Diagnostic yield in victims of sudden cardiac death and their relatives

Abstract

Aims

International guidelines recommend cardiogenetic screening in families with sudden cardiac death (SCD) if the suspected cause is an inherited cardiac disease. The aim was to assess the diagnostic yield of inherited cardiac diseases in consecutively referred SCD families.

Methods and results

In this single-centre retrospective study, we consecutively included families referred to our tertiary unit between 2005 and 2018 for screening due to SCD. Following evaluation of premortem medical records and postmortem findings for the proband, the families underwent a guideline-based screening protocol. Relatives were followed and cardiovascular events registered. In total, 304 families with 695 relatives were included. In probands, mean age at death was 39 years (75% males) and in relatives mean age at screening was 35 years (47% males). The proband-diagnosis was established through autopsy findings (n = 89), genetic analyses (n = 7), or based on premortem findings (n = 21). In the remaining 187 families with borderline/no diagnosis in the proband, screening of relatives yielded a diagnosis in 26 additional families. In total, an inherited cardiac disease was identified in 143 out of 304 families (47%). In relatives, 73 (11%) were diagnosed. Arrhythmogenic right ventricular cardiomyopathy (n = 16) was the most common diagnosis. During follow-up (mean 5.5 years), a low rate of serious cardiac events was observed (no SCD events).

Conclusion

Forty-seven percent of SCD families were diagnosed. Eleven percent of the screened relatives received a definite diagnosis and were offered treatment according to guidelines. A low rate of serious cardiovascular events was observed among SCD relatives.

Introduction

Sudden cardiac death (SCD) is defined as a sudden and unexpected non-traumatic death caused by cardiac or unknown cause.¹ A Danish nationwide study² from 2014 reported that approximately 11% of deaths in the age group 1–49 years were SCD and the annual rate of SCD in this age group was 8.6 per 100 000 persons (∼300 SCD events per year). Since the death is sudden, unexpected and often without prodromal symptoms, several different pre-symptomatic screening methods have been suggested for prevention. Due to the inherited nature of a considerable fraction of underlying diseases associated with SCD in the young,³ guidelines for detection of at-risk relatives recommend a comprehensive postmortem analysis focusing on the diagnosis in the SCD victim⁴ followed by a phenotype-driven evaluation of the relatives.⁵ Although autopsy is key, the autopsy rate in sudden unexpected death in Denmark is 68% in the age group 1–35 years, decreasing to 44% in the age group 36–49 years.

Previous studies have reported that an inherited cardiac diagnosis can be identified in 18–60% of families when performing clinical and genetic evaluation of SCD families.^5–12 Once a given diagnosis is identified, it is possible to offer disease-specific prophylactic treatment and follow-up of the affected relatives. Most of the previous studies^5–12 primarily focused on cases of sudden unexplained death syndrome (SUDS) and sudden arrhythmic death syndrome (SADS); hence, knowledge of the overall outcome of assessing consecutive families with all types of SCD is sparse. In the earlier reports, the chance of finding an inherited diagnosis in SUDS/SADS families has been estimated to 18–60%. However, the chance of finding a diagnosis in all types of SCD families remains unknown.

A few studies,^13,14 including a recent Danish study on 292 relatives (56 families), have reported a low rate of cardiac events in relatives to SCD victims when guideline-based management of SCD relatives is provided underlining the importance of a systematic approach.

In 2005, a highly specialized unit for inherited cardiac diseases was established at our institution offering guideline-based^7,8 systematic work-up for SCD families. The aim of this study was to evaluate the diagnostic yield of the systematic cardiogenetic screening in a large, consecutive population of all categories of SCD families. As a secondary aim, we investigated the rate of cardiovascular events during follow-up of a large cohort of SCD families.

