Mean Primary Care Evaluation of Mental Disorders Brief Patient Health Questionnaire (PHQ) scores by age and sex.
Prevalence of depression by age and sex.
Distribution of depression rate by age in decile distribution (10% of data in each age range) and sex.
Odds ratio (95% confidence intervals) plot for multivariable logistic regression modeling Primary Care Evaluation of Mental Disorders Brief Patient Health Questionnaire depression score of 10 or higher. Age-sex effect controlling for site, demographic factors (race, marital status, and level of education), patients' perceived economic burden, history of smoking, medical history (diabetes, hypertension, hypercholesterolemia, cerebrovascular accidents, angina, myocardial infarction, coronary artery bypass grafting, percutaneous coronary intervention, peripheral vascular disease, lung disease), and clinical status on admission (congestive heart failure, systolic blood pressure, renal failure).
Mallik S, Spertus JA, Reid KJ, Krumholz HM, Rumsfeld JS, Weintraub WS, Agarwal P, Santra M, Bidyasar S, H. Lichtman J, Wenger NK, Vaccarino V, PREMIER Registry Investigators. Depressive Symptoms After Acute Myocardial InfarctionEvidence for Highest Rates in Younger Women. Arch Intern Med. 2006;166(8):876-883. doi:10.1001/archinte.166.8.876
Depression is common in patients hospitalized with acute myocardial infarction (AMI). In the community, younger women are uniquely prone to depression. Whether younger women are also more likely to have depression during hospitalization with AMI is unknown.
A total of 2498 AMI patients (1284 patients ≤60 years; 814 women and 1684 men) were enrolled from 19 US centers in the Prospective Registry Evaluating Outcomes After Myocardial Infarction: Events and Recovery (PREMIER) study between January 2003 and June 2004. Depression was assessed at the time of hospitalization and was defined as a Primary Care Evaluation of Mental Disorders Brief Patient Health Questionnaire (PHQ) score of 10 or higher.
Younger (≤60 years) patients had higher mean PHQ scores than older patients (6.4 vs 5.0; P<.001) and women had higher mean PHQ scores than men (6.8 vs 5.2; P<.001). When stratified by both age and sex, younger women had the highest PHQ scores (8.2; P<.001 for the sex-age interaction). The prevalence of depression was 40% in women 60 years or younger, 21% in women older than 60, 22% in men 60 or younger, and 15% in men older than 60. In a logistic model adjusted for study center, race, medical history, and coronary heart disease risk factors, the odds of depression for women 60 years or younger were significantly higher than for the other sex-age groups and were 3.1 times higher than the reference group of men older than 60 years.
The prevalence of depression is high in younger women with AMI. Because depression after AMI has been associated with adverse outcomes, younger women, a high-risk group compared with men, may particularly benefit from aggressive screening and treatment of post-AMI depression.
Depression is prevalent in cardiovascular populations and is associated with a broad range of adverse outcomes. Among patients with acute myocardial infarction (AMI), depressive symptoms independently predict cardiac mortality, hospitalization, worse patient health status, and higher costs of care even in the absence of a clinical diagnosis of depression.1,2 It is important to identify AMI patients at high risk for depression, so these patients may be targeted for depression screening and treatment.
Recent studies3 have reported remarkably higher mortality rates in younger women with AMI compared with men, a finding not fully explained by medical history, clinical severity, and hospital treatments and procedures. In the community, depression is most common in women and younger adults, making this disorder particularly prevalent among younger women.4 Rates of depression in women hospitalized with AMI are 1.5 to 2 times higher than in men,5 but little is known about whether younger women with AMI are more affected by depression than other groups. If younger women with AMI have a higher prevalence of depression, this may contribute to the disproportionate mortality rates of this group. In addition to its prognostic significance, however, depression is an important illness in its own right, and prompt identification and treatment of depression in the post-AMI period is strongly indicated for all patients, regardless of their age or sex, to ensure their appropriate clinical management.
The purpose of this study was to assess differences in the prevalence of depression according to sex and age in a multicenter registry of AMI patients. We hypothesized that, among patients with AMI, depression is more common in younger women than in other demographic groups and that this difference is not explained by medical history and clinical severity characteristics.
Study participants were recruited prospectively between January 2003 and June 2004 at 19 US study sites as part of the Prospective Registry Evaluating Outcomes After Myocardial Infarction: Events and Recovery (PREMIER) study.6 The research protocol received institutional review board approval at each participating center and was monitored by each local institutional review board.
