Early white blood cell count changes and 30-day mortality in acute pulmonary embolism

Serial complete blood count (CBC) measurements are standard for acute pulmonary embolism, yet the prognostic significance of early changes remains unclear. This study investigated whether these changes enhance mortality prediction beyond baseline values and a restricted set of clinical predictors. We retrospectively evaluated ten CBC-derived measures using a 72-hour landmark in the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. The primary endpoint was all-cause mortality within 30 days of admission, following the landmark. Nested logistic models incorporated baseline values and log-ratio changes between 0-<24 and 24-<72 hours post-admission, applying Benjamini-Hochberg correction. The top-performing model underwent 1,000-resample bootstrap validation and was subsequently applied without modification to an independent hospital cohort. Among 5,556 landmark-eligible development patients, white blood cell count (WBC) demonstrated the strongest association. In 4,531 patients with 546 deaths, the adjusted odds ratio for a one standard-deviation increase in WBC change was 1.35 (95% CI, 1.22-1.48). The full model's optimism-corrected area under the receiver operating characteristic curve (AUC) was 0.766. For 266 external patients with 41 deaths, the AUC was 0.824 (95% CI, 0.736-0.899). Adding WBC change to a model already containing baseline WBC altered the AUC by 0.0035 (95% CI, -0.0262 to 0.0350), without a clear improvement in the Brier score. The calibration slope was 1.50, indicating an underestimation of risk in the highest-risk quintile. In conclusion, while early WBC change correlated with subsequent mortality, its incremental predictive value was not externally validated. Further prospective, multicenter evaluation is essential before clinical implementation.