We examined the effect of cutaneous warming on the transcutaneous oxygen measurement (TcPO2) at standard locations below the knee, on the dorsal foot, on the plantar right great toe, and on the plantar surface under the second metatarsal head of 656 diabetic and 16 nondiabetic subjects. All subjects underwent a directed medical history, physical examination, and neurovascular lower extremity evaluation and assessment of autonomic neuropathy. Associations between autonomic neuropathy and TcPO2 were tested using two-way analysis of variance and multiple linear regression. An unexpected, statistically significant fall in TcPO2 occurred with cutaneous warming from 37 degrees C to 44 degrees C on the plantar great toe surface that did not differ by presence of autonomic neuropathy, or the presence of diabetes (mean change: -8.82 mmHG, 95% confidence interval [CI]: -7.70 to -9.93). The TcPO2 measured at the plantar metatarsal head site also fell with cutaneous warming from 37 degrees C to 44 degrees C (mean change: -9.67 mmHG, 95% CI: -7.75 to -11.59). As expected, the TcPO2 increased with cutaneous warming from 37 degrees C to 44 degrees C on the dorsal foot site (mean rise:= 35.61 mmHg, 95% CI: 34.18 to 37.04). The mean TcPO2 at any site did not differ by presence of autonomic neuropathy at either 37 degrees C or 44 degrees C. We conclude that cutaneous warming leads to a paradoxical fall in TcPO2 on the plantar foot surface that does not depend on the presence of autonomic neuropathy or diabetes. Caution in interpretation of ambient versus warmed plantar TcPO2 values is recommended.