Advertisement

Individual frailty phenotype components and mortality in adults with type 2 diabetes: A UK Biobank study

Open AccessPublished:November 22, 2022DOI:https://doi.org/10.1016/j.diabres.2022.110155

      Abstract

      Aims

      This study aimed to explore associations between frailty components and mortality and rank prognostic relevance of each frailty component in predicting mortality in adults with and without type 2 diabetes (T2D).

      Methods

      We used data from the UK Biobank. Associations and prognostic discrimination of individual Fried’s frailty components and the overall frailty status with all-cause and cardiovascular (CVD) mortality were investigated using Cox proportional-hazard models and C-index in adults with and without T2D.

      Results

      In both populations the strongest association with all-cause mortality across all frailty components and overall frailty status was observed for slow walking pace (without T2D Hazard Ratio [HR] 2.25, 95 %CI: 2.12–2.38 and with T2D HR 1.95, 95 %CI: 1.67–2.28). Similarly, slow walking pace was associated with a greater risk of CVD mortality. The combination of T2D and slow walking pace had the strongest association with all-cause and CVD mortality, compared to the combination of T2D and other frailty components or overall frailty status. Slow walking pace also provided the greatest prognostic discrimination.

      Conclusion

      Slow walking pace has a stronger predictive factor for all-cause and CVD mortality compared to other frailty components and overall frailty status, especially when simultaneously present with T2D.

      Keywords

      1. Introduction

      Type 2 diabetes (T2D) is a complex metabolic disorder frequently accompanied by accelerated biological aging placing people at higher risk of age-related comorbidities, including frailty [
      • García-Esquinas E.
      • Graciani A.
      • Guallar-Castillón P.
      • López-García E.
      • Rodríguez-Mañas L.
      • Rodríguez-Artalejo F.
      Diabetes and risk of frailty and its potential mechanisms: a prospective cohort study of older adults.
      ]. Frailty is characterised as a state of increased vulnerability to stressors, resulting from the decline in physiological reserves [
      • Fried L.P.
      • Tangen C.M.
      • Walston J.
      • Newman A.B.
      • Hirsch C.
      • Gottdiener J.
      • et al.
      Frailty in Older Adults: Evidence for a Phenotype.
      ]. People with diabetes have consistently been shown to have higher rates of frailty, ranging from 13 % [
      • Hanlon P.
      • Nicholl B.I.
      • Jani B.D.
      • Lee D.
      • McQueenie R.
      • Mair F.S.
      Frailty and pre-frailty in middle-aged and older adults and its association with multimorbidity and mortality: a prospective analysis of 493000 2008;737 UK Biobank participants.
      ,
      • Hanlon P.
      • Fauré I.
      • Corcoran N.
      • Butterly E.
      • Lewsey J.
      • McAllister D.
      • et al.
      Frailty measurement, prevalence, incidence, and clinical implications in people with diabetes: a systematic review and study-level meta-analysis.
      ] to 48 % [
      • Yoon S.-J.
      • Kim K.-I.
      Frailty and Disability in Diabetes.
      ,
      • Morley J.E.
      • Malmstrom T.K.
      • Rodriguez-Mañas L.
      • Sinclair A.J.
      Frailty, sarcopenia and diabetes.
      ] which is significantly higher than 5 % to 10 % seen in people without diabetes. While the relationship between T2D and frailty is yet to be fully understood, it is increasingly acknowledged that frailty plays a critical role in diabetes outcomes [
      • Ahmad E.
      • Sargeant J.A.
      • Yates T.
      • Webb D.R.
      • Davies M.J.
      Type 2 Diabetes and Impaired Physical Function: a Growing Problem.
      ], associated with higher risk of microvascular and microvascular complications [
      • Hanlon P.
      • Fauré I.
      • Corcoran N.
      • Butterly E.
      • Lewsey J.
      • McAllister D.
      • et al.
      Frailty measurement, prevalence, incidence, and clinical implications in people with diabetes: a systematic review and study-level meta-analysis.
      ], disability [
      • Yoon S.-J.
      • Kim K.-I.
      Frailty and Disability in Diabetes.
      ], hospitalization and mortality [
      • Ida S.
      • Kaneko R.
      • Imataka K.
      • Murata K.
      Relationship between frailty and mortality, hospitalization, and cardiovascular diseases in diabetes: a systematic review and meta-analysis.
      ]. As a consequence, frailty leads to significantly higher healthcare utilization in people with T2D [
      • Chao C.-T.
      • Wang J.
      • Chien K.-L.
      • et al.
      Both pre-frailty and frailty increase healthcare utilization and adverse health outcomes in patients with type 2 diabetes mellitus.
      ]. Whilst association between frailty and T2D has been established, it is unclear how T2D and frailty interact for health outcomes.
      While there is no universally agreed definition of frailty, physical frailty phenotype developed by Fried et al. [
      • Fried L.P.
      • Tangen C.M.
      • Walston J.
      • Newman A.B.
      • Hirsch C.
      • Gottdiener J.
      • et al.
      Frailty in Older Adults: Evidence for a Phenotype.
      ] is one of the most widely used frailty concepts compromising of five physical components: slow walking pace, low activity levels, low handgrip, weight loss and exhaustion. This phenotype has been validated against various outcomes from a range of populations [
      • Avila-Funes J.A.
      • Helmer C.
      • Amieva H.
      • Barberger-Gateau P.
      • Goff M.L.
      • Ritchie K.
      • et al.
      Frailty among community-dwelling elderly people in France: the three-city study.
      ,
      • Bandeen-Roche K.
      • Xue Q.-L.
      • Ferrucci L.
      • Walston J.
      • Guralnik J.M.
      • Chaves P.
      • et al.
      Phenotype of frailty: characterization in the women's health and aging studies.
      ,
      • Romero-Ortuno R.
      • Walsh C.D.
      • Lawlor B.A.
      • Kenny R.A.
      A Frailty Instrument for primary care: findings from the Survey of Health, Ageing and Retirement in Europe (SHARE).
      ]. Furthermore, frailty phenotype is the most commonly used measure of frailty in people with diabetes. Hanlon et al. [
      • Hanlon P.
      • Fauré I.
      • Corcoran N.
      • Butterly E.
      • Lewsey J.
      • McAllister D.
      • et al.
      Frailty measurement, prevalence, incidence, and clinical implications in people with diabetes: a systematic review and study-level meta-analysis.
      ] systematic review and meta-analysis reported that 58 % of studies investigating frailty prevalence, incidence and clinical outcomes in people with diabetes used frailty phenotype.
      Measurement of frailty phenotype involves data collection of all five components and it is assumed that greater number of components is associated with higher risk of a given outcome [
      • Fried L.P.
      • Tangen C.M.
      • Walston J.
      • Newman A.B.
      • Hirsch C.
      • Gottdiener J.
      • et al.
      Frailty in Older Adults: Evidence for a Phenotype.
      ]. While the components of frailty phenotype are thought to contribute equally to the frailty score, each component have different and specific effects on adverse outcomes [
      • Alves S.
      • Teixeira L.
      • Ribeiro O.
      • Paúl C.
      Examining Frailty Phenotype Dimensions in the Oldest Old. Original Research.
      ,
      • Hoogendijk E.O.
      • van Kan G.A.
      • Guyonnet S.
      • Vellas B.
      • Cesari M.
      Components of the Frailty Phenotype in Relation to the Frailty Index: Results From the Toulouse Frailty Platform.
      ]. Additionally, it possible that these components are redundant. To date, the independent contribution of the five frailty phenotype components have not been well evaluated and it remains unclear if all components are associated with the risk of mortality in people with T2D. Exploration of these components in people with T2D may be of interest for the more pragmatic frailty screening and intervention development in routine clinical practice.
      The aim of this analysis was to investigate the association and prognostic relevance of individual frailty components and overall frailty status with risk of all-cause and cardiovascular (CVD) mortality in adults with and without T2D and to investigate the relative risk of diabetes, frailty and their combination.

