目的应用 Logistic 回归分析从围生期心肌病（PPCM）的超声指标中，筛选对 PPCM 预后有预测作用的指标。资料与方法对50例超声诊断为 PPCM 的患者行超声心动图检查，分为心功能恢复组（30例）和未恢复组（20例），在胸骨旁左心室长轴切面测量左心室舒张末期内径（LVEDd），采用双平面 Simpson 法于心尖部四腔心切面、两腔心切面测量左心室射血分数（LVEF），寻找左心室附壁血栓的证据，合并肺动脉高压时应用连续式多普勒超声技术测量三尖瓣反流峰速以估测肺动脉收缩压，建立 Logistic 回归模型，生成 ROC 曲线，评价 Logistic 模型的预报能力。结果恢复组在首次检查时 LVEDd、LVEF、左心室附壁血栓、肺动脉高压及复查时的 LVEF 与未恢复组相比，差异有统计学意义（t=－4.33、7.64、11.54， P<0.05；χ2=10.93、4.43，P<0.05），心包积液、二尖瓣反流、三尖瓣反流差异无统计学意义（χ2=2.21、0.67、3.46，P>0.05）。LVEDd、LVEF、左心室附壁血栓3个超声特征可用于 Logistic 模型的建立（χ2=5.14、11.59、14.58，P<0.05）。模型预报结果的正确率为90.0%（45/50），曲线下面积为0.945±0.030（P<0.01）。结论 Logistic回归分析可以应用于 PPCM 超声指标的筛选，LVEDd、LVEF、左心室附壁血栓能较精准地预测 PPCM 的预后。%Purpose To analyze the echocardiographic features of peripartum cardiomyopathy (PPCM) using Logistic regression, and to screen the indexes which can be used for the prognosis of PPCM. Materials and Methods Fifty patients who were diagnosed as PPCM by echocardiography were divided into recovered group (30 cases) and non-recovered group (20 cases), all the patients underwent echocardiography, left ventricular end-diastolic diameter (LVEDd) was measured in the parasternal long axis view, left ventricular ejection fraction (LVEF) was measured using Simpson biplane method in the apical four-chamber and two-chamber view to look for the evidence of left ventricular thrombosis, if complicated with pulmonary hypertension, continuous wave Doppler was used for measuring the peak velocity of tricuspid valve regurgitation, and estimating of pulmonary artery systolic pressure, Logistic regression model was established and receiver operating characteristic (ROC) curve was generated to evaluate the prediction value of Logistic regression model. Results Compared with non-recovered group, there was statistically significant difference of LVEDd, LVEF, left ventricular thrombosis, pulmonary hypertension and re-checked LVEF (t= -4.33, 7.64 and 11.54, P<0.05; χ2=10.93 and 4.43, P<0.05) in the first examination, while there was no statistical difference in pericardial effusion, mitral valve regurgitation and tricuspid regurgitation between the two groups (χ2=2.21, 0.67 and 3.46, P>0.05). Three sonographic features LVEDd, LVEF and left ventricular mural thrombus) could be used for the establishment of Logistic model (χ2=5.14, 11.59 and 14.58, P<0.05). The prediction accuracy of the model was 90.0% (45/50, P<0.001) and the area under ROC curve was 0.945±0.030 (P<0.001). Conclusion Logistic regression analysis can be applied for the screening of ultrasound index for PPCM, LVEDD, LVEF and left ventricular wall thrombus can predict the prognosis of PPCM accurately.
展开▼
