Conceived and designed the experiments: DE SR KD YN. Performed the experiments: SR AF. Analyzed the data: DE. Wrote the paper: DE YN SR MM KD. Conducted literature searches, abstracted data from published literature, contacted external authors, and wrote the Methods section: SR LF. Conducted statistical analyses and responsible for the contents of the manuscript: DE.
The authors have declared that no competing interests exist.
Acute coronary syndromes (ACS; myocardial infarction or unstable angina) can induce posttraumatic stress disorder (PTSD), and ACS-induced PTSD may increase patients’ risk for subsequent cardiac events and mortality.
Observational cohort studies that assessed PTSD with specific reference to an ACS event at least 1 month prior. We extracted estimates of the prevalence of ACS-induced PTSD and associations with clinical outcomes, as well as study characteristics. We identified 56 potentially relevant articles, 24 of which met our criteria (N = 2383). Meta-analysis yielded an aggregated prevalence estimate of 12% (95% confidence interval [CI], 9%–16%) for clinically significant symptoms of ACS-induced PTSD in a random effects model. Individual study prevalence estimates varied widely (0%–32%), with significant heterogeneity in estimates explained by the use of a screening instrument (prevalence estimate was 16% [95% CI, 13%–20%] in 16 studies) vs a clinical diagnostic interview (prevalence estimate was 4% [95% CI, 3%–5%] in 8 studies). The aggregated point estimate for the magnitude of the relationship between ACS-induced PTSD and clinical outcomes (ie, mortality and/or ACS recurrence) across the 3 studies that met our criteria (N = 609) suggested a doubling of risk (risk ratio, 2.00; 95% CI, 1.69–2.37) in ACS patients with clinically significant PTSD symptoms relative to patients without PTSD symptoms.
This meta-analysis suggests that clinically significant PTSD symptoms induced by ACS are moderately prevalent and are associated with increased risk for recurrent cardiac events and mortality. Further tests of the association of ACS-induced PTSD and clinical outcomes are needed.
In recent decades, survival rates after acute coronary syndrome [ACS; ST-segment elevation myocardial infarction (STEMI), non–ST-segment elevation myocardial infarction (NSTEMI) or unstable angina (UA)] have steadily increased
In the
It is also unclear whether ACS-induced PTSD, like other psychosocial vulnerabilities such as depression, is associated with worse survival after ACS. However, epidemiological and observational evidence suggests that PTSD due to traumatic events other than ACS is related to increased risk of incident cardiovascular disease
We sought to identify all studies that reported a valid estimate of the prevalence of ACS-induced PTSD. To be included, studies must have been observational cohorts and must have assessed PTSD with specific reference to an ACS event that had occurred at least 1 month prior to the PTSD assessment. Further, studies must have used a self-report PTSD screening instrument or clinical interview designed or specifically altered to query about only
To determine the studies to be assessed further, two authors (S.R., L.F.) independently read the, title or abstract of every record retrieved. All potentially relevant articles were investigated as full text. Where differences in opinion existed, they were resolved by consensus.
From studies describing the prevalence of ACS-induced PTSD, we abstracted PTSD rates and characteristics of the study and sample (
Source, y | PTSD Prevalence, % | PTSD Measure | Clinical Interview,Y/N | Time After ACS Event, mo | Study Location | N | Male, % | Mean Age, y | Includes Unstable Angina,Y/N |
Ayers et al,35 2009 | 16 | PDS | N | 2 | United Kingdom | 74 | 76 | 62 | N |
Bennett and Brooke,36 1999 | 11 | PDS | N | 9.24 | United Kingdom | 44 | 68 | 62.5 | N |
Bennett et al,37 2001 | 8 | PDS | N | 3 | United Kingdom | 39 | 77 | 59.7 |
N |
Bennett et al,38 2002 | 16 | PDS | N | 3 | United Kingdom | 75 | 78 |
60.4 |
N |
Chung et al,25 2007 | 31 | PDS | N | 115.8 | United Kingdom | 120 | 78 | 66.85 | N |
Doerfler et al,28 1994 | 11 | Multiple | N | 9 | United States | 27 | 100 | 59.1 | N |
Doerfler,26 1997 | 11 | PDS | N | 1 | United States | 30 | 71 |
58.9 | N |
Doerfler et al,27 2005 | 8 | PSS-SR | N | 4.5 | United States | 52 | 69 | 57.73 | N |
Edmondson et al,24 2011 | 11 | IES-R | N | 1 | United States | 247 | 55 | 60 | Y |