      2. Methods

      2.1 Study design

      This research was conducted using the UK Biobank under Application Number 52,553 and reported following STROBE guidelines (Checklist C1). The UK Biobank is a large population-based prospective study designed to investigate non-genetic and genetic determinants of health and diseases in middle age and older people in UK [
      • Sudlow C.
      • Gallacher J.
      • Allen N.
      • Beral V.
      • Burton P.
      • Danesh J.
      • et al.
      UK Biobank: An Open Access Resource for Identifying the Causes of a Wide Range of Complex Diseases of Middle and Old Age.
      ]. A total of 502,536 participants were recruited between 2006 and 2010 and attended 22 assessments centres across England, Scotland and Wales. Participants were asked to complete a touchscreen questionnaire, nurse-led interview and physical measurements. Data have been linked to hospital and mortality records. All participants gave written informed consent for data collection, analysis and linkage. Ethical approval was obtained by the North-West Research Ethics Committee (16/NW/0274).

      2.2 Assessment of baseline characteristics

      T2D status was based on a validated algorithm which consist of self-reported disease, medication and T2D noted in the medical history [
      • Eastwood S.V.
      • Mathur R.
      • Atkinson M.
      • Brophy S.
      • Sudlow C.
      • Flaig R.
      • et al.
      Algorithms for the Capture and Adjudication of Prevalent and Incident Diabetes in UK Biobank.
      ]. A previous medical history of cancer or CVD (defined as peripheral vascular disease, angina, heart attack/myocardial infarction, heart failure/ pulmonary oedema, stroke and transient ischaemic attack was self-reported during the baseline visit. Systolic and diastolic blood pressure was recorded using an automated machine. Lipid-lowering medication and antihypertensive drugs were self-reported. Low-density lipoprotein (LDL) cholesterol was measured by biochemical assays collected at baseline, using Beckman Coulter AU5800 Platform.
      Age was calculated from the date of birth and baseline assessment date. Body mass index (BMI) was calculated as (weight/heigh2) measured at the baseline. Self-reported smoking status was divided into three categories: current, non-smokers and previous smokers. Ethnicity was categorised into white and non-white as majority of the sample is white (94 %). Employment was categorised into employed, unemployed, retired and other. Alcohol intake was categorised into never/special occasion, 1–3 drinks per month, 1–4 drinks per week, daily or almost daily. Socioeconomic deprivation was based on the Townsend deprivation score which is a measure based on unemployment, household overcrowding, non-home ownership and non-car ownership with higher values indicating higher levels of deprivation [
      • Townsend P.
      Deprivation.
      ].

      2.3 Frailty phenotype

      Frailty status was assessed by Fried frailty phenotype which consists of five components related to frailty: weight loss, exhaustion, low physical activity, slow walking pace and low hangdrip. Each frailty deficits are adapted to available Biobank data (Table S1) [
      • Hanlon P.
      • Nicholl B.I.
      • Jani B.D.
      • Lee D.
      • McQueenie R.
      • Mair F.S.
      Frailty and pre-frailty in middle-aged and older adults and its association with multimorbidity and mortality: a prospective analysis of 493000 2008;737 UK Biobank participants.
      ]. Handgrip strength was assessed with Jamar J00105 hydraulic hand dynamometer. Variables related to frailty such as weight loss, exhaustion, walking pace and physical activity were self-reported using questionnaires. Physical activity (MET-hours/week) was calculated by multiplying the number of hours/week of each reported physical activity by the metabolic equivalent (MET) of that activity, summing across all activities, and categorizing the total by quintiles. Individuals with none of the indicators were characterised as robust. Those with one or two indicators were considered to represent pre-frail group and those with three or more indicators were categorised into frail group [
      • Fried L.P.
      • Tangen C.M.
      • Walston J.
      • Newman A.B.
      • Hirsch C.
      • Gottdiener J.
      • et al.
      Frailty in Older Adults: Evidence for a Phenotype.
      ]. Only UK Biobank participants with information available for all covariates and frailty components were included.

