Yu Y.H., Wu K.H., Kuo H.K. Inverse association between insulin resistance and gait speed in nondiabetic older men: results from the U.S. National Health and Nutrition Examination Survey (NHANES) 1999-2002. BMC Geriatr. 2009;9 doi: 10.1186/1471-2318-9-49. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.