Ginzburg et al,18 2003 | 16 | PTSD-I | N | 7 | Israel | 116 | 81 | 54.95 | N |
Guler et al,23 2009 | 4 | CAPS | Y | 3.25 | Switzerland | 394 | 83 | 61 | N |
Kutz et al,39 1994 | 7 | SCID | Y | 9.6 | United Kingdom | 27 | 89 | 58.2 | N |
Lukach,29 1996 | 0 | SCID | Y | 6 | United States | 70 | 69 | 59.3 | N |
O’Reilly et al,40 2004 | 7 | SCID | Y | 10.7 | United Kingdom | 28 | 89 | 58.2 | N |
Pedersen et al,43 2003 | 22 | PDS | N | 1.25 | Denmark | 112 | 71 |
60 | N |
Roberge et al,32 2010 | 4 | SCID | Y | 1 | Canada | 393 | NR | 59.22 | N |
Rocha et al,30 2008 | 4 | SCID | Y | 1.5 | United States | 25 | NR | NR | N |
Sheldrick et al,41 2006 | 24 | DTS | N | 1.5 | United Kingdom | 21 | NR | NR | N |
Shemesh et al,34 2001 | 10 | IES | N | 6 | Israel | 102 | 79 | 61.3 | N |
Shemesh et al,21 2004 | 20 | IES | N | 6 | Israel | 65 | 79 |
53 |
N |
Shemesh et al,33 2006 | 22 | IES | N | 7.5 | Israel | 65 | 90 | 58 | N |
Thompson,31 1999 | 31 | ADIS-IV | N | 6 | United States | 26 | NR | NR | N |
van Driel et al,44 1995 | 0 | SCID | Y | 24 | Holland | 18 | 56 | 61.96 |
N |
Wikman et al,64 2008 | 12 | PSS-SR | N | 12 | United Kingdom | 213 | 77 | 61 | Y |
Abbreviations: ACS, acute coronary syndrome; ADIS-IV, Anxiety Disorder Interview Schedule for
Value reported for full sample, including participants who did not complete the study.
Comprehensive MetaAnalysis (version 2, BioStat Software, Engelwood, NJ) served as the statistical platform for completing all statistical tests and associated graphic results. To summarize the prevalence findings, we computed prevalence point estimates using these formulas:
Logit Event Rate = Log [Event Rate/(1 − Event Rate)].
Event Rate Standard Error = √(1/(Event Rate · Sample Size)/(1/[(1−Event Rate) · Sample Size]),
We computed 95% confidence intervals (CIs) using this formula:
Lower Limit = Logit Event Rate − (1.96 · Logit Event Rate Standard Error).
Upper Limit = Logit Event Rate + (1.96 · Logit Event Rate Standard Error).
For meta-analytic tests of ACS-induced PTSD prevalence, in addition to the overall random-effects model, we used sensitivity analyses to assess evidence of moderator effects for the prevalence results across a number of methodological factors. These factors included the method used to assess ACS-induced PTSD, the location of the study (United States vs other), and whether studies explicitly included patients with NSTEMI and UA. Further, meta-regression analysis was used to test the association between study prevalence rates and timing of the PTSD assessment (number of months after ACS), study publication date, and sample age and sex composition. Heterogeneity assessments preceded all meta-analytic tests concerning PTSD prevalence and clinical outcomes. There was statistically significant heterogeneity, so random-effects models were used to estimate and test effects.
We calculated an aggregated point estimate for the risk ratio associated with a positive screen for PTSD on clinical outcome categories including cardiac hospitalization and a combined endpoint of cardiac re-hospitalization and all-cause mortality. Log-transformed risk ratios and 95% CIs were calculated for each study using the reported effect size and estimates of the SE of each effect drawn from data reported in the article. When articles reported multiple models, we selected the model with the highest level of covariate adjustment. For meta-analytic tests of the association of ACS-induced PTSD and clinical outcomes, there were too few studies to test for moderator effects. To address the issue of publication bias, we calculated a fail-safe N.
We initially identified 548 articles, and 2 coders agreed that 48 articles required full reading. Of these 48 potentially relevant articles plus another 8 identified from references, 24 met our criteria for inclusion (see
Note: The area of each square is proportional to the study’s weight in the meta-analysis, and each line represents the confidence interval around the estimate. The diamond represents the aggregate estimate, and its lateral points indicate confidence intervals for this estimate.
The PTSD prevalence rates varied significantly by type of assessment (
Note: The area of each square is proportional to the study’s weight in the meta-analysis, and each line represents the confidence interval around the estimate. The diamond represents the aggregate estimate, and its lateral points indicate confidence intervals for this estimate.