Methods

Study design and inclusion of index cases

In this single-centre study, we retrospectively collected data from all SCD victims >1 year of age and their relatives consecutively referred to Rigshospitalets Unit for Inherited Cardiac Diseases, Copenhagen University Hospital, Rigshospitalet, Denmark from 1 January 2005 until 1 October 2018. Hospitals (including this hospital), general practitioners, private cardiologists, and the forensic department refer to the Unit. Upon referral, the SCD victim’s age, sex, circumstances of death, and family history of SCD were registered. A comprehensive multidisciplinary investigation of the cause of death was initiated, autopsy reports and premortem data including patient files were thoroughly studied, and autopsy conclusions categorized as ‘definite’, ‘probable’, or ‘non-diagnostic or normal’. If cardiac tissue samples were available most cases were re-examined by an expert cardiac pathologist. In victims with available blood or tissue, postmortem phenotype-targeted genetic analysis was performed if the autopsy findings and/or medical history were suggestive of a specific inherited cardiac disease. The indication for a genetic analysis was discussed in a multidisciplinary team of cardiologists, pathologists, and clinical geneticists. Identified genetic variants were assessed using the ACMG 2015 guidelines.¹⁵

The study was approved by the Danish Data Protection Agency and the Danish Patient Safety Authority (3-3013-2380/1).

Definition of sudden cardiac death

We defined SCD as the sudden, natural, and unexpected death of unknown, or cardiac cause; in unwitnessed cases, the person had to have been seen alive and functioning normally <24 h before being found dead and in witnessed cases as an acute change in cardiovascular status with the time to death being <1 h. In non-autopsied cases, we used the same criteria in cases presumed to be of cardiac cause based on the circumstances relating to the death. A prior medical history was not an exclusion criterion, but the time and manner of death had to be unexpected. Sudden cardiac death victims that were resuscitated initially, but died within a few days, were also included. If no explanation of the death was found after a full postmortem analysis was performed, we used the term SUDS. Sudden arrhythmic death syndrome was defined as SUDS with normal toxicology.

Definitions of probands and relatives

A proband was defined as the index person in the family who suffered the SCD, which, due to suspected inherited cardiac cause, lead to the referral of the family to our unit. Related relatives were assessed in accordance with the principle of cascade screening; primarily first-degree relatives were screened, and additional relatives to affected relatives were also offered screening.

Screening of relatives

Upon referral, a comprehensive family screening algorithm similar to previously reported algorithms was applied.^7,8 Contact to at-risk relatives was obtained through pseudo-probands and followed a standardized protocol: a first visit with a specialized doctor or nurse, evaluation of cardiopulmonary symptoms, medical history, physical investigation, drawing of a pedigree, obtainment of different electrocardiogram (ECG) modalities (including a standard resting 12-lead ECG), and routine blood sample testing. If a definite diagnosis was established in the proband, the relatives received phenotype-specific investigations. Standard clinical investigations of relatives from families without diagnoses included transthoracic echocardiography, signal-averaged ECG, and 24-h Holter-monitoring. An exercise test was at the attending physician’s discretion, e.g. if the proband had experienced exercise-related antemortem symptoms/circumstances of death or the relative had exercise-related symptoms. Further clinical investigations depended on whether the family had an established diagnosis and could include prolonged Holter-monitoring, implantable loop recorder, drug challenges (flecainide or ajmaline), coronary computed tomography angiography, conventional coronary angiography, and/or cardiac magnetic resonance imaging. At a separate consultation, the findings were summarized, and the relatives were classified as affected, borderline affected, or unaffected. If a probable or confirmed pathogenic genetic variant had been identified in the family, cascade genetic testing was performed.

Follow-up and treatment of diagnosed families were offered according to phenotype-specific guidelines.¹⁶ If no diagnosis was established in the family, follow-up of relatives with 3–5 years intervals was offered. Phenotype/genotype-negative relatives from gene-positive families were reassured and follow-up ceased.

Cardiovascular events in relatives during follow-up

Baseline was defined as the relative’s first visit in the clinic. Relatives were followed until 1 October 2018. During the follow-up period, we registered cardiovascular events: hospitalizations/outpatient visits for cardiac symptoms, new-onset reduced left ventricular ejection fraction (LVEF) ≤45%, acute myocardial infarction, ventricular tachycardia (VT)/fibrillation, device implantation [pacemaker, implantable cardioverter-defibrillator (ICD), cardiac resynchronization therapy], ICD therapy, catheter ablation, left ventricular assist device, heart transplantation, aborted SCD, SCD, and death due to other causes. Cardiac events during follow-up were registered for all relatives regardless of whether they had received a diagnosis or not.

Statistical analysis

Descriptive statistical analyses [including mean, standard deviation (SD), range, and interquartile range (IQR)] were used. Data are presented as mean ± SD or numbers and percentage of totals. Statistical analyses were performed using Microsoft Excel^® version 1808 (Microsoft Corporation, WA, USA).

