Association between preoperative white blood cell counts and thirty-day surgical mortality after craniotomy in adult intracranial tumor patients.

ABSTRACT

Objective: White blood cell (WBC) counts has been identified as a prognostic biomarker which frequently predict adverse outcomes and mortality risk in various conditions. However, evidence for the association between WBC counts and short-term outcomes after intracranial tumor resection remains limited. This study aimed to explore associations between preoperative WBC counts and thirty-day surgical mortality after craniotomy in adult intracranial tumor patients. Methods: This retrospective cohort study performed secondary analysis of 18,049 intracranial tumor craniotomy patients from the ACS NSQIP database (2012-2015). The major exposure and outcome were preoperative WBC counts and thirty-day surgical mortality, respectively. Cox regression modeling assessed the linear association between them. Non-linear associations between them were evaluated by conducting smooth curve fitting using an additive Cox proportional hazard model in conjunction with segmented linear regression modeling. Subgroup analysis and interaction testing assessed effect modification. Sensitivity analysis evaluated result robustness. Results: The total thirty-day surgical mortality after craniotomy was 2.49% (450/18,049). The mean of preoperative WBC counts was 9.501 ± 4.402 × 10^9/L. Fully adjusted model shows that elevated preoperative WBC counts was independently associated with increased thirty-day surgical mortality (HR = 1.057, 95%CI 1.040, 1.076). Further analysis revealed a non-linear association between them below a WBC threshold of 13.6 × 10^9/L, higher WBC counts elevated thirty-day mortality (HR = 1.117; 95%CI 1.077, 1.158), while risk plateaued and no significant mortality rise occurred above this level (HR = 1.015, 95%CI 0.982, 1.050). Steroid usage status has a significant effect modification on the WBC-mortality association (P for interaction = 0.002). The non-linear WBC-mortality association was only present for non-steroid users (HR = 1.158, 95%CI 1.108, 1.210) but not steroid users (HR = 1.009, 95%CI 0.966, 1.055). The sensitivity analysis confirmed the result robustness. Conclusion: Elevated preoperative WBC counts were independently and non-linearly associated with an increased risk of thirty-day surgical mortality in adult non-steroid use patients undergoing craniotomy for intracranial tumors. As a convenient predictor, preoperative WBC data allows improved risk profiling and personalized management in adult intracranial tumor patients.