Abstract: Objective To investigate the dynamic changes over time of body temperature, procalcitonin, C-reactive protein and leukocyte count in patients after intracranial surgery and their diagnostic value for intracranial infections. Methods The clinical data of 1,308 patients with intracranial lesions treated by surgery were analyzed, and the patients were divided into infection group and non-infection group. The changes over time of body temperature, C-reactive protein, leukocyte count and mean level of leukocyte count were analyzed using generalized linear mixed model, and Logistic regression was used to predict the occurrence of infection. Results (1) The differences were statistically significant when comparing body temperature in the 2 groups on the 1st, 2nd, 3rd, 5th, 7th and 9th day after the surgery. There was no difference in leukocyte count, C-reactive protein and procalcitonin between the 2 groups within 2 days after surgery, and the differences were statistically significant when comparing these three indexes on the 3rd, 5th, 7th and 9th day after the surgery. (2) The AUC of the combined test of body temperature, leukocyte count, C-reactive protein and procalcitonin on the 2nd day was 0.802 (95% CI: 0.760 to 0.845), and the AUC could be increased to 0.915 (95% CI: 0.887 to 0.949) when the difference of the indicators between the 2nd and 3rd day were included. Conclusion The prediction accuracy can be greatly improved when the indicators and the difference of the indicators between day 2 and 3 are used for the prediction of intracranial infection, suggesting the necessity of considering both physiological indicators and their trends when diagnosing postoperative infection in clinical practice, which can also effectively improve the nursing quality of neurosurgical patients.

Key words: intracranial infection, leukocyte count, procalcitonin, C-reactive protein, prediction model