Consecutive patients admitted with AMI were eligible to participate in the PREMIER study if they were 18 years or older, had elevated cardiac enzyme levels within 24 hours of hospital admission, and demonstrated other supportive evidence of AMI, including at least one of the following: prolonged (>20 minutes) ischemic signs or symptoms, at least one electrocardiogram with ST elevation or ST depression in 2 or more consecutive leads, or other clinical evidence of AMI. Cardiac enzyme levels were considered elevated if creatine kinase was more than the upper level of normal with the creatine kinase–MB more than twice the upper level of normal on the same blood draw or troponin I or T was greater than the AMI threshold at each institution. Patients were excluded if they had been transferred to the recruiting institution from another facility more than 24 hours after their original presentation, refused or were unable to provide informed consent, or did not speak English or Spanish.
In addition to a comprehensive medical record review, patients were interviewed during the AMI admission to collect information on sociodemographic, behavioral, and psychosocial factors. Depressive symptoms were assessed in the baseline interview by means of the 9-question Primary Care Evaluation of Mental Disorders Brief Patient Health Questionnaire (PHQ).7,8 The PHQ is a validated questionnaire with a sensitivity of 88% and a specificity of 88% for major depression with a cutoff of 10 points or higher.8 Not only is its diagnostic validity comparable to clinician-administered assessments, but it also yields an index of depressive symptom severity.7 For each of the 9 depressive symptoms, patients indicated whether, during the previous 2 weeks, the symptom had bothered them “not at all,” “several days,” “more than half the days,” or “nearly every day,” yielding a score of 0 to 3. The PHQ score ranges from 0 to 27. Consistent with previous studies, we classified patients as depressed if they scored 10 or higher on the PHQ, which corresponds to a level of at least moderate depression and represents the minimum number of symptoms required for the diagnosis of major depression.9 A history of depression was defined as currently receiving medications or counseling for depression. Economic burden was defined as not taking medication owing to cost or avoiding health care due to cost. We used the Seattle Angina Questionnaire (SAQ), a 19-item questionnaire, to assess 3 components of health status: symptom burden (2-item SAQ angina frequency scale), functional status (9-item SAQ physical limitation scale), and disease-specific quality of life (3-item SAQ quality-of-life scale), based on previous studies.10,11 Responses were scored from 0 to 100, with higher scores indicating better health (ie, less symptom burden, less physical limitation, and better quality of life).10,11
All data were entered at each participating institution and transmitted to the national data coordinating center using a Web-based data entry interface. Information entered in the study forms was double checked for completeness and edited, in addition to a wide variety of range and logic checks built into the data entry system.
We compared study factors between men and women using t tests for continuous variables and χ2 or Fisher exact tests for categorical variables, as appropriate. Depression was examined both as a continuous variable (number and severity of depressive symptoms) and as a dichotomous variable (prevalence of major depression as defined by a PHQ score ≥10). Because we hypothesized that sex disparities in the prevalence of depression differ by age, we stratified patients into groups of 60 years or younger or older than 60 years, and the same cut point was used to test for the interaction between age and sex. In essence, this approach sought to explicitly describe the age-by-sex interaction of significant depressive symptoms at the time of AMI. We chose the cut point of 60 years, because this was the median age in our study population and also because women younger than this age have been previously shown to have higher rates of adverse events after AMI compared with men.3,12 To validate this approach, we conducted a sensitivity analysis to test other break points for stratification of age, using age group cutoffs of 50, 55, 60, and 65 years. Bivariate analyses were also conducted to examine the association of baseline characteristics with depression status in women and men separately.
We used a logistic regression model to compare the prevalence of depression, according to sex and age, after adjusting for key clinical variables, including study site; demographics (race, marital status, and level of education); patients' perceived economic burden; history of smoking; medical history, including diabetes, hypertension, hypercholesterolemia, peripheral vascular disease, lung disease, cerebrovascular accident, angina, MI, coronary artery bypass grafting, or percutaneous coronary intervention; and clinical status on admission, including congestive heart failure, systolic blood pressure, and renal failure. Since the degree of depression may change if the time between AMI and interview varied, we also adjusted for the time to interview in our logistic model. Logistic regression model fit and discrimination were assessed using the Hosmer and Lemeshow Goodness-of-Fit Test and the model C-statistic, respectively. All tests for statistical significance were 2-tailed with an α level of .05. Analyses were performed using SAS version 8.02 (SAS Institute Inc, Cary, NC) and R version 2.1.1 (R Foundation for Statistical Computing, Vienna, Austria).