      2.4 Assessment of outcomes

      Date of death was obtained from the National Health Service (NHS) Information Centre in England and Wales; and the NHS Central Register Scotland. End of follow-up for each participant was recorded as the date of death or censoring dates (30 September 2021 for England and Wales and 31 October 2021 for Scotland), whichever came first. Cause of death was classified using the International Classification of Diseases (ICD-10) code assigned to the underlying (primary) cause of death. We defined cardiovascular mortality using the ICD-10 codes I00–I79.

      2.5 Statistical analysis

      Summary measures were described using mean (standard deviation) or median (interquartile range) for continuous variables, categorical data were given as a count (percentage).The association of individual frailty components and frailty status with all-cause and CVD mortality was assessed with Cox proportional hazard models, separately in subjects with and without T2D. The proportional hazards assumption was tested visually with log(-log) plots. To minimise potential risk of reverse causation bias, participants who died within two years of follow-up were excluded. Results were reported as hazard ratios (HR) together with 95 % confidence intervals (95 % CI). There were two models used in this analysis: Model 1, adjusted for age and sex (base model), and model 2, adjusted for deprivation (Townsend score), ethnicity, lipid-lowering medication, antihypertensive medication, LDL cholesterol, prevalent CVD, prevalent cancer, BMI, systolic blood pressure, smoking and alcohol (maximally adjusted model). To assess whether T2D modified the association frailty components and frailty overall status with mortality outcomes, likelihood ratio tests were conducted. For descriptive purposes, we also explored joint risk association of diabetes and frailty components with the mortality outcomes, adjusted for the same two models. We used non-T2D and absence of each frailty component as the reference group. In order to further help quantify the comparative relevance of each frailty component, we used the fully adjusted HRs to calculate the population-attributable risks (PAR) expressed as a percentage. Values were calculated using Miettinen’s equation recommended for adjusted estimates, expressed as PAR = p*100 (1 – 1/HRadj)2, where p = prevalence within cases, and HRadj = the adjusted HR [
      • Al Tunaiji H.
      • Davis J.C.
      • Mansournia M.A.
      • Khan K.M.
      Population attributable fraction of leading non-communicable cardiovascular diseases due to leisure-time physical inactivity: a systematic review.
      ,
      • Mansournia M.A.
      • Altman D.G.
      Population attributable fraction.
      ]. To quantify model predictive discrimination, this analysis used the concordance index (Harrell’s C-index) for risk of all-cause and CVD mortality. Each frailty component was added individually to a base model containing age and sex and to maximally adjusted model to allow for comparison between each component. The added prognostic value of each investigated factor was quantified as the difference in C-index compared to the base model and maximally adjusted model. P < 0.05 or 95 % CI not crossing the null were taken as significant. All analysis were conducted using Stata (Statistical Software IC16.0 StataCorp LLC, College Station, TX, USA).

