Introduction
Older adults are at increased risk of both poor glycemic control and frailty (Hubbard et al., Reference Hubbard, Andrew, Fallah and Rockwood2010). Frailty is defined as an age- and chronic disease-associated decline in physiological reserve that reduces an older adult’s ability to cope with an acute stressor (Hogan et al., Reference Hogan, Maxwell, Afilalo, Arora, Bagshaw, Basran, Bergman, Bronskill, Carter, Dixon, Hemmelgarn, Madden, Mitnitski, Rolfson, Stelfox, Tam-Tham and Wunsch2017). Approximately one-quarter of adults over the age of 65 have diabetes, and this prevalence continues to increase in National Health and Nutrition Examination Survey (NHANES) data collected between 1999 and 2018 (Wang et al., Reference Wang, Li, Wang, Bancks, Carnethon, Greenland, Feng, Wang and Zhong2021). Likewise, approximately one-quarter of community-dwelling adults meet current criteria for frailty (Molloy et al., Reference Molloy, Clegg and Rockwood2021). Frailty and diabetes often co-exist (Hubbard et al., Reference Hubbard, Andrew, Fallah and Rockwood2010); in fact, a diagnosis of diabetes has been found to be associated with an increase in the 3-year risk of meeting the Cardiovascular Health Study Index (CHSI) (Fried et al., Reference Fried, Tangen, Walston, Newman, Hirsch, Gottdiener, Seeman, Tracy, Kop, Burke and McBurnie2001) for frailty (Muszalik et al., Reference Muszalik, Stępień, Puto, Cybulski and Kurpas2022). Conversely, meeting criteria for frailty prospectively increases the risk of developing diabetes (Veronese et al., Reference Veronese, Stubbs, Fontana, Trevisan, Bolzetta, De Rui, Sartori, Musacchio, Zambon, Maggi, Perissinotto, Corti, Crepaldi, Manzato and Sergi2016).
The recognition of the common coexistence of both of these chronic issues in older adults has led to the adoption of less stringent glycemic targets in the current Canadian (Harris et al., Reference Harris, Capes, Lillie, Lank, Mahon and Erickson2003), American (Diabetes Association, Reference Diabetes Association2021), and European diabetes guidelines (Sinclair et al., Reference Sinclair, Paolisso, Castro, Bourdel-Marchasson, Gadsby and Rodriguez Mañas2011). A recent systematic review of diabetes clinical practice guidelines has demonstrated a wide variety of glycemic targets recommended in the population (Christiaens et al., Reference Christiaens, Henrard, Zerah, Dalleur, Bourdel-Marchasson and Boland2021), likely due to the fact that older adults are underrepresented in trial data. The lack of clarity on glycemic targets in older adults also explains some recent work suggesting that older adults often receive overly stringent glycemic control (Mangé et al., Reference Mangé, Pagès, Sourdet, Cestac and McCambridge2021).
