Study design

We included adult unrelated patients with DCM who underwent genetic testing by NGS between 2012 and 2021. We also included patients who did not have LV dilation, provided that all other criteria for DCM were met, as well as patients with DCM in whom NGS testing was not performed due to the detection of a causative genetic variant by another method (mainly Sanger sequencing). In total, the group consisted of 280 individuals. The patients came from all over Poland, however, their geographic distribution was not even. Patients living in Mazowieckie and neighboring voivodeships predominated, while participants from distant regions were less numerous.

DNA sequencing and variant analysis

NGS was performed using different panels (details in Table 1 and Supplementary material, Tables S1 and S2). In 22 patients, pathogenic or likely pathogenic variants were detected by single‑gene testing, most often by direct Sanger sequencing of coding exons of lamin A/C (LMNA), BCL2‑associated athanogene 3 (BAG3), phospholamban (PLN), or sodium voltage‑gated channel α subunit 5 (SCN5A) including flanking intronic regions, less often by real‑time polymerase chain reaction methods or multiplex ligation‑dependent probe amplification.^17-20

The identified rare variants in cardiomyopathy‑related genes were classified according to the American College of Medical Genetics and Genomics (ACMG) criteria using the online tool GeneBe (https://genebe.net/; classification valid as of February 2025).^21-23 They were considered causative if they were pathogenic (P) or likely pathogenic (LP) and were found in 1 of the 20 genes with strong or moderate evidence of causative relationship with DCM (hereinafter referred to as DCM‑related genes, listed in Table 1).^1,8 Genes with limited or disputed association with DCM were not included.

All causative variants identified with NGS were confirmed with Sanger sequencing.

We excluded 3 patients with homozygous / compound heterozygous variants in the genes with disputed evidence for association with DCM whose cases had been published previously as genetic cardiomyopathy: 1 with a homozygous variant in the nebulin‑related anchoring protein gene,²⁴ another with variants in the 1,4-α-glucan branching enzyme 1 gene,²⁵ and the last with compound heterozygous variants in the tripeptidyl peptidase 1 gene.²⁶

In order to compare clinical characteristics and prognosis, the patients with gene‑positive DCM were divided into 3 groups of carriers of P/LP variants in: the titin (TTN) gene (group 1), the LMNA gene (group 2), and other genes (group 3). The patients with more than 1 P/LP variant were not included in either of these groups.

Clinical assessment

Medical data of all participants were retrospectively collected, including baseline clinical information at the time of the initial visit, prior medical records, and follow‑up information. We analyzed the baseline data, comprising medical and family history, 12‑lead electrocardiography (ECG), echocardiography, comorbidities, and current treatment. All records were reviewed for the occurrence of major events, such as death, heart transplantation (HTx), implantation of a LV assist device (LVAD), implantable cardioverter‑defibrillator (ICD), cardiac resynchronization therapy (CRT), pacemaker implantation, sudden cardiac arrest (SCA), or sustained ventricular tachycardia (VT).

Two composite end points were defined: severe DCM (definition in the next section) and a composite end point of cardiovascular death, HTx, or LVAD implantation.

Definitions

The diagnosis of DCM was made when the LV end‑diastolic diameter (LVEDD) exceeded 112% of the predicted value, corrected for age and body surface area according to the Henry’s formula, and when the LV ejection fraction (LVEF) was below 50%.

DCM was considered gene‑positive if P/LP variants were identified in 1 of the DCM‑related genes (Table 1).

First‑degree AVB was defined by a PR interval exceeding 200 ms on standard 12‑lead ECG. High‑degree AVB included type II second‑degree or third‑degree AVB. Atrial arrhythmias included atrial fibrillation, flutter, and paroxysmal atrial tachycardia lasting at least 30 seconds.

HF was recognized when typical symptoms were present, and was staged using the New York Heart Association (NYHA) classification (classes 1–4).

We labelled cardiomyopathy as an acute onset disease if it was diagnosed during acute hospitalization due to symptoms of HF requiring inotropic agent infusion or due to myocarditis‑like symptoms (chest pain, troponin level elevation).

SCA was defined as occurring within 1 hour of the onset of acute symptoms and reversed with cardiopulmonary resuscitation.

We used the term “severe DCM” to mark advanced, symptomatic LV systolic dysfunction (LVEF <⁠35%) persisting despite optimal medical treatment (and equivalent to the classic criteria for ICD/CRT implantation) or the occurrence of life‑threatening events. We confirmed it when one of the following occurred first: CRT/ICD implantation (persisting advanced LV dysfunction), CV death, HTx, LVAD implantation, SCA, or hemodynamically unstable VT (life‑threatening events).

DCM was considered familial if at least 1 first- or second‑degree relative was diagnosed with DCM. A family history of sudden cardiac death (SCD) was considered positive if at least 1 first‑degree relative had died suddenly before the age of 50 years.

Ethics

The study followed the Declaration of Helsinki, and was approved by the Local Bioethics Committee at the National Institute of Cardiology, Warsaw, Poland (IK.N.PIA.002.39.2153/25). All patients signed informed consent to participate in the study.

Statistical analysis

All results for categorical variables are presented as counts and percentages and for continuous variables as mean and SD or median and interquartile range (IQR). The χ² independence or Fisher exact test was used to compare the categorical variables. The differences between continuous variables were tested with the t test (for 2 independent samples, normally distributed data). Significance of differences between the means in several groups was analyzed with one‑way analysis of variance with the post hoc Tukey–Kramer test for multiple comparisons.

Survival curves were constructed using the Kaplan–Meier method. Homogeneity of the curves was verified by the log‑rank test. The multiple‑comparison method for adjusting P values of the paired tests was based on the Tukey studentized range test with Kramer approximation. Hazard ratios (HRs) with 95% CIs were calculated using the Cox proportional hazards model. The proportionality of risk assumptions was checked based on weighted Schoenfeld residuals. In survival models, the variable defining the time to the first event or the censoring time was the patient’s age. In patients without an event, the follow‑up period extended to the most recent evaluation before November 30, 2024. A P value below 0.05 was considered significant.

All hypotheses were 2‑tailed with a 0.05 type I error.

All statistical analyses were performed using SAS statistical software, version 9.4 (SAS, Cary, North Carolina, United States).

Molecular findings in the study cohort

The study cohort included 280 unrelated DCM patients. NGS was performed in 258 individuals, and it was waived in 22 cases in whom a causative genetic variant was identified by another method (14 carriers of LMNA variants, 6 of BAG3 variants, 1 of PLN, and 1 of SCN5A variant). In total, 121 different P/LP variants in one of the DCM‑related genes were identified in 130 patients (46%; Figure 1), of whom 3 carried 2 variants and 127 carried a single variant (all variants are listed in Supplementary material, Tables S3 and S4). Twenty two of these variants have not been reported in the literature so far, 15 of them have not been entered into the ClinVar database. All identified variants were heterozygous except for dystrophin (DMD) hemizygous variants in men. Ten P/LP variants were found in more than 1 patient.

Among the patients with identified P/LP variants, as many as 39% (17% of all studied patients) had variants in the TTN gene (49 in TTN only and 2 also in another DCM‑related gene). The second most frequently affected gene was LMNA, accountable for 8% of all DCM cases and 17% of gene‑positive DCM. Other genes in which variants were identified at a frequency exceeding 2% were: desmoplakin (DSP), myosin heavy chain 7 (MYH7), SCN5A, filamin C (FLNC), BAG3, and DMD. Detailed distribution of the P/LP variants in DCM‑related genes is presented in Table 2.

Clinical characteristics of the study population

Clinical characteristics of the study population at baseline evaluation are listed in Table 3. The mean (SD) age of the DCM patients was 39 (13) years, 68% were men. As many as 54% of them had familial form of DCM. In 19%, the disease had an acute onset.

