This prospective longitudinal study assessed whether baseline mean skin temperature measurements are useful in predicting the most common foot-related complications of diabetes mellitus. We evaluated the mean of baseline skin temperatures taken bilaterally from six plantar sites in 1,588 patients with diabetes. There was no difference in skin temperature based on neuropathy, foot laterality, or foot risk category or between people with and without foot deformity and elevated plantar foot pressure. Whereas people with Charcot's arthropathy had slightly but significantly higher mean temperatures (84.8 degrees +/- 3.5 degrees F versus 82.5 degrees +/- 4.7 degrees F), this was not true for those who developed ulcers or infections or who underwent amputations. The presence of vascular disease was not associated with lower skin temperatures. Mexican Americans (83.0 degrees +/- 4.6 degrees F) and blacks (83.6 degrees +/- 4.5 degrees F) had higher mean skin temperatures at baseline than did non-Hispanic whites (81.8 degrees +/- 4.6 degrees F). Baseline measurement of nonfocal mean skin temperatures is not an effective means of screening people for future events. Regular assessment of skin temperatures, using the contralateral site as a physiologic control, may be a better use of this technology.