One complicating factor is that the relationship between frailty and glycemic control in older adults with diabetes remains unclear. Some past work has demonstrated a negative (more hypoglycemic) association (Idrees et al., Reference Idrees, Zabala, Moreno, Gerges, Urrutia, Ruiz, Vaughan, Vellanki, Pasquel, Peng and Umpierrez2023; MacKenzie et al., Reference MacKenzie, Tugwell, Rockwood and Theou2020; Morita et al., Reference Morita, Okuno, Himeno, Watanabe, Nakajima, Koizumi, Yano, Iritani, Okuro and Morimoto2017; Nguyen et al., Reference Nguyen, Harris, Woodward and Chalmers2021; Yanagita et al., Reference Yanagita, Fujihara, Eda, Tajima, Yonemura, Kawajiri, Yamaguchi, Asakawa, Nei, Kayashima, Yoshimoto, Kitajima, Harada, Araki, Yoshimoto, Aida, Yanase, Nawata and Muta2018), while other work has shown a positive (more hyperglycemic) association (Aguayo et al., Reference Aguayo, Hulman, Vaillant, Donneau, Schritz, Stranges, Malisoux, Huiart, Guillaume, Sabia and Witte2019; Bilgin et al., Reference Bilgin, Aktas, Kurtkulagi, Atak and Duman2020; Chung et al., Reference Chung, Lee, Kim, Lee, Jin, Yoo, Moon and Kim2021; Fung et al., Reference Fung, Lui, Huang, Cheng, Lau, Chung, Ahmadabadi, Xie, Lee, Hui, So, Sung, King, Goggins, Chan, Järvelin, Ma, Chow and Kwok2021; García-Esquinas et al., Reference García-Esquinas, Graciani, Guallar-Castillón, López-García, Rodríguez-Mañas and Rodríguez-Artalejo2015; Hyde et al., Reference Hyde, Smith, Flicker, Atkinson, Fenner, Skeaf, Malay and Lo Giudice2019; Kong et al., Reference Kong, Zhao, Fan, Wang, Li, Bai and Mao2021; Lin et al., Reference Lin, Yu, Wu and Liu2022; Muszalik et al., Reference Muszalik, Stępień, Puto, Cybulski and Kurpas2022) between frailty and glycemic control. Other work has shown no relationship (MacKenzie et al., Reference MacKenzie, Tugwell, Rockwood and Theou2020) or a U-shaped relationship between frailty and glycemic control (Zaslavsky et al., Reference Zaslavsky, Walker, Crane, Gray and Larson2016). Abdelhafiz et al. have tried to reconcile these conflicting results by suggesting that frailty is a metabolically heterogeneous condition that lies on a spectrum between the anorexic malnourished phenotype (tending towards lower sugars) and the sarcopenic obese phenotype (tending towards higher sugars) and that body composition should be considered as a confounding factor in the assessment of older adults with diabetes (Abdelhafiz, Reference Abdelhafiz2023).
This is further obscured by the fact that there is no consensus on how to assess frailty in the population of older adults with diabetes. There are currently numerous different scales developed to assess frailty that can be loosely divided into three categories (Hogan et al., Reference Hogan, Maxwell, Afilalo, Arora, Bagshaw, Basran, Bergman, Bronskill, Carter, Dixon, Hemmelgarn, Madden, Mitnitski, Rolfson, Stelfox, Tam-Tham and Wunsch2017): (1) multidimensional physical-based measures (such as the CHSI) (Fried et al., Reference Fried, Tangen, Walston, Newman, Hirsch, Gottdiener, Seeman, Tracy, Kop, Burke and McBurnie2001), (2) judgment-based measures (such as the Rockwood Clinical Frailty Scale [RCFS]) (Clinical Frailty Scale, n.d.; Rockwood et al., Reference Rockwood, Song, MacKnight, Bergman, Hogan, McDowell and Mitnitski2005), and (3) single performance-based measures (such as gait speed) (Veronese et al., Reference Veronese, Stubbs, Volpato, Zuliani, Maggi, Cesari, Lipnicki, Smith, Schofield, Firth, Vancampfort, Koyanagi, Pilotto and Cereda2018).
In order to investigate the complex relationship between frailty, body composition, and frailty, our current study examined the association between glycemic control (HgA1C) and three commonly used frailty assessments (CHSI, RCFS, and gait speed) while accounting for measures of body composition (fat-free mass [FFM] and waist circumference [WC]). In recognition of the many scales used to assess frailty in older adults, we chose one commonly used method from each of the three categories of frailty assessment (Hogan et al., Reference Hogan, Maxwell, Afilalo, Arora, Bagshaw, Basran, Bergman, Bronskill, Carter, Dixon, Hemmelgarn, Madden, Mitnitski, Rolfson, Stelfox, Tam-Tham and Wunsch2017). We hypothesized that (1) glycemic control would show a stronger relationship with the more physical-based measure (CHSI) and (2) frailty would still show an association with glycemic control even after accounting for measures of body composition.