      3. Results

      Out of 502,536 UK Biobank participants, 395,508 had complete outcome, frailty and covariate data and were included in this study (data selection flow shown in Supplementary Figure S1). In this study, 17,252 (4.3 %) were classified as having T2D. Baseline characteristics are summarised in Table S2. Compared to those without T2D, adults with T2D were more likely to be men (45.4 % vs 63.2 %), from a non-white ethnic minority (5.0 % vs 13.4 %) and were older (median 57 years vs 62 years), more likely to live within most deprived quintile (28.2 % vs 18 %), and be unemployed (6.9 % vs 2.3 %). In those without T2D, 19,624 (5.2 %) died from all causes and 3,441 (0.9 %) from CVD over a median (IQR) follow-up of 12.5 (11.8–13.2) years. In those with T2D, 2,418 (14.0 %) died from all-causes and 664 (3.8 %) from CVD.
      Frailty prevalence was higher in T2D compared to adults without T2D (12.4 % vs 2.7 %; OR 5.23, 95 % confidence interval [CI]: 4.98–5.07). Prevalence of all frailty components prevalence were also higher in adults with T2D, with the highest OR observed for walking pace (OR 3.76, 95 %CI: 3.61–3.91) after adjusting for age and sex (Table 1).
      Table 1Frailty prevalence.
      All participantsNo T2DT2DOR (95 % CI)
      Adjusted for age and sex.
      397,254378,25617,252
      Frailty components
      Slow walking pace25,702 (6.5)22,162 (5.9)3,540 (20.5)3.76 (3.61, 3.91)
      Low physical activity79,269 (20)74,498 (19.7)4,771 (27.7)1.61 (1.55, 1.67)
      Low handgrip34,706 (8.8)30,847 (8.2)3,859 (22.4)2.93 (2.82, 3.04)
      Weight loss60,989 (15.4)55,962 (14.8)5,027 (29.1)2.52 (2.43, 2.61)
      Exhaustion45,740 (11.6)42,435 (11.2)3,305 (19.2)2.33 (2.36, 2.42)
      Frailty status
      Pre-frail167,833 (42.4)157,924 (41.8)9,909 (57.4)1.96 (1.89, 2.02)
      Frail12,444 (3.2)10,312 (2.7)2,132 (12.4)5.23 (4.98, 5.07)
      OR, Odds Ratio; T2D, Type 2 Diabetes.
      * Adjusted for age and sex.
      Results for all-cause mortality with individual frailty components and overall frailty status in people with and without T2D are shown in Table 2. All five components were significantly associated with all-cause mortality. In both populations the strongest association amongst all frailty components was observed for slow walking pace (without T2D HR 2.25, 95 %CI: 2.12–2.38 and with T2D HR 1.95, 95 %CI: 1.67–2.28) maximally adjusted models. Overall frailty status provided a risk profile similar of that for slow walking pace (without T2D HR 2.14, 95 %CI: 1.99–2.29: and with T2D HR 1.80, 95 %CI: 1.54–2.10). The strength of association between frailty component and all-cause mortality was similar across those with and without T2D in the fully adjusted model, apart from frailty status where associations were stronger in those without T2D.
      Table 2All-cause mortality in adults with and without T2D.
      Without T2DWith T2DInteraction p value
      UnadjustedFrailty indicators
      Slow walking pace3.73 (3.56, 3.89)2.67 (2.35, 3.03)0.001
      Low physical activity1.15 (1.10, 1.19)1.25 (1.12, 1.39)0.138
      Low handgrip strength2.18 (2.10, 2.28)1.96 (1.72, 2.23)0.106
      Weight loss1.10 (1.06, 1.14)0.99 (0.89, 1.09)0.041
      Exhaustion1.21 (1.16,1.27)1.23 (1.10, 1.38)0.807
      Frail status
      Pre-frail1.29 (1.25, 1.33)1.29 (1.17, 1.42)0.992
      Frail2.80 (2.64, 2.97)2.06 (1.82, 2.32)0.001
      Model 1Frailty indicators
      Slow walking pace2.93 (2.80, 3.07)2.32 (2.32, 3.00)0.050
      Low physical activity1.27 (1.22, 1.32)1.34 (1.20, 1.50)0.308
      Low handgrip strength1.35 (1.29, 1.40)1.53 (1.34, 1.75)0.395
      Weight loss1.25 (1.21, 1.31)1.15 (1.05, 1.27)0.239
      Exhaustion1.72 (1.64, 1.79)1.59 (1.43, 1.79)0.321
      Frail status
      Pre-frail1.35 (1.31, 1.39)1.43 (1.30, 1.58)0.218
      Frail2.87 (2.71, 3.05)2.45 (2.16, 2.78)0.021
      Model 2Frailty indicators
      Slow walking pace2.25 (2.12, 2.38)1.95 (1.67, 2.28)0.105
      Low physical activity1.21 (1.16, 1.26)1.21 (1.08, 1.36)0.936
      Low handgrip strength1.27 (1.22, 1.32)1.39 (1.22, 1.61)0.547
      Weight loss1.17 (1.12, 1.22)1.06 (0.96, 1.17)0.153
      Exhaustion1.39 (1.33, 1.46)1.20 (1.06, 1.36)0.148
      Frail status
      Pre-frail1.24 (1.21, 1.28)1.25 (1.13, 1.38)0.789
      Frail2.14 (1.99, 2.29)1.80 (1.54. 2.10)0.042
      Mode 1. Age, sex.
      Model 2. Age, sex, ethnicity, smoking, BMI, systolic BP, lipid-lowering medication, blood-pressure medications, LDL cholesterol, employment, alcohol, smoking, deprivation, prevalent CVD, prevalent cancer.