In the study period, 3953 patients met the eligibility criteria for participation. Of these, 2498 patients (63%) were enrolled and interviewed, including 814 women and 1684 men. Patients consenting to enroll were interviewed as soon as they were stable. Time to interview varied from 0 to 53 days from hospital arrival, with a median value of 2.0 days and a mean of 2.9 days with a standard deviation of 2.7 days. Compared with nonenrolled patients, enrolled patients were younger, more often white, more likely to have ST-elevation MI and a family history of coronary disease, and more likely to be smokers. No sex differences were found in the proportion of eligible patients who were enrolled.
In both age groups, women were more likely than men to be African American and to have a lower household income and Medicaid for insurance (all P<.001). Women in both age groups were less likely than men to be married, to have completed high school, or to be employed (Table 1).
There were age-related differences between women and men in their clinical characteristics at admission (Table 1). In both age groups, women had a higher rate of hypertension and lower SAQ scores, indicating more chest pain symptoms and poorer quality of life. Younger, but not older, women had more comorbid conditions compared with men, including diabetes, obesity, congestive heart failure, and a higher Killip classification, but they were less likely to have ST-elevation MI.
Major depression, as defined by a PHQ score of 10 or higher, was common: 22.3% of the patients were depressed. In both women and men, younger age, African American race, unmarried status, lower social support, and more unfavorable socioeconomic indicators were associated with depression (defined as a PHQ score ≥10) (Table 2). Depressed patients had more comorbidities and worse health status than nondepressed patients, including a more frequent history of AMI, congestive heart failure, and diabetes, as well as lower SAQ scores. Depressed women, but not depressed men, were more likely to be smokers and to have a history of hypercholesterolemia, whereas depressed men, but not depressed women, were more likely to have history of hypertension, revascularization procedures, and lower ejection fraction. Only approximately 27% of depressed patients (31% women and 24% men) had a history of depression. Depressed patients were more likely to be discharged with a prescription for antidepressants (18.5% vs 7.24%; P<.001) compared with nondepressed patients.
Examination of depressive symptoms as a continuous variable (mean PHQ scores) by age demonstrated that younger patients (≤60 years) had significantly higher PHQ scores compared with older patients (mean ± SD score, 6.4 ± 5.0 vs 5.0 ± 5.9; t = 6.10; P<.001), indicating more depressive symptoms. Mean PHQ scores by sex showed that women had significantly higher PHQ scores compared with men (6.8 ± 5.8 vs 5.2 ± 5.3; t = 6.68; P<.001). When data were stratified by both age and sex, women had higher mean ± SD PHQ scores compared with men in both the younger (8.2 ± 6.3 vs 5.7 ± 5.6; t = 6.23; P<.001) and older groups (5.8 ± 5.1 vs 4.5 ± 4.9; t = 4.31; P<.001). Importantly, younger women had the highest PHQ scores of all the other age and sex groups (8.2; P<.001 for the sex-age interaction) (Figure 1). The correlation coefficient for age and PHQ score was −0.118 for the overall population, −0.108 for men, and −0.207 for women.
The prevalence of major depression (PHQ score ≥10) was higher in women than in men in both younger (≤60 years) and older patients (Figure 2), but of all groups, younger women had the highest prevalence of depression (40%). When sex differences were examined according to age deciles, women had higher depression rates than men over all ranges of age, but the sex difference was most marked among patients 60 years or younger. Figure 3 demonstrates the distribution of depression rate by deciles (10% of data distribution within each age range).
In a logistic model that adjusted for study center, demographic factors, patients' perceived economic burden, behavioral factors, medical history, and clinical status on admission, the odds of depression for women 60 years or younger remained significantly higher than in the other sex-age groups and were 3.1 times higher (95% confidence interval, 2.2-4.5) than the reference group of men older than 60 years (Figure 4). Younger men (≤60 years), but not older women, had significantly higher odds of depression compared with older men (1.5; 95% confidence interval, 1.1-2.1; P=.005; and 1.4; 95% confidence interval, 1.0-2.0; P = .08; respectively). These results were almost identical after controlling for time to interview. The correlation coefficient between time to interview and depression was negligible (Pearson r = 0.0192). Sensitivity analyses that used age cutoffs of 50, 55, 60, and 65 years yielded essentially similar results. The odds of depression for younger women were similar for all age group cutoffs and remained significantly higher than the reference group of older men.