Methods
Subjects
Subjects were recruited from outpatient older adult diabetes clinics in an academic center (Vancouver General Hospital, Vancouver, Canada), and a cross-sectional, observational study design was used. These subjects were recruited as part of a larger sarcopenia study (ClinicalTrials.gov, NCT04370912) (Madden et al., Reference Madden, Feldman, Arishenkoff and Meneilly2020, Reference Madden, Feldman, Arishenkoff and Meneilly2021, Reference Madden, Feldman, Arishenkoff and Meneilly2022). Each subject was 65 years of age or older and had been referred by a primary care physician or nurse practitioner. All subjects on hemodialysis were excluded. All subjects had to have been diagnosed with diabetes for at least 5 years. Current guidelines (Diabetes Association, Reference Diabetes Association2021) were used to define the presence of diabetes. Each subject attended a single laboratory visit where all measures were taken.
Ethical issues
All subjects gave written consent, and our study protocol received approval by the Human Subjects Committee of the University of British Columbia. Our protocol adheres to the Declaration of Helsinki.
Measures of body composition
All measures were done by a single operator Boris Feldman (BF) with a HBF-510W Full-Body Composition Monitor (Omron, Seoul, Korea), which measures per cent muscle using bioelectrical impedance; FFM is then calculated by per cent muscle * subject mass. WC was measured by holding a plastic tape measure at the level of the umbilicus against the skin. Our tape measure is calibrated regularly in accordance with current guidelines (Higgins & Comuzzie, Reference Higgins, Comuzzie and Preedy2012).
Frailty measures (CHSI, RCPS, and gait speed)
The same technician obtained all measures of frailty (BF), and they were trained by a geriatric medicine specialist prior to the start of data collection Kenneth Madden (KM). Each subject had the CHSI measured during the laboratory visit. This index uses a 5-point multidimensional scale of physical measures consisting of subjective symptoms of unintentional weight loss (>10 pounds), exhaustion (self-reported), weakness (grip strength, lowest 20% by biological sex, and body mass index [BMI]), slow walking speed (slowest 20% by gender and height), and low physical activity (lowest 20% by gender) (Fried et al., Reference Fried, Tangen, Walston, Newman, Hirsch, Gottdiener, Seeman, Tracy, Kop, Burke and McBurnie2001). A dynamometer was used to measure grip strength (Sammons Preston, Nottinghamshire, UK) using the average of three readings in the dominant hand.
The RCFS is a 9-point judgment-based measure, as previously described (Clinical Frailty Scale, n.d.; Rockwood et al., Reference Rockwood, Song, MacKnight, Bergman, Hogan, McDowell and Mitnitski2005). Gait speed was measured in meters per second – subjects were told to walk at their usual pace for 6 meters from a standing start. This measure was obtained using the Kinesis system (Linus Health Europe Ltd., Dublin, Ireland). Gait speed was measured using the Kinesis Gait™ system based on internal sensors attached to the shin of both legs.
Glycemic control
Glycated hemoglobin (HgA1C) was measured in a commercial laboratory.
Statistical methods
Our outcome variable consisted of our measure of glycemic control (HgA1C). Our predictor variables in our multivariate analysis were age, biological sex, FFM, WC, and one of the three measures of frailty. We used WC as our measure of adiposity instead of our HBF-510W measures of per cent body fat in order to avoid issues of collinearity (both FFM and WC are calculated from the same body impedance measure) (Abu Khaled et al., Reference Abu Khaled, McCutcheon, Reddy, Pearman, Hunter and Weinsier1988). Our three measures of frailty consisted of CHSI, RCFS, and gait speed. CHSI was treated as an accumulation of items (with a maximum of 5). Any variables that demonstrated skewing were logarithmically transformed (base ten) prior to analyses. Issues with multicollinearity were checked with variance inflation factors for each model. The R Core software package version 4.3.1 was used for statistical analysis with a significance level of p < 0.05 (R Core Team, 2021). The format mean ± standard deviation was used to express results, and all data analysis was done in a blinded fashion.
