Association of platelet-to-lymphocyte ratio with sepsis based on MIMIC-IV database

Background/Aim. Sepsis is a potentially lethal condition that ranks among the most severe medical disorders globally and results in elevated mortality rates. The aim of this study was to examine the correlation between sepsis outcomes in patients and the platelet-to-lymphocyte ratio (PLR), including mortality and duration of intensive care unit (ICU) stay from admission, while considering relevant demographic and clinical factors. Methods. In this retrospective study, the Medical Information Mart for Intensive Care (MIMIC)-IV dataset information was used. Sepsis patients were de-identified, and their PLR values were calculated upon admission. Multivariate logistic regression models were used to assess the relationship between PLR and mortality, adjusting for confounding variables such as age, gender, comorbidities, and vital signs. Additionally, the association between PLR and the length of the ICU stay was analyzed using linear regression models. Results. This study included 4,624 sepsis patients. Higher PLR values were significantly correlated with decreased survival probabilities in the unadjusted model [Model 1 - odds ratio (OR): 0.890, 95% confidence interval (CI): 0.810-0.970, p < 0.001]. This association remained significant after adjusting for demographic factors (Model 2 - OR: 0.920, 95% CI: 0.850-0.995, p < 0.001), comorbidities and biochemical parameters (Model 3 - OR: 0.880, 95% CI: 0.800-0.960, p = 0.0273), and vital signs (Model 4 - OR: 0.860, 95% CI: 0.780-0.940, p = 0.0301). Furthermore, our analyses revealed a trend towards prolonged ICU stay with higher PLR values, although the association did not reach statistical significance. Survivors were younger (median age 63.37 vs. 70.84 years) and had lower Charlson Comorbidity Index (CCI) scores (median CCI 4.00 vs. 6.00, p < 0.001) compared to non-survivors. Conclusion. The outcomes indicate that higher PLR levels correlate with greater fatality rates in sepsis patients, underscoring its potential as a predictive biomarker. The observed trend towards pro-longed ICU stay with higher PLR warrants further investigation. Model 4, which includes demographic factors, comorbidities, biochemical parameters, and vital signs, demonstrated the strongest association between PLR and mortality, suggesting it may be the most clinically useful model for predicting sepsis outcomes. Incorporating PLR into risk assessment and therapeutic decision-making frameworks may enhance sepsis treatment and improve patient outcomes.