This was a multicenter, retrospective, observational, and longitudinal study of consecutive genetically evaluated patients with DCM recruited from inherited cardiac disease and heart failure units at 19 Spanish hospitals between 2015 and 2022. Patients underwent serial echocardiography at baseline and at 12±6 months (intermediate echocardiogram). An additional echocardiogram at the last follow-up visit was collected if available.

DCM was defined as left ventricular ejection fraction (LVEF) <50% on echocardiography in the absence of abnormal loading conditions, coronary artery disease, excessive alcohol intake, or other identifiable causes.1 Patients diagnosed in infancy (< 1 year of age) were excluded, given that the etiology, natural history, and outcomes of infant-onset cardiomyopathies can differ substantially from those in older children, adolescents, and adults.12

Participating individuals had been genetically tested using targeted next-generation sequencing (NGS) panels at participating institutions or at an accredited genetics laboratory. Although the NGS panels could differ in the number of genes, all included> 50 genes related to cardiomyopathies. Additionally, consecutive relatives with DCM (n=93) who harbored a pathogenic or likely pathogenic variant previously identified in a DCM proband through an NGS panel including> 50 cardiomyopathy-associated disease-causing genes were included to enrich the number of DCM patients with a positive genotype. Demographics, symptoms, 12-lead electrocardiogram, and transthoracic echocardiogram data were collected from clinical records at each participating center using standardized methodology.

The study was approved by the Hospital Universitario Puerta de Hierro ethics committee, which waived the requirement for informed consent. The study conformed to the principles of the Declaration of Helsinki. Data integrity was ensured by the investigators at each participating center.

Genetic variants were centrally classified as pathogenic (P), likely pathogenic (LP) or variant of unknown significance after a systematic review by a cardiologist expert in cardiovascular genetics using modified criteria of the American College of Medical Genetics and Genomics.13 A variant was considered disease-causing if it affected a DCM-related gene and was classified as P/LP. Patients harboring P/LP variants were considered”‘genotype-positive” (G+), and those harboring variant of unknown significance variants or with a negative NGS panel were considered “genotype-negative” (G-).

Genes were clustered into functional gene groups based on similar common functions, involvement in biological processes, localization to subcellular compartments, and other shared properties based on consolidated scientific evidence from the literature and available biological databases as previously described.14 Because of its specific characteristics of frequency in DCM, TTN was considered as a separate group. Functional gene groups included the following: a) structural cytoskeleton/Z-disk; b) desmosomal; c) nuclear envelope; d) motor sarcomeric; e)TTN; and f) other genes. Individuals with> 1 pathogenic or likely pathogenic variant (n=8) were excluded from the functional gene group analysis to maintain a conservative approach.

LVRR was defined as either left ventricular normalization (LVEF improvement to ≥ 50% with an absolute increase of ≥ 5% LVEF from baseline evaluation on echocardiogram) or an absolute increase in LVEF of ≥ 10%, as previously described.8,15–17 Persistent LVRR at long-term among those patients who had a favorable initial response was defined as improvement or stability (± 5%) in LVEF in the last follow-up echocardiogram compared with the mid-term evaluation.

Our primary objective was to analyze the impact of genotype on LVRR and evaluate the factors associated with LVRR.

Additionally, we evaluated the clinical impact of persistency of LVRR at long-term. Clinical outcomes were evaluated according to the presence or absence of LVRR at 12±6 months. Among patients with a positive LVRR at 12±6 months, outcomes were further assessed according to the persistence or loss of LVRR at the last follow-up. The clinical objectives considered were a composite of major adverse cardiovascular events (MACE), a composite of major ventricular arrhythmias (MVA), and a composite of end-stage heart failure (ESHF). The MACE composite objective included cardiovascular death, ESHF, and MVA. ESHF included ventricular assist device implantation for refractory heart failure, heart transplant, and ESHF-related mortality. MVA included SCD or aborted SCD, sustained ventricular tachycardia, and appropriate ICD interventions. Only appropriate ICD shocks to terminate ventricular tachycardia or ventricular fibrillation episodes were considered for the purpose of this study (anti-tachycardia pacing therapy was not considered).

All patients had planned reviews at participating centers every 6 to 12 months or more frequently if clinically indicated. A detailed flow chart of the study is presented in figure 1.

Figure 1.

Flow chart of the study. DCM, dilated cardiomyopathy; LVAD, left ventricular assist device; LVEF, left ventricular ejection fraction; LVRR, left ventricular reverse remodeling.

(0.43MB).

Continuous variables are expressed as mean±standard deviation (SD) or as median [interquartile range (IQR)], as appropriate. Groups were compared using the Student t test, the Mann-Whitney U test, the ANOVA test, or the Kruskal-Wallis test when comparing more than 2 groups. Non-continuous categorical variables were expressed as counts (percentages) and compared using the chi-square test or Fisher's exact test, as appropriate.

