Impact of Diabetes on Physical Function in Older People

Participants were recruited as part of a case-control study. We included those known to have diabetes, aged 65 years or over, from a sample of rural and urban general practices in Wales. At the time of recruitment, all of the case subjects met the World Health Organization criteria for the diagnosis of diabetes (4). The control group of nondiabetic participants was frequency matched with the case subjects for age, sex, and general practice. Matching by ethnicity was not possible because of the small numbers of non-Caucasians.

Laboratory testing was used to confirm that the entire control group had normal glucose tolerance, as evidenced by fasting plasma glucose <6.5 mmol/l and no evidence of glycosuria. All study participants were interviewed by trained research nurses; a comprehensive assessment was carried out including demographic, medical, social, and functional details with information corroborated from family members and medical records when necessary. Specific questionnaires used and reported in this study were the Barthel index, a validated scale measuring basic activities of daily living from 0 to 20 with higher scores indicating greater independence (5); the Nottingham Extended Activities of Daily Living Scale, measuring extended activities of daily living such as shopping and driving from 0 to 22, with higher scores indicating greater independence (6); and the SF-36 health status questionnaire, a generic measure of health-related quality of life, with scores ranging from 0 to 100, higher scores indicating greater satisfaction (7). Data were also collected on cognition and visual acuity, reported elsewhere (8,9).

Univariate and multivariate regressions analysis were used, with summary results presented as odds ratios (ORs) and 95% CIs. Model parameters were retained if they were clinically important potential confounders or were statistically significant at the 0.05 level in univariate analysis.

The study population consisted of 403 case subjects and 403 matched control subjects. The median (interquartile range [IQR]) age was 75 years (69–80), and 51% were female. More of the subjects with diabetes lived in a care home compared with control subjects (7 vs. 4%, P < 0.001). Participants with diabetes were on diet alone (24%), oral hypoglycemics (59%), or insulin (17%); the mean (95% CI) for A1C was 7.8% (7.6–8.0), mean random glucose was 11.2 mmol/l (10.7–11.7), and median (IQR) duration of diabetes was 7 years (3–14). In both cohorts, 51% were current or ex-smokers. Diabetic subjects had more comorbidities than control subjects (mean 2.5 vs. 1.9, P < 0.0001), especially cardiovascular disease (Table 1).

There were significant differences in the distribution of Barthel scores between cohorts (case subjects median 20 [IQR 17–20], control subjects 20 [19–20], P < 0.0001). The main differences were seen at the lower end of the Barthel score, with 4 versus 1% being highly dependant (Barthel 0–5) and 4 versus 2% being moderately dependant (Barthel 6–11).

Significant differences were seen in mobility measured by the Nottingham extended activities of daily living scale; median (IQR) for subjects with diabetes 0.5 (0–1) compared with 5 (3–6) for control subjects (P < 0.0001).

Subjects with diabetes were more likely to use a mobility aid than control subjects: 7 versus 2% used a wheelchair regularly, 7 vs. 3% a frame, and 33 vs. 26% a walking stick. Overall, 187 of 403 of case subjects (46%) compared with 125 of 403 control subjects (31%) used some form of mobility aid (P < 0.0001).

There were significant differences in the physical function domain of the SF36; the median (IQR) SF36 for physical function was 60 (30–85) in control subjects versus 40 (10–70) in those with diabetes (P < 0.0001).

We ascribed use of a mobility aid as a marker of perceived mobility limitation. In a multivariate model controlled for age, hypertension, cerebrovascular disease, COPD, cancer, osteoarthritis, and dementia, diabetes remained significantly associated with mobility limitation: OR (95% CI) 2.10 (1.50–2.83) (P < 0.001).

The main findings of this study are a reduction in physical function and health status in patients with diabetes compared with age- and sex-matched control subjects living in the same community. Along with our previous findings of impaired cognition in this cohort (8), we have provided additional evidence of the impact of diabetes on global function.

Likely explanations for lower-limb dysfunction are peripheral neuropathy and peripheral vascular disease caused by diabetes. Other important contributors to mobility limitation include age, hypertension, cerebrovascular disease, COPD, cancer, osteoarthritis, and dementia.

There is growing recognition that older people with diabetes have needs over and above those related to management of the metabolic disturbance and its complications alone, in particular relating to lower-limb mobility and function (1,016). The increase in functional limitation in diabetes seen in this study is similar to that reported in studies from the U.S. (1,012) (17,18) and Hong Kong (19); one previous study from Norway has also shown a link between lower-limb disability and quality of life (20).

There is a growing body of evidence demonstrating that physical activity, in particular resistance training (21), is a key component in the management of diabetes in older people, and it is likely that exercise will have benefits both in terms of metabolic control and improving daily function, especially mobility. Older people with diabetes are at increased risk of falls, through mechanisms such as lower-limb dysfunction, cardiovascular disease, polypharmacy, and impaired balance (22). There is robust clinical evidence supporting the role of targeted exercise programs (23) and comprehensive geriatric assessment (16,24) and interventions in falls prevention. Given the multifactorial problems facing the older person with diabetes, it would seem reasonable to recommend a comprehensive intervention, similar to that used in falls prevention programs (25).