OUP user menu

★ fast trackEditor's choice ★

Risk of cardiovascular disease in family members of young sudden cardiac death victims

Mattis Flyvholm Ranthe, Bo Gregers Winkel, Elisabeth Wreford Andersen, Bjarke Risgaard, Jan Wohlfahrt, Henning Bundgaard, Stig Haunsø, Mads Melbye, Jacob Tfelt-Hansen, Heather A. Boyd
DOI: http://dx.doi.org/10.1093/eurheartj/ehs350 503-511 First published online: 14 November 2012


Aims Descriptive and genetic studies suggest that relatives of sudden cardiac death (SCD) victims have an increased risk of several cardiovascular diseases (CVDs). Given the severe consequences of undiagnosed CVD and the availability of effective treatment, the potential for prevention in this group is enormous if they do have an increased CVD risk. This nationwide prospective population-based cohort study described the risk of CVDs in relatives of young SCD victims, compared with the general population.

Methods and Results All SCD victims aged 1–35 years in Denmark, 2000–2006, were identified (n = 470), along with their first- and second-degree relatives (n = 3073). We compared the incidence of CVD in those relatives with that in the background population using standardized incidence ratios (SIRs). The observed number of CVDs over 11 years of follow-up was 292, compared with 219 expected based on national rates [SIR 1.33, 95% confidence interval (CI) 1.19–1.50]. Risks varied significantly with age; the SIR for those <35 years was 3.53 (95% CI 2.65–4.69), compared with SIRs of 1.59 (95% CI 1.35–1.89) and 0.91 (95% CI 0.75–1.10) for those aged 35–60 years or >60 years, respectively (Phomogeneity < 0.0001). For first-degree relatives <35 years, SIRs for ischaemic heart disease, cardiomyopathy, and ventricular arrhythmia were 5.99 (95% CI 1.95–0.13.98), 17.91 (95% CI 4.88–45.87), and 19.15 (95% CI 7.70–39.45), respectively.

Conclusion CVDs co-aggregated significantly with SCD in families, with young first-degree relatives at greatest risk. Results clearly indicate that family members of young SCD victims should be offered comprehensive and systematic screening, with focus on the youngest relatives.

  • Sudden cardiac death
  • Cascade screening
  • Epidemiology
  • Cardiomyopathy
  • Ventricular arrhythmia
  • Ischaemic heart disease


Sudden cardiac death (SCD) is defined as sudden, unexpected death due to natural unknown or cardiac causes, with an acute change in cardiovascular status within 1 h of death or, in un-witnessed cases, in a person last seen functioning normally <24 h before being found dead.1,2 Sudden cardiac death in the young is often caused by cardiovascular conditions thought to be hereditary.38 An increasing number of gene mutations are known to be associated with cardiac conditions that are important causes of SCD, such as the primary arrhythmia syndromes,9,10 the cardiomyopathies,11,12 familial hypercholesterolaemia, and premature ischaemic heart disease (IHD).13,14 Up to 50% of victims of sudden arrhythmic death syndrome (SADS, SCDs remaining unexplained after standard autopsy) carry genetic mutations associated with cardiac diseases.15,16 These are predominantly mutations leading to ion-channel defects associated with lethal arrhythmias, but also mutations associated with cardiomyopathies or conduction disorders.15,17,18 Such ‘molecular autopsy’ findings imply an increased risk of cardiovascular disease (CVD) in relatives of SCD victims. These individuals are often both young and apparently CVD-free. If they are in fact at greater risk of CVD than the general population, screening and identification of affected individuals and application of widely available CVD treatments and preventive therapies would lead to significant prevention of morbidity and mortality. However, current knowledge on CVD risk in family members of SCD victims is limited and based on reports in small, highly selected case groups without control groups for comparison. Descriptive studies, primarily from tertiary referral centres, have reported that a variety of CVDs are present in families of young SCD victims,1921 and one study identified CVD in half of the family members examined.22 We tested the conclusions of previous descriptive studies in a controlled epidemiologic study to determine whether a case can be made for large-scale screening of family members of young SCD victims. We examined CVD occurrence in a nationwide cohort of persons related to young SCD victims and followed prospectively for CVD, and compared their experience with the occurrence of CVDs in the background population.