Logistic regression analysis was applied to determine the association of genotype and clinical variables with LVRR. The multivariate analysis was performed using the forward stepwise logistic regression method and a significance level of 0.05. Only patients with complete data for all covariates included in the final model were analyzed (complete-case analysis; n=636). Clinical, echocardiographic, and genetic parameters that were statistically significant in the univariate analysis were included (P <.05). The presence of negative T waves was not included due to collinearity. Treatment with SGLT2 inhibitors was not included since quadruple therapy was not well established in the inclusion period of our study, and information was not available on all patients.

The cumulative probability of an event on follow-up was estimated using the Kaplan–Meier method, and the log-rank test was used to compare survival between groups. Univariate Cox regression models were used to assess the association of LVRR status at the last follow-up with the clinical objectives (MACE, ESHF, and MVA). Analyses were conducted using Stata Statistics version 15 (StataCorp). A 2-tailed P value of <.05 was considered statistically significant.

A total of 307 individuals (43.2%) exhibited LVRR at mid-term evaluation (mean time 11.4±3.4 months), 46.6% (n=186) of patients from the genotype-negative group and 38.8% (n=121) in the genotype-positive group (P=.036). The distribution of genes according to functional gene group and LVRR is compiled in figure S1. Patients with a negative genotype and carriers of a variant in TTN showed the highest rate of LVRR (59% and 47%, respectively). In contrast, patients from other genotype groups showed a very low probability of LVRR, particularly in carriers of variants in desmosomal and nuclear envelope genes, where only 16% and 18% of patients achieved LVRR, respectively (figure 2).

Figure 2.

Central illustration. Cumulative event rates for the composite of major adverse cardiovascular events, end-stage heart failure, and major ventricular arrhythmia at the last follow-up according to response to medical treatment. “No recovery” includes patients without LVRR at mid-term; “transient recovery” includes patients with LVRR at mid-term and worsening of LVEF at last-follow-up; “persistent recovery” includes patients with LVRR at mid-term and persistent LVRR at last-follow-up. HF, heart failure; LVEF, left ventricular ejection fraction; LVRR, left ventricular reverse remodeling.

(0.64MB).

Patients with LVRR were older at diagnosis than patients without LVRR (51.7 years [IQR 42.6-63.4 years] vs 50.3 years [IQR 39.6-60.3 years]; P=.049), were more likely to be probands (91.5% vs 83.4%; P=.001), and had worse New York Heart Association functional class (table 1). In line with the finding of a lower proportion of patients with LVRR among participants with a positive genotype, a family history of DCM or SCD was more frequent in patients without LVRR. Regarding echocardiographic findings, LVEF was lower, LVEDD higher, right ventricular systolic dysfunction, and moderate or severe mitral regurgitation were more prevalent in patients who had LVRR (Table 1).

Several parameters were associated with LVRR at mid-term in the univariate analysis: TTN gene, higher age at diagnosis, proband, higher LVEDD, moderate or severe mitral regurgitation, right ventricular systolic dysfunction, worse New York Heart Association functional class, previous hospitalization due to heart failure, and treatment with mineralocorticoid receptor antagonists and SGLT2i. On the other hand, several genotypes (desmosomal, nuclear envelope, and motor sarcomeric genes) and a positive family history of DCM and SCD were negatively associated with LVRR at mid-term in univariate logistic regression. In the multivariate analysis, previous admission due to heart failure, a lower LVEF, along with either negative or TTN genotypes were associated with LVRR. Of note, a genetic variant in TTN was the strongest positive predictor for LVRR after adjustment for treatment and other clinical variables (odds ratio [OR], 2.02; 95% confidence interval [95%CI], 1.23-3.30). In contrast, genetic variants in desmosomal (OR, 0.17; 95%CI, 0.05-0.59), nuclear envelope (OR, 0.38; 95%CI, 0.14-0.98), and motor sarcomeric genes (OR, 0.43; 95%CI, 0.21-0.88) remained strongly negatively associated with LVRR in multivariate analysis (table 2). Neurohormonal treatment lost significance after adjusting for the rest of the clinical and genetic variables.

After a median follow-up since mid-term evaluation of 4.43 years [IQR 2.1-7.5 years], MACE occurred in 156 patients (21.9%), 97 (13.6%) had ESHF events, and 79 (11.1%) had MVA.