      Results for CVD mortality are presented in Table 3, also found that slow walking pace (without T2D HR 2.36, 95 %CI: 2.08–2.69: with T2D HR 3.03; 95 %CI: 2.21–4.16) was associated with a greater risk than other frailty components or overall frailty status (without T2D HR 1.94, 95 %CI: 1.65–2.29: with T2D HR 2.04, 95 %CI: 1.51–2.75). The strength of association between frailty components and status with CVD mortality was similar across individuals with and without T2D.
      Table 3Cardiovascular disease cause mortality in adults with and without T2D.
      Without T2DWith T2DInteraction p value
      UnadjustedFrailty indicators
      Slow walking pace5.17 (4.67, 5.74)4.31 (3.29, 5.64)0.219
      Low physical activity1.14 (1.03, 1.25)1.38 (1.11, 1.71)0.098
      Low handgrip strength2.54 (2.32, 2.79)2.19 (1.70, 2.83)0.297
      Weight loss1.12 (1.02, 1.23)0.91 (0.76, 1.09)0.056
      Exhaustion1.13 (1.02, 1.26)1.18 (0.96, 1.46)0.709
      Frail status
      Pre-frail1.41 (1.32, 1.51)1.57 (1.29, 1.91)0.309
      Frail3.11 (2.71, 3.58)2.40 (1.89, 3.06)0.068
      Model 1Frailty indicators
      Slow walking pace4.01 (3.61, 4.45)4.41 (3.37, 5.78)0.675
      Low physical activity1. 29 (1.17, 1.42)1.48 (1.19, 1.83)0.172
      Low handgrip strength1.54 (1.40, 1.70)1.74 (1.35, 2.26)0.895
      Weight loss1.32 (1.20, 1.45)1.09 (0.90, 1.31)0.152
      Exhaustion1.73 (1.55, 1.92)1.57 (1.26, 1.94)0.172
      Frail status
      Pre-frail1.51 (1.41, 1.62)1.78 (1.46, 2.16)0.793
      Frail3.37 (2.94, 3.88)3.03 (2.37, 3.87)0.576
      Model 2Frailty indicators
      Slow walking pace2.36 (2.08, 2.69)3.03 (2.21, 4.16)0.441
      Low physical activity1.17 (1.06, 1.29)1.27 (1.01, 1.59)0.540
      Low handgrip strength1.40 (1.27, 1.55)1.62 (1.23, 2.14)0.623
      Weight loss1.19 (1.08, 1.32)0.99 (0.81, 1.20)0.160
      Exhaustion1.23 (1.09, 1.38)1.06 (0.84, 1.35)0.552
      Frail status
      Pre-frail1.29 (1.20, 1.39)1.39, (1.14, 1.71)0.414
      Frail1.94 (1.65, 2.29)2.04 (1.51, 2.75)0.724
      Mode 1. Age, sex.
      Model 2. Age, sex, ethnicity, smoking, BMI, systolic BP, lipid-lowering medication, blood-pressure medications, LDL cholesterol, employment, alcohol, smoking, deprivation, prevalent CVD, prevalent cancer.
      The results of frailty components and frailty status in combination with T2D status with risk of all-cause mortality are shown in Fig. 1 (data used to create figure can be found in Table S3). The combination of both T2D and presence of frailty or a frailty component was associated with greater risk of all-cause mortality compared to the reference category of no-T2D and absence of frailty. The highest risk for mortality was observed in people having both T2D and slow walking pace, which was more than three times higher (HR 3.19, 95 %CI: 2.91–3.50) in the maximally adjusted model. The risk of mortality was less pronounced in those with T2D and other frailty components, including overall frailty status (HR 2.69, 95 %CI: 2.41–3.00). Slow walking pace and presence of T2D also had strongest association with CVD-mortality (HR 4.17; 95 % CI 3.49–4.99) compared to T2D and overall frailty status (HR 2.95; 95 %CI 2.38–3.66) or other frailty indicators (Fig. 2 and Table S4).
      Figure thumbnail gr1
      Fig. 1Adjusted for Model 2. Age, sex, ethnicity, smoking, BMI, systolic BP, lipid-lowering medication, blood-pressure medications, LDL cholesterol, employment, alcohol, smoking, deprivation, prevalent CVD, prevalent cancer.
      Figure thumbnail gr2
      Fig. 2Adjusted for Model 2. Age, sex, ethnicity, smoking, BMI, systolic BP, lipid-lowering medication, blood-pressure medications, LDL cholesterol, employment, alcohol, smoking, deprivation, prevalent CVD, prevalent cancer.
      Frailty components and status were associated with PAR values higher in T2D than in those without (Figure S2). Slow walking pace had the highest PAR values (A) 16.4 % for all-cause, and 27.1 % for CVD mortality (B) than other frailty components or frailty in T2D. Pre-frailty status had the highest PAR values in those without T2D (Figure S2).
      Frailty components and status also had higher C-index values in participants with T2D (Tables S5- S8). For all-cause mortality, the addition of walking pace provided the greatest improvement in prognostic discrimination compared to other frailty components or overall risk status. In maximally adjusted models, C-index change for walking pace was +0.006 (95 %CI, 0.005 to 0.008) for adults without T2D and C-index change +0.013 (95 %CI, 0.007 to 0.018) in adults with T2D. Walking pace also provided an improvement in prognostic discrimination for CVD-mortality with C-index change of +0.008 (95 %CI 0.005 to 0.010) in adults without T2D and 0.020 (95 %CI 0.007 to 0.033) in adults with T2D.