Our study cohorts were constructed based on information from a purpose-built database of deaths in young Danes and Danish national health and population registers, with follow-up also register-based. We compared the observed numbers of CVDs in cohort members who had a familial SCD, with expected numbers calculated based on the background population, using standardized incidence ratios (SIRs). To put these findings in perspective, we also examined the observed and expected number of neurological and endocrine conditions in the same cohort. In addition, we conducted the same analyses in a control cohort of persons related to young victims of death from non-cardiac causes. The study was conducted and reported in accordance with the STROBE guidelines for cohort studies.23

Data sources

Danish Civil Registration System and Danish Family Relations Database

Since 1 April 1968, the Danish Civil Registration System (CRS)24 has assigned a unique personal identification number (PIN) to each Danish resident, and regularly updates vital status and other personal information. The PIN is used in all of Denmark's population-based registers, which permits accurate linkage of individual-level information from various sources and minimizes loss to follow-up. Parent–child links registered in the CRS form the basis for the Danish Family Relations Database (DFRD). The DFRD is a unique resource containing pedigree information allowing for the identification of family members of individuals residing in Denmark. For most individuals born in 1950 or later, first-degree relatives (parents, children, and siblings) can be identified; in 90% of individuals born after 1984, second-degree relatives (half-siblings, grandparents, grandchildren, aunts/uncles, and nieces/nephews) can also be identified.

Danish National Patient Register

The Danish National Patient Register (NPR) contains information from all Danish hospitals on inpatient diagnoses assigned since 1 January 1978, as well as outpatient diagnoses from 1995 onwards. Diagnoses are registered using International Classification of Disease (ICD) codes, with ICD-8 codes used from 1977 to 1993, and ICD-10 codes used since 1994.25 We used the NPR to identify incident CVDs (ICD-10 codes for CVD overall: I00-I51; IHD: I20-I25; cardiomyopathy: I42-I43; ventricular arrhythmia: I46.0 (cardiac arrest), I47.0 (re-entry ventricular arrhythmia), I47.2 (ventricular tachycardia), and I49.0 (ventricular fibrillation and flutter)), as well as neurological (G00-G99) and endocrine conditions (E00-E90), in the relatives of SCD victims.

Database of deaths in Danes aged 1–35 years, 2000–2006

A review of all deaths in the period 2000–2006 in Danes aged 1–35 years examined 6231 deaths.8 Deaths at late ages were not reviewed. Mode and cause of death were abstracted from death scene investigations and autopsy reports. Among the 6231 deaths, there were 470 SCDs and 5707 deaths that were non-natural or natural but not sudden (NSDs). As previously reported,8 in 314 (67%) of the SCDs, an autopsy had been performed; in 178 of these, a cardiac cause of death was confirmed (autopsy-verified SCD), while 136 cases remained unexplained after autopsy and fulfilled the SADS criteria (SADS).3,22 In the 156 non-autopsied SCDs, the case for SCD was made based on death certificates and forensic investigations (non-autopsied SCD).

Study cohorts

We included all identifiable first- and second-degree relatives of the young SCD and NSD victims in our study cohorts: the SCD cohort (those related to an SCD victim) and, for sensitivity analyses, the NSD cohort (those related to an NSD victim). There was no overlap between the study cohorts.


Using the NPR, we identified incident CVD and, for sensitivity analyses, nervous system and endocrine diseases in the two study cohorts. All cohort members were followed from the date of first SCD or NSD in a relative until the first of the following: (i) CVD (or nervous system disease or endocrine disease, for the sensitivity analyses); (ii) death; (iii) emigration; (iv) labelled ‘missing’ in the CRS; or (v) 31 December 2011.

Statistical analysis

The risks of various outcomes in relatives of SCD and NSD victims was evaluated using SIRs, defined as the ratio between the observed number of, for example, CVD diagnoses in, for example, the SCD cohort and the expected number based on national rates in the general population. Expected numbers were determined using standardized rates for the Danish population based on NPR information, calculated for current age group (5-year categories), calendar period (1-year categories), and number of first- (categories: 0, 1–2, 3–4, ≥5) and second-degree relatives (categories: 0, 1–2, 3–4, 5–10, ≥11). Confidence intervals and tests of homogeneity were constructed using standard Poisson regression with log offset. However, when the expected number of events was <5, we used exact confidence intervals and exact Poisson regression based on the conditional likelihood. SIRs were compared using likelihood ratio tests. Analyses were carried out in SAS 9.2 (SAS Institute, Inc., Cary, NC, USA) and STATA 11 (Stata Corporation, College Station, TX, USA).


