心电图预测心脏再同步化治疗效果的价值

Role of electrocardiogram in predicting cardiac resynchronization therapy response

目的研究心电图预测心脏再同步化治疗（CRT）效果的价值。方法本研究为回顾性分析。人选2001年6月至2013年6月于解放军总医院住院的心力衰竭（心衰）患者92例，并分为CRT有反应组（n=64）和CRT无反应组（n=28）。回顾患者基线及随访6个月时的一般临床资料及心电图和超声心动图检测结果。CRT术后6个月左心室收缩末期容积（LVESV）较术前减少≥15％定义为CRT有反应。患者电轴变化主要为3个方向，即向前、向右以及冠状面顺时针旋转，每个方向电轴变化积1分，1〉2分定义为电轴明显变化。采用多元logistic回归分析各因素对CRT术后反应的预测价值。结果（1）两组患者基线资料的比较结果：CRT有反应组女性患者比例明显高于CRT无反应组（P〈0．01）。CRT无反应组心房颤动（房颤）和（或）心房扑动（房扑）患者比例则明显高于CRT有反应组（P〈0．01）。CRT有反应组患者术前心电图示左束支传导阻滞（LBBB）的患者比例明显高于CRT无反应组（P=0．04），平均QRS波时限高于CRT无反应组（P=0．01）。（2）两组患者随访结果的比较：CRT术后6个月随访，CRT有反应组患者纽约心脏协会（NYHA）心功能分级、6min步行距离、QRS波时限及超声心动图各参数均明显改善（P均〈0．01）。CRT无反应组患者术后QRS波时限增宽（P=0．02），LVESV增大（P〈0．01），然而NYHA心功能分级有所改善。CRT有反应组患者术前电轴指向左53例（82．8％）、向后58例（90．6％），术后变化为电轴指向右49例（76．6％）、向前30例（40．6％）。CRT无反应组患者术前电轴指向左25例（89．3％）、向后24例（85．7％），术后变化为电轴指向右17例（60．7％）、向前12例（42．9％）。术前，CRT有反应组患者多为正常电轴或电轴左偏，CRT无反应组患者多为正常电轴。术后，CRT有反应组和CRT无反应组额面电轴右偏的比例均较术前增加[分别为39．1％（25／64）比14．1％（9／64）（P=0．001），和39．3％（11／28）比21．4％（6／28）（P=0．15）]。术后，CRT有反应组和CRT无反应组电轴明显变化的比例分别为62．5％（40／64）和32．1％（9／28），两组比较差异有统计学意义（P=0．007）。（3）各因素对CRT有反应预测价值的分析结果：术前QRS波时限≥140ms（OR=4．97，95％C11．53—16．13，P=0．008）及CRT术后电轴明显变化（OR=5．1，95％CI1．67～15．5，P=0．004）是CRT有反应的独立预测因素。而术前合并房颤（OR=0．25，95％C10．08～0．80，P=0．02）是CRT无反应的独立预测因素。结论心电图在预测CRT效果方面具有重要价值，术前QRS波时限及合并房颤的情况以及术后电轴的变化均与CRT效果密切相关。

Objective To explore the role of electrocardiogram （ ECG ） in predicting cardiac resynchronization therapy （CRT） response. Methods This study retrospectively analyzed ECG of 92 CRT patients, who received CRT therapy from 2001 to 2013 in our center and were followed up for 6 months. The patients were divided into responder group （ n = 64） and non-responder group （ n = 28 ）. The baseline and 6- month data including QRS width, heart rhythm and axis variation were analyzed. The definition of responder is left ventricular end systolic volume （LVESV） reduction ≥ 15% within 6 months after CRT. After CRT therapy, the ventricular activation was changed as left to right （frontal plane） , posterior to anterior and axisehanged in a clockwise direction. The change in more than two directions was defined as prominent axis change. Logistic analysis was performed to analyze the role of ECG in predicting CRT response. Results （1） Baseline parameter eomparison between the two groups： the proportion of female and LBBB is significantly higher （P〈 0. 01; P= 0. 04）, while the proportion of atrial fibrillation/flutter （Af/AF） is significantly lower （P 〈 0. 01 ） in responder group than in non-responder group. The pre-CRT average QRS duration is much wider in responder group than in non-responder group （P = 0. 01 ）. （2） Comparison of follow-up with baseline results in two groups： NYHA heart function level, 6 minutes walking distance, QRS duration, LVEF, LVESV improved signifieantly （P 〈 0. 01 ） post-CRT in responder group. In non-responder group, the QRS duration and LVESV deteriorated signifieantly（ P = 0. 02, P 〈 0. 01 ）, while post-CRT NYHA heart function level improved significantly. In responder group, pre-CRT ECG axis of 53 patients （82. 8% ） pointed to left and 58 patients （90. 6% ） pointed to posterior; post-CRT ECG axis of 49 patients （76.6%） pointed to right and 30 patients （40. 6% ） pointed to anterior. In non-responder group, pre-CRT ECG axis of 25 patients （89. 3% ） pointed to left and 24 patients （85.7%） pointed to posterior; post-CRT ECG axis of 17 patients （60. 7% ） pointed to right and 12 patients （42. 9% ） pointed to anterior. Post-CRT, the proportion of ECG axis prominent change was significantly higher in responder than in non-responder group （62. 5% （40/64） vs. 32. 1% （9/28）, P =0. 007）. （3）Predicting value： pre-CRT QRS width ≥140 ms （ OR = 4. 97, 95% CI 1.53 to 16. 13, P = 0. 008 ） and post-CRT prominent axis ehange（ OR = 5.1,95% CI 1.67 to 15.5, P = 0. 004）were found to be independent predictors of CRT responders. Af/AF pre-CRT was associated with redueed CRT response （ OR = 0. 25,95% C10. 08 to 0. 80, P = 0. 02）. Conclusions ECG may play a role in predicting CRT response. QRS width and Af/AF before CRT and ECG axis change post-CRT could be used to predict CRT response.