Correlation between the white blood cell/platelet ratio and 28-day all-cause mortality in cardiac arrest patients: a retrospective cohort study based on machine learning.

OBJECTIVE

This study aims to evaluate the association between the white blood cell-to-platelet ratio (WPR) and 28-day all-cause mortality among patients experiencing cardiac arrest.

METHODS

Utilizing data from 748 cardiac arrest patients in the Medical Information Mart for Intensive Care-IV (MIMIC-IV) 2.2 database, machine learning algorithms, including the Boruta feature selection method, random forest modeling, and SHAP value analysis, were applied to identify significant prognostic biomarkers. Key patient characteristics, encompassing demographic data, comorbidities, hematological and biochemical indices, and vital signs, were extracted using PostgreSQL Administration Tool (pgAdmin) software. The Cox proportional hazards model assessed the impact of WPR on mortality outcomes, while Kaplan-Meier survival curves and restricted cubic spline (RCS) analysis further validated the findings. Subgroup analyses stratified the prognostic value of WPR by demographic and clinical factors.

RESULTS

WPR demonstrated the highest prognostic significance among the variables studied, showing a strong association with 28-day all-cause mortality. In the unadjusted Model 1, hazard ratios (HRs) for WPR quartiles ranged from 1.88 (95% CI: 1.22-2.90) in Q2 to 3.02 (95% CI: 2.04-4.47) in Q4 (Ptrend <0.05). Adjusted models (Models 2-4) confirmed the robustness of these associations, even after accounting for demographic and clinical covariates. Kaplan-Meier and RCS analyses revealed a significant U-shaped relationship between WPR and mortality risk. Subgroup analyses indicated that elevated WPR was particularly associated with increased mortality in males, elderly patients, married individuals, and those with chronic pulmonary disease.

CONCLUSION

WPR serves as an independent and reliable prognostic biomarker for 28-day mortality in cardiac arrest patients. Its integration into clinical decision-making may enhance the early identification of high-risk patients and guide tailored therapeutic interventions.