| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Circulation. 2003;108:1196.)
© 2003 American Heart Association, Inc.
Clinical Investigation and Reports |
From the University of Washington (T.D.R., M.C., M.S.E.), and Public Health (T.D.R., M.S.E., L.J.B., T.H.), SeattleKing County, Emergency Medical Services Division, Seattle, and the Center for Health Statistics (A.R.L.), Washington State Department of Health, Olympia, Wash.
Correspondence to Thomas Rea, Public Health, SeattleKing County, Emergency Medical Services Division, 999 Third Ave, Ste 700, Seattle, WA 98104-4039. E-mail rea123{at}u.washington.edu
Received April 15, 2003; revision received June 11, 2003; accepted June 13, 2003.
| Abstract |
|---|
|
|
|---|
Methods and Results The investigation was a retrospective cohort study of survival among persons who were discharged alive from the hospital after resuscitation from out-of-hospital SCA due to heart disease in King County, Wash, between May 1, 1976, and December 31, 2001 (n=2035). Calendar time was divided into four 5-year intervals: 1976 to 1980, 1981 to 1985, 1986 to 1990, and 1991 to 1995, and one 6-year interval, 1996 to 2001. Age-adjusted survival curves were constructed, and Cox proportional-hazards regression was used to compute hazard ratios (HRs) for the association between mortality and time period. During 11 201 person-years of follow-up, 1334 persons died. Compared with the initial time period, the HR for total mortality was 0.86 (95% confidence interval, 0.73 to 1.01) for 1981 to 1985, 0.82 (0.69 to 0.96) for 1986 to 1990, 0.66 (0.55 to 0.79) for 1991 to 1995, and 0.58 (0.47 to 0.71) for 1996 to 2001 (HR for trend=0.87 [0.84 to 0.91] for each successive time period). In analyses that assessed cardiac mortality, an even stronger temporal association was evident (HR for trend=0.79 [0.75 to 0.84]).
Conclusions Long-term survival after resuscitation from SCA improved steadily over time in this cohort. To continue this trend, future studies should identify circumstances in which proven treatments are underutilized as well as investigate new therapies that might benefit survivors of SCA.
Key Words: heart arrest epidemiology mortality survival
| Introduction |
|---|
|
|
|---|
| Methods |
|---|
|
|
|---|
2000 square miles and comprises urban, suburban, and rural components. The population of King County (excluding Seattle) grew steadily during the past 25 years, increasing from
686 500 persons in 1976 to 1 196 000 persons in 2000.12 As of 2000, 76% of the population was white, 11% Asian-American, 5% African-American, 6% Hispanic, and 1% American Indian. Eleven percent of the King County population was 65 years of age or older. Median household income was $51 300, with 8% of persons living below the poverty line. King County is served by a 2-tiered EMS response system. The first tier comprises firefighteremergency medical technicians who administer basic life support as well as defibrillation with automated external defibrillators. The second tier consists of paramedics who administer advanced cardiac life support. In cases of suspected cardiac arrest, both tiers are dispatched simultaneously. For cardiac arrest calls, emergency dispatchers offer and provide cardiopulmonary resuscitation instructions over the telephone.13
Data Collection and Definitions
Since May 1, 1976, the EMS of King County has maintained a registry of all EMS-treated SCA cases. The registry contains information regarding survival to hospital discharge and has been maintained in the Utstein format.14,15 Survival to hospital discharge and underlying etiology of arrest (ie, heart disease) were determined by review of EMS reports, death certificates, and hospital discharge records. For persons who survived to hospital discharge, a match was conducted between the EMS registry and the death records provided by the Washington State Department of Health and the National Death Index to determine whether and when a subject had died as of December 31, 2001. For each case of death, the underlying cause was recorded. Deaths were classified as cardiac or noncardiac. A death was classified as cardiac when the International Classification of Diseases 8 or 9 code was 390 to 398, 402, or 404 to 429 or when the International Classification of Diseases 10 code was I00I02, I05I09, I11I28, or I33I51.1 To assess temporal trends, calendar time period was divided into four 5-year intervals: 1976 to 1980, 1981 to 1985, 1986 to 1990, and 1991 to 1995, and one 6-year time interval, 1996 to 2001. The investigators review board approved the study.