At the time of baseline evaluation, 83% of our patients were diagnosed with HF, 21% were in the NYHA class 3 or 4. The mean (SD) baseline LVEF was 35% (12%), and 31% had LVEF below 30%. One‑fourth of the participants had ICD implanted (33% as secondary prophylaxis), and 6% received CRT.

In the entire cohort, 3% of the patients were diagnosed with muscular dystrophy, and 6% had elevated activity of serum creatine kinase at above 350 IU/l. Comorbidities were infrequent, with arterial hypertension (AH) being the most common, found in 18% of the patients.

Atrial arrhythmias and AVB were more frequent in the gene‑positive DCM (34% vs 21%; P = 0.009 and 26% vs 15%; P = 0.01, respectively). In contrast, left bundle branch block (LBBB) and AH were more common in the gene‑negative group (16% vs 34%; P = 0.001 and 13% vs 23%; P = 0.03, respectively).

The patients from the group 3 (non-TTN, non-LMNA variant carriers) were characterized by younger age at diagnosis and baseline evaluation. LMNA-variant carriers had more adverse family history, and were more likely to have arrhythmias and conduction disorders, including those requiring device implantation. AH was more common in the gene‑negative DCM and in TTN-variant carriers.

Clinical course and prognosis

Median (IQR) follow‑up was 6.2 (2.7–8.9) years.

In the study cohort, 49 patients underwent HTx and 35 died (deaths after HTx not included). Of the deaths, 34 were cardiovascular, primarily due to progressive HF; 4 were sudden. LVAD was used in 25 patients: in 17 as a bridge to HTx, and in 3 patients as long‑term treatment. Two patients died shortly after implantation from complications, and in another 3 patients, the device was explanted after improvement.

ICD was implanted in 145 patients (70 before initial assessment, the rest during follow‑up). In secondary prevention, ICD was implanted in 25 patients: in 19 after SCA (in 11 patients it was the first symptom of the disease) and in 6 following a diagnosis of sustained, hemodynamically unstable VT. CRT was used in 55 patients, in most of them (n = 52) together with a defibrillation system, and in 3 patients with pacing only.

A total of 177 participants developed severe DCM, and 89 reached the composite end point of cardiovascular death, HTx, or LVAD implantation.

The gene‑positive group had worse prognosis, both in terms of the risk of developing severe DCM (Figure 2A; HR, 1.6; 95% CI, 1.19–2.16) and a composite end point of cardiovascular death, HTx, or LVAD implantation (HR, 1.81; 95% CI, 1.19–2.77; Figure 3A), but it varied depending on the genetic background.

The prognosis in the TTN variant carriers did not differ from the gene‑negative group (HR, 0.93; 95% CI, 0.6–1.43 for developing severe DCM and HR, 1.08; 95% CI, 0.58–1.99 for cardiovascular death, HTx, or LVAD implantation).

In contrast, the risk of developing severe DCM was higher in the carriers of P/LP variants in the LMNA or other genes (Figure 2B). In comparison with gene‑negative DCM, it was 2.4‑fold higher in the group 2 (LMNA-variant carriers), and 2.3‑fold higher in the group 3 (95% CI, 1.52–3.93 and 1.54–3.33, respectively). Also, the risk of cardiovascular death, HTx, or LVAD implantation was higher in the groups 2 and 3 than in the gene‑negative group (HR, 2.97; 95% CI, 1.62–5.43 and HR, 2.02; 95% CI, 1.14–3.58, respectively; Figure 3B).

Limitations

A major limitation of the study is that it was conducted in a single tertiary referral center, and therefore it may have included more young patients, those with poorer prognosis and a burdensome family history in comparison with the general DCM population. Moreover, although the patients included came from all over Poland, the geographic distribution of their places of residence was not even. The patients living closer to the center predominated, while those from distant regions were less numerous.

Conclusions

In summary, DCM‑related genetic variants were identified in almost half of the patients from our study cohort. The patients with identified variants were characterized by a higher frequency of a positive family history, AVB or atrial arrhythmias, and a lower frequency of LBBB or AH. The prognosis was worse in the patients with variants in genes other than TTN.