Long‐term cardiac reverse remodeling after cardiac resynchronization therapy

Abstract Introduction The benefit of cardiac resynchronization therapy (CRT) in heart failure (HF) patients with reduced left ventricular ejection fraction (LVEF) have been observed in the first year. However, there are few data on long‐term follow‐up and the effect of changes of LVEF on mortality. This study aimed to assess the LV remodeling after CRT implantation and the probable effect of changes in LVEF with repeated measures on mortality over time in a real‐world registry. Methods Among our cohort of 328 consecutive CRT patients, mixed model effect analysis have been made to describe the temporal evolution of LVEF and LVESV changes over time up with several explanatory variables. Besides, the effect of LVEF along time on the probability of mortality was evaluated using joint modeling for longitudinal and survival data. Results The study population included 328 patients (253 men; 70.2 ± 9.5 years) in 4.2 (2.9) years follow‐up. There was an increase in LVEF of 11% and a reduction in LVESV of 42 mL during the first year. These changes are more important during the first year, but slight changes remain during the follow‐up. The largest reduction in LVESV occurred in patients with left bundle branch block (LBBB) and the smallest reduction in patients with NYHA IV. The smallest increase in LVEF was an ischemic etiology, longer QRS, and LV electrode in a nonlateral vein. Besides, the results showed that the LVEF profiles taken during follow‐up after CRT were associated with changes in the risk of death. Conclusion Reverse remodeling of the left ventricle is observed especially during the first year, but it seems to be maintained later after CRT implantation in a contemporary cohort of patients. Longitudinal measurements could give us additional information at predicting the individual mortality risk after adjusting by age and sex compared to a single LVEF measurement after CRT.


| INTRODUC TI ON
Cardiac remodeling occurs due to abnormal neurohormonal regulation in heart failure (HF) patients. 1,2 The principal feature of remodeling is left ventricular (LV) dilatation and deterioration of LV function. Major clinical trials have demonstrated the benefit of cardiac resynchronization therapy (CRT) in HF patients with reduced LVEF. [3][4][5][6][7][8][9] It is well established for patients with left ventricular systolic dysfunction and ventricular conduction delay under optimal medical therapy. CRT has been shown to improve functional class, reduce hospitalization and mortality among patients with symptomatic HF, as well as LV reverse remodeling. The concept of cardiac reverse remodeling was developed to explain the reduction of LV volumes and improvements in LV function observed with the use of medical therapies for HF and CRT. 10 Currently, the reduction of left ventricular end-systolic volume (LVESV) ≥15% and/or the increase in the absolute value of left ventricular ejection fraction (LVEF) ≥5% at 6-12 months have been identified as predictors of clinical prognosis and response to CRT. [10][11][12][13] Evidence suggests that the improved outcomes observed with CRT are associated with reverse remodeling. [14][15][16][17] Up to date, randomized clinical trials of CRT have documented increases in LVEF and/or reduction in LV volumes in the first year after CRT implantation. However, there are scarce data on the midterm and long-term benefit of this therapy and the sustained effect of ventricular remodeling over a prolonged follow-up period. In addition, it remains unknown the effect of LVEF on mortality over time.
Here, we evaluate the long-term cardiac reverse remodeling after CRT in patients with repeated echocardiographic measurements in a contemporary patient registry and how changes in LVEF over time, with several measurements per patient, may better predict the individual risk of mortality. We registered the baseline characteristics of all of the patients. Electrocardiographic features included QRS width and morphology. Echocardiographic parameters included LV end-diastolic (LVEDV) and end-systolic volume (LVESV), LVEF, left atrial diameter (LAD), and mitral regurgitation. For the quantification of the severity of mitral regurgitation, the ratio between the maximum regurgitant jet obtained from the flow image by color doppler and the area of the left atrium was used.

| ME THODS
In cases of eccentric mitral regurgitation, the contracted vein was used for quantification. The patients were followed up in the Heart Failure Clinic every 3 or 6 months and in the CRT-Device Clinic every 6 months. Electrocardiogram and echocardiogram were also performed at the 6-month and 2-year follow-ups and according to the discretion of the HF cardiologist. Treating cardiologists followed a specified protocol to achieve OMT. Patients with decreases in LVESV exceeding 15% and/or improvements in LVEF of more than 5% were considered to be echocardiographic responders. Patients with improvements in one category in NYHA functional class were considered to be clinical responders.

