The mortality and neurological morbidity in heatstroke have been attributed to the host's inflammatory and hemostatic responses to heat stress, suggesting that immunomodulation may improve outcome. We postulated that an experimental baboon model of heatstroke will reproduce human responses and clinical outcome to allow testing of new therapeutic strategies. Eight anesthetized juvenile baboons (Papio hamadryas) were subjected to heat stress in an incubator maintained at 44-47 degrees C until rectal temperature attained 42.5 degrees C (moderate heatstroke; n = 4) or systolic arterial pressure fell to <90 mmHg (severe heatstroke; n = 4) and were allowed to recover at room temperature. Four sham-heated animals served as a control group. Rectal temperature at the end of heat stress was 42.5 +/- 0.0 and 43.3 +/- 0.1 degrees C, respectively. All heat-stressed animals had systemic inflammation and activated coagulation, indicated by increased plasma IL-6, prothrombin time, activated partial thromboplastin time, and D-dimer levels, and decreased platelet count. Biochemical markers and/or histology evidenced cellular injury/dysfunction: plasma levels of thrombomodulin, creatinine, creatine kinase, lactic dehydrogenase, and alanine aminotransferase were increased, and varying degrees of tissue damage were present in liver, brain, and gut. No baboon with severe heatstroke survived. Neurological morbidity but no mortality was observed in baboons with moderate heatstroke. Nonsurvivors displayed significantly greater coagulopathy, inflammatory activity, and tissue injury than survivors. Sham-heated animals had an uneventful course. Heat stress elicited distinct patterns of inflammatory and hemostatic responses associated with outcome. The baboon model of heatstroke appears suitable for testing whether immunomodulation of the host's responses can improve outcome.