Author + information
- Received April 26, 2010
- Revision received August 30, 2010
- Accepted September 27, 2010
- Published online February 15, 2011.
- Melissa H. Kong, MD⁎,†,⁎ (, )
- Gregg C. Fonarow, MD‡,
- Eric D. Peterson, MD, MPH⁎,†,
- Anne B. Curtis, MD§,
- Adrian F. Hernandez, MD⁎,†,
- Gillian D. Sanders, PhD⁎,
- Kevin L. Thomas, MD⁎,†,
- David L. Hayes, MD∥ and
- Sana M. Al-Khatib, MD, MHS⁎,†
- ↵⁎Reprint requests and correspondence:
Dr. Melissa H. Kong, Duke University Medical Center, DUMC Box 31294, Durham, North Carolina 27710
The need for consistent and current data describing the true incidence of sudden cardiac arrest (SCA) and/or sudden cardiac death (SCD) was highlighted during the most recent Sudden Cardiac Arrest Thought Leadership Alliance's (SCATLA) Think Tank meeting of national experts with broad representation of key stakeholders, including thought leaders and representatives from the American College of Cardiology, American Heart Association, and the Heart Rhythm Society. As such, to evaluate the true magnitude of this public health problem, we performed a systematic literature search in MEDLINE using the MeSH headings, “death, sudden” OR the terms “sudden cardiac death” OR “sudden cardiac arrest” OR “cardiac arrest” OR “cardiac death” OR “sudden death” OR “arrhythmic death.” Study selection criteria included peer-reviewed publications of primary data used to estimate SCD incidence in the U.S. We used Web of Science's Cited Reference Search to evaluate the impact of each primary estimate on the medical literature by determining the number of times each “primary source” has been cited. The estimated U.S. annual incidence of SCD varied widely from 180,000 to >450,000 among 6 included studies. These different estimates were in part due to different data sources (with data age ranging from 1980 to 2007), definitions of SCD, case ascertainment criteria, methods of estimation/extrapolation, and sources of case ascertainment. The true incidence of SCA and/or SCD in the U.S. remains unclear, with a wide range in the available estimates that are badly dated. As reliable estimates of SCD incidence are important for improving risk stratification and prevention, future efforts are clearly needed to establish uniform definitions of SCA and SCD and then to prospectively and precisely capture cases of SCA and SCD in the overall U.S. population.
In the U.S., cardiovascular disease was the underlying cause of 1 of every 2.9 deaths occurring in 2006 (1). The proportion of these deaths that is sudden has been estimated to be as high as 50%, making sudden cardiac death (SCD) the most common cause of death in this country (2). Studies are available that provide consistent data on survival after sudden cardiac arrest (SCA) (3,4), and several randomized controlled trials offer reliable information on therapies for the prevention of SCD in highly selected, high-risk subgroups (5–11). Nevertheless, data describing the incidence of SCA and/or SCD in the overall U.S. population appear to be inconsistent, making it difficult to determine the actual magnitude of this problem. Published estimates of SCA and/or SCD incidence in the U.S. vary widely depending on the source (12). Another major issue limiting the ability to determine the true magnitude of SCD is the lack of consensus on its definition and appropriate designation. Not only are SCD and SCA often used interchangeably despite having 2 distinct definitions, an agreement on the best definition for these entities is yet to be reached. Subtleties inherent in these clinical definitions have far-reaching public health implications, as they affect the ability to meaningfully discuss these clinical outcomes, and reach any conclusions about their magnitude.
The need for consistent and current data describing the true incidence of SCA and/or SCD was highlighted during the most recent Sudden Cardiac Arrest Thought Leadership Alliance's (SCATLA) Think Tank meeting in Washington, DC, which convened a group of national experts with broad representation of key stakeholders including thought leaders and representatives from the American College of Cardiology (ACC), American Heart Association (AHA), and the Heart Rhythm Society (13). As a result, in this study, we systematically evaluate the sources of the wide range of published estimates for the incidence of SCA and/or SCD. In addition, we propose recommendations to facilitate a more reliable estimation of SCA and/or SCD incidence, which will advance our knowledge and the development of public awareness, education, and prevention.
To find estimates of the incidence of SCA and/or SCD, a systematic literature search was performed in MEDLINE using the MeSH headings, “death, sudden” OR “death, sudden, cardiac” or the terms “sudden cardiac death” OR “sudden cardiac arrest” OR “cardiac arrest” OR “cardiac death” OR “sudden death” OR “arrhythmic death.” Study selection criteria included peer-reviewed publications of primary data used to estimate the incidence of SCD in the U.S. Studies that did not report primary data and/or were not performed in a U.S. population were excluded.