Data resource and availability statement
The datasets generated during and/or analysed in the current study are available from the corresponding author upon reasonable request.
Results
Subject characteristics
From the older adult diabetes clinics, 81 subjects (28 women and 53 men) were recruited. Prior to finishing the study, one woman did not complete the protocol (27 women and 53 men). Eleven subjects identified as East Asian persons, 3 subjects identified as South Asian persons, and 66 subjects identified as White persons. Forty-six subjects were on metformin, 21 subjects were on sulfonylureas, 39 subjects were on insulin, and 37 subjects were on other oral agents, such as glitazones. Of the 80 subjects that completed the protocol, 16.3 ± 1.8 per cent were not frail (0 criteria), 63.8 ± 0.9 per cent were prefrail (1 or 2 criteria), and 20.0 ± 0.0 per cent were frail, per the CHSI. Subject characteristics stratified by biological sex are shown in Table 1.
Table 1. Subject characteristics in women and men
Abbreviations: Mean [95% confidence interval]; MT, FFM, fat-free mass; RCFS, Rockwood Clinical Frailty Scale; CHSI, Cardiovascular Health Study Index; *, p<0.05.
Prior to analysis, density plots for all variables were examined. Since none of our variables demonstrated skewing, no transformation of any of our predictor or outcome variables was required. Variance inflation factors were also checked during our multivariate analysis, and there were no multicollinearity issues (all were less than a conservative threshold of 5). The maximum variance inflation factor was 2.9 (for biological sex in our model using CHSI as our measure of frailty).
Univariate analysis (Table 2)
Only CHSI showed a significant negative association with HgA1C (p = 0.035). Age (p = 0.983), FFM (p = 0.318), WC (p = 0.914), RCFS (p = 0.842), and gait speed (p = 0.422) did not show any significant correlations with glycemic control. When we stratified our results by biological sex, only CHSI showed a significant negative association with HgA1C in men only (women, p = 0.880; men, p = 0.026).
Table 2. Univariate analysis
Abbreviations: HgA1C, glycated hemoglobin; β, standardized beta coefficient; SE, standard error; FFM, fat-free mass; WC, waist circumference; *, p<0.05; RCFS, Rockwood Clinical Frailty Scale; CHSI, Cardiovascular Health Study Index.
Multivariable analysis, glycemic control, and measures of frailty (Figure 1)
In our first model using CHSI as our measure of frailty, only CHSI (standardized β = −0.255 ± 0.121; p = 0.038) showed a significant negative association with HgA1C; age (standardized = −0.026 ± 0.126; p = 0.836), being a woman (standardized β = −0.122 ± 0.416; p = 0.770), FFM (standardized β = −0.279 ± 0.254; p = 0.274), and WC (standardized β = 0.167 ± 0.183; p = 0.364) showed no significant association.
Figure 1. Association between glycated hemoglobin and measures of frailty: With age, biological sex, fat-free mass, and waist circumference as covariates, only the Cardiovascular Health Study Index (CHSI, 5-point scale) showed a significant negative association with glycated hemoglobin (HgA1C, p = 0.038). Neither the Rockwood Clinical Frailty Scale (RCFS, 9-point scale, p = 0.518) nor gait speed (m/s, p = 0.423) showed a significant association with HgA1C. Linear regression is shown, with 95% confidence limits. Note: For the CHSI and RCFS graphs, all points have been slightly adjusted (jittered) along the x-axis to minimize overlap.
When the RCFS was used as our frailty measure, it showed no significant association with HgA1C (standardized β = −0.089 ± 0.136; p = 0.518); age (standardized β = −0.039 ± 0.130; p = 0.765), being a woman (standardized β = −0.209 ± 0.426; p = 0.625), FFM (standardized β = −0.344 ± 0.264; p = 0.197), and WC (standardized β = 0.234 ± 0.198; p = 0.242) showed no significant association.
When our performance measure (gait speed) was used to assess frailty, it also showed no significant association with HgA1C (standardized β = 0.104 ± 0.129; p = 0.423); age (standardized β = −0.018 ± 0.134; p = 0.893), being a woman (standardized β = −0.229 ± 0.425; p = 0.539), FFM (standardized β = −0.319 ± 0.260; p = 0.225), and WC (standardized β = 0.197 ± 0.188; p = 0.299) showed no significant association.
Discussion
Principal findings
In terms of the three types of frailty measures used (CHSI, RCFS, and gait speed), only CHSI showed a significant negative association with glycemic control in older men with type 2 diabetes. This negative association between frailty and glycemic control was still present even after accounting for both FFM and adiposity (WC).
Previous work
When the association between glycemic control and frailty is examined prospectively, it is quite clear that the presence of diabetes (García-Esquinas et al., Reference García-Esquinas, Graciani, Guallar-Castillón, López-García, Rodríguez-Mañas and Rodríguez-Artalejo2015) is associated with an increased risk of becoming frail. Diabetes worsens the trajectory of frailty (Aguayo et al., Reference Aguayo, Hulman, Vaillant, Donneau, Schritz, Stranges, Malisoux, Huiart, Guillaume, Sabia and Witte2019), and higher diabetes risk scores have been shown to be a risk factor for frailty in older adults (Bouillon et al., Reference Bouillon, Kivimäki, Hamer, Shipley, Akbaraly, Tabak, Singh-Manoux and Batty2013; Maggi et al., Reference Maggi, Noale, Gallina, Marzari, Bianchi, Limongi, Crepaldi and Group2004), likely due to the functional implications of diabetes complications such as peripheral neuropathy and cardiovascular events.
However, the main justification for different treatment guidelines for frail older adults is based on the association between short-term glycemic control and frailty. When one considers the association between glycemic control and frailty cross-sectionally, the previous literature is much less clear. Previous work has shown negative correlations between glycated hemoglobin and worsening frailty, suggesting that frail older adults are at increased risk of hypoglycemia in the short term. This negative association has been shown in both inpatient (Idrees et al., Reference Idrees, Zabala, Moreno, Gerges, Urrutia, Ruiz, Vaughan, Vellanki, Pasquel, Peng and Umpierrez2023; MacKenzie et al., Reference MacKenzie, Tugwell, Rockwood and Theou2020; Yanagita et al., Reference Yanagita, Fujihara, Eda, Tajima, Yonemura, Kawajiri, Yamaguchi, Asakawa, Nei, Kayashima, Yoshimoto, Kitajima, Harada, Araki, Yoshimoto, Aida, Yanase, Nawata and Muta2018) and outpatient (Morita et al., Reference Morita, Okuno, Himeno, Watanabe, Nakajima, Koizumi, Yano, Iritani, Okuro and Morimoto2017) populations; additionally, intensive glucose control (in the Action in Diabetes and Vascular Disease-PreterAx and DiamicroN Controlled Evaluation [ADVANCE] trail) has been shown to be associated with more hypoglycemia in the frail population (Nguyen et al., Reference Nguyen, Harris, Woodward and Chalmers2021). Other work, however, in community-dwelling subjects has shown a positive association between higher glycated hemoglobin levels and frailty (Aguayo et al., Reference