A longitudinal study of foot ulceration and its risk factors in community-based patients with type 2 diabetes: The Fremantle Diabetes Study

Abstract

Aims

To determine the prevalence and associates of foot ulcer, and the subsequent incidence and predictors of first-ever hospitalisation for this complication, in well-characterised community-based patients with type 2 diabetes.

Methods

Baseline foot ulceration was ascertained in 1296 patients (mean age 64 years, 48.6% male, median diabetes duration 4.0 years) recruited to the longitudinal Fremantle Diabetes Study between 1993 and 1996. Incident hospitalisation for foot ulceration was monitored through validated data linkage until end-December 2010.

Results

At baseline, 16 participants (1.2%) had a foot ulcer which was independently associated with intermittent claudication, peripheral sensory neuropathy (PSN) and diabetes duration (P ≤ 0.01). The incidence of hospitalisation for this complication in those without prior/prevalent foot ulceration was 5.21 per 1000 patient-years. This rate and other published data suggest that 1 in 7–10 foot ulcers require hospitalisation. In a Cox proportional hazards model, intermittent claudication and PSN were significant independent predictors of time to admission with foot ulceration, in addition to retinopathy, cerebrovascular disease, HbA_1c, alcohol consumption, renal impairment, peripheral arterial disease and pulse pressure (P ≤ 0.038).

Conclusions

These data confirm PSN as an important risk factor for foot ulceration but, in contrast to some other studies, peripheral arterial disease was also a major independent contributor. Associations between hospitalisation for foot ulcer and both retinopathy and raised pulse pressure suggest a role for local microvascular dysfunction, while alcohol may have non-neuropathic toxic effects on skin/subcutaneous structures. The multifactorial nature of foot ulceration complicating type 2 diabetes may have implications for its management.

Introduction

Foot ulcers contribute significantly to the morbidity, mortality and cost associated with diabetes [8]. Accurate estimates of the prevalence and incidence of foot ulceration complicating type 2 diabetes have, however, been difficult to obtain [22], [30]. Available data from largely primary care studies suggest that the point prevalence of an active ulcer is up to 1.8% [17], [25], with an annual incidence ranging from 1.0% to 4.1% [37]. Such estimates can, however, be influenced by how foot ulcers are defined and ascertained, as well as by racial/ethnic differences in propensity to foot ulceration [2], [26] and between-sample differences in the management of diabetes and associated foot disease that are strongly influenced by socio-economic and behavioural factors [30].

Similarly, risk factors for diabetic foot ulcers have been incompletely characterised [13], [14], [32]. In recent meta-analyses [14], [32], diagnostic tests and physical signs that detect peripheral sensory neuropathy (PSN) and excessive plantar pressure were consistently associated with future diabetic foot ulceration, but the role of other candidate risk factors including HbA_1c and ankle brachial index (ABI) were not. Differences in methodology including design (retrospective vs prospective), sources of patients (clinic vs community), attrition rates, durations of follow-up, definitions and ascertainment of foot ulcer, standardisation of data collection, cut-points for interpretation of diagnostic tests, and availability of potential explanatory variables for multivariate assessment of independent associates, are all barriers to effective pooling of data [14], [32]. Indeed, very few individual studies have included a full range of predictive variables in their analyses [32], a deficiency that has prompted international collaboration aimed at performing a more detailed systematic review [13].

Issues with ascertainment are highlighted by the observation that a significant proportion of diabetic patients can develop foot ulcers, including those that persist for more than three weeks, which heal without healthcare system intervention [21]. This reflects patient delay in seeking help [35], [36], especially by those who do not regard an ulcer as a serious complication of diabetes [36]. Given the problems associated with complete identification of all ulcers in a longitudinal natural history study and the inconsistent findings from previous studies, the aims of the present study were to (i) determine the prevalence and associates of foot ulceration in patients with type 2 diabetes from a large representative community-based cohort detected through active screening, and (ii) identify the incidence and predictors of all foot ulceration severe enough to require hospitalisation during up to 17 years of follow-up of the same cohort.