Admission complete blood count-derived inflammatory indices for risk stratification in hemorrhagic fever with renal syndrome: comparative performance of NLR, PLR, SII, LMR, and the neutrophil-to-red blood cell ratio

Background: Hemorrhagic fever with renal syndrome (HFRS) requires early risk stratification, yet the comparative value and reproducibility of complete blood count (CBC)-derived inflammatory indices remain unclear. We compared admission neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), lymphocyte-to-monocyte ratio (LMR), and neutrophil-to-red blood cell ratio (NRR) for identifying severe disease and 28-day mortality. Methods: We retrospectively analyzed 342 adults with serologically confirmed HFRS treated at the Second Affiliated Hospital of the Fourth Military Medical University from January 2020 to May 2026. Clinical severity was classified as mild or severe, and prognosis as 28-day survival or non-survival. Group comparisons, receiver operating characteristic (ROC) curves, exploratory Kaplan-Meier analyses, and age- and sex-adjusted single-marker logistic models were used to evaluate CBC-derived indices and conventional laboratory markers. An independent 100-patient cohort from Weinan Central Hospital was analyzed using identical endpoint definitions and a prespecified validation framework. Results: In the derivation cohort, 327 patients survived and 15 died within 28 days. For severe disease, NRR had the highest AUC among CBC-derived indices (0.750), followed by NLR (0.657), SII (0.467), LMR (0.400), and PLR (0.354). For 28-day mortality, NLR had the highest AUC (0.751), followed by NRR (0.694), LMR (0.558), SII (0.479), and PLR (0.345). After age and sex adjustment, NRR was associated with severe disease and 28-day mortality. In the validation cohort, NRR retained the highest severe-disease AUC (0.701) and remained independently associated with severe disease. For mortality, NRR showed the highest AUC (0.697), but its adjusted association was not significant, and Kaplan-Meier analysis showed no significant separation. Conclusions: NRR showed promising performance for identifying severe HFRS and was associated with mortality in the derivation cohort, but mortality findings were not confirmed externally. NRR may serve as a candidate CBC-derived adjunctive marker, requiring prospective validation.