The effect of body temperature on cerebral infarcts and thrombolytic therapy was investigated in 91 rats embolized in the right carotid territory. Hypothermia of 32 degrees C for 2 h with preembolic onset (n = 15) or hyperthermia of 39 degrees C for 2 h with postembolic onset (n = 22) was compared to normothermic controls (n = 17). After 48 h of survival, neuropathological evaluation with measurement of infarct volume was performed. Median infarct volume in percent of affected hemisphere volume was 11% (9-21, quartiles) in rats treated with hypothermia alone, compared to 46% (14-59, quartiles) in normothermic controls (P = 0.04). Hyperthermia for 2 h increased median infarct volume to 65% (37-75, quartiles). There was a positive and significant correlation between infarct volume and body temperature (R = 0.53, P = 0.0002, n = 54). Mortality rate was significantly higher among rats treated with hyperthermia compared to normothermic controls (P = 0.005). A subset of 37 rats exposed to the same temperature regimen were treated with tissue plasminogen activator (20 mg/kg i.v. during 45 min) 2 h after embolization. Judged by posttreatment carotid angiography, hyperthermic rats (n = 11) had the best degree of recanalization (P = 0.03) compared to controls (n = 17), but median infarct volume in this group was (58% (27-67, quartiles)) significantly larger (P < 0.02) than normothermic (21% (15-39, quartiles), n = 14) and hypothermic (13% (7-31, quartiles), n = 12) rats treated with thrombolytic therapy. Thrombolytic therapy following 2 h of hypothermia, could not improve the effect of hypothermia alone.(ABSTRACT TRUNCATED AT 250 WORDS)