Aguayo, Hulman, Vaillant, Donneau, Schritz, Stranges, Malisoux, Huiart, Guillaume, Sabia and Witte2019; Bilgin et al., Reference Bilgin, Aktas, Kurtkulagi, Atak and Duman2020; García-Esquinas et al., Reference García-Esquinas, Graciani, Guallar-Castillón, López-García, Rodríguez-Mañas and Rodríguez-Artalejo2015; Hyde et al., Reference Hyde, Smith, Flicker, Atkinson, Fenner, Skeaf, Malay and Lo Giudice2019; Kong et al., Reference Kong, Zhao, Fan, Wang, Li, Bai and Mao2021; Lin et al., Reference Lin, Yu, Wu and Liu2022; Muszalik et al., Reference Muszalik, Stępień, Puto, Cybulski and Kurpas2022) and higher glucose levels during continuous monitoring (Chung et al., Reference Chung, Lee, Kim, Lee, Jin, Yoo, Moon and Kim2021; Fung et al., Reference Fung, Lui, Huang, Cheng, Lau, Chung, Ahmadabadi, Xie, Lee, Hui, So, Sung, King, Goggins, Chan, Järvelin, Ma, Chow and Kwok2021). Still, other work has suggested a U-shaped relationship (Zaslavsky et al., Reference Zaslavsky, Walker, Crane, Gray and Larson2016) or no relationship (MacKenzie et al., Reference MacKenzie, Tugwell, Rockwood and Theou2020) between frailty and glycemic control.
The main explanation for the heterogeneity in the literature is that few of these studies accounted for body composition. Sarcopenia (low FFM) is associated with more episodes of hypoglycemia with diabetes treatment (Ogama et al., Reference Ogama, Sakurai, Kawashima, Tanikawa, Tokuda, Satake, Miura, Shimizu, Kokubo, Niida, Toba, Umegaki and Kuzuya2019), while adiposity (increased WC) is associated with increased insulin resistance (Goulet et al., Reference Goulet, Hassaine, Dionne, Gaudreau, Khalil, Fulop, Shatenstein, Tessier and Morais2009). Keegan et al. (Reference Keegan, Bhardwaj and Abdelhafiz2023) have suggested opposing associations between glycemic control and frailty in the low and high BMI populations, promoting some to suggest that older adults with both frailty and diabetes come from two distinct populations – a sarcopenic obese group, which results in insulin resistance and a positive association between frailty and HgA1C, and a anorexic malnourished group that has a negative association (Abdelhafiz, Reference Abdelhafiz2023). The present study is the first to our knowledge to demonstrate lower glycemic hemoglobin in older adults with diabetes with increasing frailty, even with both FFM and adiposity (WC) as covariates. We also demonstrated that the association between glycemic control and frailty was only present when frailty was assessed in a physically based measure (CHSI) as opposed to a judgment-based measure (RCFS) or single performance-based measure (gait speed).
Potential mechanisms
There are several potential explanations for the persistent negative association (after accounting for FFM and WC) between glycated hemoglobin and the CHSI observed in our study. Abdelhafiz et al. have suggested taking a more ‘multidimensional perspective’ on frailty in older adults with diabetes, in consideration of the ‘triad of impairment’ (TOI). The TOI considers cognitive and emotional impairments in addition to the physical aspects of increasing frailty in older adults with diabetes (Abdelhafiz, Reference Abdelhafiz2023). In terms of the cognitive aspect of TOI, there is a reciprocal relationship between hypoglycemia, frailty, and dementia, which can lead to a ‘vicious circle’ (Abdelhafiz, Reference Abdelhafiz2023), which could be a partial explanation for the lower HgA1C seen in our frailer subjects. Frailty syndrome is also strongly associated with malnutrition (Abdelhafiz, Reference Abdelhafiz2023), and a reduction in caloric intake might also be another partial explanation for our negative association between glycated hemoglobin and CHSI.
