Diagnostic utility of the history and physical examination for peripheral vascular disease among patients with diabetes mellitus

Background: We assessed the value of the medical history and physical examination in the diagnosis of peripheral vascular disease in diabetic subjects. Methods: We performed a cross-sectional study in 631 diabetic veteran enrollees of a general internal medicine clinic that compared data obtained from a history and clinical evaluation with the presence of severe peripheral vascular disease defined as an ankle-arm index (AAI) less than or equal to 0.5 derived from Doppler blood pressure measurement. Results: We identified 90 limbs with an AAI less than or equal to 0.5. Results presented below apply to the right leg, but do not differ from the left. Diminished or absent foot peripheral pulses (sensitivity 65%, specificity 78%), venous filling time > 20 sec (sensitivity 22%, specificity 93.9%), age > 65 years (sensitivity 83%, specificity 54%), claudication symptoms in < 1 block (sensitivity 50%, specificity 87%), and patient reported history of physician diagnosed peripheral vascular disease (PVD) (sensitivity 80%, specificity 70%) had the largest positive (or smallest negative) likelihood ratios. Capillary refill time > 5 sec or foot characteristics (absent hair, blue/purple color, skin coolness, or atrophy) conveyed little diagnostic information. Individual factors did not change disease probability to a clinically important degree. A stepwise logistic regression model identified four factors significantly (P < 0.05) associated with low AAI: absent or diminished peripheral pulses, patient reported history of PVD, age, and venous filling time. Substitution of < 1 block claudication for PVD history in this model resulted in a small reduction in model accuracy. Conclusions: Many purportedly useful historical and exam findings need not be elicited in diabetic patients suspected of having severe peripheral vascular disease, since most information related to probability of this disorder may be obtained from patient age, self-reported history of physician diagnosed PVD (or < 1 block claudication), peripheral pulse palpation, and venous filling time. (C) 1997 Elsevier Science Inc.