      4. Discussion

      T2D and frailty is a severe phenotype with a three-fold increased risk of mortality compared to those without it. This study suggests that not all five components of frailty contribute equally to the risk of mortality, with slow walking pace having the strongest association with all-cause and CVD mortality in both individuals with and without T2D. Over 1 in 5 (20.5 %) of people with T2D reported a slow walking pace, with an odds of being a slower walker 3.8 times greater than those without diabetes. Walking pace had a higher PAR in T2D than frailty status, with potentially 27.1 % of CVD mortality cases being attributable to poor physical function as measured by a slow walking pace. Walking pace also provided greater prognostic discrimination than that associated with frailty status. When combined, subjects with a slow walking pace and T2D had 3 times the risk of all-cause mortality and 4 times the risk of CVD mortality compared to those with brisk walking pace without diabetes.
      Whilst the higher risk of frailty in T2D and higher risk of death have been well-established [
      • Hanlon P.
      • Fauré I.
      • Corcoran N.
      • Butterly E.
      • Lewsey J.
      • McAllister D.
      • et al.
      Frailty measurement, prevalence, incidence, and clinical implications in people with diabetes: a systematic review and study-level meta-analysis.
      ,
      • Ida S.
      • Kaneko R.
      • Imataka K.
      • Murata K.
      Relationship between frailty and mortality, hospitalization, and cardiovascular diseases in diabetes: a systematic review and meta-analysis.
      ], to our knowledge this is the first study to explore individual components of frailty and T2D with all-cause and CVD mortality. The results of this analysis suggest that self-reported slow walking pace was the most informative frailty component which also acted as a stronger prognostic marker than overall frailty. The findings for mortality outcomes are consistent with another study in older adults which found that walking pace was the strongest predictor of chronic disability, long term nursing-home stay and injurious fall compared to other frailty components [
      • Rothman M.D.
      • Leo-Summers L.
      • Gill T.M.
      Prognostic Significance of Potential Frailty Criteria.
      ]. Studies focusing on slow walking pace have also repeatedly shown a higher risk of mortality and adverse outcomes with strongest prognostic discrimination than other lifestyle behaviours or factors [
      • Argyridou S.
      • Zaccardi F.
      • Davies M.J.
      • Khunti K.
      • Yates T.
      Walking pace improves all-cause and cardiovascular mortality risk prediction: a UK Biobank prognostic study. European Journal of.
      ,
      • Zaccardi F.
      • Franks P.W.
      • Dudbridge F.
      • Davies M.J.
      • Khunti K.
      • Yates T.
      Mortality risk comparing walking pace to handgrip strength and a healthy lifestyle: A UK Biobank study.
      ,
      • Stamatakis E.
      • Kelly P.
      • Strain T.
      • Murtagh E.M.
      • Ding D.
      • Murphy M.H.
      Self-rated walking pace and all-cause, cardiovascular disease and cancer mortality: individual participant pooled analysis of 50 225 walkers from 11 population British cohorts.
      ].
      Identifying a slow walking pace as having the strongest association with mortality in T2D has implications for assessment and management of people with T2D. It has been shown that increase in usual gait speed over 1-year period decreased risk of mortality after adjusting for multiple risk factors whereas improvement in self-reported physical function questionnaires and activities of daily living (ADL) was not associated with survival [
      • Hardy S.E.
      • Perera S.
      • Roumani Y.F.
      • Chandler J.M.
      • Studenski S.A.
      Improvement in Usual Gait Speed Predicts Better Survival in Older Adults: GAIT SPEED GAIN PREDICTS SURVIVAL.
      ]. Similar results were seen in elderly veterans where improved gait speed improved health status and physical function, which resulted in fewer disabilities and healthcare utilisation [
      • Purser J.L.
      • Weinberger M.
      • Cohen H.J.
      • Pieper C.F.
      • Morey M.C.
      • Li T.
      • et al.
      Walking speed predicts health status and hospital costs for frail elderly male veterans.
      ]. Walking pace is thought to be a subclinical indicator of physiological reserve and function and is linked to physical resilience, defined as the ability to recover from “stressors”[
      • Kim H.-J.
      • Park I.
      • Lee H.J.
      • Lee O.
      The reliability and validity of gait speed with different walking pace and distances against general health, physical function, and chronic disease in aged adults.
      ]. As such, walking pace tests form the basis for assessing the effectiveness of pulmonary programmes at improving physical function [
      • Holland A.E.
      • Spruit M.A.
      • Troosters T.
      • Puhan M.A.
      • Pepin V.
      • Saey D.
      • et al.
      An official European Respiratory Society/American Thoracic Society technical standard: field walking tests in chronic respiratory disease.
      ]. While there is an increase in studies analysing the effect of interventions targeting frailty [
      • de Labra C.
      • Guimaraes-Pinheiro C.
      • Maseda A.
      • Lorenzo T.
      • Millán-Calenti J.C.
      Effects of physical exercise interventions in frail older adults: a systematic review of randomized controlled trials.
      ], very few have evaluated the effect of interventions on components of frailty [
      • Liao Y.-Y.
      • Chen I.H.
      • Wang R.-Y.
      Effects of Kinect-based exergaming on frailty status and physical performance in prefrail and frail elderly: a randomized controlled trial.
      ]. Further research could potentially focus on more modifiable components, such as walking pace [
      • Eastwood S.V.
      • Mathur R.
      • Atkinson M.
      • Brophy S.
      • Sudlow C.
      • Flaig R.
      • et al.
      Algorithms for the Capture and Adjudication of Prevalent and Incident Diabetes in UK Biobank.
      ], and help customise interventions for diabetes management, for example, by identifying those with a slow walking pace and referring to rehabilitation type interventions to improve physical function.
      The strengths of this study includes large sample size, with data related to lifestyle, conditions and mortality. However, there are several limitations in the study. First, this study used baseline data thus we were not able to assess the impact of diabetes on frailty status over time. T2D was identified using algorithm developed by Eastwood et al [
      • Eastwood S.V.
      • Mathur R.
      • Atkinson M.
      • Brophy S.
      • Sudlow C.
      • Flaig R.
      • et al.
      Algorithms for the Capture and Adjudication of Prevalent and Incident Diabetes in UK Biobank.
      ] based on self-report physician-diagnosed T2D, which may be subject to errors as some participants may not recall their diagnosis accurately or do not want to report it [
      • Bergmann M.M.
      • Jacobs E.J.
      • Hoffmann K.
      • Boeing H.
      Agreement of self-reported medical history: comparison of an in-person interview with a self-administered questionnaire.
      ,
      • Jackson J.M.
      • DeFor T.A.
      • Crain A.L.
      • et al.
      Validity of diabetes self-reports in the Women's Health Initiative.
      ]. While authors have found that it was pragmatic and valid approach, the overall prevalence was slightly lower than the general population. Nevertheless, self-reporting have been shown to be a valid method in distinguishing diabetes [
      • Pastorino S.
      • Richards M.
      • Hardy R.
      • et al.
      Validation of self-reported diagnosis of diabetes in the 1946 British birth cohort.
      ]. Other than handgrip strength which was assessed by trained staff according to standard operating procedures, this study used self-report measures which are subject to bias. Further, UK Biobank did not aim to assess frailty specifically. However, it has been shown that self-reported indicators of frailty produce similar prevalence estimated to test based measures [

      Theou O, O'Connell M, King-Kallimanis B, O'Halloran A, Rockwood K, Kenny R. Measuring frailty using self-report and test-based health measures. Age and ageing. 02/16 2015;44doi:10.1093/ageing/afv010.

      ]. It is important to note that this study may not be applicable to older adults aged 70 and older as Biobank participant age range is 37–73 years and those from ethnic minority background. Finally, as this is an observational study, causality between frailty components and mortality cannot be confirmed. Nevertheless, this study gives novel insight into frailty and T2D in middle-aged adults.

