Practice Patterns after Return of Rare Variants Associated with Cardiomyopathy in the electronic MEdical Records and GEnomics (eMERGE) Network

Cardiomyopathies affect approximately 1 million people and contribute to over 40,000 deaths annually in the US.¹ While use of panel-based genetic testing is an effective tool following diagnosis of cardiomyopathy, the emerging genomics-first paradigm seeks to identify pre-symptomatic individuals to inform earlier detection and guide potentially life-saving interventions. However, population-based genomic screening approach may lead to increased healthcare expenditures that may be of unclear utility because of variable penetrance of pathogenic variants. Therefore, we sought to evaluate real-world practice patterns following return of pathogenic/likely pathogenic cardiomyopathy variants in electronic MEdical Records and GEnomics (eMERGE) participants.

One goal of the eMERGE Network was to examine implementation of genomic medicine. Briefly, 25,015 participants were recruited at 10 clinical sites in the US in the third phase of eMERGE with no pre-specified inclusion or exclusion criteria (ClinicalTrials.gov; Identifier: NCT03394859). Individuals with or without chronic diseases were enrolled and detailed descriptions of the study methods have been published.² The study was approved by each site’s Institutional Review Board and all participants gave informed consent. The data that support the findings of this study may be requested on the eMERGE website https://emerge-network.org/collaborate/. Many datasets summarized here will be publicly available in the dbGaP repository under phs001616.v1.p1. We included 125 participants who had results returned for pathogenic/likely pathogenic variants in 16 genes associated with cardiomyopathy (ACTC1, DSC2, DSG2, DSP, GLA, LMNA, MYBPC3, MYH7, MYL2, MYL3, PKP2, PRKAG2, TMEM43, TNNI3, TNNT2, and TPM1) that are considered clinically actionable by the American College of Genetics and Genomics (ACMG).^{3, 4} We excluded participants under the age of 18 years (n=13). We examined rates of cardiomyopathy diagnoses including prior genetic results, referrals, and testing in the preceding 5 years and 1 year post-disclosure based on abstraction and manual review of electronic health record (EHR) data.

Of 112 participants identified to have pathogenic/likely pathogenic variants in cardiomyopathy genes, 58.9% were female and 8.0% were Asian, 4.5% Black, and 3.6% Hispanic. Mean (standard deviation) age was 59.2 (15.8) years. The most common cardiomyopathy genes in which actionable variants were identified were MYBPC3 (n=34) and MYH7 (n=18) as shown in Panel A. Approximately 10% (11/112) had a pre-existing genetic cardiomyopathy diagnosis prior to study enrollment. An additional 13% (15/112) had a diagnosis of cardiomyopathy without prior genetic testing. Among the remaining 86 participants without known cardiomyopathy prior to enrollment, review of health records revealed heterogeneous patterns in post-disclosure clinical activity, such as differential utilization and timing of referrals and diagnostic imaging (Panel B). In the 1 year post-disclosure, 8.1% (7/86) had a new diagnosis of cardiomyopathy (4 hypertrophic, 1 arrhythmogenic, and 2 unspecified cardiomyopathy). While the majority were referred to a cardiologist (51% [44/86]) and/or a genetic counselor (35% [30/86]) post-disclosure, 39.5% (34/86) were not referred to either a cardiologist or genetic counselor (Panel C). Forty-five percent (39/86) had post-disclosure assessment of ventricular structure and function with either echocardiography (TTE) or cardiac magnetic resonance imaging (cMRI). Approximately 42% (36/86) had undergone TTE or cMRI for other indications pre-disclosure with repeat imaging evaluation in 19 individuals post-disclosure; 26% (22/86) had no cardiac imaging pre- or post-disclosure. Post-disclosure risk stratification for arrhythmia with a Holter was performed in 24% (21/86). No differences in age, sex, or race/ethnicity were observed in those with or without post-disclosure testing.

This observational study demonstrates that return of cardiomyopathy variants contributed to new cardiomyopathy diagnoses and facilitated meaningful clinical activities. Among those with a pre-existing cardiomyopathy diagnosis, we identified actionable genetic variants that can inform family-based cascade screening. However, practice patterns were highly variable. This may have been, in part, due to available data from cardiac testing in the 5-years prior to disclosure for other indications. Alternatively, comfort and knowledge regarding discussion of genetic results may have driven post-disclosure clinical activity. In a recent survey of primary care physicians, over 60% reported lack of comfort in discussion of genetic information and were identified to have no or minimal knowledge for when and how to incorporate genomic information into clinical practice.⁵

Limitations include lack of standardized participant follow-up across sites to define genotype-phenotype relationships and inability to identify reasons for lack of clinical action post-disclosure. However, referral and testing data were manually adjudicated from the EHR. There is the potential to miss clinically relevant variants, such as variants of uncertain significance (VUS) that lack robust clinical evidence to support pathogenicity. However, we based our actionable return on expert consensus annotation of variants and that return of VUS as a secondary finding is not advised by the ACMG.⁴ Additionally, gene-specific analysis is limited based on the small sample size. However, given the rare nature of these pathogenic variants, we provide a large, representative sample from across the US. Lastly, reliable information on family history and cascade testing following disclosure was not uniformly available.

In summary, we demonstrate differences in practice patterns following disclosure of actionable variants in cardiomyopathy genes. Further study is needed to demonstrate if genomics-enhanced screening approaches may facilitate earlier diagnostic and therapeutic interventions in asymptomatic/pre-symptomatic patients and their first-degree family members.

Figure. Pathogenic or likely pathogenic variants according to gene (A) and heterogeneous patterns (B) as well as varying percent of clinical follow-up (C) after disclosure of incidental findings in genes associated with CM.

In Panel B, circle indicates clinical consultation visit and square represents diagnostic testing; Red color indicates post-disclosure referral, diagnostic testing, or new diagnosis of cardiomyopathy.

Abbreviations include: CV: cardiovascular; TTE: transthoracic echocardiogram; cMRI: cardiac magnetic resonance imaging; HCM: hypertrophic cardiomyopathy; ROR: return of incidental finding of rare pathogenic or likely pathogenic variant results (abbreviations not defined here are gene names)