Familial Sudden Death Is an Important Risk Factor for Primary Ventricular Fibrillation
A Case-Control Study in Acute Myocardial Infarction Patients
Background— Primary ventricular fibrillation (VF) accounts for the majority of deaths during the acute phase of myocardial infarction. Identification of patients at risk for primary VF remains very poor.
Methods and Results— We performed a case-control study in patients with a first ST-elevation myocardial infarction (STEMI) to identify independent risk factors for primary VF. A total of 330 primary VF survivors (cases) and 372 controls were included; patients with earlier infarcts or signs of structural heart disease were excluded. Baseline characteristics, including age, gender, drug use, and ECG parameters registered well before the index infarction, as well as medical history, were not different. Infarct size and location, culprit coronary artery, and presence of multivessel disease were similar between groups. Analysis of ECGs performed at hospital admission for the index STEMI revealed that cumulative ST deviation was significantly higher among cases (OR per 10-mm ST deviation 1.59, 95% CI 1.25 to 2.02). Analysis of medical histories among parents and siblings showed that the prevalence of cardiovascular disease was similar between cases and controls (73.1% and 73.0%, respectively); however, familial sudden death occurred significantly more frequently among cases than controls (43.1% and 25.1%, respectively; OR 2.72, 95% CI 1.84 to 4.03).
Conclusions— In a population of STEMI patients, the risk of primary VF is determined by cumulative ST deviation and family history of sudden death.
Received December 7, 2005; de novo received February 27, 2006; revision received June 1, 2006; accepted June 16, 2006.
Sudden cardiac death (SCD) remains one of the most prevalent modes of death in industrialized countries. Sudden death in the adult population most often occurs in the setting of coronary artery disease.1–3 In more than half of such cases, coronary heart disease has not previously been recognized clinically, and SCD occurs as its first symptom.3 In a large fraction of cases, SCD is caused by primary ventricular fibrillation (VF), ie, VF in the setting of acute ischemia and in the absence of heart failure.1 The incidence of primary VF is reported to be stable or declining during recent decades.4,5 In any case, mortality and morbidity due to primary VF remain overwhelming.
Editorial p 1134
Clinical Perspective p 1145
Despite many attempts, the majority of earlier studies failed at identifying risk factors for primary VF. Conventional cardiovascular risk factors are not predictive of the suddenness of coronary death, because the cardiovascular risk profile of middle-aged sudden death victims completely concurs with that of coronary artery disease patients without VF.6 A large ischemic area is not associated with an increased probability of VF.7 Site of infarction has been reported as a potential risk factor, because 2 relatively small studies showed that anterior myocardial wall infarction was more prevalent among primary VF patients than acute myocardial infarction patients without VF.7,8
Two important epidemiological studies introduced the concept of a genetic risk factor for SCD. In a population-based case-control study, a family history of myocardial infarction or SCD was, after correction for all common risk factors, positively associated with the risk of SCD.9 A major limitation of this retrospective study is that the control group consisted of healthy volunteers identified from the community and not from patients who had coronary artery disease. In the Paris Prospective Study, selectively performed in men, among whom 118 cases of sudden death occurred, parental sudden death has been found to be an independent risk factor for sudden death.10 In the latter study, a parental history of sudden death increased the relative risk of SCD among French men to 1.8 after correction for conventional cardiovascular risk factors; SCD of both parents increased the relative risk of SCD to 9.4.
Although SCD in the adult population most often occurs in the setting of coronary artery disease, VF remains an undifferentiated consequence of many arrhythmogenic mechanisms combined with various external triggers, such as ischemia, reperfusion, autonomic state, or electrolyte disturbances; a final common pathway for all malignant arrhythmias is highly unlikely.11 In other words, there is a broad variety of SCD phenotypes. Until now, the 2 earlier epidemiological studies on the genetic predisposition of SCD included all SCD mechanisms.
In the present case-control study in patients with an acute, first ST-elevation myocardial infarction (STEMI), before reperfusion therapy, we aimed at selectively including primary VF victims in an attempt to further analyze risk factors of this specific SCD phenotype. We demonstrate that in this selected patient group, the risk of primary VF is determined by cumulative ST deviation and sudden death among parents and siblings.