      5. Conclusion

      While frailty phenotype is a multicomponent model, assessing all five components may be a barrier for routine clinical practice and may not be necessary within diabetes management as not all provide additional information on frailty state. A pragmatic simple measure of slow walking pace identifies more people whilst being a stronger predictor of mortality outcomes than did overall frailty status. Additionally, this can easily be implemented in routine care. Further research is required to establish whether identifying slow walkers and referring rehabilitation style exercise programmes can improve longer-term outcomes within the management of T2D.

      Declaration of Competing Interest

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Acknowledgments

      The study has been conducted using the UK Biobank Resource (Application Number 52553).

      Funding

      The research was supported by the NIHR Leicester Biomedical Research Centre, which is a partnership between University Hospitals of Leicester NHS Trust, Loughborough University and the University of Leicester. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.

      References

        • García-Esquinas E.
        • Graciani A.
        • Guallar-Castillón P.
        • López-García E.
        • Rodríguez-Mañas L.
        • Rodríguez-Artalejo F.
        Diabetes and risk of frailty and its potential mechanisms: a prospective cohort study of older adults.
        J Am Med Dir Assoc. 2015; 16: 748-754https://doi.org/10.1016/j.jamda.2015.04.008
        • Fried L.P.
        • Tangen C.M.
        • Walston J.
        • Newman A.B.
        • Hirsch C.
        • Gottdiener J.
        • et al.
        Frailty in Older Adults: Evidence for a Phenotype.
        J Gerontol: Series A. 2001; 56: M146-M157
        • Hanlon P.
        • Nicholl B.I.
        • Jani B.D.
        • Lee D.
        • McQueenie R.
        • Mair F.S.
        Frailty and pre-frailty in middle-aged and older adults and its association with multimorbidity and mortality: a prospective analysis of 493000 2008;737 UK Biobank participants.
        Lancet Public Health. 2018; 3: e323-e332https://doi.org/10.1016/S2468-2667(18)30091-4
        • Hanlon P.
        • Fauré I.
        • Corcoran N.
        • Butterly E.
        • Lewsey J.
        • McAllister D.
        • et al.
        Frailty measurement, prevalence, incidence, and clinical implications in people with diabetes: a systematic review and study-level meta-analysis.
        Lancet Healthy Longevity. 2020; 1: e106-e116
        • Yoon S.-J.
        • Kim K.-I.
        Frailty and Disability in Diabetes.
        Ann Geriatr Med Res. 2019; 23: 165-169https://doi.org/10.4235/agmr.19.0036
        • Morley J.E.
        • Malmstrom T.K.
        • Rodriguez-Mañas L.
        • Sinclair A.J.
        Frailty, sarcopenia and diabetes.
        J Am Med Dir Assoc. Dec 2014; 15: 853-859https://doi.org/10.1016/j.jamda.2014.10.001
        • Ahmad E.
        • Sargeant J.A.
        • Yates T.
        • Webb D.R.
        • Davies M.J.
        Type 2 Diabetes and Impaired Physical Function: a Growing Problem.
        Diabetology. 2022; 3: 30-45https://doi.org/10.3390/diabetology3010003
        • Ida S.
        • Kaneko R.
        • Imataka K.
        • Murata K.
        Relationship between frailty and mortality, hospitalization, and cardiovascular diseases in diabetes: a systematic review and meta-analysis.
        Cardiovasc Diabetol. 2019; 18: 81https://doi.org/10.1186/s12933-019-0885-2
        • Chao C.-T.
        • Wang J.
        • Chien K.-L.
        • et al.
        Both pre-frailty and frailty increase healthcare utilization and adverse health outcomes in patients with type 2 diabetes mellitus.
        Cardiovasc Diabetol. 2018/09/27; 17: 130https://doi.org/10.1186/s12933-018-0772-2
        • Avila-Funes J.A.
        • Helmer C.
        • Amieva H.
        • Barberger-Gateau P.
        • Goff M.L.
        • Ritchie K.
        • et al.
        Frailty among community-dwelling elderly people in France: the three-city study.
        J Gerontol A Biol Sci Med Sci. 2008; 63: 1089-1096
        • Bandeen-Roche K.
        • Xue Q.-L.
        • Ferrucci L.
        • Walston J.
        • Guralnik J.M.
        • Chaves P.
        • et al.
        Phenotype of frailty: characterization in the women's health and aging studies.
        J Gerontol A Biol Sci Med Sci. 2006; 61: 262-266
        • Romero-Ortuno R.
        • Walsh C.D.
        • Lawlor B.A.
        • Kenny R.A.
        A Frailty Instrument for primary care: findings from the Survey of Health, Ageing and Retirement in Europe (SHARE).
        BMC Geriatr. 2010; 10
        • Alves S.
        • Teixeira L.
        • Ribeiro O.
        • Paúl C.
        Examining Frailty Phenotype Dimensions in the Oldest Old. Original Research.
        Front Psychol. 2020; : 11https://doi.org/10.3389/fpsyg.2020.00434
        • Hoogendijk E.O.
        • van Kan G.A.
        • Guyonnet S.
        • Vellas B.
        • Cesari M.
        Components of the Frailty Phenotype in Relation to the Frailty Index: Results From the Toulouse Frailty Platform.
        J Am Med Dir Assoc. 2015; 16: 855-859https://doi.org/10.1016/j.jamda.2015.04.007
        • Sudlow C.
        • Gallacher J.
        • Allen N.
        • Beral V.
        • Burton P.
        • Danesh J.
        • et al.
