Left Ventricular Hypertrabeculation

Definition

Left ventricular hypertrabeculation — taught for three decades as "left ventricular noncompaction cardiomyopathy" — describes a ventricle with prominent trabeculae, deep intertrabecular recesses and a thin compacted epicardial layer. The terminology itself is the argument: hypertrabeculation names a morphological observation, noncompaction names an abandoned embryological mechanism, and the evidence has moved decisively toward the first. It is a phenotypic trait that appears across genotypes and loading states, not a cardiomyopathy in its own right — the European Society of Cardiology's position, in direct opposition to the American Heart Association's classification of it as a primary cardiomyopathy (sources/trabeculation-dcm-circ-2026, rating: high).

Key Concepts

Two ways to measure it, one conclusion

• The two agree substantially (84.8%; Cohen κ 0.638) but not completely, and neither is a gold standard. The Petersen ratio is limited by manual measurement, interobserver variability and the impossibility of capturing three-dimensional complexity; fractal analysis is more reproducible but over-reads apical thrombus and apically displaced papillary muscles as trabeculae. (sources/trabeculation-dcm-circ-2026)
• Both methods gave the same prognostic answer — none. Reporting hypertrabeculation is therefore not a reason to choose a more aggressive imaging or follow-up strategy.

Prevalence is high, and it is highest where prognosis is best

In 1,160 genotyped patients with DCM, hypertrabeculation clustered by functional gene group in an order that inverts clinical intuition: (sources/trabeculation-dcm-circ-2026)

• The genotypes that kill — nuclear envelope, cytoskeleton — are the ones that do not produce hypertrabeculation. Any cohort assembled by imaging for noncompaction is therefore enriched for the milder genotypes, which is a plausible explanation for why older series found such high complication rates: they were selected on a trait, not on a disease.
• The same gradient appears outside DCM: hypertrabeculation is reported in other cardiomyopathies, neuromuscular disease, congenital heart disease, trained athletes, pregnancy (reversible, de novo) and healthy individuals — up to 43% in one healthy cohort — and MESA found no progressive remodelling over 9.5 years.

It is not prognostic in DCM

• Over a median 5.1 years in the Circulation 2026 cohort: embolism 3.2%, advanced heart failure 5.3%, major ventricular arrhythmia 11.7%. No end point differed by hypertrabeculation status after adjustment for genotype, LGE, NYHA class and LVEF (sources/trabeculation-dcm-circ-2026).
• The univariate analysis pointed the wrong way — hypertrabeculation appeared protective (MVA HR 0.64; AHF HR 0.54) — because hypertrabeculated patients were younger, had higher LVEF (38.2% vs 34.9%) and less LGE (24.7% vs 33.7%). This is confounding by indication in reverse: the trait marks the milder ventricle.
• What does predict outcome: nuclear envelope genotype (HR 4.4 for MVA), LGE (HR 3.1), lower LVEF — the same variables that stratify DCM without any reference to trabeculation. See concepts/VA-Risk-Stratification-DCM.
• No genotype showed a hypertrabeculation-attributable risk difference. The single nominal signal (fewer AHF events in TTN carriers with hypertrabeculation, P=0.024) is one of 21 comparisons, in a TTN group where 46 hypertrabeculated patients had zero events.

The anticoagulation question

Prophylactic anticoagulation in hypertrabeculation with reduced LVEF but sinus rhythm has been recommended on the basis of case series from the 1990s and 2000s reporting thromboembolism in 22–29%. The contemporary rate is an order of magnitude lower: (sources/trabeculation-dcm-circ-2026)

• Adjusted HR for hypertrabeculation: 1.50 (95% CI 0.75–3.00) in the whole cohort and 1.89 (0.66–5.46) in the sinus-rhythm/LVEF ≤40% subgroup. Neither excludes a moderate effect; neither supports treating everyone.
• Embolism in this cohort was largely an atrial fibrillation phenomenon: 15 of 37 (40.5%) had documented AF before the event, and 6 more (16.2%) had AF first identified at or within 6 months of it. AF carried HR 3.26 (1.47–7.23) and each 1% fall in LVEF HR 1.04 (1.01–1.08).
• The position that follows: anticoagulate for AF, not for trabeculation. This is consistent with HFrEF trial data (WARCEF, COMPASS, CORONA/GISSI-HF), where prophylaxis in sinus rhythm reduced ischaemic events at the cost of major bleeding and is not guideline-recommended.

