|Home | About | Journals | Submit | Contact Us | Français|
Syncope is a common reason for emergency department (ED) visits, and patients are often admitted to exclude syncope of cardiovascular origin. Population based data on patterns and predictors of cardiac outcomes may improve decision-making. Our objective was to identify patterns and predictors of short-term cardiac outcomes in ED patients with syncope. Administrative data from an integrated health system of 11 Southern California EDs were used to identify cardiac outcomes after ED presentation for syncope from 1/1/02-12/31/05. Syncope and cause of death were identified by ICD-9 codes. Cardiac outcomes included cardiac death and hospitalization or procedure consistent with ischemic heart disease, valvular disease or arrhythmia. Predictors of cardiac outcomes were identified through multivariate logistic regression. There were 35,330 adult subjects who accounted for 39,943 ED visits for syncope. The risk of cardiac outcome sharply decreased following the 7 days post syncope. A 7-day cardiac outcome occurred in 893 cases (3%). Positive predictors of 7-day cardiac outcomes included age ≥ 60, male gender, congestive heart failure, ischemic heart disease, cardiac arrhythmia, and valvular heart disease. Negative predictors included dementia, pacemaker, coronary revascularization, and cerebrovascular disease. There was an age-dependent relationship between 7-day cardiac outcomes and arrhythmia and valvular disease, with younger patients (age <60 years) experiencing greater risk of an event compared to their same age counterparts. In conclusion, ED decision-making should focus on risk of cardiac event in the first 7-days after syncope and special attention should be given to younger patients with cardiac comorbidities.
Emergency department (ED) evaluation of syncope may benefit from improved epidemiologic understanding of patterns and predictors of short-term, cardiac events. Previous studies have examined risk factors at 1 year following an episode of syncope, a timeframe not ideal for decision making in the acute care setting1-4. Recent cohort studies identifying predictors of short-term (7-30 days) events after syncope are relatively small (n=444-791)2,5-8 and reported prediction models have limited stability and generalizability.9,10 In this retrospective cohort study, we describe patterns and predictors of short-term cardiac outcomes after emergency department visits for syncope. We studied a population-based, managed care cohort receiving care from a regional, integrated health system. Cardiac outcomes included cardiac death, and hospitalizations and procedures consistent with a diagnosis of arrhythmia, ischemic heart disease, and valvular heart disease.
Kaiser Permanente Southern California is an integrated health system which provides comprehensive care to 3.1 million members throughout Southern California. Health care is delivered at 12 medical centers and over 100 outpatient clinics. At the time of the study, 11 health system EDs were available to members. All members have similar health care benefits, including coverage of emergency services both within and outside the health system. Electronic administrative databases track all health care encounters within the health system. A claims reimbursement system tracks health care provided at outside facilities. Detailed information on diagnoses and procedures are available regardless of setting. Laboratory, pharmacy and other specialized databases provide information on clinical care. All members are assigned a unique medical record number which is used for data linkage.
Study subjects were members of Kaiser Permanente Southern California with at least 1 emergency department visit for syncope from 1st Jan 2002 to 31st Dec 2005. Subjects were restricted to age 18 years and older due to the different nature of syncope in children.11 ED visits both within and external to the health system were included. A subject had to be a member of the health plan at the time of the ED visit; however, no minimum health plan enrollment period was required.
Syncope was identified by ICD-9 code 780.2 “Syncope and Collapse” among all ED diagnoses of both Kaiser and non-Kaiser facilities. Cases with multiple diagnoses in addition to syncope were included. To validate the accuracy of these codes, blinded physician chart review was performed on consecutive ED visits (n=100) with and without a diagnosis code consistent with syncope. Compared to physician chart review, ICD-9 codes demonstrated a positive predictive value of 92% and a negative predictive value of 100%.
Demographic information on date of birth, age, sex and race were obtained from administrative databases. Preexisting comorbid conditions were used to classify risk in subanalyses. We obtained information on major comorbid conditions related to syncope using data available from the health plan within the observation period and prior to the ED visit. Case-identification criteria for diabetes included a combination of inpatient and outpatient diagnosis and procedural codes, medications and laboratory tests.12 Identification of other comorbidities was based solely on diagnosis and procedural codes. History of arrhythmia was based on ICD-9 codes signifying ventricular tachycardia, ventricular fibrillation/flutter, type II mobitz heart block, anomalous atrioventricular excitation, paroxysmal atrial tachycardia, atrial flutter, atrial fibrillation, or sinoatrial node dysfunction. A subject was noted to have a prior history of syncope if there was an ED visit for syncope within 30 days preceding the index ED visit.
