急性白血病患者并发肛周感染的风险因素Logistic回归模型构建

Risk factors for perianal infection in patients with acute leukemia and the construction of logistic regression model

目的 研究急性白血病患者并发肛周感染的风险因素,建立Logistic回归模型,并分析其预防措施。方法 回顾性选取2018年1月-2023年1月简阳市人民医院急诊科收治的314例急性白血病患者为研究对象,根据是否并发肛周感染分为感染组(31例)和未感染组(283例)。分析急性白血病患者肛周感染的现况,采用单因素和多因素Logistic回归分析法分析急性白血病患者并发肛周感染的风险因素,并建立风险因素模型;绘制受试者工作特征(ROC)曲线分析风险因素模型对急性白血病患者并发肛周感染的预测价值。结果 急性白血病患者并发肛周感染率为9.87%(31/314);有痔及肛裂、中性粒细胞恢复时间≥14 d、血清白蛋白<30 g/L是急性白血病患者并发肛周感染的风险因素(P<0.05),基于此而构建的风险因素模型是Logit(P)=-10.420+有痔及肛裂史×0.898+中性粒细胞恢复时间≥14 d×1.212+血清白蛋白<30 g/L×0.463;该模型Logit(P)的ROC曲线下面积(AUC)为0.835,敏感度为97.10%,特异度为66.08%。结论 急性白血病患者并发肛周感染率较高,有痔及肛裂、中性粒细胞恢复时间≥14 d、血清白蛋白<30 g/L是急性白血病患者并发肛周感染的风险因素,据此构建的风险因素模型对急性白血病患者并发肛周感染具有优良的预测价值,临床上可针对以上风险因素给予针对性的预防措施,以降低急性白血病患者并发肛周感染的风险,改善预后。

OBJECTIVE To study the risk factors for perianal infection in acute leukemia patients and construct the logistic regression model for the analysis of the preventive measures. METHODS Clinical data of 314 patients with acute leukemia admitted to Jianyang City People′s Hospital from Jan 2018 to Jan 2023 were retrospectively recruited as research subjects, and were divided into the infected group(31 cases) and uninfected group(283 cases) according to whether they perianal infection occurred. The status of perianal infection in acute leukemia patients was analyzed, and the risk factors for perianal infection in acute leukemia patients were analyzed by univariate analysis and multivariate logistic regression analysis;the risk factor model was established, and its predictive value was analyzed by receiver operating characteristic(ROC) curve. RESULTS The perianal infection rate of acute leukemia patients was 9.87%(31/314). History of hemorrhoids anal fissures, recovery time of neutrophil ≥14 d, serum albumin <30 g/L were risk factors for perianal infection in acute leukemia patients(P<0.05). The established risk factor model was established as follows: Logit(P)=-10.420+ history of hemorrhoids and anal fissures ×0.898+ neutrophil recovery time ≥14 d ×1.212+serum albumin <30 g/L ×0.463. The area under ROC curve(AUC) was 0.835, with the sensitivity and the specificity of 97.10% and 66.08%, respectively. CONCLUSION The perianal infection rate in acute leukemia patients was high, and its risk factors were history of hemorrhoids anal fissures, recovery time of neutrophil ≥14 d and serum albumin <30 g/L. The established risk factor model has good predictive value for perianal infection in acute leukemia patients, and preventive measures can be taken to reduce the risk of perianal infection in acute leukemia patients and improve the prognosis.