| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
REVIEW ARTICLES |
From the Department of Rehabilitation Psychology, Institute of Psychology, University of Freiburg, Germany (J.B., M.S.); and the Department of Psychosomatic Medicine, University of Goettingen, Germany (C.H.-L.).
Address correspondence and reprint requests to Jürgen Barth, PhD, University of Freiburg, Institute of Psychology, Department of Rehabilitation Psychology, Freiburg, Germany. E-mail: mail{at}juergen-barth.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
OBJECTIVE: To quantify the impact of depressive symptoms (eg, BDI, HADS) or depressive disorders (major depression) on cardiac or all-cause mortality. We analyzed the strength of the relationship, the time dependency, and the differences in studies using depressive symptoms or a clinical diagnosis as predictors of mortality.
METHOD: English and German language databases (Medline, PsycInfo, PSYNDEX) from 1980 to 2003 were searched for prospective cohort studies. Sixty-two publications were identified. The inclusion criteria were met by 29 publications reporting on 20 studies. A random model was used to estimate the combined overall effect as crude odds ratios (OR) or adjusted hazard ratios (HR [adj]).
RESULTS: Depressive symptoms increase the risk of mortality in CHD patients. The risk of depressed patients dying in the 2 years after the initial assessment is two times higher than that of nondepressed patients (OR, 2.24; 1.37–3.60). This negative prognostic effect also remains in the long-term (OR, 1.78; 1.12–2.83) and after adjustment for other risk factors (HR [adj], 1.76; 1.27–2.43). The unfavorable impact of depressive disorders was reported for the most part in the form of crude odds ratios. Within the first 6 months, depressive disorders were found to have no significant effect on mortality (OR, 2.07; CI, 0.82–5.26). However, after 2 years, the risk is more than two times higher for CHD patients with clinical depression (OR, 2.61; 1.53–4.47). Only three studies reported adjusted hazard ratios for clinical depression and supported the results of the bivariate models.
CONCLUSIONS: Depressive symptoms and clinical depression have an unfavorable impact on mortality in CHD patients. The results are limited by heterogeneity of the results in the primary studies. There is no clear evidence whether self-report or clinical interview is the more precise predictor. Nevertheless, depression has to be considered a relevant risk factor in patients with CHD.
Key Words: depression, • coronary heart disease, • mortality, • meta-analysis, • depressive symptoms, • risk factor.
Abbreviations: AMI = acute myocardial infarction;; AP = angina pectoris;; BDI = Beck Depression Inventory;; CABG = coronary artery bypass graft;; CHD = coronary heart disease;; CI = confidence interval;; DIS = Diagnostic Interview Schedule;; DS = Zerssen Self-Rating Scale;; DSM = Diagnostic and Statistical Manual of Mental Disorders;; ECG = electrocardiogram;; f/u = follow-up period;; GMS = Global Mood Scale;; HADS = Hospital Anxiety and Depression Scale;; HDL = high-density lipoprotein;; HPA = hypothalamic–pituitary–adrenocortical axis;; HR (adj) = adjusted hazard ratio;; IL = interleukin;; LVEF = left ventricular ejection fraction;; Medline = database of the U.S. National Library of Medicine;; MI = myocardial infarction;; MD = major depression;; OR = odds ratio;; PSE = Present State Examination;; PsycInfo = database of the American Psychological Association;; PSYNDEX = database of the Center for Psychological Information and Documentation at the University of Trier, Germany;; PTCA = percutaneous transluminal coronary angioplasty;; RR = relative risk;; SBP = systolic blood pressure;; SCID = Structured Clinical Interview for DSM;; SDS = Zung Self-Rating Depression Scale;; TNF = tumor necrosis factor.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Several prospective studies on healthy people have demonstrated the predictive role of depression or depressive symptoms in the development of CHD. The results of two recent metaanalyses support the hypothesis that depression is a risk factor for the development of CHD (5,6). The risk of becoming inflicted with CHD was 60% higher in depressed patients (relative risk [RR], 1.64; confidence interval [CI], 1.29–2.08). It is noteworthy that according to these analyses, clinical depression proved to be a substantially better predictor for the development of CHD in initially healthy people (RR, 2.69; CI, 1.63–4.43) than depressive symptoms (RR, 1.49; 1.16–1.92).
