Advertisement

Antidepressant Drug Compliance: Reduced Risk of MI and Mortality in Depressed Patients

      Abstract

      Background

      The long-term risk of myocardial infarction (MI) associated with use of antidepressants is uncertain, especially for nontricyclic antidepressants. The present study uses a national Veterans Affairs cohort to test whether antidepressants increase or decrease risk of MI and all-cause mortality.

      Methods

      US Department of Veterans Affairs patient records were analyzed to identify a cohort free of cardiovascular disease in fiscal years 1999 and 2000, aged 25-80 years, who had an International Classification of Diseases, Ninth Revision, Clinical Modification code indicating an episode of depression (n=93,653). Incident MI and all-cause mortality were modeled in patients who received 12 weeks or more of antidepressant pharmacotherapy as compared with 0-11 weeks during follow-up. Age-adjusted Cox proportional hazard models were computed before and after adjusting for baseline sociodemographics and time-dependent covariates.

      Results

      Receipt of 12 or more weeks of continuous antidepressant therapy was associated with significantly reduced rates of incident MI across classes of antidepressants: selective serotonin reuptake inhibitor (SSRIs) (hazard ratio [HR] 0.48; 95% confidence interval [CI], 0.44-0.52), serotonin-norepinephrine reuptake inhibitors (SNRIs) (HR 0.35; 95% CI, 0.32-0.40), tricyclic antidepressants (TCAs) (HR 0.39; 95% CI, 0.34-0.44), and “Other” (HR 0.41; 95% CI, 0.37-0.45). Risk of all-cause mortality also was decreased with receipt of 12 weeks of pharmacotherapy with all classes of antidepressants (SSRI, SNRI, TCA, Other), with HRs ranging from 0.50 to 0.66.

      Conclusions

      Across classes of antidepressants, 12 weeks of pharmacotherapy appears to be safe in terms of MI risk. Although the mechanism for this association remains uncertain, it is possible that compliance with pharmacotherapy for depression reflects compliance with cardiovascular medications. It also is possible that a direct drug effect or improved depressed mood may attenuate the risk of MI in depressed patients.

      Keywords

      Depression is associated with increased risk of coronary heart disease in healthy individuals, and with cardiac morbidity and mortality in individuals with established coronary heart disease.
      • Frasure-Smith N.
      • Lespérance F.
      Recent evidence linking coronary heart disease and depression.
      • Nicholson A.
      • Kuper H.
      • Hemingway H.
      Depression as an aetiologic and prognostic factor in coronary heart disease: a meta-analysis of 6362 events among 146,538 participants in 54 observational studies.
      • Rugulies R.
      Depression as a predictor for coronary heart disease A review and meta-analysis.
      • Scherrer J.F.
      • Xian H.
      • Bucholz K.K.
      • et al.
      A twin study of depression symptoms, hypertension, and heart disease in middle-aged men.
      • Scherrer J.F.
      • Virgo K.S.
      • Zeringue A.
      • et al.
      Depression increases risk of incident myocardial infarction among Veterans Administration patients with rheumatoid arthritis.
      • Scherrer J.F.
      • Chrusciel T.
      • Zeringue A.
      • et al.
      Anxiety disorders increase risk for incident myocardial infarction in depressed and nondepressed Veterans Administration patients.
      Treatment of depression has not resulted in reduced morbidity and mortality in clinical trials.
      • Joynt K.E.
      • O'Connor C.M.
      Lessons from SADHART, ENRICHD and other trials.
      Sample size can be a hindrance to addressing this question. Use of medical record claims data has shown promise in overcoming small sample size issues and recently was used to show that venlafaxine reduced risk of sudden cardiac death.
      • Martinez C.
      • Assimes T.
      • Mines D.
      • Dell'Aniello S.
      • Suissa S.
      Use of venlafaxine compared with other antidepressants and the risk of sudden cardiac death or near death: a nested case-control study.
      • Antidepressants do not increase the risk for incident myocardial infarction in patients without preexisting heart disease.
      • Compliance with acute-phase antidepressant therapy is necessary to detect a reduced risk for myocardial infarction and all-cause mortality.
      • The mechanism of this association is uncertain but may include general adherence to health-promoting behaviors, reduced depression, or direct drug effects.
      While tricyclic antidepressants (TCAs) are associated with increased risk of adverse cardiovascular outcomes,
      • Rosenberg L.B.
      • Whang W.
      • Shimbo D.
      • Shah A.
      • Shapiro P.A.
      • Davidson K.W.
      Exposure to tricyclic antidepressants is associated with an increased risk of incident CHD events in a population-based study.
      • Hippisley-Cox J.
      • Pringle M.
      • Hammersley V.
      • et al.
      Antidepressants as risk factor for ischaemic heart disease: case-control study in primary care.
      the selective serotonin reuptake inhibitor (SSRI) antidepressants are neutral or associated with reduced risk for myocardial infarction (MI).
      • Glassman A.H.
      • O'Connor C.M.
      • Califf R.M.
      • et al.
      Sertraline treatment of major depression in patients with acute MI or unstable angina.
      • Swenson J.R.
      • Doucette S.
      • Fergusson D.
      Adverse cardiovascular events in antidepressant trials involving high-risk patients: a systematic review of randomized trials.
      • Meier C.R.
      • Schlienger R.G.
      • Jick H.
      Use of selective serotonin reuptake inhibitors and risk of developing first-time acute myocardial infarction.
      • Schlienger R.G.
      • Fischer L.M.
      • Jick H.
      • Meier C.R.
      Current use of selective serotonin reuptake inhibitors and risk of acute myocardial infarction.
      • Sauer W.H.
      • Berlin J.A.
      • Kimmel S.E.
      Selective serotonin reuptake inhibitors and myocardial infarction.
      • Sauer W.H.
      • Berlin J.A.
      • Kimmel S.E.
      Effect of antidepressants and their relative affinity for the serotonin transporter on the risk of myocardial infarction.
      • Maurer-Spurej E.
      Serotonin reuptake inhibitors and cardiovascular diseases: a platelet connection.
      • Serebruany V.L.
      • Glassman A.H.
      • Malinin A.I.
      • et al.
      Platelet/endothelial biomarkers in depressed patients treated with the selective serotonin reuptake inhibitor sertraline after acute coronary events: the Sertraline AntiDepressant Heart Attack Randomized Trial (SADHART) Platelet Substudy.
      Despite growing use of serotonin-norepinephrine reuptake inhibitors (SNRIs), little is known about the potential cardiovascular risks associated with this relatively new class of antidepressants. How different classes of antidepressant pharmacotherapy affect cardiovascular risk has not been satisfactorily answered.
      We conducted this study to investigate whether antidepressants increase or decrease risk of incident MI and all-cause mortality that may be due to treatment effects, drug effects, or compliance. The study utilized a retrospective cohort design with sufficient power to simultaneously adjust for multiple cardiovascular risk factors, psychiatric comorbidity, and markers of depression severity and health service utilization.