This study was designed as a multicenter, case-control study in acute STEMI patients who underwent a percutaneous coronary intervention (PCI); only patients of white ethnicity were included. Inclusion took place between April 2001 and August 2005 in 6 heart centers in the Netherlands; procedures are in accordance with the ethical standards of the national ethics committee on human research and with the Helsinki Declaration of 1975, as revised in 1983.
Cases were defined as survivors of VF, which had to be registered, that occurred within the first 12 hours of an acute and first STEMI. Patients with VF during or after PCI were not eligible. Control patients were matched across centers and on a group level with respect to age, gender, and infarct size. Although the ratio of case and control patients differed between centers, cases from the different centers and controls from different centers were similar with respect to baseline and infarct characteristics and family history. Because early reperfusion limits the opportunity of developing VF, potential control patients undergoing PCI within 2 hours after onset of complaints were not included. This time interval was based on the observation that >90% of cases developed VF within 2 hours after onset of complaints. Signed informed consent was available for all patients included in the present study.
Exclusion criteria for both patient groups were an actual non-STEMI, prior myocardial infarction, and age <18 or >80 years. Patients with congenital heart defects, known structural heart disease, or severe comorbidity, electrolyte disturbances, trauma, surgery, or CABG within 4 weeks were not selected. Class I and III antiarrhythmic drug use constituted an exclusion criterion.
Data on the family health history were collected verbally by trained interviewers after PCI. To minimize recall bias, case and control patients had to complete a questionnaire mailed to them 8 weeks after initial inclusion; the investigator collecting these data was unaware whether the patient was a control or case. In case of conflicts between verbal and written information, patients were contacted for clarification; if necessary, medical files were explored. Family history included data on the prevalence among parents and siblings of coronary artery disease and SCD, defined as acute death within 1 hour of symptoms in persons younger than 80 years.1 We also scored successful resuscitation, or aborted sudden death, among parents and siblings.
Data collection also covered age, gender, cardiovascular risk factors, body mass index (BMI), drugs, general and cardiac medical history in the preceding 12 months, interval between onset of complaints and PCI, an earlier episode of angina in the 24 hours preceding the infarction, culprit artery and number of diseased coronary arteries, and infarct size (based on maximal creatine kinase-MB rise and its time to peak). From the ECG produced at the moment of hospital admission for the index STEMI, cumulative ST deviation was calculated. Hypertension, hypercholesterolemia, and diabetes mellitus were scored on the basis of previous diagnosis and initiation of therapy.
Every patient was asked for ECGs made well before the index event. Maximal effort was taken to collect such ECGs, from which heart rate, electrical axis, PQ interval, QRS width, and corrected interval (QTc) were measured.
Differences in continuous clinical variables, infarct characteristics, and family characteristics between case and control patients were tested with an independent t test (with or without correction for unequal variances) when data were normally distributed or a Mann-Whitney U test otherwise. Values are presented as mean±SD or median and first and third quartile (Q1, Q3), respectively. Differences in categorical clinical, infarct, and family characteristics were compared with a Fisher exact test, and values are presented as percentages. Owing to the fact that families with SCD victims (from both cases and controls) were marginally larger, (univariate and multiple) conditional logistic regression was used to model the association between the occurrence of VF and family history of SCD and other potential risk factors (conditioning variable: family size). To minimize the effect of sample size reduction due to missing ST-deviation information, 2 multivariate models were used that differed only with respect to inclusion of ST deviation as a predictor. Statistical analysis was performed with SAS software (version 9; Cary, NC), and probability values <0.05 were considered to be statistically significant.
The authors had full access to the data and take full responsibility for its integrity. All authors have read and agree to the manuscript as written.
Characteristics of both patients groups are summarized in Table 1. In total, 330 cases and 372 controls were included. Groups were not different with respect to age, gender, general medical history, and drug use. Cardiovascular risk profiles were similar except for hypercholesterolemia and BMI. Controls and cases were different with respect to cardiac history, because 5.6% and 10.0%, respectively, had consulted a cardiologist in the last 12 months before the index event. The most prevalent complaint was chest pain (patients with previous myocardial infarction or structural heart disease were not eligible). Surprisingly, there were 7 versus 0 patients with atrial fibrillation during the 12 months preceding the index event among cases and controls, respectively. In a small subset of cases and controls, historical ECGs made under baseline conditions could be recollected. Such baseline ECGs were not different with respect to heart rate, axis, PQ interval, QRS width, and QTc.
