Long-Term Outcomes in Adults With Excessive Left Ventricular Trabeculation

Graphical Abstract

What is the clinical question being addressed?

What are the time-specific overall and major adverse cardiovascular event-free survival estimates in adult cohorts historically labeled as having left ventricular noncompaction?

What is the main finding?

In cohorts historically labeled as left ventricular noncompaction, estimated 10-year overall survival was 70.8% across 14 studies (n = 1,469), and major adverse cardiovascular events-free survival was 46.5% across 10 studies (n = 2,008). These unadjusted estimates were imprecise beyond 4 years and do not establish risk attributable to trabeculation.

Left ventricular (LV) noncompaction (LVNC) is a unique cardiomyopathy marked by prominent trabeculations and deep intertrabecular recesses, whose diagnostic criteria have proven difficult to agree on. Its clinical course ranges from asymptomatic to heart failure, arrhythmias, sudden cardiac death, and thromboembolic events. LVNC displays a heterogeneous phenotype, overlapping with normal variants and other cardiomyopathies, complicating diagnosis and risk stratification.¹ Contemporary nomenclature is also evolving, with increasing recognition that “LVNC” may be better conceptualized in many patients as “LV hypertrabeculation”, a phenotypic trait that can occur in isolation or overlap with other cardiomyopathies, rather than a uniformly distinct cardiomyopathic entity.² The long-term prognosis of LVNC is unclear, as prior studies are limited by small sample sizes, heterogeneous populations, and lack of agreement on diagnostic criteria.³ Importantly, although prior studies have described survival in LVNC, long-term estimates for major adverse cardiovascular events (MACE) remain poorly defined, leaving uncertainty regarding the cumulative morbidity and mortality burden in adults historically classified as LVNC. Despite these limitations, we performed a reconstructed Kaplan-Meier (KM) meta-analysis to assess long-term all-cause mortality and MACE in LVNC patients.

This study was conducted in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines and registered at PROSPERO (CRD420261401080).⁴ Studies reporting time-to-event outcomes in adult patients with LVNC were systematically identified using the source studies' LVNC definitions, based on heterogeneous echocardiographic or cardiac magnetic resonance criteria. Participants were not reclassified using contemporary cardiomyopathy criteria. The primary endpoint was all-cause mortality, and the secondary endpoint was MACE. Given the heterogeneity in MACE definitions across studies, subgroup analyses were performed based on whether stroke was included within the composite endpoint. Individual patient data were reconstructed from digitized KM curves using the curve approach, implemented through a validated 2-stage method.⁵ This incorporated extracted survival probabilities and numbers at risk, and via the IPDfromKM package in R. Reconstructed data sets from individual studies were merged to generate pooled survival estimates. Accuracy of reconstructed data was confirmed by visual inspection and quantitative error assessment, demonstrating high concordance with the original published curves. Pooled KM curves were generated to estimate long-term survival probabilities up to 10 years. Survival estimates with corresponding 95% CIs were reported. Subgroup analyses were performed for MACE based on endpoint definitions. Sensitivity analyses using a leave-one-out approach were conducted to evaluate the robustness of pooled estimates. All analyses were performed using R (R Foundation for Statistical Computing, version 4.2.1).

Among 21 unique studies, including 2,714 patients, the weighted mean age was 45.4 years, 58.5% were males, and the weighted mean LV ejection fraction (LVEF) was 40.1% (Supplemental Tables 1 to 3). A total of 14 studies (n = 1,469 LVNC patients) were included in the analysis of all-cause mortality, that is, overall survival. The pooled KM curve demonstrated a progressive decline in survival over the 10-year follow-up period. Survival probability was 94.5% (95% CI: 93.3%-95.7%) at 1 year and 90.7% (95% CI: 89.1%-92.3%) at 2 years. This decline continued over time, reaching 78.2% (95% CI: 75.0%-81.4%) at 6 years and 70.8% (95% CI: 64.3%-78.0%) at 10 years (Figure 1A). For MACE, a total of 10 studies (n = 2,008) were included. Event-free survival declined more steeply compared with all-cause mortality. MACE-free survival was 88.9% (95% CI: 87.5%-90.3%) at 1 year and 82.8% (95% CI: 81.0%-84.6%) at 2 years, decreasing to 60.9% (95% CI: 58.2%-63.9%) at 6 years and 46.5% (95% CI: 42.5%-51.0%) at 10 years (Figure 1B). Subgroup analysis demonstrated similar early MACE regardless of endpoint definition. In studies including stroke within the composite endpoint (n = 1,281), survival was 82.6% (95% CI: 80.4%-84.9%) at 2 years and declined to 46.7% (95% CI: 41.8%-52.1%) at 10 years. In studies excluding stroke (n = 727), survival was 83.1% (95% CI: 80.2%-86.1%) at 2 years, decreasing to 57.3% (95% CI: 52.7%-62.3%) at 6 years and 50.7% (95% CI: 45.2%-57.0%) at 10 years. Sensitivity analysis using a leave-one-out approach demonstrated consistent findings, with pooled KM estimates remaining stable across iterations and within corresponding CIs, supporting the robustness of the results. At 10 years, only 33 patients remained at risk for overall survival and 73 for MACE-free survival.

