Routine blood-derived immune-inflammatory indices and early hematoma expansion in spontaneous intracerebral hemorrhage: A retrospective cohort study

Background: To examine the associations between routine blood-derived immune-inflammatory indices and early hematoma expansion (HE) in patients with spontaneous intracerebral haemorrhage (sICH), and to describe their apparent discriminatory performance. Methods: This single-centre retrospective cohort included 286 patients with sICH. HE was defined as a hematoma-volume increase of <span style="font-family: Symbol; font-size: 16px;">³</span>6 mL or <span style="font-family: Symbol; font-size: 16px;">³</span>33% on follow-up CT. NLR, PLR, SII, SIRI, and AISI were calculated from routine blood tests. Index-specific multivariable logistic regression models adjusted for age, admission NIHSS score, and baseline hematoma volume. ROC analyses summarised apparent discrimination; formal AUC-difference testing and internal or external validation were not performed. Results: HE occurred in 72 patients (25.2%). After adjustment, SII (OR 1.15, 95% CI 1.03-1.29), SIRI (OR 1.27, 95% CI 1.04-1.56), and AISI (OR 2.06, 95% CI 1.46-2.91) remained associated with HE, whereas NLR and PLR did not. Apparent AUCs were 0.731 for SIRI, 0.724 for SII, and 0.701 for AISI. Adding SII to the clinical model yielded an apparent AUC of 0.846 versus 0.812 for the basic model (DAUC=0.034); this difference was not formally tested. Conclusion: Higher SII, SIRI, and AISI were statistically associated with HE after adjustment, but these observational findings do not establish causation. Individual indices showed moderate or limited discrimination; AISI had a lower AUC than SII and SIRI. The numerical AUC increase after adding SII is an exploratory within-cohort finding and does not establish clinical incremental utility. Prospective studies and independent validation are needed.