The primary outcome was a 7-day cardiac outcome occurring after an ED visit for syncope. Cardiac outcomes included cardiac death, and hospitalization or procedure consistent with an arrhythmia, ischemic heart disease, and valvular heart disease. Mortality and cause of death data were identified through linked California vital statistics files. A death was classified as cardiac in origin if the ICD9 cause of death code indicated ischemic heart disease, arrhythmia, cardiac valve disease, or congestive heart failure (CHF).
We defined arrhythmic events as either a hospitalization with primary discharge code consistent with arrhythmia, or procedure codes consistent with insertion or revision of a cardiac pacemaker or an implantable cardioverter defibrillator (AICD). Ischemic heart events included hospitalization with a primary discharge diagnosis of myocardial infarction or unstable angina, or procedure codes consistent with coronary artery bypass graft surgery (CABG) or percutaneous transluminal coronary angioplasty (PTCA). Finally, we defined valvular heart events as hospitalization with a primary discharge diagnosis of valve disease, or procedure codes consistent with valve replacement or revision.
Analyses were conducted to identify the risk of a cardiac outcome following an ED visit for syncope. For subjects with multiple ED visits for syncope, counts over time were determined using the first visit within the study period. The association of pre-existing comorbidities and a cardiac outcome at 7 days was examined by Fischer's exact test. Hazard functions were prepared for age groups to examine the absolute probability of an event as a function of time after syncope. Risk by age was examined for ages 18-39, 40-59, 60-79, and 80 plus years. Cause of death was examined using ICD-9 and ICD-10 codes from death certificates. Reliability of ICD grouping to identify cardiac death was assessed by physician chart review (blinded to ICD cause of death code) of 60 inpatient deaths occurring after syncope (kappa=0.7).
Hazard plots were examined for 7, 30, 180 and 365 days and suggested that excess cardiac risk was concentrated in the first 7 days after syncope, the time frame used for subsequent risk modeling. Logistic regression results were used to help guide subgroup analysis of hazard functions, such as the risk relationship between age groups and having a cardiac comorbidity. Multiple ED visits by the same subject were controlled for in the regression analyses. Coefficient standard errors were adjusted for subject correlation (i.e., clustering) using robust variance estimates. Predictors of 7-day cardiac outcomes were identified through multivariate logistic regression. Univariate analyses suggested a step increase in risk at 60 years of age, and subsequent models dichotomize age at this threshold. Interactions were tested between age and cardiac co-morbidities. The final model used statistically (i.e., P<0.05) and clinically significant covariates and interactions to identify the predictors of a 7-day cardiac outcome.
Sensitivity analyses were conducted to evaluate the robustness of our model. Additional models were constructed with the same predictors and distinct outcomes of cardiac death, atherosclerotic event, or arrhythmic event. The 3 models were qualitatively similar to the combined outcomes model. Valvular events were not modeled because of too few events to perform reliable modeling. Our model also evaluated 30-day combined cardiac outcomes. The predictors in this modified model were also very similar in their significance and direction of effect. To improve interpretability of our findings, we present only the results of the combined 7-day cardiac events model.
All analyses were conducted at Kaiser Permanente Southern California's Department of Research using SAS 9.1 (SAS Institute, Cary, NC). The study protocol was reviewed and approved by the Institutional Review Board of Kaiser Permanente Southern California and the University of California Los Angeles.
Over the 4-year observation period, there were 35,330 subjects who accounted for 39,943 ED visits for syncope. There were 893 7-day cardiac outcomes, representing an event rate of 2.5% (Table 1). Ninety % of subjects had 1 visit for syncope; the number of syncope visits ranged from 1 to 12 per subject. Table 2 shows the types of cardiac events at 7 days. Of the 893 subjects who experienced a cardiac event, the majority of outcomes were caused by arrhythmias (63%). There were several subjects who experienced multiple categories of cardiac outcomes.
Figures 1 and and22 illustrate the absolute probability of cardiac event after ED visit for syncope. There is a marked increase in risk within the first 3 days, with return to baseline risk across all age groups following the first 7 days.
Logistic regression identified demographic characteristics and comorbidities that are most strongly associated with a cardiac outcome following an ED visit for syncope. Table 3 shows results for a 7-day cardiac outcome with adjustment for interaction terms between age and the comorbidities of CHF, arrhythmia, valvular disease, and MI. Cardiac outcomes were significantly positively associated with age ≥ 60 (OR 3.8, 95%CI 2.9-5.0), male gender (OR 1.5, 95%CI 1.3-1.7), congestive heart failure (CHF) (OR 2.0, 95%CI 1.1-3.5), and ischemic heart disease (OR 3.7, 95%CI 2.1-6.5). We noted significant negative interactions between age and arrhythmia, and age and valvular heart disease. When compared to the same age group, we found that the risk of developing a cardiac outcome is greater for those that are younger (ages 18-60) with an arrhythmia compared to their older (ages 60+) counterparts. For patients with a history of valvular disease, the age-stratified odds ratios were significant in the positive direction for younger patients only. Factors that appeared protective against cardiac outcomes include history of cardiac re-vascularization, pacemaker or AICD, cerebrovascular disease, and dementia.