Results

Characteristics of probands

In total, 304 SCD probands were included, mean (± SD) age at death was 38.9 years (±13.6), 75% males. One hundred and seven (35%) SCD victims had a family history of SCD before the family was referred to our unit. Forty-four victims (15%) had syncope prior to the SCD event. The circumstances of SCD were evenly distributed between sleep, rest (e.g. recumbent or sitting), light (e.g. walking), and intense exercise (e.g. running, biking, swimming, heavy lifts, etc.) but was unknown in 28% of the cases. Results are summarized in Table 1.

Diagnostic yield in probands

An autopsy was performed in 220 probands (72%) and led to the establishment of a definite autopsy diagnosis in 91 (41%), a probable autopsy diagnosis in 47 (21%), and no diagnosis in 82 (37%) of the probands. Combining SCD victims with definite and probable autopsy conclusions the most common diagnoses were arrhythmogenic right ventricular cardiomyopathy (ARVC, 25%), ischaemic heart disease (IHD, 16%), and hypertrophic cardiomyopathy (HCM, 10%) (Table 1). In the non-autopsied group (A. Kjerrumgaard et al., unpublished data, 2019), a total of 22 probands were diagnosed either based on premortem medical records, emergency cardiac work-up (i.e. ECG monitoring, echocardiography, coronary artery catheterization or myocardial biopsy), or genetic testing (Figure 1).

Genetic testing was performed in 32% (n = 29) of probands with a definite autopsy diagnosis; in 19% (n = 9) with a probable autopsy diagnosis; in 15% (n = 12) with a non-diagnostic autopsy diagnosis, and in 5% (n = 4) of non-autopsied probands. A definite or probable pathogenic genetic variant was identified in 4, 1, 3, and 2 probands, respectively (Table 2).

Thus, the combined diagnostic yield based on review of pre-existing medical information, autopsy findings, and genetic analyses was the establishment of a definite diagnosis in 117 of 304 victims (38%). When including probands with probable autopsy findings the diagnostic yield was 163 (54%). In the seven genetically diagnosed probands, the positive genetic test identified a definite diagnosis in four cases and changed the diagnosis retrieved from the autopsy in three cases (Figure 1). The genetic variants identified in the probands are presented in Supplementary material online, Table S1.

Characteristics of relatives

We included 695 relatives, mean age 35 years (SD ± 17), 329 (47%) males (Table 3). The majority (564; 81%) were first-degree relatives and mean number of relatives included per proband was 2.3 (range 1–10). In total, 224 (32%) reported cardiac symptoms at their first visit. The prevalence of comorbidities was low; the most frequent cardiovascular comorbidity was hypertension (13%). Baseline characteristics on relatives are summarized in Table 3. The mean follow-up period was 5.5 years (SD 3.7; IQR 2.3–8.1). Clinical data on the work-up of relatives is summarized in Supplementary material online, Table S2.

Diagnostic yield in relatives

Clinical work-up of the 695 relatives yielded a definite clinical diagnosis in 65 (9%; 17 gene-positive) relatives. Additional eight relatives were genotype-positive but phenotype-negative. Combining the clinical and genetic yields, 73 relatives (from 49 families) out of the 695 (11%) were diagnosed with an inheritable cardiac disease. There was a predominance of cardiomyopathies: ARVC (22%), dilated cardiomyopathy (DCM, 13%), and HCM (10%) followed by the primary arrhythmia disorders: long QT syndrome (LQTS, 9%), Brugada syndrome (BrS, 9%), and short QT syndrome (SQTS, 8%). The diagnostic yield in the isolated group of relatives to SUDS/SADS victims was 10% (23 out of 227 relatives). In addition to the 73 relatives with a definite diagnosis, 77 (11%) relatives had borderline findings suggestive, but not diagnostic, of an inherited cardiac disease. Affected relatives were offered pharmacotherapy, invasive, or device-based therapy according to guidelines (n = 46; 7%). Results are summarized in Figure 2.

Overall diagnostic yield in families

In 187 families, no definite diagnosis in the proband was obtained after evaluation of clinical, genetic and autopsy findings. However, as a result of the cardiogenetic work-up of the relatives, an additional 34 relatives from 26 families received a definite inherited cardiac diagnosis (Figure 1). Hence, the overall diagnostic yield in the families was 143 out of 304 families (47%), i.e. screening of the relatives increased the diagnostic yield in the families by 22%. The diagnostic yield in families with SUDS/SADS probands was 16 out of 75 families (21%). Most relatives obtaining a definite diagnosis came from families with a proband with either definite or non-diagnostic autopsy findings (Figure 1). Numbers and percentages of the diagnostic yield in different SCD families are summarized in Table 4.