Statistical Analysis
Follow-up time started at the date of hospital discharge after the SCA and extended to the date of death or the end of the follow-up period (December 31, 2001). Age-adjusted survival curves were constructed for each time period. Cox proportional-hazards regression was used to compute hazard ratios (HRs) for the association between total mortality and time period for 1, 5, and 10 years and total duration of follow-up after adjustment for age at hospital discharge and gender.16 The earliest time period served as the referent time period. Analyses were also stratified by gender and age at hospital discharge (>65 and
65 years of age). Comparable analyses assessed the association between cardiac mortality and time period and noncardiac mortality and time period. Sensitivity analyses evaluated whether grouping the time period differently or whether coding as cardiac the relatively few cases for whom cause of death was missing changed the results. The proportional-hazards assumption was tested and confirmed. Analyses were done with commercially available software (STATA 7.0, STATA Corp).
| Results |
|---|
|
|
|---|
Among the 2035 persons in the cohort, average age at hospital discharge increased over time (61.7 for 1976 to 1980, 63.6 for 1981 to 1985, 64.4 for 1986 to 1990, 64.1 for 1991 to 1995, and 65.0 for 1996 to 2001; test for trend P=0.002). No temporal trend was evident with regard to gender (proportion who were female for the 5 time periods was 21.9%, 25.0%, 23.8%, 25.7%, and 25.8%, respectively; test for trend P=0.3).
During the 11 201 person-years of follow-up, 1334 persons died. Survival improved steadily over time (Figure 1). Compared with the 1976 to 1980 time period, the overall HR of total mortality for the 1996 to 2001 time period was 0.58 (95% confidence interval [CI], 0.47 to 0.71), with mortality decreasing, on average, 13% for each successive time period. Results were similar for 1-, 5-, and 10-year follow-up periods. In sensitivity analyses, results were comparable when the time period was grouped as five 4-year and one 6-year time periods or as seven 3-year and one 5-year period or when modeled as a continuous (1-year) variable. Exclusion of age from the Cox model modestly attenuated the HRs associated with time period (HR for trend without age adjustment=0.91 [95% CI, 0.87 to 0.95]).
|
The temporal pattern of improved survival over time was evident among men (HR=0.87 [95% CI, 0.83 to 0.92] for each successive time period) and women (HR=0.88 [95% CI, 0.81 to 0.96] for each successive time period), as well as for persons <65 years of age (HR=0.84 [95% CI, 0.79 to 0.91] for each successive time period) and persons
65 years of age (HR=0.89 [95% CI, 0.84 to 0.94] for each successive time period).
In analyses that assessed the relation between time period and cardiac mortality, an even stronger temporal association was evident, with cardiac mortality decreasing, on average, 21% with each successive time period (Figure 2). When the 36 cases with a missing cause of death were coded as a cardiac cause, the results were similar (HR=0.81 [95% CI, 0.77 to 0.86]). In contrast, there was no evidence of a temporal trend for noncardiac mortality (Figure 3).
|
|
| Discussion |
|---|
|
|
|---|
Mortality for the entire cohort regardless of time period was greatest during the first year after hospital discharge and in part likely reflects the greater early mortality of those with neurological impairment and the elevated risk of recurrent cardiac ischemic or arrhythmic events that occur early after a cardiac event.12,17,18 After the first year, the mortality rate declined for all time periods. Indeed, among the more recent time periods, the mortality rate approached that of the clinical trial populations that evaluated implantable cardioverter defibrillators versus pharmacotherapy among survivors of near-fatal arrhythmias, suggesting that SCA survivors in the community are receiving care that might be comparable to that of clinical trial populations.1921
In this investigation, survival improved steadily over time. The reason for this improvement was likely multifactorial. SCA often represents the critical confluence of ischemic and arrhythmic risk factors.22 Numerous therapies have been developed during the past 25 years to prevent or treat 1 or both of these components.11,22 Although temporal trends regarding health behaviors and treatments are generally unavailable for the specific clinical population of SCA survivors, most evidence suggests that beneficial therapies for persons with various forms of heart disease have increased over time. For example, the use of coronary revascularization as either bypass surgery or percutaneous intervention, both associated with improved survival among persons resuscitated from SCA,23,24 has increased several-fold during the 1980s and 1990s, with ongoing improvements in approach and technique.11,25,26 Similarly, the use of pharmacotherapies that have been demonstrated to reduce mortality in various heart disease populations, such as aspirin, lipid-lowering medications, ß-blockers, angiotensin-converting enzyme inhibitors, thrombolytics, and amiodarone, has also increased over time.2731 Temporal trends in some health behaviors such as smoking have improved over time, while trends in others such as exercise or diet are less certain.3234 Finally, among persons aged 65 years or greater who are admitted to the hospital with a diagnosis of ventricular arrhythmia in the United States, the use of implantable cardioverter defibrillators has increased substantially from the mid 1980s through the mid 1990s.4 The temporal trend of decreased mortality observed in this cohort is consistent with temporal patterns of improved survival and decreased morbidity after myocardial infarction, attributed potentially in part to an increased use of pharmacological and interventional therapies.28,35 Importantly, the majority of deaths even in the more recent time periods in the current investigation were due to heart disease, highlighting the need to continue to improve treatments directed toward heart disease for this population.