The mean number of months between the event and PTSD assessment across studies was 10.24; however, 1 study
Later publication date was related to lower estimates of PTSD prevalence and accounted for 5% of the variance in estimates (
We compared prevalence rates by whether studies were conducted in the United States (7 studies)
The mean age of the participants varied from 53 to 67 years. A younger mean age was significantly related to a greater prevalence rate in the 21 studies in which data on age were reported (standardized meta-regression weight = .25;
Two studies explicitly included patients with NSTEMI or UA (N = 460; prevalence = 12%; 95% CI = 9%–15%), and their aggregated prevalence estimate did not differ significantly from those of studies that did not explicitly include NSTEMI or UA patients (
Source, y | N | PTSD, % | Mean Follow-up, y | Study Location | Unstable Angina, Y/N | Clinical Outcome | PTSD Assessmentfor Prediction | Magnitude of Association | Covariates Included in Final Model |
Edmondson et al,24 2011 | 247 | 11 | 2 | United States | Y | MACE/ACM | IES-R score >33 | HR, 1.4 (95% CI,.54–3.46) | Age, sex, Charlson comorbidity index, GRACE, LVEF, BDI |
Shemesh et al,21 2004 | 65 | 20 | 1 | Israel | N | Cardiac re-hospitalization | IES response consistent with |
OR, 2.7 (95% CI,1.2–8.8) | none |
von Känel et al,22 2011 | 213 | 12 | 2.8 | Switzerland | Y | MACE | Per 10 points on PDS | HR, 1.4(95% CI,1.07–1.88) | Age, sex, hypertension |
Abbreviations: ACM, all-cause mortality; ACS, acute coronary syndrome; BDI, Beck Depression Inventory; CI, confidence interval;
Using the most conservative estimates available (
Note: The size of the box associated with each study’s estimate represents the precision of the estimate, and the line represents the confidence interval around the estimate.
In recent years, cardiologists and the broader medical community have become increasingly aware that psychological disorders, particularly depression, are common in patients with ACS and are associated with adverse clinical outcomes
To our knowledge, this is the first meta-analytic review of ACS-induced PTSD. We found a 12% prevalence of clinically significant ACS-induced PTSD among ACS patients and, based on a small number of studies, a doubling of risk of mortality and recurrent cardiac events among ACS patients with PTSD symptoms. In previous qualitative literature reviews,
Although the overall prevalence of ACS-induced PTSD was 12%, individual study prevalence estimates ranged from 0% to 32%, suggesting a considerable degree of heterogeneity. As expected, prevalence estimates based on screening questionnaires were greater than estimates based on clinical interviews. Also, similar to within-study findings that younger age is associated with greater likelihood of ACS-induced PTSD, younger mean sample age was related to greater prevalence estimates across studies. Finally, later publication date was associated with lower prevalence rates, though the association was modest, which may reflect reduced psychological trauma due to advances in treatment or the fact that the definition of ACS has broadened in recent years to include less severe disease. However, since we found that the inclusion of patients with less severe ACS (ie, NSTEMI and UA) was unrelated to prevalence estimates, advances in treatment may well be responsible for slightly better post-ACS psychological outcomes in recent years. Timing of PTSD assessment, study location, and sex of the sample were unrelated to prevalence estimates.
Our analyses indicated that ACS-induced PTSD symptoms are associated with an approximate doubling of risk for recurrent cardiac events or mortality. Two of the 3 studies included in the analysis
Important caveats to this review include the fact that only studies with relatively small sample sizes have addressed ACS-induced PTSD, many of which were marginally powered to detect associations with outcomes over relatively short intervals. Furthermore, no study to date has been sufficiently powered to test for potential mechanisms or effect modifiers of the relationship between ACS-induced PTSD and adverse clinical outcomes. Thus, although this meta-analysis was able to quantify more precisely the magnitude of the increased risk of clinical events due to ACS-induced PTSD symptoms, a clear need for additional research remains.
Caveats aside, there is now considerable evidence for a link between PTSD due to various life stressors and subsequent heart disease
The present review has several limitations. First, since we did not include unpublished articles or articles from non–peer-reviewed journals, we are more likely to have excluded negative findings
We were able to locate only one PTSD treatment study in ACS patients: a phase I safety trial of cognitive behavioral therapy for ACS-induced PTSD
While little research exists on treatment of ACS-induced PTSD, a growing body of literature suggests risk factors for PTSD diagnosis and symptom severity after ACS. These risk factors include intense fear
In conclusion, across published studies concerning the prevalence of ACS-induced PTSD and its associations with clinical outcomes, a few general conclusions can be drawn: (1) ACS is a traumatic event for many, and ACS-induced PTSD is relatively common, with approximately 12% experiencing clinically significant PTSD symptoms and 4% meeting full diagnostic criteria for the disorder, and (2) based on a small number of studies, clinically significant symptoms of ACS-induced PTSD appear to double patients’ risk of recurrent cardiac events and mortality. These results identify areas requiring further research; in particular, studies testing treatments for ACS-induced PTSD are needed.