      Methods

      For this retrospective cohort study, data were obtained from inpatient and outpatient International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnoses, Current Procedural Terminology codes, Pharmacy Benefits Management records, and vital status files maintained by the Veterans Health Administration beginning in Fiscal Year (FY) 1999—the first year that national data are considered complete. These data are maintained by the Veterans Health Administration Office of Information at the Austin Information Technology Center (www.virec.research.va.gov/datasources name/medical-sas-datasets).

      Depressed Cohort

      From all patients using US Department of Veterans Affairs (VA) health care in 1999 and 2000 (n=1,380,433), we used ICD-9-CM codes to exclude patients with at least one primary or secondary diagnosis of heart disease and cerebrovascular disease. Depressed patients must have had either one inpatient ICD-9-CM code for depression in FY2000 or at least 2 outpatient ICD-9-CM codes for depression in a 12-month span, with at least one code occurring in FY2000, which began on October 1, 1999. This algorithm of one inpatient visit or 2 outpatient visits in a 12-month span was used to define all other psychiatric conditions. This algorithm has 99% positive predictive value for depression diagnoses in administrative claims data.
      • Solberg L.I.
      • Engebretson K.I.
      • Sperl-Hillen J.M.
      • Hroscikoski M.C.
      • O'Connor P.J.
      Are claims data accurate enough to identify patients for performance measures or quality improvement? The case of diabetes, heart disease, and depression.
      We excluded patients with psychotic disorders and bipolar disorder to reduce risk of misclassification of depression. Patients also must have been between the ages of 25 and 80 years at the beginning of follow-up to allow for variability in risk of MI. To establish that our patients were regular users of VA health care, a patient must have had at least one outpatient visit in FY1999 and FY2000. Patients were excluded if an MI occurred within 1 month of follow-up. Details of the creation of the cohort have been previously reported.
      • Scherrer J.F.
      • Chrusciel T.
      • Zeringue A.
      • et al.
      Anxiety disorders increase risk for incident myocardial infarction in depressed and nondepressed Veterans Administration patients.
      From this cohort we excluded those who did not receive 12 weeks' follow-up from baseline (n=868) and patients who only had a diagnosis of dysthymia (n=2091), resulting in 93,653 depressed patients eligible for the present analyses.

      Antidepressant Use

      According to VA guidelines,
      Office of Quality and Performance (10Q).
      patients experiencing a new episode of depression should receive 12 weeks of antidepressants. Of depressed patients, 78.7% received guideline concordant care, 8.4% received no antidepressants, and 12.9% received 1-11 weeks of antidepressant therapy. All antidepressants available during the retrospective period were included if prescribed at the minimum effective dose used to treat depression.
      • Robinson R.L.
      • Long S.R.
      • Chang S.
      • et al.
      Higher costs and therapeutic factors associated with adherence to NCQA HEDIS antidepressant medication management measures: analysis of administrative claims.
      Patients were considered treated with a specific antidepressant if they received 12 or more continuous weeks of treatment with the same drug and had refilled at least one prescription. Patients were considered not treated if they received only 0-11 weeks of any drug. Polypharmacy was possible but is beyond the scope of the present report. All antidepressant use data were based on the days supply variable from the Pharmacy Benefits Management records.

      Outcome Variables

      Incident MI during the period of October 1, 2000-September 30, 2007 was identified if patient records contained ICD-9-CM codes 410-411. The duration of follow-up was calculated as the time from October 1, 2000 to the diagnosis of MI. Conversely, if coding for MI was not present, the follow-up time was calculated as the time from October 1, 2000 to the last recorded visit (date of last inpatient or outpatient ICD-9-CM code).
      All deaths during the period of October 1, 2000-September 30, 2007 were obtained from the VA Vital Status File, which tracks deaths by incorporating information from the Beneficiary Identification and Records Locator Subsystem Death File created by the Veterans Benefits Administration, the Medical SAS Inpatient Datasets that track mortality and death dates that occur during a hospital stay, and the Social Security Administration Death Master File.

      Covariates

      Conditions identified by ICD-9-CM code and known to be risk factors for heart disease included: hypertension, hyperlipidemia, type II diabetes (also identified by diabetic medication including insulin) and obesity (also identified by body mass index), alcohol abuse/dependence, and nicotine dependence/smoking history. Conditions must have been present before MI or censoring.
      Sociodemographics at baseline included year of birth, sex, race, insurance status, and marital status. We adjusted for insurance status to account for the possibility that private insurance may be linked to use of care outside the VA, thus reducing our ability to detect all health care encounters.

      Cardiovascular Screening and Interventions

      Variability in rates of screening and preventive interventions for heart disease by depression status could confound our results, so we constructed variables for these potential confounders using Current Procedural Terminology codes and procedural ICD-9-CM codes. Our cardiac screening variable contained tests such as electrocardiogram, stress test, and echocardiography. Cardiac procedures included catheter placement, stent placement, and angioplasty. We controlled for lipid-lowering and vasodilator prescriptions.

      Depression Severity

      Depression severity was measured by hospitalization and the fifth-digit ICD-9-CM indicator of severity, as well as accounting for posttraumatic stress disorder (PTSD), anxiety disorder unspecified, and panic disorder, which were found to be significantly associated with MI in our previous work.
      • Scherrer J.F.
      • Chrusciel T.
      • Zeringue A.
      • et al.
      Anxiety disorders increase risk for incident myocardial infarction in depressed and nondepressed Veterans Administration patients.

      Analytic Design

      Bivariate analyses included computing t tests for continuous variables and chi-squared tests for categorical variables. We first investigated the relationship among MI, depression, and pharmacotherapy while controlling for age only. Due to a nonlinear relationship among age, MI, depression, and pharmacotherapy, both a linear and a quadratic age term were included in all multivariate models. Survival models were computed to analyze time to MI and all-cause mortality. Hazard ratios for incident MI and all-cause mortality were estimated using Cox proportional hazards models with time-dependent covariates. Sociodemographics were modeled from their status at baseline, and the time-dependent covariates fixed occurrence before incident MI. Separate multivariate models were computed for each drug class and for each individual antidepressant. We adjusted for covariates described above. Analyses were performed using SAS version 9.1.3 (SAS Institute, Cary, NC) with alphas set at 0.05. Two-tailed tests were used to allow for both risk factors and protective effects. The PROC PHREG procedure was used to compute Cox proportional hazards models. Month was the unit of time for survival analyses.
      This project was approved by the Institutional Review Board of the St. Louis VA, Washington University, and Saint Louis University.

