Depression and incident heart disease in VA patients

BACKGROUND: Among older Americans, depression is an independent risk factor for incident heart disease. The present paper will measure the association between depression and heart disease in Veterans Administration (VA) patients 35-70 years of age. This study will demonstrate the utility of the VA's electronic medical record system for identifying baseline depression and incident heart disease, overcoming some limitations of the extant literature by including male and female patients spanning a broad age range that is geographically representative of the U.S. METHODS: Using a convenience cohort of nationally distributed VA patients free of cardiovascular disease in 1999, we tested for an association between depression in fiscal year 1999 and incident heart disease between 2000 and 2005. We analyzed data from the electronic inpatient and outpatient files and the Pharmacy Benefits Management (PBM) files. The VA's inpatient and outpatient files contain International Classification of Diseases 9th Revision Clinical Modification (ICD-9-CM) diagnosis codes, as well as limited patient demographic characteristics common to both files, including age, gender and race. The PBM contains detailed data on all prescribed inpatient and outpatient medications. Cases of depression had a VA outpatient record of depression and a PBM record of anti-depressant prescription. Incident heart disease was defined by the presence of outpatient or inpatient ICD-9-CM codes: 410-429.9 in fiscal years 2000 through 2005. The resulting patient cohort consisted of 10,975 non-depressed controls and 1,333 cases. Risk ratios were computed to test for associations between baseline depression and incident heart disease within 5 age cohorts. Logistic regression was used to measure the association between baseline depression and incident heart disease after controlling for age, gender, race, hyperlipidemia, hypertension and diabetes. RESULTS: Risk ratios ranged from 0.94 to 1.33. Baseline depression in 1999 was significantly associated with incident heart disease for age cohort 40-49, (RR=1.3; 95%CI:1.1-1.6), age cohort 50-59 (RR=1.2; 95%CI:1.1-1.4) and age cohort 60-69 (RR=1.2; 95%CI:1.1-1.4). Depression remained significantly associated (OR=1.38; 95%CI:1.2-1.6) with incident heart disease after controlling for covariates. DISCUSSION: Our results are novel because we demonstrate an association between depression and incident heart disease in a VA population, and we provide new evidence that depression contributes to early onset heart disease. Results demonstrate the utility of using the VA's electronic patient data that have been reported as an important resource for epidemiology (Boyko et al. 1998).