Outcomes according to LVRR at mid-term are presented in figure 3 and table S4. MACE occurred in 121 (30.0%) patients among those without LVRR and in 35 (11.4%) patients among those with LVRR. The hazard ratio (HR) for MACE in patients without LVRR was 2.88 (95%CI, 1.92-4.33; P <.001) compared with LVRR peers. ESHF occurred in 80 (19.8%) patients in the group without LVRR and in 17 (5.5%) in the group with LVRR (HR, 3.80; 95%CI, 2.25-6.42; P <.001). MVA occurred in 60 (14.9%) in the group without LVRR and in 19 (6.2%) in the group with LVRR (HR, 2.60; 95%CI, 1.50-4.50; P <.001).

Figure 3.

Outcomes in patients with LVRR vs without LVRR. Cumulative event rates for composite (A) major adverse cardiovascular events, (B) end-stage heart failure, and (C) major ventricular arrhythmia at last follow-up. 95%CI, 95% confidence interval; LVRR, left ventricular reverse remodeling.

(0.43MB).

Among the 284 patients who exhibited LVRR at mid-term and who underwent long-term evaluation, 73 patients (25.7%) experienced LVEF worsening; 21.6% (n=37) of genotype-negative and 31.9% (n=36) in genotype-positive (P=.054). Maintenance of LVRR or deterioration of LV function at long-term according to genotype functional groups is presented in figure S2.

Demographic, clinical, and imaging characteristics are shown in table 3. Patients with LVEF worsening were more often males (78.1% vs 64.5%; P=.03) and had a higher LVEF at mid-term (48.0% vs 45.6%; P=.049). Other clinical characteristics and echocardiographic parameters were similar between those who maintained LVRR and those who did not. Notably, LVEF decline occurred despite patients remaining on neurohormonal therapy.

Outcomes and events according to LVRR persistence at long-term are presented in table S5. After a median follow-up of 4.54 years [IQR 2.80-7.53 years], 33 patients had MACE (11.6%), 16 (5.6%) had ESHF events, and 18 (6.3%) MVA. MACE occurred in 18 (24.7%) patients in the worsening LVEF group and in 15 (7.1%) patients in the persistent LVRR group. The HR for MACE was 2.34 (95%CI, 1.17-4.71; P=.017) for patients with LVEF worsening compared with the persistent-LVRR group. ESHF occurred in 13 (17.8%) patients in the worsening LVEF group and 3 (1.4%) patients in the persistent LVRR group (HR, 7.82; 95%CI, 2.20-27.83; P <.001). MVA occurred in 9 (12.3%) patients in the worsening LVEF group and 9 (4.3%) patients in the persistent LVRR group (HR, 2.12; 95%CI, 0.79-5.67; P=.139) (figure 4, table S6).

Figure 4.

Outcomes in patients with persistent LVRR vs those with worsening LVEF among patients with an initial positive LVRR. Cumulative event rates for (A) major adverse cardiovascular events, (B) end-stage heart failure, and (C) major ventricular arrhythmia at last follow-up. 95%CI, 95% confidence interval; LVEF, left ventricular ejection fraction; LVRR, left ventricular reverse remodeling.

(0.47MB).

Our study adds to the available body of data to consider formulating the indications and timing for ICD implantation in DCM. We identified a subgroup of patients (negative genotype and carriers of TTN variants) who exhibit a high rate of LVRR at mid-term and a low rate of clinical events when LVRR occurs. Our data suggest that it may be reasonable to wait longer than the recommended 3-month period to assess treatment response and decide on device implantation in these patients. In contrast, other genetic clusters were associated a very low rate of LVRR at mid-term and with a greater risk of worsening LVEF at long term, even after an initial recovery. In these patients, ICD could be implanted earlier given the high chances of an unfavorable remodeling response. Finally, our study also highlights the need for close follow-up in DCM patients, even after an initial recovery, given the high rate of worsening LVEF and the increased rate of cardiovascular events observed in patients with LVEF deterioration.

The limitations of the study include its observational nature and retrospective design. Neurohormonal therapy was initiated and up-titrated according to routine clinical practice. Given the inclusion period of the study, most patients were not receiving SGLT2 inhibitors. Differences in baseline clinical profiles among genotypes may have led to differences in treatment intensity, potentially influencing LVRR outcomes. Main DCM genes were evaluated in all cases, but the genes included in NGS target panels varied between centers and over time, reflecting the evolving knowledge of DCM genetics in recent years. Although this is the largest cohort of genotyped DCM patients with serial echocardiographic evaluations reported to date, the limited number of patients belonging to certain gene groups restricts our ability to reach conclusions about these patients, especially regarding long-term follow-up, where the number of patients is limited. Furthermore, for the purpose of this study, at least two echocardiograms were required throughout follow-up. Therefore, patients who died or required a transplant before the second echocardiogram did not qualify for the study, which may introduce survivor bias. Finally, participating centers were specialized inherited cardiac diseases and heart failure units; therefore, findings might not be extrapolated to other settings.