Our study confirms that depression is common in AMI patients and that overall women have a higher prevalence of depression than men. For the first time, however, we found that younger women (≤60 years) have a remarkably higher rate of depression at the time of AMI compared with the other demographic groups. We found that 40% of these women had a PHQ score of 10 or higher, indicating at least moderate depression. Even after adjusting for various demographic, behavioral, medical history, and clinical factors and time to interview, the odds of depression for women 60 years or younger remained significantly higher than for the other groups and were more than 3 times higher than for the reference group (men >60 years of age). In contrast, older women had similar odds of depression compared with younger and older men.
The reasons why younger women have such a high prevalence of depression after AMI are not known. It is possible that sociodemographic factors, comorbid conditions, and health status might explain sex differences in rates of depression after AMI. Indeed, we found that depressed patients and women, particularly younger women, were more likely to be African American, with more unfavorable socioeconomic status compared with nondepressed patients. Depressed patients also had more comorbidity, worse health status, and a higher Killip classification, indicating more symptoms and greater severity of ventricular dysfunction. However, even after adjustment for these factors, younger women remained significantly more likely to be depressed than the other demographic groups.
It has been suggested that women are more likely to recall or acknowledge depression compared with men13 and that they focus on interpersonal relationships in a ruminative way that could be depressogenic.14 Although these characteristics could potentially explain higher rates of depression in women, they should not explain why younger women have higher rates of depression than older women. Differential sex roles and exposure to social and environmental stressors, such as poverty, lower level of education, responsibilities both at work and home, single parenthood, and caring for children and aging parents, could theoretically contribute to a higher preponderance of depression in younger women, who may have greater exposure to these stressors compared with other groups.15
Some authors speculate that estrogen cyclicity may predispose women to the neurotoxic processes of stress hormones, thus playing a role in increased sensitivity to stress and depression in younger women.16 Recently, the metabolic syndrome has been linked to depression.17 Although depression may lead to insulin resistance and the metabolic syndrome through the neurohormonal activation typical of depressive mood,18 the metabolic syndrome may also predispose individuals to depression. Although we adjusted for body mass index and other risk factors that are part of the metabolic syndrome, this may not have completely accounted for the neuroendocrine effects of insulin resistance on emotion and behavior.
It is important to recognize depression after AMI. In addition to being an important illness in its own right, depression during hospitalization with AMI confers 3 to 5 times higher adjusted odds of death by 6 months after AMI after controlling for a wide range of clinical predictors of mortality, and its prognostic role is at least equivalent to that of established prognostic indicators, such as left ventricular dysfunction and a history of AMI.2,5 Depressed mood predicts mortality even in the absence of a clinical diagnosis of major depression, with a clear dose-response relationship between severity of depressive symptoms and mortality risk.19 In addition to mortality, depression is associated with longer hospital stays and predicts worse symptomatic, psychological, and social outcomes at 3 and 12 months.20
Younger women, the group with the highest levels of depression in our study, also have higher rates of adverse events and mortality after AMI compared with men of similar age.3 These outcome differences are only partially accounted for by coexisting conditions, clinical characteristics, and early management.3 Although the current literature has been unable to provide a complete explanation for these outcome differences, no study has evaluated the potential role of depression. Given the high rate of depression in younger women, future studies should evaluate whether the effects of depression on health outcomes differ by sex and whether depression explains the higher adverse outcomes after AMI in women compared with men.
Depression was largely unrecognized and untreated in our study. Approximately 73% of depressed patients did not have a history of depression at the time of hospitalization with AMI, suggesting that depression was either unrecognized or developed for the first time after AMI. However, only 18% of the depressed patients were discharged with prescriptions for antidepressants. Although treatment of depression with cognitive behavioral therapy was not associated with reducing mortality following AMI in the Enhancing Recovery in Coronary Heart Disease trial, in a post hoc analysis antidepressant use was associated with a significantly lower risk of the main study end point (death or recurrent AMI).21 Secondary analyses of a safety study of the antidepressant sertraline in patients with acute coronary syndromes showed nonsignificant reductions in mortality and cardiac end points, but the sample size was inadequate to evaluate such outcomes.21,22 Despite these inconclusive results, depression in patients with AMI remains a serious condition and deserves treatment. Patients with moderate-severe depressive symptoms may warrant treatment or at least close follow-up and monitoring of their mood, with initiation of treatment for depression if symptoms persist.