This large retrospective cohort study using clinically obtained data reveals 3 novel insights into patterns and predictors of short-term cardiac events and mortality after an ED visit for syncope. First, there is a marked risk of a cardiac outcome within the first 3 days of an ED visit, and risk returns to baseline after the first 7 days. Second, positive predictors of cardiac outcomes within 7-days of an ED visit include male gender, age ≥ 60, and cardiac co-morbidity. A history of a pacemaker, AICD, or a cardiac revascularization procedure appears to be protective. Finally, we found a negative interaction effect between age and arrhythmia and age and valvular heart disease. This finding suggests that when compared to older patients, younger individuals with arrhythmias or valvular heart disease have a greater, age-stratified odds ratio of developing a cardiac outcome. This is the first description of the predictors of short-term cardiac outcomes after syncope in a large cohort of patients. Previous studies have shown that cardiac etiologies of syncope are common and can result in increased short-term mortality but these studies are limited by cohort size.7,13-15
Our data demonstrates that a majority of cardiac events and mortality occur in the first 3 days of the ED visit, suggesting that when decisions are made regarding disposition, physicians should take into account that the highest risk occurs in the first 3 days. We also found that following the first 3 days until 7 days following the ED visit, patients are at increased risk of developing a cardiac outcome, suggesting that patients discharged home following a negative evaluation should be closely monitored as outpatients.
We observed an event rate of 2.5 %, a rate lower than what has been previously noted of 8-11%.7,9,10,16 This could be due to the younger age of our cohort and to the better health status of managed care patients. In addition, prior studies evaluating serious events in 7 days were conducted in academic centers that could have sicker patients. 7,9,10,16
We identified multivariate predictors of 7-day cardiac outcomes after syncope. Our findings of male gender and age-related risk have been confirmed by other studies.2,4,5,7,8,15 Prior studies have also found cardiac co-morbidity to be associated with an increased risk of morbidity and mortality.1,2,4,8,13 What is most compelling and novel in our findings is the considerable age-dependent risk associated with arrhythmia and to a lesser extent with valvular disease. We found the relative age-stratified risk for arrhythmia much greater for younger patients (ages 18-59: OR 7.3, 95%CI 4.4-11.9) compared to older patients (ages 60+: OR 3.2, 95%CI 2.7-3.8) and to a lesser extent for valvular disease. These findings suggest that younger patients with a history of arrhythmia or valvular disease who experience syncope may be at increased risk of a cardiac event. When evaluating young patients with syncope, special attention must be given to those with a history or signs concerning for an arrhythmia or valvular disease.
Finally, we found that the risk of developing a cardiac outcome was reduced in patients with history of cardiac re-vascularization, pacemaker or AICD, cerebrovascular disease, and dementia. Presumably, an intervention provides benefit by protecting against potentially fatal arrhythmias or coronary ischemia. Prior cerebrovascular disease may indicate increased risk for a neurologic rather than a cardiac cause of syncope. The lower cardiac risk associated with dementia may be attributed to poor recall ability resulting in difficulty distinguishing syncope from other conditions such as generalized weakness, seizure, or an unwitnessed fall.
Our study has some limitations. The demographic characteristics of this managed care population are similar to the surrounding population in Southern California. However, compared to syncope ED visits in a nationally representative sample17, our study cohort has a higher proportion of younger, non-white patients, and Hispanics. Care processes in this integrated health system may be different from other settings.7,9,10 Therefore, the generalizability of our findings will need to be assessed in other settings. In addition, we collected comorbidity information from existing administrative data which limited our ability to assess clinically important elements such as symptoms and exam findings.2,4,7,8,15,18 Addition of such data could potentially improve the ability to predict short term cardiac death and events.
Dr. Gabayan is supported by a Greater Los Angeles Veteran's Affairs Health Services Research and Development Fellowship. This research was also supported by grants from the American Geriatrics Society Dennis Jahnigen Career Development Award (20051687, B. Sun) and the UCLA National Institutes of Aging K12 Mentored Clinical Scientist Development Program in Geriatrics (AG 01-004, B. Sun).
Grant/Support Information: The Greater Los Angeles Veteran's Affairs Health Services Research and Development Fellowship (G.Gabayan), American Geriatrics Society Dennis Jahnigen Career Development Award (B.Sun) and UCLA National Institutes of Aging K12 Mentored Clinical Scientist Development Program in Geriatrics (B.Sun)
Study performed at the Department of Research and Evaluation, Kaiser Permanente Southern California, 100 S. Los Robles Ave, 2nd Fl, Pasadena, California 91101
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.