Clinical Research Of Mechanical Remodeling And Electrical Remodeling In Cardiac Resynchronization Therapy

Part Ⅰ. Evaluation of Effectiveness and Type of therapeutic Response to Cardiac Resynchronization TherapyBackground Approximately30-50%of recipients lack of response to CRT. Non-response rate ranged greatly between different clinical trials and might attribute to variation in response definition and time point for evaluation. Additionally, process of reverse remodeling is time-dependency. Thus, improvement in clinical status and reversal of anatomy might possess differential time courses.Purpose We attempted to explore the difference and time course of clinical response and echocardiographic response to CRT. The time-dependent alteration in LVEF was also determined according to therapeutic response to echocardiography.Methods Data of patients with regular in-hospital visits within the first two years post-CRT were retrospectively reviewed. All patients had baseline QRS duration≥120ms, LVEF<35%, NYHA class-Ⅳ on optimal medical therapy. Clinical response was defined as improvement≥1NYHA class (baseline NYHA Ⅲ-Ⅳ) or stabilized in NYHA class Ⅱ (baseline NYHA Ⅱ) at follow-up. Echocardiographic response was defined as absolute improvement in LVEF by≥10%from baseline. Super-response was described as absolute value of LVEF by≥50%at follow-up. All patients completed visits at the time point of3,6,12, and24months after CRT. A portion of subjects had visits at36and48months. Response rate of clinical and echocardiography was calculated at each time point within the first two years. The time-dependent alteration in LVEF was depicted according to echocardiographic responses.Results Overall of62patients included into this study,8with RVP upgrading to CRT,30with CRT-P,18with CRT-D for primary prevention, and14with CRT-D for secondary prevention. Eleven patients had ischemic cardiomyopathy (ICM),42had LBBB,6recipients had history of atrial fibrillation. Median history for symptomatic heart failure was4years (interquartile1-8years). All patients had regular follow-up within the first two years. Twenty-four patients had visits at36months and another11had visits at48months. Response rate of clinical and echocardiography at the time point of3,6,12, and24months was77.4%,82.3%,83.9%,83.9%and41.9%,58.1%,59.7%,66.1%, respectively. Early after implant, response rate of echocardiography was rather low, which increased progressively over time with a non-response rate of33.9%at2years. Clinical response occurred early after operation and maintained in relatively high level during long-term follow-up.11non-responders,27responders, and24super-responders were identified by maximal improvement in LVEF during visits. Wide variation and alteration in LVEF was observed in responders. In contrast, progressive improvement in LVEF observed in super-responders with occasionally small depression. Conclusions The discordance between clinical response and echocardiographic response was observed in patients underwent CRT. Improvement in functional status occurred early post-CRT, while recovery in echocardiography progressively proceeded over time. Responders might suffer worsened heart function after early amelioration. Super-responders experienced sustainable improvement in LVEF during long-term follow-up. Part Ⅱ. Value of Neurohormones in Prediction of Hyper-response to Cardiac Resynchronization TherapyBackground Activation of neurohormones is a prominent feature in heart failure. Endothelin (ET) and big endothelin-1(big ET) are proved to be associated with cardiovascular remodeling, disease progression, as well as suboptimal prognosis. However, it is still unclear whether plasma ET relates to CRT responses. We speculated that circulating big ET may be negatively associated with therapeutic response to CRT.Purpose We sought to determine the predictive value of big ET-1and NT-proBNP to super-response to CRT.Methods Data of consecutive patients with CRT implantation between January2011and December2012in our hospital were respectively collected. All patients had baseline QRS>120ms, LVEF<35%, and NYHA class Ⅱ-Ⅳ on standard medical therapy. Exclusion criteria included acute myocardial infarction; unstable angina pectoris; renal dysfunction (serum creatinine>1.6mg/dl); persistent atrial fibrillation without atrioventricular junction ablation. Super-response was defined as absolute LVEF≥50%at1year follow-up. Comparisons