The effect of participation in cardiac rehabilitation on medical costs was determined by measuring hospitalization charges for cardiac admissions over a 3-year period in 580 post-coronary event patients (58% after coronary bypass surgery, 42% after myocardial infarction), of whom 230 entered a cardiac rehabilitation program and 350 did not. Baseline left ventricular ejection fraction was similar in entrants and nonentrants (59.9% vs 59.5%). Over the 1 to 46-month follow-up period (mean 21 months), per capita hospitalization charges for participants in cardiac rehabilitation were $739 lower than charges for nonparticipants ($1197 +/- 3911 vs $1936 +/- 5459, p = 0.022). This was due to both a lower incidence of hospitalizations and lower charges per hospitalization. Inasmuch as groups differed with regard to age, sex, diagnostic category, and smoking status, data were adjusted for these baseline differences by means of analysis of covariance. Rehospitalization charges remained significantly higher in nonparticipants (p = 0.015). Because physician charges were not measured, the cost differential between groups is underestimated. Results of this study show an association between participation in comprehensive cardiac rehabilitation and lowered cardiac rehospitalization costs in the years after an acute coronary event.