Why the older literature looked so alarming

• Early series were small, referred for echocardiography, often did not report whether AF was present, and were conducted in decades when AF was frequently untreated or treated with antiplatelet therapy alone. The embolic events attributed to the trabeculae were frequently AF events with an unmeasured arrhythmia.
• Modern cohorts converge on 3–15% — and this cohort, with central CMR adjudication and AF as a time-dependent covariate, lands at the bottom of that range.

Contradictions / Open Questions

• The AHA and ESC disagree on what this entity is. The AHA classifies LV hypertrabeculation as a primary cardiomyopathy; ESC 2023 calls it a phenotypic trait. This is not a semantic dispute — it determines whether a young athlete or a pregnant woman with trabeculae but normal function receives a lifetime diagnosis, exercise restriction and family screening, or a note in the imaging report. (sources/trabeculation-dcm-circ-2026)
• The strongest evidence against the "distinct cardiomyopathy" model comes from a cohort that cannot fully test it. The nuclear-envelope group contained a single hypertrabeculated patient (who had an event), and the desmosomal group seven. "No interaction by genotype" is an absence of evidence in exactly the genotypes where the answer matters most. (sources/trabeculation-dcm-circ-2026)
• The embolic confidence interval is compatible with real harm. An adjusted HR of 1.89 in sinus rhythm with LVEF ≤40% and a CI reaching 5.46 means this study shows the absolute event rate is low, not that trabeculation is inert. A patient with additional risk factors could still cross a threshold no trial has defined.
• Whether hypertrabeculation is reactive or primary remains unresolved. The reactive hypothesis — LV dilatation stretches the wall and makes existing trabeculae visible — predicts more hypertrabeculation in more remodelled ventricles. This cohort found the opposite (higher LVEF, smaller LVEDV), which argues against it but does not establish a developmental cause.
• The MYBPC3 signal may be an artefact of misclassification. The authors concede some of the 11 MYBPC3 carriers may have had end-stage HCM rather than DCM. Since MYBPC3 shows 54.6% hypertrabeculation, the highest-prevalence motor sarcomeric result could partly reflect burnt-out HCM entered into a DCM cohort.
• No one has studied what happens to trabeculation over time in DCM. Whether it appears, resolves or persists with reverse remodelling — and whether that trajectory carries information — is not addressed by a single-baseline CMR study.
• Pregnancy makes trabeculation appear and disappear (reversible de novo trabeculations are described in pregnancy), which is difficult to reconcile with a fixed developmental defect and has practical consequences for anyone imaged for the first time while pregnant. See concepts/Pregnancy-in-Genetic-Cardiomyopathy.
• The 98% White cohort cannot answer the ancestry question. Hypertrabeculation is reported to be commoner in Black individuals, and this study's genotype gradients and prevalence may not transfer — a gap the authors name directly.

Connections

• Related to entities/DCM — the phenotype in which hypertrabeculation is commonest and best studied
• Related to concepts/VA-Risk-Stratification-DCM — the null result that leaves genotype, LVEF and LGE as the real stratifiers
• Related to entities/TTN — 37.7% prevalence in TTN-DCM
• Related to entities/MYH7 — 71.4% prevalence, highest of any individual gene
• Related to entities/LMNA — 5% prevalence, lowest, in the genotype with the highest arrhythmic risk
• Related to entities/FLNC — 7.1% prevalence
• Related to entities/MYBPC3 — 54.6%, possibly contaminated by end-stage HCM
• Related to concepts/Late-Gadolinium-Enhancement — LGE outperforms trabeculation as a prognostic marker
• Related to concepts/Pregnancy-in-Genetic-Cardiomyopathy — reversible de novo trabeculation in pregnancy
• Related to entities/HCM — the LVOTO phenotype that can mimic hypertrabeculated DCM at end stage

Sources

• sources/trabeculation-dcm-circ-2026