The effects of physical treatment on induced fever in humans

PURPOSE: Initial treatments for fever include the amelioration of underlying causes and administration of antipyretic medications. However, patients who fail these treatments are often actively cooled, which may be counterproductive because decreasing skin temperature increases the thermoregulatory core target temperature. Cooling may also provoke metabolic and autonomic stress and thermal discomfort. SUBJECTS AND METHODS: We studied 9 subjects, each on 3 days. Fever was induced each day with 100,000 IU/kg of interleukin-2 administered intravenously (elapsed time zero). Randomly assigned treatments were 1) control (a cotton blanket), 2) cooling (forced air at 15°C), or 3) self-adjust (forced-air warming adjusted to comfort). Treatments were maintained for 3 to 8 elapsed hours. RESULTS: Peak core temperatures (mean ± SD) were 38.4 ± 0.5°C on the control day, 38.1 ± 0.5°C on the cooling day, and 38.5 ± 0.4°C on the self-adjust day. Integrated core temperatures were 6.0 ± 1.6°C • h on the control day, 5.7 ± 2.2°C • h on the cooling day, and 6.4 ± 1.2°C • h on the self-adjust day. Neither peak nor integrated core temperatures differed significantly on the 3 days. Shivering was common on the cooling day but otherwise rare. Oxygen consumption was normal on the control and self-adjust days but increased 35% to 40% during cooling (P = 0.0001). Mean arterial pressure and plasma norepinephrine and epinephrine concentrations were significantly greater during cooling (P<0.05). On a self-reported thermal comfort scale, the subjects were miserable during cooling and significantly more comfortable on the self-adjust than control day (P<0.05). CONCLUSION: We conclude that active cooling should be avoided in unsedated patients with moderate fever, because it does not reduce core temperature but does increase metabolic rate, activate the autonomic nervous system, and provoke thermal discomfort.