We could identify one or more relatives for 98.5% (463 of 470) of the SCD victims and 98.4% (5617 of 5707) of the NSD victims. There were 1629 young (<35 years of age at the start of follow-up) SCD cohort members and 3073 SCD cohort members overall. We identified 40 450 relatives for the NSD victims. Table 1 provides descriptive details for the SCD and NSD victims with identifiable relatives, and for the SCD and NSD cohorts, including details on incident outcomes. In the main part of the Results section, we present results from the SCD cohort on risk of CVD. Risk of CVD in the NSD cohort as well as risks of endocrine and nervous system diseases in both the SCD and NSD cohorts are reported in the Sensitivity analysis section, with full details given in Supplemental material online.

View this table:
Table 1

Descriptive details for victims of sudden cardiac death and victims of non-sudden deaths (upper part) and their relatives (lower part)

We followed the 3073 SCD cohort members for 22 281 person-years and observed 292 incident CVDs. The number of CVDs expected based on the rates in the background population was 219. The SIR for any CVD given an SCD in a relative was 1.33 (95% CI 1.19–1.50). The risk of IHD was comparable with the overall estimates, while risks for cardiomyopathy and ventricular arrhythmia were 2–3 times higher (Table 2).

View this table:
Table 2

Risk by age of cohort member

Effects of age and gender

Although CVD risk was significantly increased in the SCD cohort as a whole, compared with the background population, risks of both CVD overall and CVD subtypes were significantly higher among younger SCD cohort members. Among the 1629 SCD cohort members aged <35 years, the risk of any CVD was increased >3-fold, this was significantly higher than the SIR for any CVD in the older part of the SCD cohort (P-value for homogeneity <0.0001) (Figure 1). In SCD cohort members <35 years of age, the risks for subgroups of CVD were elevated 6-fold or more (Table 2). For cardiomyopathy and ventricular arrhythmia, the rates in those aged 35–60 years were also higher than expected, with risks significantly elevated by 3- to 4-fold. In contrast, for SCD cohort members aged >60 years, the observed and expected number of cardiomyopathies and ventricular arrhythmias were comparable. The rates of IHD in all SCD cohort members aged 35 years or above were comparable with the background rates.

Figure 1

Risk of any cardiovascular disease in relatives of young sudden cardiac death victims. Standardized incidence ratios with 95% confidence intervals of any cardiovascular disease in 3073 first- and second-degree relatives of 463 victims of sudden cardiac death (aged 1–35 years, Denmark from 2000 to 2006). The cohort of relatives was followed from the time of death until cardiovascular disease to the end of 2011. Observed and expected numbers of any cardiovascular disease are given.

Analyses for any CVD and CVD subtypes stratified by gender showed no significant differences in SIRs by gender of SCD cohort member, SCD victim, or both (data not shown). Finally, when we examined whether the effect of age varied by gender, in analyses stratifying by gender of the SCD cohort member, SCD victim, or both, we saw the same pattern as described above for both men and women, with the highest risks in young relatives and the lowest risks in the oldest relatives (data not shown).

Degree of kinship

The SIR for any CVD given an SCD in a first-degree relative was higher than the SIR given an SCD in a second-degree relative [1.49 (95% CI 1.29–1.72) vs. 1.11 (95% CI 0.91–1.35), Pdifference = 0.016] (Table 3). Sudden cardiac death cohort members <35 years of age with an SCD in a first-degree relative had a 4-fold increase in risk of any CVD (Table 3), while those with an SCD in a second-degree relative had more than a 2-fold increased risk (Pdifference = 0.076). We found no significant difference in SIRs by degree of kinship for IHD and cardiomyopathy, either overall or in the youngest cohort members (Table 3). In contrast, estimates for ventricular arrhythmia were significantly higher given SCDs in first-degree relatives (Table 3). Figure 2 shows the SIRs for young first-degree relatives (<35 years) of SCD victims for any CVD and CVD subtypes.

View this table:
Table 3

Risk by degree of relation

Figure 2

Risk of any cardiovascular disease and subtypes in young first degree relatives. Standardized incidence ratios with 95% confidence intervals of any cardiovascular disease and subtypes in sudden cardiac death cohort members aged <35 years at diagnosis, with a first-degree relative suffering sudden cardiac death. Observed and expected numbers of each condition are given.