Although speculative, there might be several reasons why our negative association between glycated hemoglobin and the CHSI only reached statistical significance in men. Men generally have higher muscle mass than women (Lindle et al., Reference Lindle, Metter, Lynch, Fleg, Fozard, Tobin, Roy and Hurley1997); since muscle is the largest glucose sink in the body (Koh, Reference Koh2016), sarcopenia in frail older men might theoretically have a more significant impact on glycemic control than in women. Also, community-dwelling frail older men tend to be at higher risk of malnutrition due to poorer dietary habits than older women (Chang, Reference Chang2017), which might partially explain the negative association between glycated hemoglobin and the CHSI seen in our male subjects.
Clinical implications
Many diabetes organizations have reasonably advocated for less intensive glycemic control in frail older adults with diabetes, and this has led to changes in glycemic targets in this vulnerable population (Diabetes Association, Reference Diabetes Association2021; Harris et al., Reference Harris, Capes, Lillie, Lank, Mahon and Erickson2003; Sinclair et al., Reference Sinclair, Paolisso, Castro, Bourdel-Marchasson, Gadsby and Rodriguez Mañas2011). Some have suggested that this is too simplistic and that the heterogeneous nature of the frailty syndrome suggests that the treatment for diabetes should be much different in the anorexic malnourished frail older adult with diabetes as compared to the sarcopenic obese older adult with diabetes (Abdelhafiz, Reference Abdelhafiz2023). Our results have shown that even when you account for body composition (both FFM and WC) increasing frailty tends to be associated with lower glycated hemoglobin levels, suggesting that less-intensive control should be advocated in all frail older adults with diabetes. Our study also demonstrated that the interaction (or lack thereof) between glycemic control and frailty is quite different depending on how frailty is assessed, showing that much more work is needed to understand exactly how glycemic targets should be altered for different portions of the frail, older adult population with diabetes.
Limitations
All our subjects were recruited from a diabetes clinic specifically for older adults; the frailer nature of this patient population means that some of our findings might not extrapolate to the general population. Our clinics get large numbers of emergency department referrals for patients that lack a primary care provider. Since men often do not have a family physician at much higher rates than women, our subject pool had a higher proportion of men. Although our dataset has information on the presence of diabetic neuropathy, nephropathy, and cardiovascular disease, we do not have access to the rates of diabetic retinopathy. Our analysis was cross-sectional in nature – given that both frailty and glycemic control are dynamic, future work should examine how the trajectories of these indicators change and relate over time. There are many different methods of assessing frailty (Hogan et al., Reference Hogan, Maxwell, Afilalo, Arora, Bagshaw, Basran, Bergman, Bronskill, Carter, Dixon, Hemmelgarn, Madden, Mitnitski, Rolfson, Stelfox, Tam-Tham and Wunsch2017), and we were only able to study three of them – however, the fact we chose a commonly used method from each category of frailty assessment (Hogan et al., Reference Hogan, Maxwell, Afilalo, Arora, Bagshaw, Basran, Bergman, Bronskill, Carter, Dixon, Hemmelgarn, Madden, Mitnitski, Rolfson, Stelfox, Tam-Tham and Wunsch2017) is a strength of our study.
Conclusions
Even after accounting for muscle mass (FFM) and adiposity (WC), we demonstrated a negative association between glycated hemoglobin and increasing frailty in older adults with diabetes. In terms of the three types of frailty measures used (CHSI, RCFS, and gait speed), only CHSI showed significant negative association with glycemic control in older adults with type 2 diabetes. This suggests that more work needs to be done on how to assess frailty in older adults with diabetes and to more fully understand the exact interaction between frailty and glycemic control in this vulnerable population.
Data availability statement
The datasets generated during and/or analysed in the current study are available from the corresponding author upon reasonable request.
Author disclosures
All authors declare no conflicts of interest in this paper. The sponsors (the Allan M. McGavin Foundation and VGH Innovation and Translation Grant) had no role in the study design, analysis, or writing of the paper.
Acknowledgments
This work was supported by the Allan M. McGavin Foundation and a VGH Innovation and Translation Grant, neither of whom had any involvement in the study or in the preparation of the manuscript.
Open access