between super-responders and non-super-responders were conducted by independent sample’s t test or Chi-square test. Pearson correlation analysis or Spearman correlation analysis was performed to detect factors correlated with plasma big ET-1and super-response as appropriate. The best cutoff point for big ET-1to predict super-responder was determined by receptor operating characteristic (ROC) curve. A forward stepwise regression model was build to determine independent predictors to hyper-responders.Results Overall84consecutive patients were screened and74entered into final analysis (52male, mean age,61±12years). Four patients with CRT replacement,4with persistent atrial fibrillation, and another2with renal dysfunction were excluded. At one year follow-up,15patients (20.3%) were recognized as super-responders and none of them with ischemic cardiomyopathy. At a median of13months follow-up (10-29months),7patients died and2got heart transplantation, none of the9subjects was super-responder. Compared with non-super-responders, super-responders had less male gender (33%vs.80%, p=0.001), more LBBB (87%vs.58%, p=0.041), lower big ET-1(pmol/L)(0.4±0.2vs.1.0±0.9, p=0.034) and serum creatinine (mg/dl)(0.88±0.14vs.1.04±0.25, p=0.003). A positive correlation was revealed between big ET-1and NYHA class (r=0.301, p=0.009), NT-proBNP (r=0.290, p=0.014), LAD (r=0.270, p=0.020), and LVEDD (r=0.329, p=0.004), but a negative correlation with super-response to CRT (r=-0.387, p=0.001). The best cutoff point for big ET-1to predict super-responder was0.5pmol/L (AUC0.778,95%CI0.657-0.899, p=0.001) with a sensitivity of71%and specificity of67%. Multivariate analysis revealed that big ET-1(<0.5pmol/L) was an independent super-response predictor (OR8.7,95%CI1.9-38.8, p=0.005), besides female gender (OR7.5,95%CI1.8-32.1, p=0.007) and LBBB (OR6.6,95%CI1.1-40.5, p=0.041).Conclusions Plasma big ET-1correlated positively with the severity of heart failure and the extent of adverse remodeling before CRT. Plasma big ET-1(≤0.5pmol/L) could predict super-response independently. Patients with lesser extent of remodeling may be more accessible to CRT. There is no relationship between plasma NT-proBNP and hyper-response to CRT. Part III. Changes in Native QRS Duration and Its Relationship to Mechanical Remodeling in Cardiac Resynchronization TherapyBackground Cardiac resynchronization therapy is an electrical method to restore electrical synchrony and rectify mechanical dyssynchrony. It has been demonstrated that QRS narrowing after initiation of biventricular pacing suggested reversal of electrical asynchrony and predicted favorable response to CRT. However, the changes in native QRS duration has been less investigated, particularly with scant date concerning its relationship with CRT response and improvement in echocardiography.Purpose We sought to determine the changes in native QRS duration before and after CRT, and explore its relationship to CRT response as well as improvement in echocardiography.Methods Data of74patients with completed ECG records pre-CRT, post-CRT and at follow-up were reviewed. All patients had QRS>120ms, LVEF≤35%, LVEDD>55mm, NYHA functional class II-IV, and with a de novo CRT device implantation. Those with right ventricular pacing upgraded to CRT, pacer dependency, and persistent atrial fibrillation were excluded. Data of visits were obtained beyond the first three months post-implantation. In cases with more than one follow-up, data of the last visit was analyzed. Response was defined as absolute improvement in LVEF≥10%from baseline. ECG parameters including pre-QRSd, post-QRSd, unpaced-QRSd, AQRSd and native AQRSd were collected. Comparisons were conducted between responders and non-responders and between patients with native QRS shortening (native AQRSd>0ms) and without native QRS shortening (native AQRSd≤0ms).Results A total of74patients (48male, mean age,61±9years) were included in this study,35with TLBBB and17with ischemic cardiomyopathy (ICM). At a median follow-up of13months (6to36months),47patients were identified as responders. Compared with responders, non-responders had more ICM (37%vs.15%, p=0.044), less TLBBB (33%vs.55%, p=0.092), much shorter pre-QRSd (156±25ms vs.167±23ms, p=0.055). QRS duration was significantly shortened in responders (from167±23ms to157±20ms, p=0.006), but prolonged in non-responders (from156±25ms to165±16ms, p=0.112) after initiation of resynchronization therapy. Native QRS duration was slightly abbreviated in responders while broadened in non-responders (4±19ms vs.