        UK Biobank: An Open Access Resource for Identifying the Causes of a Wide Range of Complex Diseases of Middle and Old Age.
        PLoS Med. 2015; 12: e1001779https://doi.org/10.1371/journal.pmed.1001779
        • Eastwood S.V.
        • Mathur R.
        • Atkinson M.
        • Brophy S.
        • Sudlow C.
        • Flaig R.
        • et al.
        Algorithms for the Capture and Adjudication of Prevalent and Incident Diabetes in UK Biobank.
        PLoS One. 2016; 11: e0162388https://doi.org/10.1371/journal.pone.0162388
        • Townsend P.
        Deprivation.
        J Social Policy. 1987; 16: 125-146
        • Al Tunaiji H.
        • Davis J.C.
        • Mansournia M.A.
        • Khan K.M.
        Population attributable fraction of leading non-communicable cardiovascular diseases due to leisure-time physical inactivity: a systematic review.
        BMJ Open Sport & Exercise Med. 2019; 5: e000512https://doi.org/10.1136/bmjsem-2019-000512
        • Mansournia M.A.
        • Altman D.G.
        Population attributable fraction.
        BMJ. 2018; 360k757https://doi.org/10.1136/bmj.k757
        • Rothman M.D.
        • Leo-Summers L.
        • Gill T.M.
        Prognostic Significance of Potential Frailty Criteria.
        J Am Geriatr Soc. 2008; 56: 2211-2216https://doi.org/10.1111/j.1532-5415.2008.02008.x
        • Argyridou S.
        • Zaccardi F.
        • Davies M.J.
        • Khunti K.
        • Yates T.
        Walking pace improves all-cause and cardiovascular mortality risk prediction: a UK Biobank prognostic study. European Journal of.
        Prev Cardiol. 2020/07/01 2019,; 27: 1036-1044https://doi.org/10.1177/2047487319887281
        • Zaccardi F.
        • Franks P.W.
        • Dudbridge F.
        • Davies M.J.
        • Khunti K.
        • Yates T.
        Mortality risk comparing walking pace to handgrip strength and a healthy lifestyle: A UK Biobank study.
        Eur J Preventive Cardiology. 2021; 28: 704-712
        • Stamatakis E.
        • Kelly P.
        • Strain T.
        • Murtagh E.M.
        • Ding D.
        • Murphy M.H.
        Self-rated walking pace and all-cause, cardiovascular disease and cancer mortality: individual participant pooled analysis of 50 225 walkers from 11 population British cohorts.
        Br J Sports Med. 2018; 52: 761https://doi.org/10.1136/bjsports-2017-098677
        • Hardy S.E.
        • Perera S.
        • Roumani Y.F.
        • Chandler J.M.
        • Studenski S.A.
        Improvement in Usual Gait Speed Predicts Better Survival in Older Adults: GAIT SPEED GAIN PREDICTS SURVIVAL.
        J Am Geriatr Soc. 2007; 55: 1727-1734
        • Purser J.L.
        • Weinberger M.
        • Cohen H.J.
        • Pieper C.F.
        • Morey M.C.
        • Li T.
        • et al.
        Walking speed predicts health status and hospital costs for frail elderly male veterans.
        J Rehabil Res Dev Jul-Aug. 2005; 42: 535
        • Kim H.-J.
        • Park I.
        • Lee H.J.
        • Lee O.
        The reliability and validity of gait speed with different walking pace and distances against general health, physical function, and chronic disease in aged adults.
        J Exerc Nutrition Biochem. 2016; 20: 46-50https://doi.org/10.20463/jenb.2016.09.20.3.7
        • Holland A.E.
        • Spruit M.A.
        • Troosters T.
        • Puhan M.A.
        • Pepin V.
        • Saey D.
        • et al.
        An official European Respiratory Society/American Thoracic Society technical standard: field walking tests in chronic respiratory disease.
        Eur Respir J. 2014; 44: 1428-1446
        • de Labra C.
        • Guimaraes-Pinheiro C.
        • Maseda A.
        • Lorenzo T.
        • Millán-Calenti J.C.
        Effects of physical exercise interventions in frail older adults: a systematic review of randomized controlled trials.
        BMC Geriatr. 2015; 15
        • Liao Y.-Y.
        • Chen I.H.
        • Wang R.-Y.
        Effects of Kinect-based exergaming on frailty status and physical performance in prefrail and frail elderly: a randomized controlled trial.
        Sci Rep. 2019/06/27; 9: 9353https://doi.org/10.1038/s41598-019-45767-y
        • Eastwood S.V.
        • Mathur R.
        • Atkinson M.
        • Brophy S.
        • Sudlow C.
        • Flaig R.
        • et al.
        Algorithms for the Capture and Adjudication of Prevalent and Incident Diabetes in UK Biobank.
        PLoS One. 2016; 11: e0162388
        • Bergmann M.M.
        • Jacobs E.J.
        • Hoffmann K.
        • Boeing H.
        Agreement of self-reported medical history: comparison of an in-person interview with a self-administered questionnaire.
        Eur J Epidemiol. 2004; 19: 411-416https://doi.org/10.1023/b:ejep.0000027350.85974.47
        • Jackson J.M.
        • DeFor T.A.
        • Crain A.L.
        • et al.
        Validity of diabetes self-reports in the Women's Health Initiative.
        Menopause. Aug 2014; 21: 861-868https://doi.org/10.1097/gme.0000000000000189
        • Pastorino S.
        • Richards M.
        • Hardy R.
        • et al.
        Validation of self-reported diagnosis of diabetes in the 1946 British birth cohort.
        Prim Care Diabetes. Oct 2015; 9: 397-400https://doi.org/10.1016/j.pcd.2014.05.003
      1. Theou O, O'Connell M, King-Kallimanis B, O'Halloran A, Rockwood K, Kenny R. Measuring frailty using self-report and test-based health measures. Age and ageing. 02/16 2015;44doi:10.1093/ageing/afv010.