OBJECTIVE To investigate changes in both core and peripheral skin-surface temperatures during and after application of a unilateral leg pneumatic tourniquet in adult patients. METHODS Prospective, observational clinical study. METHODS University hospital. METHODS 21 ASA physical status I and II adult patients scheduled for elective leg orthopedic surgery with lumbar epidural anesthesia. METHODS Rectal and fingertip skin-surface temperatures were recorded every minute after steady-state lumbar epidural anesthesia was established. RESULTS Significant (p < 0.05) increases in both rectal and fingertip temperatures were observed during tourniquet application for 91 +/- 6 minutes from 36.5 +/- 0.14 degrees C to 37.0 +/- 0.17 degrees C and from 32.6 +/- 0.79 degrees C to 35.5 +/- 0.44 degrees C, respectively. In contrast, both rectal and fingertip temperatures progressively decreased following tourniquet release; significant (p < 0.05) decreases in the rectal and fingertip temperatures were observed 6 and 5 minutes after tourniquet release, respectively. Decreases (approximately maximum) in the rectal and fingertip temperatures 15 minutes after tourniquet release were 0.25 +/- 0.05 degrees C and 1.26 +/- 0.26 degrees C, respectively. In each case, changes in fingertip temperature were approximately six times greater than those in the rectal temperature. CONCLUSIONS Limb tourniquets appear to cause thermal perturbations during epidural anesthesia. The progressive increases in core temperature during tourniquet application presumably resulted from constraint of metabolic heat to the core thermal compartment, and the greater increases in the skin-surface temperature during tourniquet application appear to represent vasodilation in response to the core hyperthermia. On the other hand, redistribution of body heat and the efflux of hypothermic venous blood from the tourniqueted area into systemic circulation following tourniquet deflation probably decreased the core temperature, which might switch off the thermoregulatory vasodilation, leading to the decreases in skin-surface temperature. Recognition of these thermal perturbations are useful in diagnosing intraoperative thermal perturbations.