-11±19ms, p=0.001). Patients with native QRS narrowing (native△QRSd>0ms) presented with notably geometrical reversal as assessed by LVEF (20%±11%vs.10%±10%, p=0.000) and LVEDD (14±11mm vs.4±10mm, p=0.000). Positive correlation was revealed between native△QRSd and pre-QRSd, TLBBB, CRT response and alteration in echocardiography, but with the strongest correlation with△QRSd (r=0.458, p=0.000) and changes in LVEF (r=0.415, p=0.000). Multivariate analysis indicated that native△QRSd was the sole independent factor of ECG in association to response to CRT.Conclusions Modification on electrical remodeling occurred instantly after the onset of biventricular pacing therapy. QRS shortening on biventricular pacing ECG reflected the direct effect imposed by CRT. Alteration in native QRS duration associated with changes in paced QRS duration post-CRT. Abbreviation in native QRS duration was a reflection of reversal of underlying electrical abnormalities which associated with CRT response independently and correlated with anatomical reversal positively. Electrical disorders proceeding progressively might underlay the mechanism for lack of response to CRT. Part IV. Changes in Fragmented QRS Complex and Its Relationship to Response to Cardiac Resynchronization TherapyBackground Fragmentation of QRS complex (fQRS), known as various deflection on QRS morphology, is an indication of inhomogeneous ventricular activation and discoordination of intraventricular contraction. More leads with fQRS on surface ECG associated with poorer responses to CRT. However, until recently, there is no report concerning changes in fQRS post-CRT. We speculated that changes in fQRS would correspond to the therapeutic response to CRT.Objectives We sought to investigate the changes in fQRS after a period of biventricular pacing and its relationship to response to CRT.Methods Data of patients with naive ECG recorded at follow-up were reviewed. All patients had QRS≥120ms, LVEF<35%,LVEDD≥55mm, NYHA functional class II-IV, and with a de novo CRT device implantation. Those with right ventricular pacing upgrading to CRT, pacer dependency, and persistent atrial fibrillation were excluded. Data of visits were obtained beyond the first three months post-implantation. Response to CRT was defined as absolute improvement in LVEF≥10%from baseline or by improvement>1NYHA class and without heart failure hospitalization. Changes in QRS duration post-CRT (AQRS) and changes in fQRS (AfQRS) were analyzed to response to CRT.Results A total of75patients (48male, mean age,61±9years) were included in this study,36(48%) had TLBBB and17(23%) had ischemic cardiomyopathy (ICM). At a median follow-up of13months (6-36months),57patients had response to CRT. In comparison with responders, non-responders had much shorter pre-QRSd (151±26ms vs.167±23ms, p=0.013). At follow-up, responders had significant improvement in NYHA class (from2.8±0.6to1.7±0.6, p<0.001), LVEF (from26%±6%to43%±11%, p<0.001), and LVEDD (from71±9mm to61±11mm, p<0.001). After the onset of resynchronization therapy, QRS duration was prolonged from151±26ms to168±16ms (p=0.033) with a prolongation of17±31ms in non-responders. However, QRS duration was shortened from167±23ms to158±19ms (p=0.003) with an average abbreviation of9±22ms in responders. fQRS in>1lead was found in19patients at baseline and in25patients at follow-up. Eleven of12patients (91.7%) with reduction in fQRS had response to CRT and8of12patients (66.7%) with increase in fQRS were non-responders. There was a positive correlation between AfQRS and TLBBB, QRS pre-CRT and AQRS, with the strongest correlation with CRT response (r=0.403, p=0.000).△QRS also had a positive correlation with TLBBB and CRT response, but with the strongest correlation to QRS pre-CRT (r=0.747, p=0.000). A weak and positive relationship existed between△fQRS and△QRS (r=0.206, p=0.076). Reduction of fQRS in≥1lead associated with response to CRT in high specificity (95%) but in low sensitivity (19%).Conclusions Number of leads with fQRS could be changed corresponding to the therapeutic response to CRT. Reduction in fQPS lead associated with favorable response to CRT in high specificity but in low sensitivity. Regression of fQRS might be a marker of reversal of electrical asynchrony in resynchronization therapy.