Outcomes in neonatal pulmonary atresia with intact ventricular septum. A multiinstitutional study.

A total of 171 neonates with pulmonary atresia and intact ventricular septum were entered into a prospective multiinstitutional study between January 1, 1987, and January 1, 1991. Treatment was not assigned randomly but was selected by the responsible physicians. The Z-value (standard deviation units) of the diameter of the tricuspid valve was less than -2 in 52% of patients and less than -4 in 26%; it was highly correlated with right ventricular cavity size (which was small in 90% of patients and was severely reduced in 54%). Coronary artery-right ventricular fistulas were present in 45% of patients, and right ventricular dependency was severe in 9%; diameter (Z-value) of the tricuspid valve was negatively correlated (P < 0.0001) with the prevalence of both. Survival was 81% at 1 month after the first intervention and 64% at 4 years; the hazard function (instantaneous risk of death) declined rapidly after the initial procedure but remained appreciable for 24 months. Multivariable analysis showed small diameter of the tricuspid valve, severe right ventricular coronary dependency, birth weight, and the date and type of initial procedure to be risk factors for time-related death. Subsequent procedures were performed in 51% of patients. Among patients undergoing an initial procedure that did not include a systemic-pulmonary artery shunt, only 49% did not receive a shunt in the subsequent 1 month; small size of the tricuspid valve was the only risk factor for receiving the subsequent shunt. Ninety-eight percent of living patients whose initial procedure did not include a transannular patch were free of a subsequently placed transannular patch within 1 month, but only 45% were free of it 3 years later; no risk factors were identified. Eighteen percent of living patients had received a one-ventricle repair within 3 years, and 32% had received a two-ventricle repair; the remainder (50%) had incompletely separated pulmonary and systemic circulations. The only patient-specific risk factor for not receiving a two-ventricle repair was the Z-value of the tricuspid valve.