Prognostic value of inflammatory markers for in-hospital mortality in intensive care patients with acute ischemic stroke: a retrospective observational study based on MIMIC-IV

BackgroundAcute ischemic stroke (AIS) is a primary cause of death and disability worldwide. Four markers that can be readily determined from peripheral blood, namely, the systemic immune-inflammation index (SII), neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and total bilirubin, were measured in this study. We examined the relationship between the SII and in-hospital mortality after AIS and evaluated which of the above four indicators was most accurate for predicting in-hospital mortality after AIS. MethodsWe selected patients from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database who were aged >18 years and who were diagnosed with AIS on admission. We collected the patients' baseline characteristics, including various clinical and laboratory data. To investigate the relationship between the SII and in-hospital mortality in patients with AIS, we employed the generalized additive model (GAM). Differences in in-hospital mortality between the groups were summarized by the Kaplan-Meier survival analysis and the log-rank test. The receiver operating characteristic (ROC) curve analysis was used to assess the accuracy of the four indicators (SII, NLR, PLR, and total bilirubin) for predicting in-hospital mortality in patients with AIS. ResultsThe study included 463 patients, and the in-hospital mortality rate was 12.31%. The GAM analysis showed a positive correlation between the SII and in-hospital mortality in patients with AIS, but the correlation was not linear. Unadjusted Cox regression identified a link between a high SII and an increased probability of in-hospital mortality. We also found that patients with an SII of >1,232 (Q2 group) had a considerably higher chance of in-hospital mortality than those with a low SII (Q1 group). The Kaplan-Meier analysis demonstrated that patients with an elevated SII had a significantly lower chance of surviving their hospital stay than those with a low SII. According to the results of the ROC curve analysis, the in-hospital mortality of patients with AIS predicted by the SII had an area under the ROC curve of 0.65, which revealed that the SII had a better discriminative ability than the NLR, PLR, and total bilirubin. ConclusionThe in-hospital mortality of patients with AIS and the SII were positively correlated, but not linearly. A high SII was associated with a worse prognosis in patients with AIS. The SII had a modest level of discrimination for forecasting in-hospital mortality. The SII was slightly better than the NLR and significantly better than the PLR and total bilirubin for predicting in-hospital mortality in patients with AIS.