Sudden cardiac death subtype

The three subtypes of SCD (autopsy-verified SCD, SADS, and non-autopsied SCD) were unevenly distributed among the SCD victims: 37.6% were not autopsied, 33.5% were autopsied, and 29% were SADS victims. Examining risk of CVD overall, we found a significant difference in SIRs by subtype of SCD, with the highest SIR in SCD cohort members with a relative whose SCD was verified by autopsy (Table 4). Although tests for differences in the effect of SCD subtype on risk of CVD subgroups were not significant due to low power, SIRs—particularly in cohort members <35 years of age—appeared to be dramatically increased given autopsy-verified SCD (Table 4).

View this table:
Table 4

Risk by subtypes of sudden cardiac death

Time since sudden cardiac death

The enhanced risk of CVD and CVD subtypes in the SCD cohort was highest during the first year after SCD in a relative (Table 5). In persons aged <35 years, risks in the first year were dramatically increased, with an 11-fold increase in risk of any CVD and as much as a 40-fold increase in risk of ventricular arrhythmias (Table 5).

View this table:
Table 5

Risk by time since sudden cardiac death in relative

Sensitivity analysis

We conducted two sensitivity analyses to test if (i) any early death (i.e. NSDs) in a relative would increase the risk of CVD and (ii) if risks of non-cardiac conditions (endocrine and neurological) would also be elevated in those related to a young victim of SCD/NSD. The risk of any CVD was significantly lower in the NSD cohort than in the SCD cohort (Pdifference < 0.0001) (see expanded results in Supplemental material online for details). In addition, the risks of neurological and endocrine diseases were generally comparable among persons related to an SCD victim and persons related to an NSD victim (see expanded results in Supplemental material online for details). However, in the first year after SCD in the SCD cohort, we did observe two cases of muscular dystrophy, where only 0.1 was expected, corresponding to an SIR for muscular dystrophy of 22.9 (2.77–82.7) in those related to an SCD victim.


In this nationwide cohort study, we followed a cohort of persons related to young (1–35 years) SCD victims for up to 11 years and compared the risk of CVD in this cohort with the risk in the background population. To our knowledge, population-based findings of this kind have not been reported before. Regardless of cohort member age, we found that among first-degree relatives of an SCD victim, the overall CVD risk rose 50%, compared with the background population. For cardiomyopathies and ventricular arrhythmias, risks increased by up to 400%, while risk of IHD was only modestly affected. In relatives aged <35 years (first- and second-degree combined), the risk was elevated 3-fold for any CVD, 6-fold for IHD, and more than 10-fold for cardiomyopathies and ventricular arrhythmias. For cardiomyopathies and ventricular arrhythmias, risks were elevated 3- to 4-fold in those aged 35–59 years, but not in those aged ≥60 years.

Our findings support that SCD (or the underlying CVDs) has a large hereditary component, i.e. relatives are also at risk for CVD.19,21,22 Ischaemic heart disease is generally thought to be polygenic, with significant interplay between genes and environment. The exception is IHD caused by familial hyperlipidaemia, which is typically a monogenic-dominant trait presenting at an early age. Only in the young (<35 years) did the risk of IHD in our cohort appear to be affected by familial SCDs, and interestingly deaths in both first- and second-degree relatives increased this risk 6-fold. The underlying genetics of both the cardiomyopathies and ventricular arrhythmias are thought to be monogenic, most often with dominant traits. Consistent with this, we found that the rates of cardiomyopathy and ventricular arrhythmia were higher than expected in all age groups among relatives of SCD victims, and we found increased SIRs for cardiomyopathy even in second-degree relatives.

Males were predominant both among SCD victims and their relatives, and the literature suggests that males may have a higher risk of SCD,8 as well as of, for example, IHD. However, we found no evidence of a gender effect (our SIRs did not differ according to the gender of either the SCD victim or the cohort member), and our findings do not support gender-specific differences in evaluation of the relatives of SCD victims.

Interestingly, there were two cases of muscular dystrophy in the SCD cohort. Muscular dystrophies are rare conditions in which severe cardiac involvement is often seen; in some instances, cardiac manifestations lead to the diagnosis. Most muscular dystrophies have genetic causes, and our observation of two such cases could suggest an increased risk of muscular dystrophy among the relatives of SCD victims, or vice versa.