      Results

      The mean (SD) age of patients was 51.5 (11.4) years, and 14.1% of the sample was female. The majority of patients were white (77.8%), not married (57.1%), and had only VA health care coverage (67.3%). Comparing patients with and without incident MI during follow-up (Table 1), those who had an MI were significantly older (55.3 years vs 51.4 years, P <.0001), more often white (82.1% vs 77.6%, P <.0001), and less likely to be female (P <.0001). Patients who had an MI were significantly more likely to have diagnoses of diabetes (37.2% vs 26.6%, P <.0001), hypertension (76.1% vs 63.8%, P <.0001), hyperlipidemia (63.1% vs 59.5%, P <.0001), and obesity (60.8% vs 58.8%, P <.05). Patients experiencing an MI were less often diagnosed with alcohol or drug dependence (42.9% vs 44.6%, P <.05). Patients experiencing an MI compared with those who did not were screened for cardiovascular disease less often (63.1% vs 66.8%, P <.0001), had more cardiac procedures (4.1% vs 2.5%, P <.0001), and were prescribed more vasodilators (21.6% vs 10.3%, P <.0001) and lipid-lowering drugs (50.1% vs 47.7%, P <.01). Those experiencing an MI were higher utilizers of VA health care compared with patients who did not experience an MI (P <.0001). Lastly, fewer patients experiencing an incident MI received 12 weeks of any antidepressant (75.0% vs 78.9%, P <.0001) and each of the individual antidepressant drug classes (all P <.05).
      Table 1Distribution of Sociodemographic Variables, Cardiovascular Risk Factors, Health Care Utilization, Psychiatric Comorbidity, Depression Severity and Antidepressant Use among 93,653 VA Patients 25-80 Years of Age Free of Heart Disease at Baseline (1999-2000) by Incident Myocardial Infarction (MI) during Follow-up (October 1, 2001-September 30, 2007)
      TotalMINo MI
      Sociodemographics
       Age (mean, SD)51.5 (11.4)55.3 (10.1)51.4 (11.4)
      P <.0001;
       Female13,241 (14.1)279 (7.5)12,962 (14.4)
      P <.0001;
       Race
        White72,812 (77.7)3047 (82.1)69,765 (77.6)
      P <.0001;
        Nonwhite17,958 (19.2)638 (17.2)17,320 (19.3)
        Unknown2883 (3.1)27 (0.7)2856 (3.2)
       Marital status
        Married38,927 (41.6)1576 (42.5)37,351 (41.5)
        Not married
      Includes divorced, widowed, separated, and never married.
      53,455 (57.1)2080 (56.0)51,375 (57.1)
        Unknown1271 (1.4)56 (1.5)1215 (1.4)
       Health care coverage
        VA only63,033 (67.3)2473 (66.6)60,560 (67.3)
        Other30,620 (32.7)1239 (33.4)29,381 (32.7)
      Anxiety disorders
       Posttraumatic stress disorder31,431 (33.6)1269 (34.2)30,162 (33.5)
       Anxiety disorder unspecified28,821 (24.4)927 (25.0)21,894 (24.3)
       Panic disorder5727 (6.1)232 (6.3)5495 (6.1)
      Cardiovascular risk factors
       Diabetes25,297 (27.0)1381 (37.2)23,916 (26.6)
      P <.0001;
       Hypertension60,239 (64.3)2825 (76.1)57,414 (63.8)
      P <.0001;
       Hyperlipidemia55,896 (59.7)2341 (63.1)53,555 (59.5)
      P <.0001;
       Nicotine dependence44,671 (47.7)1761 (47.4)42,910 (47.7)
       Alcohol/drug dependence41,729 (44.6)1591 (42.9)40,138 (44.6)
      P <.05.
       Obesity55,180 (58.9)2256 (60.8)52,924 (58.8)
      P <.05.
      Cardiovascular screening and interventions
       Cardiac screen62,410 (66.6)2342 (63.1)60,068 (66.8)
      P <.0001;
       Cardiac procedure2367 (2.5)153 (4.1)2214 (2.5)
      P <.0001;
       Vasodilator Rx10,040 (10.7)800 (21.6)9240 (10.3)
      P <.0001;
       Lipid-lowering Rx44,780 (47.8)1858 (50.1)42,922 (47.7)
      P <.05.
      Depression severity
       Depression NOS52,364 (55.9)2118 (57.1)50,246 (55.9)
       Depression mild3123 (3.3)111 (3.0)3012 (3.4)
       Depression moderate16,995 (18.2)654 (17.6)16,341 (18.2)
       Depression severe21,171 (22.6)829 (22.3)20,342 (22.6)
       Hospitalization for depression during follow-up9069 (9.7)361 (9.7)8708 (9.7)
       Health care utilization
      Clinic stops per month, categorized by quartile values in the entire sample.
        Lowest 25%22,930 (24.5)384 (10.3)22,546 (25.1)
      P <.0001;
        25th-50th percentile23,396 (25.0)731 (19.7)22,665 (25.2)
        50th-75th percentile23,689 (25.3)1085 (29.2)22,604 (25.1)
        Highest 25%23,638 (25.2)1512 (40.7)22,126 (24.6)
      Antidepressant use
       12 weeks of any antidepressant73,784 (78.8)2783 (75.0)71,001 (78.9)
      P <.0001;
       12 weeks of any selective serotonin reuptake inhibitors58,714 (62.7)2153 (58.0)56,561 (62.9)
      P <.0001;
       12 weeks of any serotonin-norepinephrine reuptake inhibitors21,330 (22.8)634 (17.1)20,696 (23.0)
      P <.0001;
       12 weeks of any tricyclics9950 (10.6)348 (9.4)9602 (10.7)
      P <.05.
       12 weeks of any “Other”31,468 (33.6)1074 (28.9)30,394 (33.8)
      P <.0001;
      low asterisk P <.0001;
      low asterisklow asterisk P <.05.
      Includes divorced, widowed, separated, and never married.
      Clinic stops per month, categorized by quartile values in the entire sample.
      We examined the distribution of covariates by class of antidepressants. These data are not shown but are available from the authors upon request. No statistical tests were performed on these data because the classes of antidepressants were not mutually exclusive (eg, patients could have used more than one antidepressant). The most common antidepressants received for 12 weeks or longer were SSRIs (n=58,714), followed by prescriptions for “Other” (n=31,468), SNRIs (n=21,330), and TCAs (n=9950). Whites, married patients, and patients with only VA health care coverage were more likely to receive 12 or more weeks of antidepressants. For all cardiovascular risk factors, patients were more likely to receive antidepressant treatment for 12 or more weeks. This pattern also was observed for patients who had cardiovascular screening and interventions, patients with anxiety disorders, moderate and severe depression, hospitalization, and greater overall health care use.