In our study, it was not feasible to obtain a clinical diagnosis of major depression, and since the PHQ measures present-state depressive symptoms, we do not know if these depressive symptoms are secondary to the AMI or began before the coronary event. However, the PHQ has high sensitivity and specificity for major depression, and multiple studies have demonstrated that even minimal symptoms of depression measured 2 to 15 days after AMI, no matter what their cause or natural history might be, are associated with adverse outcomes after AMI, even in the absence of a diagnosis of major depression.1 Selection bias must also be considered because we were only able to include individuals who could provide informed consent and undergo an interview. Approximately 36% of the eligible patients were not enrolled, and older patients were less likely to be enrolled than younger patients, a finding consistent with previous studies.5 However, no sex differences were apparent in the proportion of eligible patients who were enrolled. Although we do not have data on whether refusal rate differed by depression status, most likely depressed patients were also less likely to participate; therefore, these exclusions should result in underrepresentation of women with depression in our study and therefore an underestimation of the true association between female sex and depression. Finally, since this is an observational study, unmeasured confounding is a possibility. However, we attempted to control for potential confounding as much as possible by adjusting for a wide array of measured demographic, social, and clinical factors.
Our study found that 2 in 5 women 60 years or younger with AMI are depressed at the time of hospitalization. This rate was significantly higher when compared with older women and men. Although the reasons for these differences need further study, based on our results, clinicians should be aware that younger women have a high susceptibility for being depressed after AMI. Although screening for depression is warranted in all AMI patients, screening should be particularly aggressive in younger female patients with AMI.
Correspondence: Susmita Mallik, MD, MPH, MS, Department of Medicine, Division of General Medicine, Emory University School of Medicine, Room 473, Faculty Office Building, 49 Jesse Hill Jr Dr, Atlanta, GA 30303.
Accepted for Publication: October 24, 2005.
Author Contributions: Drs Mallik and Spertus had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
PREMIER Registry Investigators:Mid America Heart Institute, Kansas City, Mo: John Spertus, MD, MPH, Carole Decker, RN, PhD, Philip Jones, MS, Kimberly Reid, MS; Baptist Health System, Little Rock, Ark: Gary Collins, MD; Barnes Jewish Hospital/Washington University, Saint Louis, Mo: Richard Bach, MD; Beth Israel–Deaconess Medical Center/Harvard University, Boston, Mass: David Cohen, MD, MSc; Denver General Health System, Denver, Colo: Frederick Masoudi, MD, MSPH; Denver VA Medical Center, Denver: John Rumsfeld, MD, PhD; Duke University, Durham, NC: Eric Peterson, MD, MPH; Emory University Atlanta, Ga: Susmita Mallik, MD, MPH, MS, Viola Vaccarino, MD, PhD, William S. Weintraub, MD; Henry Ford Medical Center, Detroit, Mich: Sanjaya Khanal, MD, Jane Jie Cao, MD, MPH; Kaiser Permanente, Denver: David Magid, MD, MPH; MeritCare, Fargo, ND: Wallace Radke, MD, Mohamed Rahman, MD; Sentara Health System (both Sentara and Sentara Lee Hospitals), Norfolk, Va: John E. Brush, Jr, MD; Stanford University/Palo Alto VA Medical Center, Palo Alto, Calif: Paul Heidenreich, MD; Swedish Medical Center, Seattle, Wash: Timothy Dewhurst, MD; Truman Medical Center and the University of Missouri–Kansas City: Annette Quick, MD; University of Alabama, Birmingham: John Canto, MD; University of Colorado Health System, Denver: John Messenger, MD; Yale University, New Haven, Conn: Harlan Krumholz, MD, SM.
Financial Disclosure: Drs Spertus, Krumholz, and Rumsfeld have received consulting fees or honoraria from CV Therapeutics and United Healthcare, and Dr Rumsfeld has also received consulting fees or honoraria from Pfizer. Drs Spertus and Krumholz have received research grants from CV Therapeutics, and Dr Vaccarino received consulting fees from CV Therapeutics. Dr Rumsfeld received the VA Health Services Advanced Research Career Development Award (ARCD-98342-2). Dr Spertus holds intellectual property rights to the Seattle Angina Questionnaire.
Funding/Support: This study was supported by the Emory University General Clinical Research Center, Atlanta, Ga (MO1-RR00039), and by a grant from CV Therapeutics, Palo Alto, Calif. Dr Mallik is supported by American Heart Association Scientist Development Award 0630084N and grant K12RR17643 from the National Institutes of Health. Dr Vaccarino is supported by grants K24HL077506, R01 HL68630, and R01 AG026255 from the National Institutes of Health; Dr Rumsfeld by a VA Advanced Health Services Research Career Development Award (VA ARCD 98341-2); and Dr Lichtman by a CDC Career Development Award (K01-DP000085-01).
Role of the Sponsor: The grants and sponsors had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; in the preparation of the data; or in the preparation, review, or approval of the manuscript.