Of the 13,649 abstracts that resulted from the comprehensive search strategy, the vast majority were excluded for the reasons detailed in Figure 1. Thirty-five full-length, peer-reviewed, publications were selected for full review, which yielded 6 studies providing primary data used to derive an estimated annual incidence of SCD in the U.S. (14–19). The study by Chugh et al. (17) published in 2004 provides primary data for a count estimate of SCD; however, an estimated annual incidence of SCD in the U.S. is only presented in a subsequent paper published by Chugh et al. (20) in 2008. As such, we included the 2004 study and cited additional data from the 2008 study. Review of expert consensus documents including the ACC/AHA/European Society of Cardiology (ESC) 2006 Guidelines for Management of Patients with Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death and bibliographies identified 2 additional sources for the estimate of the incidence of SCA and/or SCD; however, both were excluded owing to lack of primary data (2,21). Given the wide range of estimated incidences of SCA and/or SCD cited in the literature despite the small number of primary sources providing such estimates, to obtain a sense of the degree of data dissemination in the medical literature we used the Web of Science's Cited Reference Search tool to determine the number of times each of these “primary sources” has been cited.
Estimates of SCA and/or SCD incidence
Despite a very broad search strategy, only 6 publications have reported estimates of the incidence of SCA and/or SCD in the U.S. derived from primary data (Table 1) (14–19). Estimations of SCA and/or SCD incidence ranged widely from 180,000 to >450,000 (17,18). The oldest study was published in 1989, using National Center for Health Statistics (NCHS) and death certificate data from 1980 to 1985 (14). The most recent study was published in 2008 using data collected between May 1, 2006, and April 30, 2007, from 10 North American sites (8 U.S. sites) that was subsequently extrapolated to the U.S. population (19).
Methods of estimation differed among studies. Two studies extrapolated estimates of national annual incidence of SCA and/or SCD based on smaller community-based studies (16,17). One study used registry data from 8 different sites in the U.S. to extrapolate a national annual incidence (19). The 3 remaining studies used national level data to provide estimates (14,15,18); however, 1 of these studies reported only data from the 40 states that reported data to the NCHS (14).
Differences in definitions
The stated definitions of SCA and/or SCD varied across the 6 studies (Table 1). Three studies included time constraints in their case definitions for SCD (15,17,18). Four sources included a geographical location of the event as part of the case definition (14,16,18,19). Although 2 studies specifically defined SCD as death attributable to ischemic or coronary heart disease (14,15), the other studies broadened their criteria to include death from a cardiac or cardiovascular etiology (16–19). One study included “survivors of cardiac arrest” in their case definition of SCD (17), and the most recent study also specifically examined out-of-hospital cardiac arrest and included deaths as well as patients who survived to hospital discharge (19).
Differences in case ascertainment
Reflecting the variation in definitions of SCA and SCD among these primary sources, the derivation methodologies for each study were also different (Table 2) (22). The age cut-offs for the studied populations were highly variable. Two studies used an age cut-off of ≥25 years (14,15); 1 study used an age cut-off of ≥35 years (18); and another study used age ≥20 years (16). The 2 most recent studies did not specify any age criteria (17,19).
Five studies used death certificate data in addition to other survey or report data including the NCHS, the National Mortality Followback Survey, U.S. Bureau of the Census data, and emergency medical services reports (14–18). In contrast, the study by Nichol et al. (19) assesses out-of-hospital cardiac arrest incidence and outcomes and did not use death certificates to determine mortality, but used case-fatality rates from a registry. When vital status data were missing, the authors used multiple imputation methods to determine mortality rates (19). Only 2 studies used prospective case ascertainment (17,19).
Dissemination of data from primary sources
Despite the markedly heterogeneous SCA and/or SCD equivalent definitions and methods of estimation, all of these papers have been cited as sources of estimates for the incidence of SCA and/or SCD in the U.S. Four of the 6 source papers identified by our MEDLINE search were also referenced in the ACC/AHA/ESC 2006 guidelines (12,14–16,18); however, the guidelines cited 2 additional studies as primary data sources for the wide ranging estimates of SCD incidence in the U.S. (Table 3) (21,23,24). As these citations were selected by the Writing Committee to Develop Guidelines for Management of Patients with Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death, they represent expert consensus on the best source documents providing information on the incidence of SCA and/or SCD. The oldest reference was published in 1989 (14), and the most current source was published in 2004 (21); however, 2 of the sources cited by the guidelines were not primary derivations of the estimated incidence of SCA and/or SCD (2,21).
These documents have been widely cited in the published body of peer-reviewed literature (Table 4). The most often cited source was published almost a decade ago using U.S. vital statistics and death certificate data from 1989 to 1998 (18), but has been cited 400 times from March 2002—most recently in November 2009. The next most cited source was published in 1993 and does not provide primary data from which an estimated incidence of SCD was derived, but instead states that “estimates are based on published epidemiologic and clinical data” (2). Finally, the third most often cited source was published in 1989 and was based on death certificates and data from the NCHS collected from 1980 to 1985—30 years ago (14).