      Rate of incident MI in depressed patients receiving 12 or more weeks of an antidepressant was modeled separately for each drug class: SSRIs, SNRIs, TCAs, and Other (Table 2). The risk (ie, hazard ratio [HR]) of MI for patients receiving 12 weeks or more of SSRIs was 0.48 (95% confidence interval [CI], 0.44-0.52), for SNRIs, HR=0.35 (95% CI, 0.31-0.39), for TCAs, HR=0.39 (95% CI, 0.34-0.44), and for “Other,” HR=0.41 (95% CI, 0.37-0.45). Cardiovascular risk factors, including diabetes, hypertension, hyperlipidemia, nicotine dependence, and obesity were all significantly associated with increased rate of incident MI, with HRs ranging from 1.10 (SSRIs and obesity) to 1.67 (Other drugs and hypertension). Substance dependence was significantly associated with increased rate of MI for all drug classes except TCAs (HR range 1.12-1.19).
      Table 2Hazard of Incident Myocardial Infarction among 93,653 Depressed Veterans Administration Patients Who Received 12 or More Weeks of Continuous Antidepressant Pharmacotherapy as Compared with Depressed Patients Who Did Not Receive Such Treatment (Models Computed Separately for Each Drug Class)
      Adjusted for sociodemographics, anxiety disorders, cardiovascular risk factors, cardiovascular screening and interventions, and depression severity.
      Selective Serotonin Reuptake Inhibitors (n=58,714)Serotonin-Norepinephrine Reuptake Inhibitors (n=21,330)Tricyclics (n=9950)Other (n=31,468)
      12 weeks of treatment0.48 (0.44-0.52)0.35 (0.31-0.39)0.39 (0.34-0.44)0.41 (0.37-0.45)
      low asterisk Adjusted for sociodemographics, anxiety disorders, cardiovascular risk factors, cardiovascular screening and interventions, and depression severity.
      Screening for cardiovascular disease was associated with a significant increase in rate of incident MI for all antidepressant classes (HR range 1.29-1.44), while cardiac intervention was not significantly associated with MI in any model. Prescription of vasodilators was associated with incident MI in all models (HR range 2.38-2.59), while prescription of lipid-lowering drugs was associated only with an increased risk in patients receiving SSRIs (HR 1.15) and patients receiving “Other” antidepressants (HR 1.14). Greater health care utilization was significantly associated with increased rate of incident MI in all models, with greater utilization having a stronger association with MI. There was no significant association between markers of severity of depression and risk of incident MI, with the exception of panic disorder, which was significantly associated with MI among those receiving “other” classes of antidepressants.
      Rate of all-cause mortality among depressed patients receiving 12 weeks of antidepressant pharmacotherapy was modeled separately for each drug class (Table 3). After adjusting for all covariates, receiving 12 weeks of antidepressant therapy for any of the drug classes (SSRIs, SNRIs, TCAs, and other drugs) was associated with decreased risk of all-cause mortality (HR range 0.50-0.66). Cardiovascular risk factors associated with increased rate of all-cause mortality included: diabetes, hypertension, nicotine dependence, and substance dependence (HR range 1.28-1.80). Hyperlipidemia and obesity were associated with decreased risk of all-cause mortality (HR range 0.74-0.92). Cardiac screening was associated with all-cause mortality for all drug classes (HR range 1.08-1.12). Patients receiving prescriptions for vasodilators were at significantly increased risk of all-cause mortality (HR range 1.28-1.34), while receipt of a prescription for a lipid-lowering drug was associated with decreased risk of all-cause mortality (HR range 0.65-0.68). Health care utilizers in the 25th-75th percentile had a reduced rate of all-cause mortality compared with those in the lowest percentile, while utilizers in the highest 25th percentile were at increased risk of mortality for all drug classes (HR range 1.25-1.35). Hospitalization for depression was associated with a decreased rate of all-cause mortality (HR range 0.84-0.86), as was having a comorbid diagnosis of PTSD (HR range 0.81-0.86).
      Table 3Hazard of All-cause Mortality among 93,653 Depressed Veterans Administration Patients Who Received 12 or More Weeks of Continuous Antidepressant Pharmacotherapy as Compared with Depressed Patients Who Did Not Receive Such Treatment (Models Computed Separately for Each Drug Class)
      Adjusted for sociodemographics, anxiety disorders, cardiovascular risk factors, cardiovascular screening and interventions, and depression severity.
      Selective Serotonin Reuptake Inhibitors (n=59,179)Serotonin-Norepinephrine Reuptake Inhibitors (n=21,678)Tricyclics (n=10,043)Other (n=31,863)
      12 or more weeks continuous treatment0.62 (0.59-0.65)0.51 (0.48-0.54)0.66 (0.62-0.71)0.50 (0.47-0.53)
      low asterisk Adjusted for sociodemographics, anxiety disorders, cardiovascular risk factors, cardiovascular screening and interventions, and depression severity.
      As shown in Table 4, after adjusting for all covariates, receiving 12 weeks of any individual antidepressant was associated with decreased risk of incident MI, with HRs ranging from 0.33-0.46.
      Table 4Hazard of Myocardial Infarction among 93,653 Depressed Veterans Administration Patients Who Received 12 or More Weeks of Continuous Antidepressant Pharmacotherapy as Compared with Depressed Patients Who Did Not Receive Such Treatment (Models Computed Separately for Each Antidepressant Drug)
      ExposureNumber on DrugOn Drug for 12 Weeks (Yes/No) HR for MI (CI)
      Less than 12 weeks antidepressants19,869Ref
      SSRI
       Citalopram21,3000.38 (0.34-0.42)
       Fluoxetine18,6170.43 (0.39-0.48)
       Paroxetine12,7480.44 (0.39-0.50)
       Sertraline22,9390.46 (0.42-0.51)
      SNRI
       Venlafaxine12,8070.34 (0.30-0.39)
       Mirtazapine11,0760.33 (0.28-0.37)
      TCA
       Amitriptyline35230.39 (0.32-0.47)
       Doxepin20270.39 (0.30-0.50)
       Nortriptyline33940.35 (0.28-0.43)
      Other
       Bupropion20,0520.34 (0.30-0.38)
       Nefazodone48250.45 (0.38-0.54)
       Trazodone12,6880.40 (0.35-0.45)
      CI = confidence interval; HR = hazard ratio; MI = myocardial infarction; SNRI = serotonin-norepinephrine reuptake inhibitor; SSRI = selective serotonin reuptake inhibitor; TCA = tricyclic antidepressant.
      Sensitivity analysis indicated that results did not differ when the cohort was limited to patients with only VA health care coverage.

