Abstract Background and aims Intracerebral haemorrhage (ICH) causes about 28 % of strokes but almost 44 % of stroke-related deaths, and current prognostic tools rely on clinical or imaging features. Inflammatory biomarkers such as the neutrophil-to-lymphocyte ratio (NLR) captures the dynamic secondary injury processes. Our primary aim was examining whether admission NLR independently predicts mortality in spontaneous ICH. Methods We retrospectively collected data at two centers, UK and Italy (in progress). 210 of 568 patients with non-traumatic ICH were analysed. All variables were recorded at admission and again 24 and 48 hours later. Stabilised inverse-probability weights derived from a regression model of death on baseline covariates were used to reduce confounding. Unweighted and weighted comparisons were performed and time-to-death associations were estimated using weighted Cox-models. Results Data analysed from the UK center showed that the cohort (mean age 66.8 ± 16.9 years; 113 males, 97 females) had baseline imbalances that were mitigated by weighting. In unweighted analyses, non-survivors had lower admission red-cell counts (median 4.23 vs 4.60 × 10¹2 L-¹) and haemoglobin (133 vs 141 g L-¹) and slightly higher NLR; only the red-cell measures remained significant after adjustment. Weighted medians still showed higher NLR (4.24 vs 3.95) and platelet-to-lymphocyte ratio (PLR; 219 vs 147) among those who died. Each unit increase in admission NLR (hazard ratio HR 1.08; p=0.0108) and PLR (HR 1.004; p=0.0018) independently increased mortality risk, whereas red-cell count at 48 hours was inversely associated with death. Conflict of interest Ayesha Ansari: Nothing to disclose Figure 1 - belongs to Background and aims
Ansari et al. (Fri,) studied this question.