There are 2 main findings for this study: 1) definitions of SCA and SCD are not standardized across the medical community; and 2) the true incidence of SCA and/or SCD in the U.S. population remains unclear.
The widely ranging estimates published in the few studies that have tackled this question highlight the lack of clarity about the true annual incidence of SCA and/or SCD in the U.S. By extension, discrepancies in published estimates reflect inconsistencies in the medical community's knowledge and understanding of SCD.
Multiple and sometimes discordant case definitions are used to describe SCA, which is often incorrectly used interchangeably with SCD in the literature. To date, there is no standardized, universal definition used uniformly across the large numbers of studies of this clinical entity. Confusion surrounding the definition of “sudden cardiac death” starts with the fundamentally complex nature of the entity we are attempting to capture within the constraints of the English lexicon, namely, what counts as “sudden”? What qualifies as “cardiac”? Do “sudden” and “cardiac” refer to “death” and/or “resuscitated death”? These definitional issues have been recognized, and numerous definitions have been put forth over the last half century; however, no single, “all-purpose” definition has emerged in light of the myriad, often multifactorial, etiologies of SCD (25,26). Rather than propose a universal definition of SCD, the World Health Organization Scientific Group on SCD wrote in 1985 that it is “more meaningful to define the specific characteristics surrounding cardiac arrest, and register and collect data in a standardized way, than to try to define the word ‘sudden’ in the context of death” (25). According to the ACC/AHA/ESC 2006 guidelines, SCA is defined as “death from an unexpected circulatory arrest, usually due to a cardiac arrhythmia occurring within an hour of the onset of symptoms, in whom medical intervention (e.g., defibrillation) reverses the event” (12); and SCD is defined as “death from an unexpected circulatory arrest, usually due to a cardiac arrhythmia occurring within an hour of the onset of symptoms” (12).
There are a number of possible reasons that the annual incidence of SCA and/or SCD in the U.S. remains unclear. To begin, there are few studies focused on answering this question, and among this handful of studies, most are old, dependent on death certificate data, and not prospective. Across studies there was significant variation in case definitions of SCD and methods of case ascertainment.
Instead of prospective case ascertainment, most of the available estimates are based on retrospective death certificate-based methodology. Death certificate data do not account for survived SCA and have been demonstrated in validation studies to overestimate the incidence of SCD (27–29). The study providing the highest estimate of annual SCD incidence (>450,000) used only death certificate-based retrospective surveillance (18) compared with the more recent prospective study that provides a range that included the lowest estimate (180,000 to 250,000) and employed multiple methods of case ascertainment (17). Additional difficulties stemming from these estimates' dependence on death certificate data are the lack of uniform designation of cause of death on the certificate (30). On death certificates, SCD is often listed as the primary cause of death even when the patient had multiple comorbidities that may have directly contributed to death. Like death certificate data, other sources of retrospective case ascertainment such as vital statistics and census data also misclassify and can overestimate SCD (30). Finally, the prospective, community-based studies, used data from first-responders, which results in the inherent exclusion of unwitnessed SCA and/or SCD.
From a public health perspective, the ability to reliably estimate the incidence of SCA and/or SCD in the U.S. is important because it informs our efforts to prevent SCD, to improve risk stratification techniques, and to direct future research that may save lives. On the basis of our findings from this review of the literature, we recommend 3 actionable next steps toward addressing our knowledge and data gaps:
1. Establish precise, uniform definitions of SCA and of SCD, to be agreed upon and adopted by the major stakeholders in the medical/scientific community. SCA and SCD should have 2 distinct definitions that should not be used interchangeably, as SCA includes a subset of potentially lethal events that are reversed with medical intervention. We support the definitions put forth by the ACC/AHA/ESC 2006 guidelines that state: 1) sudden cardiac arrest is “an unexpected circulatory arrest, usually due to a cardiac arrhythmia occurring within an hour of the onset of symptoms, in whom medical intervention (e.g., defibrillation) reverses the event” (12); and 2) sudden cardiac death is “death from an unexpected circulatory arrest, usually due to a cardiac arrhythmia occurring within an hour of the onset of symptoms” (12).
2. Perform larger, multicenter, prospective studies in the general population as well as across different geographic regions and races, using the above definitions to facilitate optimal case ascertainment. The Resuscitation Outcomes Consortium (ROC) is a clinical research network of 11 sites and 1 central coordinating center in addition to its companion prospective, multicenter, observational registry of the 11 sites (ROC Epistry-Cardiac Arrest) established to evaluate out-of-hospital cardiac arrest and resuscitation outcomes in the participating geographic regions (19). One potential approach for achieving more reliable national estimates of SCA and SCD could be based on an expansion of the ROC and ROC-ECA design, establishing a national SCA and SCD research network composed of 40 to 50 emergency medical services agencies with a central receiving institution selected to be nationally representative.