      Discussion

      Depressed VA patients free of heart disease at baseline receiving 12 or more weeks of continuous antidepressant pharmacotherapy were at decreased risk of both incident MI and all-cause mortality compared with patients receiving 0-11 weeks of treatment. This effect remained after controlling for covariates. Results were consistent across drug classes (SSRI, SNRI, TCA, Other) and across individual antidepressants. We have previously demonstrated that this depressed cohort, as compared with a nondepressed cohort, is at risk of incident MI.
      • Scherrer J.F.
      • Chrusciel T.
      • Zeringue A.
      • et al.
      Anxiety disorders increase risk for incident myocardial infarction in depressed and nondepressed Veterans Administration patients.
      Thus, the effect of medication is associated not only with decreased risk of MI, but appears to alleviate the elevated risk of MI attributable to depression. In analyses subsequent to the present study, we found that depressed patients receiving 12 weeks of antidepressants were at significantly reduced risk of MI when compared with nondepressed patients (HR = 0.83; 95% CI, 0.79-0.87).
      The mechanisms for this decrease in risk of MI and mortality are uncertain. Potential explanations include reduced depression due to antidepressant pharmacotherapy, a direct effect of the drugs themselves, or a function of compliance such that patients who use 12 or more weeks of antidepressants also comply with other prescriptive health behaviors. Such a pattern has been observed in the human immunodeficiency virus literature.
      • Horberg M.A.
      • Silverberg M.J.
      • Hurley L.B.
      • et al.
      Effects of depression and selective serotonin reuptake inhibitor use on adherence to highly active antiretroviral therapy and on clinical outcomes in HIV-infected patients.
      Because SSRIs' protective effects are generally attributed to effects on platelet aggregation, patients unable to tolerate these medications may not benefit from this effect, or patients intolerant to SSRIs may be physiologically different.
      Our findings for SSRIs are consistent with previous studies.
      • Glassman A.H.
      • O'Connor C.M.
      • Califf R.M.
      • et al.
      Sertraline treatment of major depression in patients with acute MI or unstable angina.
      • Swenson J.R.
      • Doucette S.
      • Fergusson D.
      Adverse cardiovascular events in antidepressant trials involving high-risk patients: a systematic review of randomized trials.
      • Meier C.R.
      • Schlienger R.G.
      • Jick H.
      Use of selective serotonin reuptake inhibitors and risk of developing first-time acute myocardial infarction.
      • Schlienger R.G.
      • Fischer L.M.
      • Jick H.
      • Meier C.R.
      Current use of selective serotonin reuptake inhibitors and risk of acute myocardial infarction.
      • Sauer W.H.
      • Berlin J.A.
      • Kimmel S.E.
      Selective serotonin reuptake inhibitors and myocardial infarction.
      • Sauer W.H.
      • Berlin J.A.
      • Kimmel S.E.
      Effect of antidepressants and their relative affinity for the serotonin transporter on the risk of myocardial infarction.
      With the exception of venlafaxine,
      • Martinez C.
      • Assimes T.
      • Mines D.
      • Dell'Aniello S.
      • Suissa S.
      Use of venlafaxine compared with other antidepressants and the risk of sudden cardiac death or near death: a nested case-control study.
      no large cohort studies have investigated the association between SNRIs and “other” antidepressants on cardiovascular disease. Our results suggest that these drugs do not put patients at risk of MI and mortality. However, older men using antidepressants may be at increased risk of mortality associated with antidepressants,
      • Ryan J.
      • Carrier I.
      • Ritchie K.
      • et al.
      Late-life depression and mortality: influence of gender and antidepressant use.
      and antidepressant use in patients with heart failure increased mortality.
      • Sherwood A.
      • Blumenthal J.A.
      • Trivedi R.
      • et al.
      Relationship of depression to death or hospitalization in patients with heart failure.
      It is possible that our findings are limited to younger patients free of diagnosed cardiovascular disease at baseline.
      Most studies have found that TCAs are associated with increased risk of adverse outcomes.
      • Rosenberg L.B.
      • Whang W.
      • Shimbo D.
      • Shah A.
      • Shapiro P.A.
      • Davidson K.W.
      Exposure to tricyclic antidepressants is associated with an increased risk of incident CHD events in a population-based study.
      • Hippisley-Cox J.
      • Pringle M.
      • Hammersley V.
      • et al.
      Antidepressants as risk factor for ischaemic heart disease: case-control study in primary care.
      The present finding of a reduced risk for MI in TCA users is inconsistent with the literature that has demonstrated cardiotoxic effects of TCAs associated with arrhythmia. Given that our findings were unchanged when examining long-term use (2 years or more of antidepressant prescriptions, data not shown), the evidence for the mechanism of effect in the present study leans toward improvement in depression, compliance, or the possibility that patients free of cardiovascular disease are not at increased risk when taking TCAs.