Despite the empiric evidence that depression increases the risk of cardiovascular morbidity and mortality, there is no common accepted model that describes the underlying mechanisms (7,8). Under discussion are both "direct" and "indirect" pathways. "Direct" influences of depression on physiological factors may lead to atherosclerosis or coronary events. More indirectly, depression leads to an increase in classic coronary risk factors, which in turn may cause coronary heart disease. Finally, there may be some underlying background factors influencing the risk for both depression and coronary heart disease. The psychobiologic pathways involve at least three mechanisms. First, depression is associated with autonomic imbalance and activation of the HPA axis (9). Second, depression may lead to dysregulation of immunologic mechanisms (eg, proinflammatory cytokines such as interleukins [IL-1, IL-6] or tumor necrosis factor [TNF]), which are associated with an increased risk of CHD (10–12). Third, coagulation abnormalities and vascular endothelial dysfunction are thought to play an etiologic role in the development or the progression of atherosclerosis in depressed people. High white blood cell counts, fibrinogen, and raised platelet activation contribute to a prothrombotic state, thrombus formation, and myocardial ischemia (13–15).
Indirect pathways refer to psychosocial and behavioral mediators, which correlate with depression and CHD. Depression is associated with poor health behavior, maladaptive coping style, social isolation, and chronic life stress (16). Behavioral risk factors such as smoking, low physical activity, a poor diet, and the failure to adhere to medical recommendations mediate the relationship of depressive disorders with CHD (17,18). Low levels of perceived emotional support and social isolation are related to both CHD and depression (19–21). Psychosocial stressors are known to be predictors of depression in patients with CHD and are also known to be predictors of CHD and the prognosis in CHD patients (22,23).
Some narrative reviews have been published that indicate the impact of depression after a cardiac event (22,24–30). The reviews report depression as having a negative influence on outcome criteria such as cardiac morbidity, cardiac mortality, or total mortality. None of the reviews analyzes the impact of depression on mortality in CHD patients systematically. The objective of our meta-analysis was thus to fill this gap by indicating the extent of this impact. We hypothesized that depressive symptoms and clinical depression have a negative impact on survival in patients with CHD. The following questions were examined: To what extent do depressive symptoms or clinical depression increase mortality in CHD patients? Do effect sizes differ when they are computed for short-term, medium-term, or long-term follow-up periods? Is the impact of depression still present in studies that adjusted for known risk factors?
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Definition of Coronary Heart Disease
The patients included in this meta-analysis sustained an initial CHD event (MI, coronary artery bypass graft [CABG], percutaneous transluminal angiography [PTCA]) or angiographically validated CHD. PTCA and CABG may be diagnosed clearly, whereas the diagnosis of acute MI needs further definite criteria. According to the World Health Organization definition, two of the following indicators are necessary: 1) ischemic chest pain lasting 
20 minutes; 2) modified enzyme patterns (elevated peak creatine phosphokinase); or 3) changes in electrocardiogram (ECG). Trials studying patients with angina pectoris but without a CHD diagnosis were excluded. Studies on populations suffering from cardiovascular diseases in general, which also include cerebrovascular diseases, were excluded (33).
Definition of Depression
Either clinical depression or depressive symptoms had to be assessed at baseline. Clinical depression could be assessed by a standardized clinical interview. Depressive symptoms were to be measured with standardized psychometric scales.
