Objective To explore the predictive effect of fragmented QRS complex (f-QRS） on the decline of left ventricular ejection fraction (LVEF） in patients after pacemaker implantation.

Methods One hundred and nine cases of sick sinus syndrome or third degree atrioventricular block with dual chamber pacemaker (DDDR mode） or single chamber pacemaker (VVIR mode） were retrospectively analyzed. The patients underwent 12 lead ECG and echocardiography at admission. The pacing ECG, echocardiography and the cumulative pacing ratio of the right ventricle were followed up after the operation. The patients whose absolute value of LVEF decline (ΔLVEF） ≥ 10% at the last follow-up were in the LVEF decline group, totaling 46 cases; The patients with ΔLVEF ＜ 10% were divided into LVEF not decreased group, totaling 63 cases. F-QRS on ECG was recorded as f-QRS（+） in the two groups; No f-QRS was recorded as f-QRS（–）. Univariate analysis was used to screen the factors that may increase the risk of postoperative LVEF decline. Multivariate Cox regression was used to screen the independent risk factors associated with postoperative LVEF decline in patients with pacemaker implantation.

Results there were statistically significant differences in preoperative f-QRS, postoperative f-QRS, preoperative self QRS duration (IQRSd）, the number of preoperative f-QRS leads, the proportion of right ventricular pacing, hypertension, atrial fibrillation, pacing mode, postoperative LVEF, postoperative lad, and postoperative LVEDd between patients with decreased LVEF and patients without decreased LVEF (P ＜ 0.05）. After pacemaker implantation, LVEF decreased, lad increased, and LVEDd increased in all patients (P ＜ 0.01）. Preoperative f-QRS, the number of pre-operative f-QRS leads, preoperative IQRSd, and hypertension were independent predictors of LVEF decline after pacemaker implantation (P ＜ 0.05）. There were significant differences in preoperative iqrs, pacing mode, right ventricular pacing ratio, postoperative LVEF, postoperative lad, and postoperative LVEDd between the f-QRS（+） group and the f-QRS（–） group (P ＜ 0.01）; The log rank test showed that the average time of LVEF decline in the overall patients, preoperative f-QRS（+） group and preoperative f-QRS（–） group was statistically significant (P ＜ 0.01）. Preoperative f-QRS lead number ＞ 1 and preoperative IQRSd ＞ 122 MS were the best cut-off values for predicting postoperative LVEF decline in patients with pacemaker implantation.

Conclusions Long term right ventricular pacing can increase the risk of postoperative LVEF decline. Patients with pre-operative f-QRS（+） develope LVEF decline earlier after surgery; F-QRS can predict the decline trend of LVEF and evaluate the risk of cardiac function decline earlier.