Owing to diverse clinical protocols within the participating centers, comparable data sets could not be collected in all patients; exact numbers are shown in Table 2. Anginal symptoms within 24 hours preceding the STEMI occurred in equal numbers in cases and controls.
ECGs made at hospital admission and before PCI could be gathered from 193 cases and 211 controls. Maximal cumulative ST deviation was significantly higher in case patients (Table 2). Because the high-energy electrocardioversion necessary for termination of VF is known to induce temporary ST changes, we also examined a subgroup of case patients from whom ECGs taken during admission and before electrocardioversion were available (n=75). In this group, too, ST deviation remained an independent risk factor for VF, with an adjusted OR of 1.49 (95% CI 1.10 to 2.02) per 10-mm increase in ST deviation. In the case group, there was a nonsignificant trend toward more anterior wall acute myocardial infarctions. Multivessel disease was equally present in cases and controls. The median time interval between onset of complaints and PCI was 159 minutes (Q1 to Q3 120 to 219 minutes) among cases and 185 minutes (Q1 to Q3 150 to 270 minutes) among controls (P<0.001); this is a consequence of inclusion criteria, because in the case group, patients were only included after at least 120 minutes of ischemic time. This time difference did not affect infarct size, for which cases and controls were matched.
Recall bias with respect to family history could have been present. A written questionnaire was sent to patients 8 weeks after the initial verbal data acquisition. In 16 patients, conflicts appeared that could be resolved later. In Table 3, family histories among parents and siblings are summarized. In some instances, loss of contact with family members prohibited data retrieval; these data were scored as missing. The median number of parents and siblings was similar in cases and controls. Prevalence of cardiovascular diseases among parents and siblings was not different between cases and controls. The percentage of cases with 1 or more cases of sudden death among parents and siblings was significantly higher than among controls (43.1% versus 25.1%, respectively). Likewise, aborted sudden death among parents and siblings occurred in 3.8% of cases and 1.5% of controls.
Table 4 summarizes the results of univariate and multivariate analysis for potential primary VF risk factors. SCD among parents and siblings significantly increased the risk of primary VF, with an adjusted OR of 2.72. Notably, the OR for family history of SCD remained significant when ST deviation was included in the second multivariate model (3.30; 95% CI 1.91 to 5.68).
For each 10-mm increase in ST deviation, the odds ratio of VF was 1.59. Thus, predicted ORs (95% CIs) for patients with 20-, 30-, or 40-mm ST deviation compared with patients with 10-mm ST deviation were 1.59 (95% CI 1.25 to 2.02), 2.51 (95% CI 1.55 to 4.07) and 3.99 (95% CI 1.93 to 8.22), respectively. In addition, hypercholesterolemia showed a significant OR for VF. Earlier consultation of a cardiologist and BMI were of borderline significance after multivariate analysis.
To correct for potential confounding factors, center was included in the multivariate models. As a consequence of the considerable difference in the ratio of included numbers of cases and controls across centers, overall probability values and the ORs for the individual centers (data not shown) reached statistical significance.
The most intriguing finding of this case-control study among patients with a first acute, transmural myocardial infarction is that SCD among parents and siblings is such a strong predictor of primary VF. We defined strict selection criteria that aimed at identifying cases with a unique SCD phenotype, ie, primary VF in the setting of a first, transmural myocardial infarction, before reperfusion therapy. In the 2 earlier, pioneering studies by Friedlander et al9 and Jouven et al,10 cases comprised all pathophysiological variants of SCD. To optimally analyze whether an additional, or profibrillatory, risk factor is present during acute myocardial infarction, we selected control patients among first-STEMI patients. This is in contrast to the study by Friedlander et al,9 in which controls were selected among healthy people from the community; in that study, a family history of either myocardial infarction or sudden death was positively associated with the risk of sudden death.9 In the study by Jouven et al,10 combined maternal and paternal sudden death increased the relative risk for SCD by 9.4. Because of the low numbers of such patients in the present study (both parents had died suddenly in only 4 patients’ families), this observation could not be reproduced.
