Abstract
OBJECTIVE: We compared the reproducibility and accuracy of conventional clinical examination of the diabetic foot to monofilament examination. We also sought to simplify the monofilament examination by reducing it to fewer touch points. METHODS: In a cross-sectional study at 10 centers in the United States, Canada, and Switzerland, general internists and residents performed a structured history and physical examination for neuropathy on the feet of diabetic patients. Independent examination by two observers included monofilament sensation, pinprick, vibration, position sense, and ankle reflexes. MAIN RESULTS: A total of 304 patients were examined by at least one practitioner, and 200 received duplicate examinations. Monofilament examination and ankle reflexes had the best reproducibility, with moderate agreement (kappa = 0.59); pinprick, position, and vibration sense had fair agreement (kappa = 0.28-0.36). No component of the history or physical examination, singly or in aggregate, was both sensitive and specific for identifying a patient with an abnormal monofilament examination. A simplified monofilament examination using only 4 sites per foot (total 8 sites) detected 90% of patients with an abnormal 16-site monofilament evaluation. CONCLUSIONS: Conventional clinical examination had low reproducibility and correlated poorly with monofilament examination for the identification of the at-risk patient. The Semmes-Weinstein monofilament examination, a reproducible, valid, and generalizable test of foot sensation, is recommended as the screening procedure of choice for examining diabetic feet.