Literature Search and Data Sources
Databases in English and German were searched for relevant studies published between 1980 and 2003. The databases we used were MEDLINE (U.S. National Library of Medicine), PsycInfo (American Psychological Association), and PSYNDEX (a German database of the Center for Psychological Information and Documentation at the University of Trier, Germany). The search strategy used both free-term searches and MeSH term searches. We used the combination of 1) "depression" or "affective disorder;" 2) "coronary disease" or "myocardial ischemia;" and 3) "mortality" or "death" as subject headings or search terms (see the Appendix for detailed information). The search was not restricted by publication language or by publication type. All findings were downloaded and stored in the reference database program EndNote 6.2. The search was complemented by crosschecking references listed in narrative reviews (see "Introduction") and in a recently published systematic review by the senior author (34).
Study Selection and Data Extraction
The search results were assessed by the second author for eligibility. This consisted of scanning the titles and abstracts. Sixty-two eligible papers remained and went into the coding process. Each eligible study was coded according to standardized criteria (35): 1) citation of reference; 2) inclusion and exclusion criteria; 3) description of the patient sample (size of sample, age, sex, subgroups); 4) time frame of measurement of mortality (short-term as >3 months and 
6 months, medium term as >6 months and 2 years; long-term as >2 years); 5) variables measured at baseline (clinical depression/depressive symptoms, cardiac event, cardiac status); 6) type of outcome (cardiac mortality or total mortality); 7) statistics (adjusted hazard ratio [HR], odds ratio [OR]) or raw values; and 8) adjustment for known confounding risk factors (eg, age, sex, physical illness, smoking, hyperlipidemia, hypertension).
In the end, 20 studies published in 29 papers met the inclusion criteria. Thirty-three papers had to be excluded (see the section "Excluded Studies" in the Appendix for reasons of exclusion).
Data Management and Data Analysis
Data were extracted independently by two reviewers (J.B., M.S.). Differences were solved by discussion and ended in one final coding. Two figures are necessary to compute ORs or adjusted HRs in the meta-analysis. One is an estimate of the effect sizes and the other is the standard error of the effect size. The effect size was computed by using the OR or adjusted HR as shown in formula 1. As standard error, we used raw values as in formula 2 or estimated the standard error with the CI reported in the study. 
|
|
|
|
|
|
|
|
a, b, c, d: reported raw data in a cross table
Data management and data analysis were performed with Review Manager 4.2, provided by the Cochrane Collaboration (www.cochrane.org). A random effects model was used to pool the studies. The studies were weighted by the inverse variance method as described in the manual of the review manager (see www.cochrane.org/resources/handbook/section8.pdf). Summary statistics were reported as adjusted HR or OR with a CI of 95%. Values greater than 1 indicate an unfavorable impact of depression or depressive symptoms on mortality. The chi-square value tests for statistically significant heterogeneity among trials; p values lower than .05 indicate heterogeneity; additionally higher I2 values indicate greater variability among trials than would be expected by chance alone (range, 0–100%) (36). The results are clustered for three follow-up periods: short-term (>3 months and 6 months), medium-term (>6 months and 2 years), and long-term (>2 years). Funnel plots of all outcome measures can be found in the Appendix (Figures 1a, 2a, and 3a).
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
|
Type of Participants
Most of the studies (n = 15) included patients with myocardial infarction (studies 3, 5, 8–10, and 12–20). Two studies included patients after coronary bypass surgery (studies 2 and 6), one study included patients after PTCA (study 1), and three studies included patients with a variety of diagnoses (MI, CABG, PTCA) or angiographically validated CHD (studies 4, 7, and 11). Most of the patients in the studies were male. One study reported only on male patients (study 15). The mean age ranged from 53.6 to 74.5 years (the minimum was 19 and the maximum 90).
Measurement of Depression
Four studies measured clinical depression (studies 4–5, 14, and 19), 13 depressive symptoms (studies 1–2, 6–8, 10–13, 15–17, and 20), and two both clinical depression and depressive symptoms (studies 3 and 9). One study assessed depression by way of self-report or clinical diagnostic interview and set up an index for depression (study 18). This study was coded as measuring depressive symptoms.