Cardiovascular adverse events during follow-up

No SCD events were reported during follow-up, but seven relatives died of non-cardiac causes (Table 5). A 73-year ARVC relative experienced aborted SCD related to an ST-elevation myocardial infarction. Ventricular fibrillation/sustained VT occurred in additional five relatives. Non-sustained VT was documented in 38 relatives, 16 (2%) relatives had a primary prophylactic ICD implanted, and three relatives received appropriate therapy (shock-therapy or anti-tachycardia pacing). A catheter ablation procedure was performed in 11 relatives. Seven relatives developed new-onset of systolic dysfunction with reduced LVEF. During follow-up, 106 relatives (15%) were referred to the hospital due to cardiac symptoms. Acute myocardial infarction occurred in four relatives (three males and one female); all were over 55 years old, hence, the events could not fully be explained by an inherited cardiac disease.

In addition to the 65 relatives phenotypically and/or genetically diagnosed at baseline, eight relatives developed a clinical phenotype during follow-up (5.5 years): ARVC (n = 2), LQTS (n = 2), BrS (n = 1), HCM (n = 1), DCM (n = 1), and progressive cardiac conduction disease (PCCD, n = 1). All eight relatives were younger than 55 years (mean age at diagnosis: 45.6 years), came from autopsied SCD probands with five being SUDS/SADS cases and had no signs or symptoms of cardiac disease at the initial work-up. However, five of the eight relatives came from families where an inherited cardiac diagnosis had already been identified. Affected relatives were treated according to guidelines.

Discussion

In this hitherto largest study of consecutive families referred to screening for inherited cardiac diseases due to SCD in the family, the overall diagnostic yield of inherited cardiac diseases in 304 SCD families was 47%; and 11% of 695 relatives were found to be phenotype- and/or genotype-positive. Cardiomyopathies were the most frequent diagnoses followed by primary arrhythmia disorders. The diagnosed relatives were offered phenotype-specific treatment including pharmacological therapy, invasive treatment, or device implantation. Apart from cardiac hospital referrals, we experienced a low number of serious cardiac events in SCD relatives.

Our total diagnostic yield among all types of SCD families was 47%, and the yield among SUDS/SADS families was 21%. These findings are in agreement with other similar studies^5–12 which have reported yields of 18–60% in SUDS/SADS families. The proportion (11%) of relatives diagnosed in this study is a bit lower than previous reports showing 13–33% diagnostic yields among relatives.^5,8,9 The reasons for these variable diagnostic yields among families and relatives are many and include differences in study design/inclusion criteria, different age groups, variable autopsy rates, and definitions of affected as a clinical and/or genetic diagnosis, variable number of relatives screened per family, and inclusion of borderline diagnoses in diagnostic yields. Including relatives with borderline diagnoses would in our study have resulted in a yield of 54% among the entire group of SCD families, 28% among the SUDS/SADS families, and 22% among all SCD relatives. However, our cohort represents a wider, consecutive, and more heterogeneous study population than previous reports and yields may not be directly comparable. So, the overall diagnostic yield is strongly context/referral-dependent, and our 47% estimate should be interpreted as the diagnostic yield of inherited cardiac disease in a real-world SCD clinic.

The diagnoses found in the probands, in the present study, were mostly cardiomyopathies followed by IHD and aortic disease. In contrast to previous studies,^1,9 we did not find IHD as the most common explained cause of SCD. In addition to geographical variation and age differences in the study populations referral bias likely exists, e.g. families with a cardiomyopathy diagnosis are more likely to be referred for evaluation. Our findings suggest a possible underdiagnosis of factors associated with premature IHD, e.g. familial hypercholesteraemia, supported by only 3% of relatives in our study had a significantly elevated low-density lipoprotein cholesterol. Aortic disease represents an important cause of SCD consistent with previous studies^1,2 and underlines the importance of presymptomatic screening of relatives as the disease is generally asymptomatic until a complication occurs. Myocarditis has also been reported as a common cause of SCD in the young,² but due to our study design focused on the hereditability of SCD no autopsied cases existed in our cohort. Among relatives to SUDS probands, we identified phenotypes of primary electrical diseases (LQTS, SQTS, BrS, PCCD, and catecholaminergic polymorphic VT). The favourable outcome of early treatment of patients with these diagnoses underlines the value of clinical screening of relatives. Systematic molecular autopsy would have led to the diagnosis in some of these cases^9,17 but the yield of genetic screening in e.g. BrS is as low as 20%¹⁸ suggesting that most families would have remained undiagnosed based on only genetic testing of the SCD victim. In our study, a genetic variant supported clinical findings or isolated lead to a diagnosis in 10 probands and 25 relatives (Table 2). Genetic testing evolved during the study period and gene panels were continuously updated. Larger gene panels have increased diagnostic yields but at the likely cost of a lower signal-to-noise ratio.¹⁹

