QRS Shortens after Left Ventricular Assist Device Implantation

Abstract

Left ventricular assist device (LVAD) has profound effects on multiple levels, unloading left ventricle (LV) and sometimes even resulting in complete cardiac recovery. We hypothesized that unloading of the heart by LVAD is accompanied by significant electrocardiography (ECG) changes, such as a change in QRS width. Therefore, we analyzed the records of all patients implanted with LVADs at the University of Kentucky from November 2005 to May 2016. Patients with native depolarization and those with paced ventricular rhythms were included. Overall, 140 patients implanted with continuous flow LVADs had pre- and postimplant ECGs. Twenty-seven patients were excluded because they alternated (between native and paced QRS). A total of 113 patients, mean age 52.1 ± 71 years, 83% men, 81.3% with HeartMate 2 (Thoratec Corporation Pleasant, CA), and the rest with Heartware (Medtronic Framingham, MA), were analyzed. Paired t-test was used to compare the QRS duration at baseline versus the following time points (pairwise comparison): 1 week, 1 month, 3 months, 6 months, 1 year, and 2 years. Although all 140 patients had ECG at baseline, some patients had follow-up ECGs only at some of the prespecified time points. For example, when we compared the QRS at the baseline and 1 week after the implant, we included only the patients who had ECGs at both time points. This explains why there are some variations of the baseline QRS duration when compared with the different time points. Overall, including patients with both native and paced QRS, its duration decreased from 129.1 ± 37.8 ms at baseline to 112.5 ± 33.7 ms in 1 week, from 119.4 ± 32.7 ms to 107.0 ± 28.9 ms in 1 month, from 131.7 ± 41.3 ms to 117.3 ± 34.3 ms in 3 months, from 135.0 ± 37.5 ms to 118.2 ± 29.0 ms in 6 months, and from 129.6 ± 37.8 ms to 113.4 ± 27.2 ms in 1 year, and from 125.6 ± 35.7 to 117.5 ± 47 in 2 years. All changes during the follow-up period of 1 week to 1 year were significant compared with the baseline (pre-implant) QRS duration (Figure 1).Figure 1.: Box and whisker plots illustrating the relationship between QRS duration at baseline (pre-implant) and on follow-up at 1 week (A), 1 month (B), 3 months (C), 6 months (D), 1 year (E), and 2 years (F) among the whole study cohort. The five horizontal lines defining the box and whiskers represent the 10th, 25th, 50th, 75th, and 90th percentiles, from bottom to top, excluding outliers that are shown as circles. Baseline values vary because for pairwise comparison, only patients with the electrocardiography (ECG) available for specific time points were analyzed.In patients with native rhythm (Figure 2A), QRS duration decreased from 99.9 ± 17.1 ms pre-implant to 85.9 ± 14.9 ms (p < 0.001) at 1 week, from 103.6 ± 23.6 to 92.6 ± 19.7 (p = 0.006) at 1 month, 101.3 ± 19.8 to 94.2 ± 18.9 (p = 0.140) at 3 months, 103.5 ± 16.2 to 95.5 ± 12.1 (p = 0.073) at 6 months, 101.5 ± 18.1 to 95.8 ± 14.3 (p = 0.298) at 1 year, and 97.5 ± 12.2 to 77.3 ± 23.4 (p = 0.032) at 2 years.Figure 2.: Box and whisker plots illustrating the relationship between baseline QRS duration and on follow-up at 1 week (a), 1 month (b), 3 months (c), 6 months (d), 1 year (e), and 2 years (f) in patients with native (A) and paced (B) QRS. The five horizontal lines defining the box and whiskers represent the 10th, 25th, 50th, 75th, and 90th percentiles, from bottom to top, excluding outliers that are shown as circles and extreme outliers that are shown as asterisks.For patients with a paced rhythm, reduction in QRS duration was seen from 161.8 ± 25.8 at baseline to 142.2 ± 21.8 (p < 0.001) at 1 week, 151.1 ± 24.5 to 136.0 ± 21.8 (p = 0.084) at 1 month, 168.9 ± 27.2 to 145.6 ± 26.6 (p = 0.106) at 3 months, 163.0 ± 27.1 to 138.4 ± 24.1 (p = 0.004) at 6 months, 160.0 ± 28.6 to 132.5 ± 25.1 (p = 0.022) at 1 year, and no change at 2 years (Figure 2B). The only variable correlating with QRS duration was LV end-diastolic dimension at baseline, 3 months, and 6 months, with r = 0.39, 0.61, and 0.77 and p = 0.001, 0.06, and 0.02, respectively. To our knowledge, this is the largest retrospective study with the longest follow-up on the effects of LVAD unloading and its subsequent changes to QRS duration. Such ECG changes may be considered as evidence of LV unloading. In addition, this is the first study that included patients with paced QRS. In one prior study, the QRS duration was evaluated in a very small group of patients exclusively with nonischemic cardiomyopathy, with limited time of observation (up to 6 months), and almost only supported with first-generation pulsatile flow LVADs which are no longer in use.1 In another instance, the ECG changes were studied after a very short time (1 week) on LVAD support.2 Xydas et al3 studied 36 LVAD patients for about 100 days and found that decrease of both QRS duration and QT interval occurred in parallel with improvement of LV ventricular ejection fraction and LV end-diastolic diameter, myocyte size, and collagen deposition. In patients with intrinsic as well as with paced rhythm, LVAD implantation resulted in significant shortening of QRS duration, consistent from as early as 1 week after the implant until 1 year or 2 years of follow-up. This may represent the effect of LVAD on unloading the heart with favorable remodeling of the left ventricle.