Huang Lihua, Wu Jun, Luo Jiao, Gu Wei
Department of Infection Disease, The First Affiliated Hospital of Dali University, Dali, Yunnan, People's Republic of China.
Department of Ophthalmology, Dali Bai Autonomous Prefecture People's Hospital, Dali, Yunnan, People's Republic of China.
Int J Gen Med. 2025 Apr 9;18:2033-2045. doi: 10.2147/IJGM.S518644. eCollection 2025.
To explore the risk factors for the severity of hemorrhagic fever with renal syndrome (HFRS) and construct a nomogram model.
A retrospective analysis was conducted on the data of 191 patients diagnosed with HFRS at the First Affiliated Hospital of Dali University between January 1, 2013, and September 30, 2024. Based on whether severe disease occurred, the patients were divided into a severe HFRS group (n=42) and a mild HFRS group (n=149). The clinical data of the two groups were compared, and after eliminating the influence of collinearity, LASSO-Logistic regression analysis was used to screen for factors influencing the severity of HFRS. Additionally, a nomogram model was constructed to predict the severity of HFRS.
Compared with the mild HFRS group, patients in the severe HFRS group had a prolonged length of stay, increased usage rates of Continuous Renal Replacement Therapy (CRRT) and ventilators, and an elevated 30-day mortality rate (<0.001). Procalcitonin (PCT, OR= 0.86), Albumin (ALB, OR: 0.86), Platelet count-to-Albumin ratio (PAR, OR: 0.64), and pleural effusion (OR: 4.49) were identified as independent risk factors for severe HFRS. The Area Under Curve (AUC) of the nomogram model was 0.890. The Hosmer-Lemeshow test result was χ²=2.92, =0.94, and in combination with the Calibration curve, it indicated a good fit between the calibration curve and the ideal curve. Most of the Decision Curve Analysis (DCA) curves of the nomogram model were above the two extreme lines, suggesting that using this model to predict severe HFRS patients could clinically benefit those with severe HFRS, demonstrating the clinical practicality of the nomogram model.
PCT, ALB, PAR, and pleural effusion are risk factors for the severity of HFRS. The constructed nomogram model exhibits good discriminatory power, fit, and clinical practicality, enabling early identification of patients with severe HFRS in southwestern China.
探讨肾综合征出血热（HFRS）严重程度的危险因素并构建列线图模型。
对大理大学第一附属医院2013年1月1日至2024年9月30日确诊为HFRS的191例患者的数据进行回顾性分析。根据是否发生重症疾病，将患者分为重症HFRS组（n = 42）和轻症HFRS组（n = 149）。比较两组的临床资料，在消除共线性影响后，采用LASSO-Logistic回归分析筛选影响HFRS严重程度的因素。此外，构建列线图模型以预测HFRS的严重程度。
与轻症HFRS组相比，重症HFRS组患者住院时间延长，连续性肾脏替代治疗（CRRT）和呼吸机使用率增加，30天死亡率升高（<0.001）。降钙素原（PCT，OR = 0.86）、白蛋白（ALB，OR：0.86）、血小板计数与白蛋白比值（PAR，OR：0.64）和胸腔积液（OR：4.49）被确定为重症HFRS的独立危险因素。列线图模型的曲线下面积（AUC）为0.890。Hosmer-Lemeshow检验结果为χ² = 2.92，P = 0.94，结合校准曲线，表明校准曲线与理想曲线拟合良好。列线图模型的多数决策曲线分析（DCA）曲线位于两条极端线之上，表明使用该模型预测重症HFRS患者对临床重症HFRS患者有益，证明了列线图模型的临床实用性。
PCT、ALB、PAR和胸腔积液是HFRS严重程度的危险因素。构建的列线图模型具有良好的鉴别能力、拟合度和临床实用性，能够早期识别中国西南部的重症HFRS患者。