Three studies used DSM-III-R criteria to assess clinical depression (studies 3–4 and 18) without any specification of the diagnostic procedure. One study assessed clinical depression by using DSM-IV criteria with a specific depression interview developed for this purpose (study 5). The Diagnostic Interview Schedule (DIS) was used in two studies (studies 9 and 14), one study measured clinical depression with the Schedule for Affective Disorders and Schizophrenia (study 19), and another one used a modified SCID (study 3).
The Beck Depression Inventory was used most widely to measure depressive symptoms (studies 3, 6, 9, 12, 16, 18, and 20). Other diagnostic instruments were the Zung Self-Rating Depression Scale (SDS; studies 1 and 20), the Hospital Anxiety and Depression Scale (HADS; studies 11 and 17), and the Millon Behavioral Health Inventory (studies 7 and 8). One study used the Center of Epidemiologic Studies Depression Scale (CES-D; study 2), one used three items from the psychosocial questionnaire of the ASSET Study (study 13), and one used the Zerssen Self-Rating Scale (study 15).
Study Outcome
The outcome of the studies was either cardiovascular mortality, reported in 12 studies (studies 1–2, 6–10, 12, 15–16, 18, and 20), or all-cause mortality, reported in seven studies (studies 3–5, 11, 13–14, and 17). The length of follow up varied from 3 to 4 months (short follow up in studies 1, and 18–19) to 10 years (study 20). Six studies reported data for 5 years or longer (studies 2, 7–8, 10–11, and 20).
Statistics
Eleven studies reported crude ORs (studies 2–4, 7–8, 13–14, and 16–19). In three studies, the authors reported HRs adjusted for known risk factors (smoking, age, sex, hypertension, hyperlipidemia, diabetes mellitus, body mass index) or sociodemographic characteristics and cardiac parameters at baseline (low left ventricular ejection fraction, previous AP, Killip Class, nonQ-wave-MI) (studies 5, 11–12, and 16). Six studies reported crude and adjusted values (studies 1, 6, 9–10, 15, and 20). Multivariate ORs or univariate HRs were not provided in a sufficient number of studies to justify their metaanalytic aggregation.
Effect of Depressive Symptoms
The results are presented as crude estimated effects of depressive symptoms on mortality (Figure 1) and after adjustment for known risk factors (Figure 2). The nonadjusted OR of 2.24 (CI, 1.37–3.60) we found in short- and medium-term follow-up studies supports our hypothesis that depressed patients have higher rates of mortality. The results from long-term follow-up studies also indicate higher mortality but show a lower OR of 1.78 (CI, 1.12–2.83). The effect size for long-term mortality is based on heterogeneous results (p = .0002; I2 = 71.3%), whereas short- and medium-term effects are based on more homogeneous results in primary studies (p = .09; I2 = 45.7%).
|
|
After adjustment for known cardiac risk factors, depressive symptoms still show a significant impact on mortality (HR [adj], 1.76; CI, 1.27–2.43). This estimation is also based on heterogeneous results (p = .002; I2 = 71.4%). We were not able to cluster the results depending on follow-up length because of the low number of primary studies.
Effect of Clinical Depression
The effect of clinical depression was assessed predominantly in the short and medium term with unadjusted ORs. Only three studies reported an adjusted HR. Figure 3 shows a nonsignificant short-term effect of clinical depression on mortality (OR, 2.07; CI, 0.82–5.26; I2 = 64.8%). The OR resulting from medium-term studies is significant (OR, 2.61; CI, 1.53–4.47). The result for the medium-term follow-up period is based on homogenous effects in the primary studies (p = .89; I2 = 0%). Contrary to the hypothesis that the prognostic impact of initial depression declines over time, we find a higher OR in studies with longer follow-up periods. As mentioned previously, only three studies performed an adjusted risk analysis. In the short term, such an analysis results in an adjusted HR of 4.29 (CI, 3.14–5.86) (47). Connerney et al. (43) also reported a significant adjusted HR of 2.31 (CI, 1.17–4.56) after 12 months. In the only published study with a follow-up period of more than 2 years, Carney et al. (42) reported an adjusted HR of 2.4 (CI, 1.2–4.7).
