马尔尼菲蓝状菌病预后危险因素Logistic回归分析及预后模型的建立

Logistic Regression Analysis of Prognostic Risk Factors for Talaromycosis Marneffei and Establishment of Prognostic Model

目的研究艾滋病(AIDS)合并马尔尼菲蓝状菌病(TSM)患者预后的危险因素,并建立预后判断模型。方法回顾性收集2011年1月-2014年1月在广西柳州市人民医院收治的100例TSM患者的临床资料,运用非条件Logistic回归分析建立预后模型。用独立的TSM资料进行评判预后模型的预测价值。结果 100例患者中好转组64例,恶化组36例,恶化组中23例死亡,病死率23.00%(23/100)。Logistic回归分析结果显示白细胞、血小板、谷草转氨酶和CD4^+T淋巴细胞计数作为独立的危险因素可用于预后判断模型的建立。对所构建模型预测的灵敏度、特异性、阳性预测值、阴性预测值和总正确率分别为88.00%、84.00%、84.62%、87.50%和86.00%。结论白细胞、血小板、谷草转氨酶、CD4^+T淋巴细胞计数作为独立的危险因素可用于预后判断模型的构建。本研究中所构建的预后模型能够较准确的预测AIDS合并TSM患者的短期预后。

Objective To study prognostic risk factors and establish a prognostic model for tients with AIDS combined with talaromycosis marneffei （TSM）. Methods pa- The clinical data of 100 inpatients with TSM who had been treated in The People＇s Hospi- tal of Liuzhou from January 2011 to January 2014 were retrospectively analyzed. Non- condition logistic regression analysis was used to develop a prognostic model, the pre- dictive value of which was evaluated using independent TSM data. Results Of 100 patients, there were 64 in recovery group and 36 in deterioration group, of which 23 died, with a fatality rate of 23.00%. Logistic regression analysis indicated that white blood cells, platelets, glutamic oxalacetic transaminase （GOT） and CD4＋T lympho- cyte count as independent risk factors could be used to establish the prognostic mod-el. The sensitivity and specificity of prediction, positive predictive value, negative predictive value and total accuracy of prediction of the model were 88.00%, 84. 00%, 84.62%, 87.05% and 86.00%, respectively. Conclusion White blood cell, platelets, GOT, CD4＋T lymphocyte count as independent risk factors can be used to establish the prognostic model, thereby predicting the short-term prognosis of patients with AIDS combined with TSM.