Discussion
Main findings
This is the first study to systematically evaluate the risk factors in developing LV dysfunction in patients with LV summit VA. In the present study, we found that:(1) The incidence of LV dysfunction with VA originated from LVS was 28.8%; (2) PVC QRS duration and AEAD were independent predictors of LV dysfunction; (3) After ablation, the LV systolic function could be improved in patients with LVS VA, and (4) PVC QRS duration and baseline LVEF before ablation were two predictors for patients with LV function recovery after ablation.
Prevalence and incidence of LV dysfunction in patients with LVS VA
The prevalence of PVC-induced cardiomyopathy has been reported at 7% among patients with a PVC burden of more than 10% 24; however, it is likely underestimated 4. Clinical studies have reported a diagnosis of PVC-induced cardiomyopathy from 9% to 30% of patients referred for ablation of PVC 6, 23, 25-27. Nevertheless, most of the previous studies were heterogeneous on PVC origin, and Yamada et al. reported that PVC-induced cardiomyopathy in 19.2% of patients referred for ablation of PVC originated from RVOT16. In our study, with all 139 patients, the PVC originated from LV summit referred for ablation, LV dysfunction was recognized as 28.8%.
Prediction of LV dysfunction in patients with LVS VA
Most patients with PVC-induced cardiomyopathy have very frequent PVCs; however, the PVC burden alone does not reliably predict whether cardiomyopathy will be induced. Baman et al. demonstrated a PVC burden of 24% best predicted those with and without cardiomyopathy6. Reported cutoff numbers vary from 16% to 26%6, 7, 24; however, PVC-induced cardiomyopathy has been reported in patients with a PVC burden of only 4% 28, and normal heart function is often seen in patients with a high PVC burden. Similarly, in the present study with PVC originating from LV summit, the mean PVC burden before ablation was 20.6%, and it was not associated with PVC-induced cardiomyopathy. LV dysfunction could be found as low as 5% of PVC burden, and normal LV function could be noted in PVC burden as high as 55% in this large cohort. The result means that patients’ characteristics and PVC features play more critical roles in the pathophysiology of PVC-induced cardiomyopathy than PVC burden.
Patients with more prolonged exposure to PVCs or an asymptomatic status have a higher risk of developing PVC-induced cardiomyopathy in asymptomatic status 12. Patients without symptoms may have a higher probability of prolonged exposure to PVCs before they are disclosed. Of the 139 patients in our study population, only 2 (1.4%) presented without symptoms, and both had normal LV systolic function. Because most patients had symptoms, we could not conclude the association between asymptomatic status and PVC-induced cardiomyopathy in patients with VA originated from LVS.
PVC QRS duration, with a cut-ff level of >150ms best separated patients with and without PVC induced cardiomyopathy, reported by Yokokawa et al 9. QRS duration was still be found to be the predisposing factor for LV dysfunction from the present study9. The result was in line with previous studies with PVCs originating from various locations throughout both ventricles. The proposed mechanisms included ventricular dyssynchrony, asymmetrically increased wall thickness, and work overload in the late activated regions, all contributing to further myocardium remodeling4.
As previously mentioned, most studies focusing on PVC-induced cardiomyopathy were heterogeneous on PVC origin; however, there were some reports demonstrated that an epicardial origin was independently associated with PVC-induced cardiomyopathy 9, 13, 26.
Epicardial PVCs are shown to have longer QRS duration than other PVCs9, maybe due to the paucity of Purkinje fibers in the epicardium. The initial part of the wavefront progresses slowly through the myocardial wall until reaching the Purkinje system at the subendocardium. This slow transmural activation is reflected as the slow onset of the QRS on the surface electrocardiogram 29and prolonged transmural activation by measuring the AEAD [epi-endo]. To the best of our knowledge, this is the first study to demonstrate the relationship between PVC-induced cardiomyopathy and AEAD [epi-endo], a novel parameter associated with LV dysfunction in patients with LVS VA, which reflects the depth of intramural foci. The longer AEAD [epi-endo] might reflect superficial epicardial foci, resulting in a longer activation time difference between epicardial and endocardial exit 22. Although we enrolled all patients with VA originating from LVS, the wider QRS and larger AEAD [epi-endo] might indicate VA foci close to the epicardial surface, causing a long transmural activation time and LV dyssynchrony.
Catheter ablation of LVS VA and the induced LV dysfunction
Catheter ablation of PVCs has been reported to have an acute success rate of 80%-94%, with a complication rate of up to 5.6%30-32. However, the outcomes for catheter ablation of LVS VAs were diverse and the success rate was lower than the outcomes for PVC ablation originated other than LVS, ranging from 22% to 100% for acute procedural success and from 23% to 100% for freedom from VA recurrences 14, 18, 33, 34. In the present study, ablation of LVS VAs was effective, with a high acute success rate (92.8%). Also, in patients with LV dysfunction, the decreased LVEF improved from 37.5 ± 9.3% to 48.5 ± 10.2% (P < 0.001), indicating the reversible phenomenon of PVC-induced cardiomyopathy, which was comparable to previous studies showing that after successful ablation of the PVCs originated from various locations, there was a mean improvement of LVEF from 10%-15% 13, 26, 31, 35, 36.
In the present study, 50% of patients recovered LV systolic function. For patients without recovery, we found that longer VA QRS duration and poorer LVEF before ablation were two independent factors in predicting irreclaimable LV dysfunction. Our study was in accordance with previous study 37. Combining with the poor LVEF as another prediction, the results echoed our postulation that patients with longer PVC QRS duration may have more severe and irreversible underlying LV substrate abnormalities, which was an indicator rather than a cause for LV cardiomyopathy.
Clinical Implications
LVS has been demonstrated to be an essential anatomic focus for the origin of VA. According to the present study, the incidence of LV dysfunction in patients with frequent VA from LVS PVC is high (28.8%). A significant improvement in LV systolic function in patients with LVS VA-induced cardiomyopathy could be achieved after successful ablation. The discrepancy of activation (between endocardium and epicardium) and the QRSd were the only predictors for LV dysfunction; hence, patients with longer QRS duration should be advocated for an earlier intervention to eliminate the VA.
Study limitations
There were several limitations for the present study. First, the present results were obtained from relatively small study samples and retrospective in nature, besides, no control group was included in the present study. The prospective, randomized study with large sample size to validate the results is mandated. Second, we did not collect cardiac magnetic resonance (CMR) data, and therefore, myocardial fibrosis before the ablation could not be analyzed and limiting us to detect significant CMR predictors of the development of PVC induced cardiomyopathy. Third, the VA duration was proved to be a predictor for VA induced cardiomyopathy 12, and which was not collected in the present study.