A genetic predisposition implies a higher recurrence rate of VF in the setting of an acute myocardial infarction among survivors of primary VF. This has already been demonstrated by early studies from the 1970s, an era in which preventive regimens were not available.12 Likewise, in the more recent Antiarrhythmics Versus Implantable Defibrillators (AVID) registry, the prognosis of patients surviving VF due to a transient cause, which mainly consisted of ischemia, was extremely poor.13 An earlier study by Schwartz et al14 suggested that the poor prognosis of VF survivors was confined to patients with an anterior-wall STEMI, because their prognosis was much worse than that of VF survivors with an inferior-wall STEMI. Even within a single SCD phenotype, as designed in the present study, the range of underlying pathophysiological mechanisms and candidate genes remains sizeable. Multiple pathways, such as cardiac ion channels or gap junctions, but also extracardiac mechanisms, such as the autonomic nervous system, could all be involved.15
The second independent risk factor for primary VF identified in the present study is cumulative ST deviation. After multiple regression analysis, this appeared to be independent of time to PTCA, which was somewhat longer in the control group (as a consequence of inclusion criteria), and of infarct size. Infarct sizes estimated by peak creatine kinase-MB levels and time to peak were not different between groups. Whether this increased ST deviation is a hallmark of an electrophysiological mechanism underlying the predilection to VF, such as increased heterogeneity in repolarization, or increased injury current remains to be elucidated. It appears unlikely that electrocardioversion that induced temporary ST deviation acted as an important confounder, because ST deviation was still an independent risk factor among cases with ECG registrations before electrocardioversion.
Angina during the 24 hours preceding the acute infarction, which could be a hallmark of ischemic preconditioning, was not different, which suggests that this protective mechanism does not play a role in this patient group. Infarct location and culprit coronary artery were not identified as independent, significant risk factors. Recently, Gheeraert et al7 showed that an occlusion of the left coronary artery (left anterior descending coronary artery and right circumflex artery) was associated with a greater risk of VF than an occlusion of the right coronary artery. Probably, differences in inclusion criteria and smaller sample size in the latter study explain these differences. However, other observations concur, because both studies showed that infarct size and the extent of coronary artery disease are not associated with VF risk.
Patients with earlier cardiac consultation in the 12 months preceding the present study showed a trend toward an increased risk of primary VF. In 7 cases, atrial fibrillation was diagnosed in the 12 months preceding the index event compared with 0 in the control group. Earlier infarcts or any other signs of structural heart disease prohibited inclusion in the present study; whether the occurrence of atrial fibrillation among cases is another sign of predisposition to arrhythmias remains purely speculative.
In the present study in infarct patients, there obviously was a high prevalence of cardiovascular risk factors. We do not have a clear explanation for the observation that hypercholesterolemia was significantly less common (and BMI marginally significantly lower) in cases than in controls. Whether this represents a true “protective effect” remains highly questionable, because, to the best of our knowledge, a direct effect of BMI and hypercholesterolemia on arrhythmogenesis has not been described.
Recall bias with respect to family history could have been introduced among cases; therefore, a written questionnaire was sent out 8 weeks after the initial verbal data acquisition. In 16 patients, conflicts appeared that could be resolved later. A major restriction of the present study was the selection of VF survivors who reflected &10% of the total primary VF population; 90% of SCD victims do not reach the hospital alive.16 In animal experiments, VF in acute myocardial ischemia clearly occurs in peaks in time (phase 1a, 1b, and 2). If we assume that such peaks occur in humans, it seems likely we largely missed the early peaks (phase 1a and 1b), because adequate medical attendance is more likely to be present in patients with a delay between onset of chest pain and cardiac arrest. However, it is not clear whether such peaks occur in humans; in the present study population, temporal peaks could not be discriminated.