3. Integrate multiple source methods for case ascertainment to capture more cases and rule out those cases that do not meet strict criteria according to the definitions determined above. Using multiple sources to confirm cases of SCA and/or SCD will help to minimize dependence on death certificate data.
It is hoped this initial framework will help to put in place the necessary elements enabling us to fill in data gaps on this important public health issue. Establishing definitional agreement in the medical community and in the literature will facilitate the more reliable evaluation and comparisons of therapies, utilization, and outcomes, which will ultimately enhance our strategies for risk stratification and prevention of SCD. Given the paucity of studies providing current data on the annual incidence of SCD in the U.S., these findings highlight the need for additional contemporaneous studies to be performed prospectively and to integrate multiple methods of case ascertainment to determine the true incidence of SCD in the U.S. Consideration should be given to create a national surveillance system to assess the characteristics, treatments, and outcomes of patients with SCA; facilitate efforts to improve the quality of resuscitation and SCA patient care; and implement research findings and evidence-based recommendations into clinical practice.
The true incidence of SCA and/or SCD in the U.S. remains unclear, with a wide range in available estimates. To advance knowledge and understanding of the epidemiology of SCD and improve risk stratification and prevention, there needs to be a focused effort to create a standard definition of both SCA and SCD and to expand national research efforts to precisely capture cases of SCA and SCD prospectively and more broadly across communities.
For a list of participants in the SCATLA meeting and recent participants in Duke's Center for the Prevention of Sudden Cardiac Death, please see the online version of this article.
Systematic Review of the Incidence of Sudden Cardiac Death in the United States
Dr. Kong serves on an advisory board to Medtronic, receives research funding from Biotronik, and was supported by a Ruth L. Kirschstein-National Research Service Award National Institutes of Health (NIH) grant (5-T32-DK-007731-15). The contents of this work are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. Dr. Fonarow receives research funding from the National Heart, Lung, and Blood Institute; serves as a consultant for Novartis, Medtronic, and St. Jude Medical; and receives honoraria from GlaxoSmithKline, St. Jude Medical, and Medtronic. Dr. Peterson has received research support from Bristol-Myers Squibb/Sanofi, Merck/Schering-Plough, Lilly, and Johnson & Johnson. Dr. Curtis receives research funding from Medtronic and St. Jude Medical; consulting, Speakers' Bureau, and/or fellowship support from Medtronic, St. Jude Medical, Biotronik, and Sanofi-Aventis; and has served on the advisory board of Biosense Webster. Dr. Hernandez receives research funding from Johnson & Johnson and honorarium from Medtronic. Dr. Sanders is coinvestigator on a Medtronic Inc. research grant. Dr. Hayes serves on an advisory board for St. Jude Medical, Boston Scientific, and Pixel Velocity, and receives speaking honoraria from St. Jude Medical, Boston Scientific, Medtronic, Sorin Medical, Biotronik, and Blackwell-Futura. Dr. Al-Khatib has received research funding and speaking fees from Medtronic and Biotronik. All other authors have reported they have no relationships to disclose.
- Abbreviations and Acronyms
- American College of Cardiology
- American Heart Association
- European Society of Cardiology
- National Center for Health Statistics
- sudden cardiac arrest
- sudden cardiac death
- Received April 26, 2010.
- Revision received August 30, 2010.
- Accepted September 27, 2010.
- American College of Cardiology Foundation
- Lloyd-Jones D.,
- Adams R.J.,
- Brown T.M.,
- et al.
- Goldberg R.J.,
- Gorak E.J.,
- Yarzebski J.,
- et al.
- Connolly S.J.,
- Gent M.,
- Roberts R.S.,
- et al.
- Kuck K.H.,
- Cappato R.,
- Siebels J.,
- Ruppel R.
- Zipes D.P.,
- Camm A.J.,
- Borggrefe M.,
- et al.
- Gillum R.F.
- Escobedo L.G.,
- Zack M.M.
- Chugh S.S.,
- Jui J.,
- Gunson K.,
- et al.
- Zheng Z.J.,
- Croft J.B.,
- Giles W.H.,
- Mensah G.A.
- American Heart Association
- Myerburg R.J.,
- Castellanos A.
- ↵(1985) WHO Scientific Group on Sudden Cardiac Death Technical Report Series 726 (World Health Organization) (World Heatlh Organization, Geneva), p 28.
- Fox C.S.,
- Evans J.C.,
- Larson M.G.,
- Kannel W.B.,
- Levy D.