Despite increasing experience with the surgical treatment of tricuspid atresia, controversy exists regarding the early and late results of the various surgical options. Between January 1, 1967, and September 1, 1982, 92 patients underwent 147 operations for tricuspid atresia. Eighty-five patients underwent 1 or more palliative operations (108 procedures), with 8 hospital deaths (9%; confidence limits [CL], 6 to 14%). Thirty-eight patients underwent 44 classic (Blalock-Taussig or Gore-Tex) shunts, with 3 early (7%; CL, 3 to 13%) and 3 late deaths. Actuarial survival at 5 years was 78%, but only 56% were alive and free from having to undergo further operation at 5 years. Thirty-seven patients underwent a Fontan procedure, with 5 hospital deaths (14%; CL, 8 to 22%). Of these 37 patients, 34 (92%) had a nonvalved connection between the right atrium and the right ventricular infundibulum or the pulmonary artery. Incremental risk factors for hospital mortality after the Fontan procedure in patients with tricuspid atresia as well as other cardiac anomalies include young age (p = 0.0003), diagnosis other than tricuspid atresia (p = 0.03), and complex associated procedures (p = 0.02). During the year 1983, hospital mortality was 7% (1 out of 14; CL, 1 to 22%) for the Fontan procedure without complex additional procedures. Actuarial survival after a Fontan procedure was 71% at 5 years, with 3 patients requiring late reoperation. Of the survivors, 96% are in New York Heart Association Functional Class I or II. The Fontan procedure without a valve offers good intermediate survival, good functional results, and few reoperations. In patients with diminished pulmonary blood flow, a classic shunt also provides good palliation and good intermediate survival, but there is a necessity for additional operations in many patients within 5 years.