There are missing data for the parameters summarized in Table 2. Owing to differences in clinical protocols for primary PTCA patients used in the various participating centers, not all data were retrievable. Also, in many instances, ECGs made just before acute PCI were not archived. Close inspection of the data supported the presumption that the missingness mechanism was likely “missing at random.” Another potential limitation is the fact that the ratio of cases and controls across centers was different. Nevertheless, case and control patients were similar between centers with respect to all relevant parameters. The introduction of center into the multivariate models obviously resulted in a probability value lower than 0.05 for center; however, there was no effect on the other risk factors. Therefore, we consider it very unlikely that the difference in ratio of cases and controls across centers constitutes an important confounding factor.
The present observation that familial sudden death is such a strong predictor for primary VF constitutes another argument for further research into its genetic basis.9,10 Although progress in genomics is rapid, researchers from the National Human Genome Research Institute recently discussed that genetic analysis of common, multifactorial, and probably multigenetic diseases, such as primary VF, is so complex that we need many years of research before we can completely understand the underlying genetic background. Meanwhile, those researchers advocate the use of family history as a clinically valuable tool for individualized risk stratification.17 We conclude that at present, family history, which acts as an integrator of complex interactions of multiple genes and environmental factors, is a key predictor of primary VF in acute myocardial infarction patients.
Sources of Funding
The Netherlands Heart Foundation (2001-D019) and The Leducq Foundation (grant 05-CVD) are gratefully acknowledged for financial support.
Zipes DP, Wellens HJJ. Sudden cardiac death. Circulation. 1998; 98: 2334–2351.
Thompson CA, Yarzebski J, Goldberg RJ, Lessard D, Gore JM, Dalen JE. Changes over time in the incidence and case-fatality rates of primary ventricular fibrillation complicating acute myocardial infarction: perspectives from the Worcester Heart Attack Study. Am Heart J. 2000; 139: 1014–1021.
Friedlander Y, Siscovick DS, Weinmann S, Austin MA, Psaty BM, Lemaitre RN, Arbogast P, Raghunathan TE, Cobb LA. Family history as a risk factor for primary cardiac arrest. Circulation. 1998; 97: 155–160.
Jouven X, Desnos M, Guerot C, Ducimetière P. Predicting sudden death in the population: the Paris Prospective Study I. Circulation. 1999; 99: 1978–1983.
Janse MJ, Wit AL. Electrophysiological mechanisms of ventricular arrhythmias resulting from myocardial ischemia and infarction. Physiol Rev. 1989; 69: 1049–1169.
Wyse DG, Friedman PL, Brodsky MA, Beckman KJ, Carlson MD, Curtis AB, Hallstrom AP, Raitt MH, Wilkoff BL, Greene HL; AVID Investigators. Life-threatening ventricular arrhythmias due to transient or correctable causes: high risk for death in follow-up. J Am Coll Cardiol. 2001; 38: 1718–1724.
Schwartz PJ, Vanoli E, Stramba-Badiale M, De Ferrari GM, Billman GE, Foreman RD. Autonomic mechanisms and sudden death: new insights from analysis of baroreceptor reflexes in conscious dogs with and without myocardial infarction. Circulation. 1988; 78: 969–979.
Sudden cardiac death remains one of the most prevalent but difficult to predict modes of death. Ventricular fibrillation (VF) that occurs in the setting of acute myocardial ischemia and in the absence of heart failure is assumed to underlie most cases of sudden cardiac death. To identify independent risk factors for VF during myocardial ischemia, a case-control study was performed among patients with a first ST-elevation myocardial infarction (STEMI) and without preexisting structural heart disease. A total of 330 primary VF survivors (cases) and 372 controls were included. Baseline characteristics, including cardiovascular risk factors and ECG parameters before the index infarction, were not different. Cumulative ST-segment elevation was significantly greater before VF (cases) than in controls, whereas infarct size was similar between groups. Infarct location, culprit coronary artery, and presence of multivessel disease were similar between groups. Interestingly, a history of sudden death among parents and siblings was significantly more frequent among cases than controls, whereas familial cardiovascular disease was equally distributed. Thus, when myocardial ischemia occurs, whether VF follows is determined in part by inherited factors. Further research into these factors may suggest new preventive strategies for reducing the occurrence of sudden death.