To identify and quantify independent physiological risk factors for foot lesions in diabetic patients.
There were 352 patients enrolled in a 1-year randomized controlled trial aimed at reducing risks for lower-extremity pathology through patient education and system interventions. Inclusion criteria were as follows: being age 40 years or over, being at or above ideal body weight, and having been diagnosed with NIDDM. Participants were predominantly African-American (76%), elderly (mean 60 years of age), indigent (77% with annual income < $10,000), or women (81%) who had diabetes for 10 years. Prospective multivariate modeling used baseline clinical signs (e.g., blood pressure, dermatological characteristics, and neuropathic measures) and laboratory values (e.g., lipid profiles and measures of glycemic control) to predict foot lesions rated using the Seattle Wound Classification.
When controlling for intervention effects, only measures of neuropathy (monofilament testing [odds ratio {OR} 2.75, 95% CI 1.55–4.88] and thermal sensitivity testing [2.18, 1.13–4.21]) predicted wounds classified 1.2 (minor injury), but investigation of wounds rated at least 1.3 (nonulcerated lesions) indicated baseline wounds (13.41), 3.19–56.26), monofilament abnormalities (5.23, 2.26–12.13), and low HDL (1.63, 1.11–2.39) as predictors. Although fungal dermatitis, dry cracked skin, edema, ingrown nails, microalbuminuria, fasting blood glucose, and hemoglobin A1c were candidates for one or both of the multivariable models (P < 0.3), they were not significant multivariate predictors.
Lesions may be preventable with aggressive screening for peripheral neuropathy and abnormal lipids. Also, these results provide empirical support for the commonly held belief that foot lesions prospectively predict future wounds.
