Abstract: Objective To explore the predicative value of body temperature for mortality risk of children with sepsis in PICU. 
Methods A total of 152 children with sepsis were included and divided into three groups based on the body temperature at 24h before and after PICU admission: ultrahyperpyrexia(39.5-41 ℃) group，fever(37.5-39.5 ℃) group and apyrexia(＜37.5 ℃) group. Pediatric critical illness score(PCIS), mortality rate(overall and at 7 d after admission), inflammation and immune function, multiple organs function were compared.They were also divided into the survival group and the death group according to condition at 7 d after PICU admission, and temperature distribution and PCIS score were compared. The risk factors of 7-day mortality was analyzed by Logistic regression analysis. 
Results Of 152 children with sepsis,the number of children that died within 7 d accounted for 82.2%(37/45). The 7-day mortality of the three groups were 36.4%, 16.7% and 42.9%, respectively, with the lowest mortality rate in the fever group(P＜0.05). In the ultrahyperpyrexia group, 11 children received blood purification, accounting for 50.0%; In the fever group, 34 children received blood purification, accounting for 33.3%; In the apyrexia group, 3 children received blood purification, accounting for 10.7%. There was significant difference in the proportion of children receiving blood purification among the three groups(χ2=9.242, P=0.010). There was no significant difference in the proportion of children receiving blood purification between ultrahyperpyrexia group and fever group(χ2 =0.109, P=0.741). The proportion of children receiving blood purification in the ultrahyperpyrexia group was higher than that in the apyrexia group, and the difference was statistically significant(χ2 =9.432, P=0.002). The proportion of children receiving blood purification in the fever group was higher than that in the apyrexia group, and the difference was statistically significant(χ2=5.520, P=0.019). There was no significant difference in the related inflammatory indexes［white blood cells(WBC), C-reactive protein(CRP), procalcitonin(PCT) and naive lymphocytes］ among three groups(P＞0.05). However, there was significant difference in the lymphocyte count of the three groups at 48-72 h(P＜0.05), and the values of apyrexia group and ultrahyperpyrexia group were lower than those of fever group(P＜0.05). The abnormal rates of CK, CK-MB, Cr and LDH in the three groups were not statistically significant(P＞0.05). The abnormal rates of ALT in the ultrahyperpyrexia group and fever group were lower than those in the apyrexia group(P＜0.05). All patients were divided into survival group and death group according to 7-day survival. In the death group, the proportion of children without fever was the highest, up to 40.5%; In the survival group, the proportion of children with fever was the highest, up to 76.5%(P＜0.001). The PCIS score of children in the survival group was higher than that in the death group, and the absolute lymphocyte count at 48-72 h was higher than that in the death group(P＜0.05). There was no significant difference in CRP and PCT between two groups(P＞0.05). Multivariate Logistic regression analysis showed that PCIS score in children and the absolute lymphocyte count at 48-72 h were the risk factors of death(P＜0.05). 
Conclusion The proportion of septic children with fever is high in PICU, and children with fever have a higher survival rate than those with ultrahyperpyrexia or with apyrexia. However, body temperature is not the independent risk factor for mortality.  

Key words: sepsis, body temperature, mortality