Most diagnoses identified in the relatives were related to the diagnosis in the proband; however, in a few cases, there were discrepancies between the diagnosis established in the relative and in the proband. This could partially be explained by phenotypic plasticity and co-occurrence of multiple inherited diseases in the same family. However, postmortem clinical and genetic investigations do not always provide definitive results and many acquired conditions may phenocopy inherited cardiac disease.²⁰ This underlines the complexity of diagnosing inheritable cardiac diseases in clinical practice. Furthermore, when comprehensively screening a large number of individuals, incidental findings should be expected, e.g. bicuspid aortic valve in relatives of a proband with autopsy findings suggestive of ARVC. Furthermore, we found cardiomyopathies such as HCM and DCM in relatives to SUDS probands suggesting that the SCD victims had suffered a lethal ventricular arrhythmia despite the absence of structural disease at autopsy.

The guideline-based approach for screening and for treatment of affected relatives used in our clinic was associated with a low number of serious cardiac events during follow-up but hospital referrals for cardiac causes were common. However, our data reflect the disease course with guideline-based treatment and should not be interpreted as the natural history of the disease for these relatives. Nonetheless, we detected no SCD events and only one aborted SCD event unrelated to the inherited disease in the family. An ICD device was implanted in 16 relatives and three received appropriate ICD therapy indicating that SCD was likely prevented. New-onset left ventricular failure was uncommon and occurred in seven relatives. Interestingly, 15% of relatives were referred or hospitalized due to cardiac symptoms, but this endpoint also included relatives where a cardiac condition subsequently was excluded, suggesting possibly increased anxiety in this population of SCD relatives. Previous studies have also found low rates of cardiac events in SCD relatives^13,14 indicating a good prognosis. Therefore, it has been suggested to cease follow-up of SUDS relatives if their initial evaluation is normal. We found that only eight relatives (8 out of 695; 1%) developed a phenotype during follow-up and five of the eight relatives came from families where an inherited cardiac diagnosis had already been identified. Hence, our study underlines that very few relatives develop a clinical phenotype during follow-up and suggests that asymptomatic, unaffected relatives from families without established diagnoses may be excluded from follow-up under consideration of factors such as age at evaluation.

The study originates from a single centre, was retrospective in design, and not all data were available for all variables. Not all SCD victims were autopsied leading to lack of important information on the cause of death. Re-evaluation by an expert cardiac pathologist was not feasible in all SCD victims, which potentially limited the robustness of some diagnoses in borderline cases. Furthermore, several SCD victims in our cohort did not have blood/suitable tissue for DNA extraction hampering the possibility of genetic testing. In terms of potential bias, when comparing with the national autopsy-rate in SCD victims,² this study reported a rather high autopsy-rate (72%), which could indicate a degree of referral bias. Lastly, we cannot rule out that some relevant probands were not referred for family evaluation.

Conclusion

In this large-scale study of all categories of consecutive SCD families, we found a diagnostic yield of inherited cardiac disease in 47% of SCD victims. Evaluation of family members resulted in a 22% increase in diagnostic yield. A definite diagnosis was established in 73 relatives (11%) enabling prevention. During 5.5 years of follow-up, we found a low rate of serious cardiovascular events in early treated and in unaffected relatives indicating a good overall prognosis in SCD relatives. Our findings emphasize the importance of structured cardiac work-up in SCD families to identify and offer early treatment of affected relatives. Finally, our findings indicate that follow-up of unaffected relatives might be limited to selected cases.

Funding

This work was supported by the Rigshospitalets Research Foundation, Copenhagen University Hospital, and the A.P. Moeller Foundation.

Conflict of interest: none declared.

References

Author notes