|
In short- and medium-term follow-up studies, the effect sizes of the adverse impact of depressive symptoms were slightly higher for cardiac mortality (OR, 2.88; CI, 1.47–5.63) than for total mortality, as shown in Figure 1. The pooled estimate was still homogeneous (p = .10; I2 = 51.7%). The effect sizes of the adjusted model were also slightly stronger (HR [adj], 2.07; CI, 1.31–3.27) than the results shown in Figure 2. The fact that we limited the data set to studies with an outcome of cardiac mortality contributed only marginally to the homogeneity of the underlying results. On the whole, the data set was still heterogeneous (p = .01; I2 = 66.4%).
In the sensitivity analysis that tested the effects of clinical depression on cardiac mortality, only two studies with a short-term follow up remained in the data set. One of these studies found a nonsignificant negative result (study 19), and the other showed a very strong influence (study 9). For studies with a medium-term follow up, the OR remained almost the same (OR, 2.62; 1.51–4.53), even after the exclusion of one study (study 4); this pooled effect size is based on homogenous results (p = .73; I2 = 0%).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
The risk of mortality is at least two times higher in the short and medium term for patients suffering from CHD and comorbid clinical depression. These results were based on bivariate statistics. Analyses with adjusted Cox regression models confirmed these effect sizes.
Heterogeneity in primary studies was quite an issue in this meta-analysis, especially for studies that looked at depressive symptoms rather than clinical depression. Interestingly, even when depressive symptoms were measured with the same instrument (eg, BDI), different effects emerged. Not even the length of the follow-up period seemed to contribute to making the basis of the underlying studies more homogeneous. It was not sufficient to limit the analysis to cardiac mortality as an outcome to reduce heterogeneity either. One possible explanation for the heterogeneity of the adjusted analyses may be the selection of risk factors, which varied greatly from study to study. One possible solution to this problem would be to pool and reanalyze the original data of all included studies.
A closer look at the studies with negative or null effects (studies 13, 16, and 19), which were the ones that contradicted the total effect, revealed no shared study characteristics. There was not any indication that methodologic biases could have affected the results of these three studies either. This is at least true of two of the three studies (studies 16 and 19). In the study by Jenkinson et al. (study 13), only the assessment of depressive symptoms by three unvalidated self-report items is quite questionable and may explain the null result.
Likewise, the closer inspection of the four studies with large effect sizes indicated no lack of methodologic quality (studies 8–9, 15, and 18). Although the assessment of depressive symptoms in the study by Denollet et al. may be questioned (study 8), the other studies used widely accepted scales for measuring depressive symptoms. The use of a clinical interview or self-report to index depression in the study by Romanelli et al. was certainly a disadvantage, but it does not explain the high ORs.
As mentioned previously, our meta-analysis found that depression has an adverse effect after a first manifestation of CHD. This raises the question as to whether patients who are depressed after a cardiac event already had more depressive symptoms before the event. The data in our analysis cannot answer this question. Retrospective data from CHD patients about earlier episodes of major depression showed that one of four patients had a lifetime diagnosis (50). Ongoing research is investigating the effect of affectivity before the clinical manifestation of CHD (32), which would confirm the unfavorable effect.
