Ruize Ma, Yanli Xu, Shuwen Ding, Ruihua Zhang, Jingxia Wang, Zhouling Jiang, Ling Lin, Zhihai Chen
Early differentiation of severe fever with thrombocytopenia syndrome (SFTS) and hemorrhagic fever with renal syndrome (HFRS) remains difficult in endemic settings. In this single-center retrospective comparative study, we analyzed admission-time demographic, clinical, and laboratory data from 330 hospitalized patients with laboratory-confirmed SFTS (n = 220) or HFRS (n = 110) admitted between May 2023 and November 2025. Compared with HFRS, SFTS showed an older age distribution, female predominance, more frequent neurological symptoms, and a predominantly hematologic phenotype with modest hepatic differences, whereas HFRS showed more prominent urinary manifestations and greater inflammatory and renal abnormalities. Abnormality-based and normalized organ-domain burden analyses supported the same phenotypic separation, with hematologic, renal, and inflammatory domains showing the clearest exploratory differences. Two differentiation models were developed as proof-of-concept tools using the same complete-case data set and internally evaluated by full-process nested cross-validation. The basic and extended models achieved out-of-fold AUCs of 0.975 and 0.971, respectively; the paired difference was -0.004 (95% CI, -0.014 to 0.005; p = 0.404). Adding CRP, albumin, and LDH did not provide consistent incremental predictive value. These single-center findings provide a proof-of-concept framework requiring independent external validation and recalibration before potential clinical application.