Non-linear association between neutrophil-to-lymphocyte ratio and 90-day mortality in patients with pneumonia receiving glucocorticoids alone or in combination with other immunosuppressants: A retrospective cohort study.

objectiveThe relationship between neutrophil-to-lymphocyte ratio (NLR) and 90-day mortality in patients with pneumonia receiving glucocorticoids alone or in combination with other immunosuppressants has not been fully verified. We aimed to explore the influence of NLR on 90-day mortality in this specific population.

methodsThis study utilized the data set from the Dryad database, involving 696 participants diagnosed with pneumonia who were receiving glucocorticoids alone or in combination with other immunosuppressants. Data on demographics, vital signs, laboratory results, and comorbidities were collected to assess the link between NLR and 90-day mortality. Multivariable Cox hazard regression analyses and smooth curve fitting were employed to assess the independent association between NLR and 90-day mortality. A two-piecewise linear regression model was used to examine the nonlinear association between NLR and in-hospital mortality. Receiver-operating characteristic curves (ROC) and area under the curves (AUC) were used to assess the ability of different biomarkers to predict the 90-day mortality in patients with pneumonia.

resultsIn total, 696 patients with pneumonia were included in this study. There were 332 individuals (47.7%) aged 18-59 years and 364 (52.3%) aged 60-99 years; 52.6% were male. The 90-day mortality rate across the study population was 26.1%. A non-linear association was noted between NLR and 90-day mortality, with an inflection point at approximately 16.475. On the left side of the inflection point, the hazard ratio was 1.145(95% confidence interval [CI]: 1.091-1.2, p < 0.001). On the right side of the inflection point, the hazard ratio was 1.0057(95% CI:0.9923-1.0192; p = 0.406), reflecting a lack of statistical significance. Similar patterns were observed in subgroup analyses, with significant interaction effects noted for age and smoking status. Furthermore, the ROC curve analysis revealed that NLR was the optimal biomarker for predicting the 90-day mortality with an AUC of 0.714 (95% CI:0.670-0.757). Using 9.34 as the cutoff value of NLR, the sensitivity was 69.8%, and the specificity was 67.7%.

conclusionsA nonlinear correlation between NLR and 90-day mortality was identified in pneumonia patients undergoing glucocorticoid treatment. The NLR value of 16.475 represented the optimal threshold for predicting the 90-day mortality, after exceeding the threshold,90-day mortality tended to stabilize. The findings suggest that NLR is a practical and useful biomarker for predicting the 90-day mortality in this population.