In contrast, a study from a European community reported no change in survival when 2 approximate 10-year time periods from 1980 to 1998 were compared.36 An explanation for this contrast might be differences between the 2 cohorts with regard to health behaviors or medical care or differences in initial EMS resuscitation; short-term survival to hospital discharge remained higher but stable over time in King County, whereas short-term survival was lower but appeared to improve over time in the European experience.36,37
The temporal trend of improved long-term survival was evident among men and women as well as older and younger persons discharged from the hospital after SCA. The finding suggests that in total, each demographic group is benefiting from improvements in therapies, although the distribution of specific therapies might differ by gender or age.38,39 Although this temporal improvement in survival across various demographic groups is encouraging, it might incur more cost. Among Medicare beneficiaries resuscitated for ventricular fibrillation cardiac arrest, initial hospital cost associated with care of survivors also rose steadily during the 10-year period from 1986 to 1995, primarily due to the increased number of procedures.4
This study has several limitations. We were unable to assess temporal changes in various underlying health conditions, health behaviors, and treatments that would provide a better understanding of the temporal pattern of improved survival observed in this cohort. Although the temporal trends in all-cause mortality are certain, the temporal trend of cause-specific mortality (ie, cardiac) might be subject to bias if cause-of-death coding changed systematically over time. It is unlikely, however, that such a bias would account for the marked differences in temporal trends observed between cardiac and noncardiac causes of death. The investigation did not assess neurological or functional status of the survivors. However, neurological status of survivors is generally satisfactory.40 Moreover, poor neurological status typically predicts poor outcome, making it unlikely that the survivors during the more recent time periods had poorer neurological status than did survivors from the earlier time periods. Although the results might be representative of the experience of other communities, the findings report 1 communitys experience and might not necessarily be generalizable. These limitations should be balanced against the strengths of the study: the large size of the cohort, the substantial duration and extent of follow-up, and the community-based composition of the cohort.
In this cohort of persons resuscitated and discharged alive from the hospital after resuscitation from SCA, long-term survival increased substantially and steadily over time due to a temporal trend of decreasing cardiac mortality. To continue this trend, future studies should identify circumstances or clinical groups in which proven effective treatments are underutilized as well as investigate new therapies that might benefit the survivor of SCA.
| Acknowledgments |
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
K. Bibbins-Domingo, P. Coxson, M. J. Pletcher, J. Lightwood, and L. Goldman Adolescent Overweight and Future Adult Coronary Heart Disease N. Engl. J. Med., December 6, 2007; 357(23): 2371 - 2379. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. P. Pell, M. Corstorphine, A. McConnachie, N. L. Walker, J. C. Caldwell, A. K. Marsden, N. R. Grubb, and S. M. Cobbe Post-discharge survival following pre-hospital cardiopulmonary arrest due to cardiac aetiology: temporal trends and impact of changes in clinical management Eur. Heart J., February 2, 2006; 27(4): 406 - 412. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. S. Abella, N. Sandbo, P. Vassilatos, J. P. Alvarado, N. O'Hearn, H. N. Wigder, P. Hoffman, K. Tynus, T. L. Vanden Hoek, and L. B. Becker Chest Compression Rates During Cardiopulmonary Resuscitation Are Suboptimal: A Prospective Study During In-Hospital Cardiac Arrest Circulation, February 1, 2005; 111(4): 428 - 434. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. S. Abella, J. P. Alvarado, H. Myklebust, D. P. Edelson, A. Barry, N. O'Hearn, T. L. Vanden Hoek, and L. B. Becker Quality of Cardiopulmonary Resuscitation During In-Hospital Cardiac Arrest JAMA, January 19, 2005; 293(3): 305 - 310. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Circulation Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 2003 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |