The hypothesis that forced-air warming preserves core temperature better than circulating-water mattresses was tested in: (a) 16 adults undergoing major maxillofacial surgery, including radical node resection and flap reconstruction; (b) 53 adults undergoing hip arthroplasty, having approximately 25% of their body surface area available for warming; (c) 20 infants undergoing minor maxillofacial surgery; and (d) 10 young children undergoing pelvic or femoral osteotomies. Patients having each type of surgery were randomly assigned to forced-air warming (approximately 40 degrees C) or conductive warming using a full-length circulating-water mattress at 40 degrees C. Forced-air warming was applied to the legs of the adults undergoing maxillofacial surgery and to one arm, the shoulders, and the neck in the adults undergoing hip arthroplasty; a U-shaped, tubular forced-air cover was positioned around the pediatric patients. Core temperatures increased in all patients given forced-air warming and decreased or remained constant in those without active warming. Furthermore, we needed to decrease the temperature of the warmer from high to medium (approximately 37 degrees C) in most patients assigned to forced-air warming to prevent hyperthermia. After 15 h of anesthesia, rectal temperatures in the adults undergoing maxillofacial surgery were 3.4 degrees C higher in the forced-air group (P < 0.01). After 4 h of anesthesia, esophageal temperatures had increased 0.8 +/- 0.5 degrees C in the patients warmed with forced-air and decreased 0.8 +/- 0.3 degrees C in those warmed by circulating-water mattresses (P < 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)