A positive autopsy (especially with signs of cardiomyopathy or IHD) might lead to increased medical awareness among relatives and more intense screening for CVD. When we stratified by SCD subtype in the full cohort, we found significant differences in the risks of CVD overall, with the greatest risk associated with autopsy-verified SCD. However, when we looked only at younger cohort members, all subtypes of SCD were associated with increased risks of CVD, and the differences in strength of association were not significant. When we looked at CVD subtypes, only the risks of cardiomyopathy differed significantly by SCD subtype for the youngest cohort members, hinting that having a positive autopsy is important in prompting evaluation of relatives for cardiomyopathy.

It could be speculated that increased medical awareness conferred by a positive autopsy could have led to confounding by indication, resulting in inflated SIRs. There is certainly some increased medical awareness, as is evident from our finding that CVD risk was highest in the first year after an SCD in the family. In some regions of Denmark, focused screening of the relatives of SCD victims has been undertaken since 2006, which likely accounts for some of the increased risk observed in the first year. However, the fact that all risks remained elevated beyond the first year clearly indicates that there is a real excess of CVD in SCD families.

Inherited CVDs have increasingly gained attention over the past decades, with the continued discovery of underlying genetic mechanisms for a number of CVDs,14 the arrhythmic syndromes, and the cardiomyopathies in particular.15,26 Furthermore, the most important CVD entity, IHD, is also a condition that, especially in the young, often has a genetic background.27 This focus on inherited disease has led to mounting awareness in the cardiac community of the importance of systematic screening of high-risk families, and to some extent also of relatives of victims of cardiac death, for CVDs.28 SADS clinics and centres for inherited cardiac diseases have been established in many countries.29 However, these policy actions have not been based on large, well-controlled, population-based studies because until now, such studies did not exist. Our findings support these actions and lead to suggestions as to which age groups to include in such screening programmes and how to handle different types of SCDs. Our results show that screening for the cardiomyopathies and ventricular arrhythmia might benefit those up to age 60 years, whereas when familial IHD is suspected as the cause of SCD, the youngest relatives are likely to benefit most from screening. Furthermore, in general, the strongest case is made for screening first-degree relatives, although for IHD and the cardiomyopathies, screening of young second-degree relatives may also be warranted. Whether SCD is autopsy confirmed or not should probably not dictate the level of medical attention paid to the victim's relatives.

We lacked information on behavioural risk factors for CVD, as the Danish registers do not contain information on, for example, body mass index, smoking, and alcohol use for the entire population, if at all. However, these factors contribute in a major way only to risk of IHD and some subtypes of CM, whereas our findings were consistent across all CVD types. Furthermore, these risk factors are thought to play only a minor role in younger persons, and as our most dramatic results are based on cohort members <35 years of age, we believe that this potential limitation had only a limited impact, if any, on our findings. We had no detailed information on the underlying causes of the ventricular arrhythmia, but in the young the most likely aetiologies include channelopathies, cardiomyopathies, primary ventricular fibrillation (VF) with infarction, and idiopathic VF.

Major strengths of this study include the number of ‘affected’ families we were able to identify, a lack of selection bias due to the review of all relevant death certificates in a given period, and the use of registers to identify and follow-up the study cohorts.

Furthermore, the sensitivity analyses have strengthened the study in two important ways: first, the finding of an elevated CVD risk among relatives of SCD victims but not among relatives of NSD victims suggests that the elevated risk of CVD in SCD families is not due to general characteristics or behaviour in families in response to the tragic loss of a young person, but rather is related to the specific type of early death and its underlying causes. Secondly, the fact that the risks of neurological and endocrine conditions were generally comparable in the SCD and NSD cohorts suggests that the elevated CVD risk in SCD families does not just reflect increased medical attention following a devastating event but a real association reflecting familial clustering of SCD and CVD.


In this large population-based cohort study with up to 11 years of follow-up per person, we showed significant increases in risk of CVD, the cardiomyopathies, and ventricular arrhythmia in particular, in persons related to SCD victims. Risk increases were even more dramatic for first-degree relatives and persons <35 years of age. Since the cardiovascular conditions on which we focused are treatable, early identification of at-risk persons is potentially a life-saving action. Our findings are the first of their kind and support the initiation of cascade screening in families experiencing a SCD, with customization of screening based on the underlying condition suspected to have caused the SCD and family member ages.


This work was supported by grants from the Danish Heart Foundation (10-04-R78-A2799-22615 and the Lundbeck Foundation (287/06).

Conflict of interest: none declared.


View Abstract