Figure 1.

Reconstructed Kaplan-Meier analysis of pooled outcomes in patients With Left Ventricular Noncompaction

(A) All-cause mortality; (B) MACE-free survival. MACE = major adverse cardiovascular events.

Despite the challenges with LVNC diagnostic criteria and the studies available for review, this reconstructed KM meta-analysis of patients with LVNC demonstrates substantial long-term adverse outcomes, in patients with LVNC, with progressive declines in overall survival and MACE-free survival over 10-years. Overall survival and MACE-free survival declined progressively over time, with a more pronounced reduction in MACE-free survival compared with all-cause mortality, with less than half of patients remaining free from MACE at 10 years. These findings highlight the importance of longitudinal risk assessment in LVNC and suggest that the course of the disease is associated with a substantial burden of morbidity and mortality. Future studies should prioritize standardized reporting of individual MACE components. Sudden cardiac death and malignant ventricular arrhythmias are particularly important targets, as improved characterization of their incidence and predictors may directly inform implantable cardioverter-defibrillator risk stratification and management decisions in patients with LVNC or LV hypertrabeculation. Several limitations should be acknowledged. Various definitions of LVNC across studies, with diagnosis based on different echocardiographic or cardiac magnetic resonance criteria, and MACE definitions were not uniform. Recent expert consensus has proposed the term excessive LV trabeculation in place of LVNC, reflecting evolving concepts regarding myocardial development; however, the present analysis uses the terminology and diagnostic criteria reported in the original studies included in this meta-analysis.² These estimates describe outcomes in heterogeneous historical cohorts, not the natural history of isolated hypertrabeculation. In the recent multicenter dilated cardiomyopathy study, LV hypertrabeculation was not associated with worse outcomes; LVEF, late gadolinium enhancement, and genotype were important prognostic factors.⁶ Our unadjusted analysis cannot distinguish the contribution of underlying cardiomyopathy, systolic dysfunction, or myocardial fibrosis from that of trabeculation. The cohort-level mean LVEF of 40.1% further limits extrapolation to individuals with isolated hypertrabeculation and normal ventricular function. These findings should not be interpreted as supporting trabeculation-specific treatment or surveillance. In addition, as this analysis is based on reconstructed data from observational studies, residual confounding cannot be excluded. Finally, there was no pooled comparator without hypertrabeculation, and merging reconstructed observations did not adjust for between-study differences. Survival estimates beyond 4 years are imprecise because of diminishing numbers at risk; 10-year estimates should therefore be interpreted cautiously. Further prospective studies with standardized definitions and long-term follow-up are warranted to better characterize outcomes in this population.

Funding support and author disclosures

Dr Martinez has received consulting fees for his roles with Major League Soccer and the National Basketball Players Association and as a clinical cardiologist for various Major League Baseball teams (paid to employer), National Collegiate Athletic Association teams (paid to employer), the Minnesota Vikings (paid to employer), and the National Hockey League; is an advisor for Cardiac Insight; has received speaker fees from Bristol Myers Squibb and Cytokinetics for hypertrophic cardiomyopathy (HCM) therapies (both relationships have ended); and is the national leader for phase 3 HCM-related treatment trials for Braveheart and Bristol Myers Squibb. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.