Although the Heath-Edwards classification has been used for more than 25 years to evaluate pulmonary vascular changes, its potential to predict the long-term course after successful ventricular septal defect closure has not been proved. Operative lung biopsy slides obtained at the time of closure from 57 infants and children who had been among the first survivors (between 1954 and 1960) of such surgery were graded in blinded fashion according to the Heath-Edwards classification system, and the resultant biopsy grade of each was compared with the eventual long-term outcome of each child. In 53 (93%) of the 57 cases, the Heath-Edwards system correlated well with long-term clinical or hemodynamic status of the patient. Grade IV changes were predictive of a usually fatal outcome, while grade I and II changes were generally benign. Death from pulmonary vascular disease occurred, however, in four children whose operative lung biopsy had been classified as either grade I or II, indicating that the focal nature of the higher Heath-Edwards grades, or human error, must be considered when this classification system is employed.