目的

探究影响胃食管反流病（GERD）患者发生心房颤动的因素，并建立预测GERD患者发生心房颤动风险的临床模型。

方法

收集2015 年3 月至2017 年12 月新疆维吾尔自治区人民医院268 例明确诊断为GERD 患者的资料，根据病情将患者分为GERD 合并心房颤动组和GERD 组，运用单因素及多因素Logistic 回归分析明确GERD 患者发生心房颤动的影响因素，运用nomogram方法建立预测GERD 患者患心房颤动的临床预测模型，并根据预测模型建立评分量表，之后以参与患者的75%为建模组其余25%患者为验证组对所见模型进行外部验证，同时绘制受试者工作特征（ROC）曲线，计算ROC 曲线下面积（AUC），并应用决策曲线分析法（DCA）进一步对所建量表进行内部验证。

结果

在268 例GERD 患者中，有35 例合并心房颤动，运用单因素及多因素Logistic回归模型分析相关指标对GERD 患者发生心房颤动的影响，分析得出年龄、体质量指数（BMI）、患有冠心病或心力衰竭、患有睡眠呼吸暂停等可明显影响GERD 发生心房颤动（P＜0.05），而性别、民族、糖尿病、吸烟及高血压的影响并不显著（P＞0.05）。再进行趋势性检验时发现，随着年龄及BMI 逐渐升高，患者发生心房颤动的风险是逐渐增加的（P＜0.05）。随后建立了临床预测模型，运用内部验证联合外部验证，建模组患者的AUC 是0.6858，而外部验证组的AUC 是0.6473，2 组无明显的统计学差异（P＞0.05），该模型的C 指数为0.7041。

结论

年龄、BMI、患有冠心病或心力衰竭、患有睡眠呼吸暂停等可明显影响GERD 发生心房颤动，同时我们初步构建了切实有效的临床模型去预测GERD 患者发生心房颤动。

Objective

To explore the influencing factors of atrial fibrillation in patients with gastroesophageal reflux disease （GERD）, and to establish a clinical model to predict the risk of atrial fibrillation in patients with GERD.

Methods

We collected the data of 268 patients diagnosed with GERD from 2015 to 2017 in the People's Hospital of Xinjiang Uygur Autonomous Region. According to the patients' condition, we divided them into GERD with atrial fibrillation group and GERD group. The influencing factors of atrial fibrillation in patients with GERD were identified by univariate regression and multivariate regression. A clinical prediction model for predicting atrial fibrillation in GERD patients was established by nomogram method, and a scoring scale was established according to the prediction model.After that, 75% of the participating patients were used as the modeling group and the remaining 25% as the verification group, the observed model was verified externally. Meanwhile, ROC curve was drawn, AUC was calculated, and DCA curve was drawn to further verify the built scale internally.

Results

Among 268 patients with GERD, 35 patients were complicated with atrial fibrillation. Univariate and multivariate logistic regression models were used to analyze the influence of related indexes on atrial fibrillation in patients with GERD. It was found that age, BMI, coronary heart disease or heart failure, and sleep apnea could significantly affect atrial fibrillation in GERD （P＜0.05）, but gender, nationality, diabetes, smoking and hypertension had no significant influence （P＞）. Further trend test showed that the risk of atrial fibrillation increased gradually with the increase of age and BMI （P for trend＜0.05）. Then a clinical prediction model was established. The AUC of the patients in the modeling group was 0.6858, while that in the external verification group was 0.6473. There was no significant difference between the two groups （P＞0.05）, and the C index of the model was 0.7041.

Conclusion

Age, BMI, coronary heart disease or heart failure, sleep apnea, etc. can significantly affect the occurrence of atrial fibrillation in GERD. At the same time, we initially constructed a practical and effective clinical model to predict atrial fibrillation in patients with GERD.