      Strengths

      Because of the large number of data resources available, we could model the effect of antidepressants while accounting for the diagnosis of depression and its severity. In addition, the duration of follow-up and sample size permitted examination of a large number of antidepressants and use over a long period of time, presumably long enough to detect risk of MI and mortality if it existed in this population of antidepressant users. Our study overcomes limitations of projects that analyze only pharmacy data relying on the use of antidepressants as an indicator of depression status.

      Limitations

      It is possible that less severe cases of depression are undetected; therefore, the present results may not generalize to undiagnosed depression. Misclassification of covariates may confound our results if covariates were systematically under- or overdiagnosed by depression or MI status. We have previously reported
      • Scherrer J.F.
      • Chrusciel T.
      • Zeringue A.
      • et al.
      Anxiety disorders increase risk for incident myocardial infarction in depressed and nondepressed Veterans Administration patients.
      consistency between our nicotine-dependence variable and the 23% to 33% prevalence of smoking reported from surveys of the general veteran population.
      • Eisen S.A.
      • Griffith K.H.
      • Xian H.
      • et al.
      Lifetime and 12-month prevalence of psychiatric disorders in 8,169 male Vietnam War era veterans.
      • McKinney W.P.
      • McIntire D.D.
      • Carmody T.J.
      • Joseph A.
      Comparing the smoking behavior of veterans and nonveterans.
      ICD-9-CM codes for MI have very high agreement (>99%) with written medical records in the VA,
      • Kashner T.M.
      Agreement between administrative files and written medical records: a case of the Department of Veterans Affairs.
      but it is acknowledged that incident MI may be a recurrence or exacerbation of a preexisting cardiovascular condition first recorded in non-VA records. Absence of a diagnosis during the 2-year baseline period is a reasonable length of time to consider patients free of clinically significant cardiovascular disease before MI. Receipt of care for acute MI among VA patients often occurs in emergency situations in a non-VA hospital;
      • Wright S.M.
      • Daley J.
      • Fisher E.S.
      • Thibault G.E.
      Where do elderly veterans obtain care for acute myocardial infarction: Department of Veterans Affairs or Medicare?.
      however, based on information from the VA Health Services Research and Development Ischemic Heart Disease Quality Enhancement Research Initiative, VA patient records would include these emergency events.
      The purpose of receiving antidepressants may have been for multiple disorders, including PTSD. Sensitivity analyses, which removed comorbid anxiety disorders from our models, did not change results. Effects of antidepressants are the same regardless if they are prescribed for depression alone or for comorbid disorders. Our goal was to determine whether 12 or more weeks of antidepressant use was a risk factor or protective for incident MI and mortality, in which case the exact reason for prescription is not critical. While we can't be certain that depression was alleviated, nor can we measure drug plasma levels, at a minimum, all patients had a diagnosis of depression, and we can conclude that if therapy reduced depression symptoms, this may account for reduced risk of MI and mortality. In addition to our observation that 12 weeks or more use of antidepressants reduced risk, we computed post hoc analyses to determine if a dose-response effect existed for long-term use. Results indicate an average of a 3% reduction in risk for every additional month on antidepressants across drug class.
      These data suggest that 12 weeks of treatment with antidepressants is associated with reduced risk of MI and mortality. This finding was consistent across individual drugs, drug classes, and outcomes. Compliance with antidepressant treatment potentially reduces risk for MI and death. Potential mechanisms include direct effects of antidepressants and alleviation of depression, and compliance with antidepressants also may reflect compliance with cardiovascular medications.

