We studied the yearly occurrence of isolated cardiovascular malformations in liveborns for the period 1981-1988 in Maryland and the District of Columbia. We used Poisson regression to fit yearly prevalence at birth as a function of time for all cardiovascular malformation case types combined and for three diagnostic subgroups. We give results separately by diagnostic method and race (white/nonwhite), with categories combined when appropriate. Time trends were similar for both races except in pulmonic stenosis diagnosed by echocardiography. The yearly prevalence for all case types, all races, and all diagnostic methods combined increased from 28 to 43 per 10,000 livebirths over the 8 years, due to an increase in the rate among infants diagnosed by echocardiography. For every diagnostic-type-by-race group studied, there were increases in the prevalence for those diagnosed by echocardiography and decreases (or no change in one group) for those diagnosed by cardiac catheterization and/or surgery. Although there may have been an increase in actual occurrence of cardiovascular malformation over our study period, it is also possible that wider use of a new echocardiography technology resulted in increased detection.