In this meta-analysis, we differentiated between studies with self-report measures and those with clinical interviews. Self-report measures are very useful for screening patients, planning, and evaluating psychotherapeutic or drug interventions in clinical practice. Additionally, we know that clinical depression and depressive symptoms (eg, BDI 10) share 60% to 80% of common variance (50,67). A clinical diagnosis cannot be made by self-report questionnaires, and this may be a disadvantage. On the other hand, self-rating scales may be more sensitive in detecting subthreshold disorders, and we found that the low cutoff of 10 in the BDI is sufficient for indicating an increased mortality risk. Hence, for epidemiologic purposes, it is difficult to demonstrate the superiority of one or the other measure of depression. Accordingly, we found no clear prognostic difference between studies that defined depression through self-report and those that did so with a clinical interview.
Limitations
Our review included only published studies with sufficient data. We did not include abstracts because they cannot give reliable information on inclusion and outcome criteria. The adjusted OR reported in one publication of the EPPI study (48) could not be included in our analysis, because it appeared inappropriate to mix results of logistic regression and Cox regression models in one meta-analysis. Like in every meta-analysis, we have to take into account a publication bias. We found some hints for a publication bias in studies assessing depressive symptoms (see funnel plots in the Appendix, Figures 1a, 2
b, and 3a). The funnel plot of studies assessing the impact of depressive disorders seems appropriate. It is possible that small nonsignificant studies of depressive symptoms are not submitted because they have lower chances of being accepted. We did not assess the quality characteristics of each study as recommended for randomized, controlled studies (68). Therefore, we did not use a sum score for study quality either (for limitations of sum scores, see (69)). Some results of this meta-analysis are based on heterogeneous effects of the primary studies. Therefore, the pooled estimates were not an accurate summary estimate in all reported results.
Future Research
Missing Studies
There are sufficient data on the influence of depressive symptoms on total or cardiac mortality in the long term. Results on the long-term impact of clinical depression are rather limited. Data from the ENRICHD trial (study 5) give initial hints on this topic, but further research is necessary to obtain more stable effects on the impact of depressive disorders on mortality in CHD patients.
Heterogeneity of End Points
Further research should report cardiac events, cardiac mortality, and total mortality as end points. Many studies mixed these criteria by reporting results on cardiac prognosis. An aggregated index is a good way to deal with insufficient statistical power. On the other hand, the comparison, reanalysis, and aggregation of these studies are problematic because of heterogeneity in the end points.
Adjustment of Risk Factors
In reporting data, it would be helpful to get results of nonadjusted ORs and adjusted HRs. Some results might reach the level of significance only before or after adjustment. For transparency, it would be helpful to obtain both results. Another point is the selection of risk factors. Authors focused mainly on somatic parameters that may increase physical distress and decrease cardiac functioning. On the other hand, established risk factors like smoking or lipids should also be controlled for the prognosis of the CHD. Initial cessation of smoking is quite common in CHD patients. However, on the other hand, approximately 50% of those who stop smoking begin smoking again within 1 year after the event. Recent epidemiologic studies show evidence for a link between smoking behavior and depression (70). To get more reliable information on the impact of depression on CHD, it would be helpful to adjust the analysis for risk factor information, not only in the initial phase, but also during the follow up.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
#2 myocardial infarction (205,412 records)
#3 angina pectoris (37,630 records)