| Statistical analysis
Qualitative variables are expressed as frequency and percentages.
Quantitative variables are presented as mean (SD) or as a median (interquartile range) as appropriate. The chi-square test X2 or U of Mann-Whitney was used to compare qualitative variables. P ≤ .05 was considered statistical significance.

| Patient characteristics
The study population included 328 patients (253 men and 75 women; mean age: 70.2 (9.5) years) who were consecutively implanted with a CRT device at our institution. The mean follow-up duration was 4.2 (2.9) years. Of the 328 patients, only 122 (37.2%) were on triple neurohormonal therapy at the time of implantation. The baseline characteristics are listed in Table 1.

| Baseline echocardiographic characteristics
The mean baseline LVEF was 28%. In our sample more than half of the patients had severe mitral regurgitation at the time of the CRT implant due to severe dilatation of the LV (182, 55.5%) and 13 (4.0%) did not present mitral regurgitation prior to CRT.

| Echocardiographic characteristics at follow-up
The echocardiogram was performed during the first year and then Changes are observed during the first year in both sexes, but more marked in women than in men. It is also observed in patients with nonischemic cardiomyopathy. However, the response to CRT in patients with ischemic HF is unchanged in the long term. Patients in NYHA III class, and with LV lead located in posterior and lateral veins have better reverse remodeling of LV ( Figures S1-S5). The increases in LVEF greater than 5% and reductions in LVESV greater than 15% were associated with a reduction in mortality risk ( Figure 2). Table 2  The smallest increase in LVEF was observed in patients with ischemic etiology, longer QRS and location of the VI lead in a nonlateral vein of coronary sinus. Figure 3 shows the changes of the LVEF and LVESV along time after adjusting by potential confounding variables.

| Joint modeling analysis
A survival prediction model has been carried out over time adjusted on age, sex, and changes of LVEF over time using the joint modeling methodology. Table 3 shows the relative risks with their corresponding confidence intervals. The analysis showed that the increase in LVEF after CRT was associated with higher survival rates. Male sex and older age were associated with a worse prognosis. In a sub-analysis of CARE HF trial, CRT induced sustained LV reverse remodeling in the long-term follow-up with the most marked effects occurring within the first 3-9 months. This effect may contribute to an improvement in morbidity and mortality. These beneficial effects were observed even in ischaemic patients and patients with very severe cardiac dysfunction. 24 MIRACLE study investigated serial changes in LV size and function by using longitudinal data analysis in a consecutive cohort of patients with long-term follow-up.

| D ISCUSS I ON
These changes were more pronounced at 6 months and remained during the long-term, but they became much less pronounced.
Besides, patients with an uneventful survival demonstrated a greater decrease in the LVESV compared with patients with adverse events.
Factors associated with less reverse remodeling were ischemic etiology, male sex, and QRS duration <140 ms. 25 Similarly, the reverse LV remodeling of CRT in a sub-analysis of REVERSE trial was sustained over 5 years. The functional and LV remodeling improvements were maximal after 2 years and were accompanied by very low mortality and HF hospitalization, although the largest change in LV volumes was noted in the first year with further remodeling. 26 We

| LI M ITATI O N S
The retrospective nature of this analysis conducted at a single center is a potential weakness. As a consequence, the patient sample size was limited. This is an observational registry with their inherent limitations (eg selection bias, unmeasured bias), and thus associations may be confounded by unmeasured variables. Several unmeasured confounders or details about physician or patient decision-making might not be available in our collection data protocol and could account for some of the reported findings. Also, there may have been appropriate contraindications to adjunctive pharmacotherapy that were nor collected. In addition, during the echocardiographic followup, patients who died are excluded from the analysis and this can be a positive bias because only surviving patients were included in the analysis. Finally, long-term outcomes could be modified by many circumstances that might not be available or controlled in the follow-up protocol of our center. As such, the results presented in this analysis should be considered hypothesis-generating and deserve confirmation in other registries and clinical trials.

| CON CLUS IONS
Reverse remodeling of the left ventricle is observed especially in the first year after CRT implantation and this effect is maintained a long time in a contemporary cohort of patients. Longitudinal LVEF measurements throughout the follow-up could predict better the individual mortality risk adjusting for potential confounding variables.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interests for this article.