      References

        • Frasure-Smith N.
        • Lespérance F.
        Recent evidence linking coronary heart disease and depression.
        Can J Psychiatry. 2006; 51: 73-77
        • Nicholson A.
        • Kuper H.
        • Hemingway H.
        Depression as an aetiologic and prognostic factor in coronary heart disease: a meta-analysis of 6362 events among 146,538 participants in 54 observational studies.
        Eur Heart J. 2006; 27: 2763-2774
        • Rugulies R.
        Depression as a predictor for coronary heart disease.
        Am J Prev Med. 2002; 23: 51-61
        • Scherrer J.F.
        • Xian H.
        • Bucholz K.K.
        • et al.
        A twin study of depression symptoms, hypertension, and heart disease in middle-aged men.
        Psychosom Med. 2003; 65: 548-557
        • Scherrer J.F.
        • Virgo K.S.
        • Zeringue A.
        • et al.
        Depression increases risk of incident myocardial infarction among Veterans Administration patients with rheumatoid arthritis.
        Gen Hosp Psychiatry. 2009; 31: 353-359
        • Scherrer J.F.
        • Chrusciel T.
        • Zeringue A.
        • et al.
        Anxiety disorders increase risk for incident myocardial infarction in depressed and nondepressed Veterans Administration patients.
        Am Heart J. 2010; 159: 772-779
        • Joynt K.E.
        • O'Connor C.M.
        Lessons from SADHART, ENRICHD and other trials.
        Psychosom Med. 2005; 67: S63-S66
        • Martinez C.
        • Assimes T.
        • Mines D.
        • Dell'Aniello S.
        • Suissa S.
        Use of venlafaxine compared with other antidepressants and the risk of sudden cardiac death or near death: a nested case-control study.
        BMJ. 2010; 340: c249
        • Rosenberg L.B.
        • Whang W.
        • Shimbo D.
        • Shah A.
        • Shapiro P.A.
        • Davidson K.W.
        Exposure to tricyclic antidepressants is associated with an increased risk of incident CHD events in a population-based study.
        Int J Cardiol. 2010; 145: 124-125
        • Hippisley-Cox J.
        • Pringle M.
        • Hammersley V.
        • et al.
        Antidepressants as risk factor for ischaemic heart disease: case-control study in primary care.
        BMJ. 2001; 323: 666-669
        • Glassman A.H.
        • O'Connor C.M.
        • Califf R.M.
        • et al.
        Sertraline treatment of major depression in patients with acute MI or unstable angina.
        JAMA. 2002; 288: 701-709
        • Swenson J.R.
        • Doucette S.
        • Fergusson D.
        Adverse cardiovascular events in antidepressant trials involving high-risk patients: a systematic review of randomized trials.
        Can J Psychiatry. 2006; 51: 923-929
        • Meier C.R.
        • Schlienger R.G.
        • Jick H.
        Use of selective serotonin reuptake inhibitors and risk of developing first-time acute myocardial infarction.
        Br J Clin Pharmacol. 2001; 52: 179-184
        • Schlienger R.G.
        • Fischer L.M.
        • Jick H.
        • Meier C.R.
        Current use of selective serotonin reuptake inhibitors and risk of acute myocardial infarction.
        Drug Saf. 2004; 27: 1157-1165
        • Sauer W.H.
        • Berlin J.A.
        • Kimmel S.E.
        Selective serotonin reuptake inhibitors and myocardial infarction.
        Circulation. 2001; 104: 1894-1898
        • Sauer W.H.
        • Berlin J.A.
        • Kimmel S.E.
        Effect of antidepressants and their relative affinity for the serotonin transporter on the risk of myocardial infarction.
        Circulation. 2003; 108: 32-36
        • Maurer-Spurej E.
        Serotonin reuptake inhibitors and cardiovascular diseases: a platelet connection.
        Cell Mol Life Sci. 2005; 62: 159-170
        • Serebruany V.L.
        • Glassman A.H.
        • Malinin A.I.
        • et al.
        Platelet/endothelial biomarkers in depressed patients treated with the selective serotonin reuptake inhibitor sertraline after acute coronary events: the Sertraline AntiDepressant Heart Attack Randomized Trial (SADHART) Platelet Substudy.
        Circulation. 2003; 108: 939-944
        • Solberg L.I.
        • Engebretson K.I.
        • Sperl-Hillen J.M.
        • Hroscikoski M.C.
        • O'Connor P.J.
        Are claims data accurate enough to identify patients for performance measures or quality improvement? The case of diabetes, heart disease, and depression.
        Am J Med Qual. 2006; 21: 238-245
      1. Office of Quality and Performance (10Q).
        in: FY 2007, Q3 Technical Manual for the VHA Performance Measurement System. Office of Quality and Performance, Veterans Health Administration, Department of Veteran Affairs, Washington, DC2007: 14-21
        • Robinson R.L.
        • Long S.R.
        • Chang S.
        • et al.
        Higher costs and therapeutic factors associated with adherence to NCQA HEDIS antidepressant medication management measures: analysis of administrative claims.
        J Manag Care Pharm. 2006; 12: 43-54
        • Horberg M.A.
        • Silverberg M.J.
        • Hurley L.B.
        • et al.
        Effects of depression and selective serotonin reuptake inhibitor use on adherence to highly active antiretroviral therapy and on clinical outcomes in HIV-infected patients.
        J Acquir Immune Defic Syndr. 2008; 47: 384-390
        • Ryan J.
        • Carrier I.
        • Ritchie K.
        • et al.
        Late-life depression and mortality: influence of gender and antidepressant use.
        Br J Psychiatry. 2008; 192: 12-18
        • Sherwood A.
        • Blumenthal J.A.
        • Trivedi R.
        • et al.
        Relationship of depression to death or hospitalization in patients with heart failure.
        Arch Intern Med. 2007; 167: 367-373
        • Eisen S.A.
        • Griffith K.H.
        • Xian H.
        • et al.
        Lifetime and 12-month prevalence of psychiatric disorders in 8,169 male Vietnam War era veterans.
        Mil Med. 2004; 169: 896-902
        • McKinney W.P.
        • McIntire D.D.
        • Carmody T.J.
        • Joseph A.
        Comparing the smoking behavior of veterans and nonveterans.
        Public Health Rep. 1997; 112: 212-217
        • Kashner T.M.
        Agreement between administrative files and written medical records: a case of the Department of Veterans Affairs.
        Med Care. 1998; 36: 1324-1336
        • Wright S.M.
        • Daley J.
        • Fisher E.S.
        • Thibault G.E.
        Where do elderly veterans obtain care for acute myocardial infarction: Department of Veterans Affairs or Medicare?.
        Health Serv Res. 1997; 31: 739-754

      Linked Article

      • Antidepressant Compliance and Incident Myocardial Infarction
        The American Journal of MedicineVol. 124Issue 7
        • Preview
          Scherrer et al1 report that compliance with 12 or more weeks of continuous antidepressant therapy was associated with significantly reduced rates of incident myocardial infarction and mortality across all classes of antidepressants in a large nationwide cohort of patients from the Veterans Administration. Unfortunately, the data used to measure exposure to antidepressants in this study do not really reflect compliance or patient adherence to medication. Instead, they reflect prescriber patterns.
        • Full-Text
        • PDF