#4, #1, or #2 or #3 (267,536 records)
#5 affective disorder (11,517 records)
#6 depression (256,320 records)
#7 depressive mood (1533 records)
#8 depressive symptoms (19,577 records)
#9, #5, or #6 or #7, or #8 (265,290 records) #10, #4, and #9 (16,683 records) #11 mortality (491,759 records) #12, #10, or #11 (6038 records) 
6038 hits in Medline, Premedline, BIOSIS, and Journals@Ovid
Search History in PsycInfo (1980–2003)
#1 TI CHD or AB CHD or MJ CHD (1531 records)
#2 TI myocardial infarction or AB myocardial infarction or MJ myocardial infarction (1531 records)
#3 TI angina pectoris or AB angina pectoris or MJ angina pectoris (247 records)
#4 (S1 or S2 or S3) (3048 records)
#5 TI affective disorder or AB affective disorder or MJ affective disorder (12,235 records)
#6 TI depression or AB depression or MJ depression (89,077 records)
#7 TI depressive mood or AB depressive mood or MJ depressive mood (611 records)
#8 TI depressive symptoms or AB depressive symptoms or MJ depressive symptoms (7176 records)
#8 (S5 or S6 or S7 or S8) (97,783 records)
#9 TI mortality or AB mortality or MJ mortality (7053 records)
#10 (S8 and S9) (81 records)
81 hits in PsycInfo
Search History in PSYNDEXplus–Lit.& AV (1980–2003)
#1 coronary heart disease (99 records)
#2 myocardial infarction (249 records)
#3 angina pectoris (41 records)
#4, #1, or #2 or #3 (368 records)
#5 affective disorder (159 records)
#6 depression (7031 records)
#7 depressive mood (112 records)
#8 depressive symptoms (291 records)
#9, #5, or #6, or #7, or #8 (7171 records)
#10 mortality (461 records)
#11, #4, and #9 (43 records)
#12, #10, and #11 (2 records) 2 hits in PSYNDEX
|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
CONFLICT OF INTEREST
J.B. is a principal investigator in an intervention study on depressed CHD patients. C.H.L. did epidemiologic research on the topic of this review. One study was included in the meta-analysis.
Received for publication April 7, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. C. Huffman, L. T. Park, C. A. Welch, A. A. Nierenberg, J. L. Januzzi Jr., and S. R. Pomerantz Case 14-2008 -- A 78-Year-Old Man with Anergia and Anhedonia Associated with Cardiovascular Surgery N. Engl. J. Med., May 8, 2008; 358(19): 2051 - 2059. [Full Text] [PDF]
|
|
|
|
 |
|
 L. Wouts, R. C. O. Voshaar, M. A. Bremmer, J. K. Buitelaar, B. W. J. H. Penninx, and A. T. F. Beekman Cardiac Disease, Depressive Symptoms, and Incident Stroke in an Elderly Population Arch Gen Psychiatry, May 1, 2008; 65(5): 596 - 602. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Dickens, L. McGowan, C. Percival, B. Tomenson, L. Cotter, A. Heagerty, and F. Creed New Onset Depression Following Myocardial Infarction Predicts Cardiac Mortality Psychosom Med, May 1, 2008; 70(4): 450 - 455. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Lett, S. Ali, and M. Whooley Depression and Cardiac Function in Patients With Stable Coronary Heart Disease: Findings From the Heart and Soul Study Psychosom Med, May 1, 2008; 70(4): 444 - 449. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Rutledge, S. E. Linke, M. B. Olson, J. Francis, B. D. Johnson, V. Bittner, K. York, C. McClure, S. F. Kelsey, S. E. Reis, et al. Social Networks and Incident Stroke Among Women With Suspected Myocardial Ischemia Psychosom Med, April 1, 2008; 70(3): 282 - 287. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-K. Wong, E.W. Tang, P. Herbison, B. Birmingham, L. Barclay, and S.Y.F. Fu Pre-existent depression in the 2 weeks before an acute coronary syndrome can be associated with delayed presentation of the heart attack QJM, February 1, 2008; 101(2): 137 - 144. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Vaccarino, C. McClure, B. D. Johnson, D. S. Sheps, V. Bittner, T. Rutledge, L. J. Shaw, G. Sopko, M. B. Olson, D. S. Krantz, et al. Depression, the Metabolic Syndrome and Cardiovascular Risk Psychosom Med, January 1, 2008; 70(1): 40 - 48. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Zisook, L. Montross, J. Kasckow, S. Mohamed, B. W. Palmer, T. L. Patterson, S. Golshan, I. Fellows, D. Lehman, and E. Solorzano Subsyndromal Depressive Symptoms in Middle-Aged and Older Persons with Schizophrenia Am J Geriatr Psychiatry, December 1, 2007; 15(12): 1005 - 1014. [Abstract] [Full Text] [PDF]
|
|
