To review and quantify the impact of depression on the development of coronary heart disease (CHD) in initially healthy subjects. Cohort studies on depression and CHD were searched in MEDLINE (1966-2000) and PSYCHINFO (1887-2000), bibliographies, expert consultation, and personal reference files. Cohort studies with clinical depression or depressive mood as the exposure, and myocardial infarction or coronary death as the outcome. Information on study design, sample size and characteristics, assessment of depression, outcome, number of cases, crude and most-adjusted relative risks, and variables used in multivariate adjustments were abstracted. Eleven studies met the inclusion criteria. The overall relative risk [RR] for the development of CHD in depressed subjects was 1.64 (95% confidence interval [CI]=1.29-2.08, p<0.001). A sensitivity analysis showed that clinical depression (RR=2.69, 95% CI=1.63-4.43, p<0.001) was a stronger predictor than depressive mood (RR=1.49, 95% CI=1.16-1.92, p=0.02). It is concluded that depression predicts the development of CHD in initially healthy people. The stronger effect size for clinical depression compared to depressive mood points out that there might be a dose-response relationship between depression and CHD. Implications of the findings for a broader bio-psycho-social framework are discussed.