From Jan. 1, 1961, through Dec. 31, 1987, 530 patients underwent an intracardiac operation that included a tricuspid valve procedure. The tricuspid valve was repaired in 351 patients (66%) and replaced in 179 (34%). Mean age was 56.9 years. Risk factors associated with tricuspid valve replacement included tricuspid stenosis (p = 0.02), jugular venous distention (p = 0.04), previous operation (p = 0.05), and angiographic severity of tricuspid valve incompetence (p less than 0.001). There were 78 hospital deaths (15%). Risk factors for hospital death included previous operation (p = 0.03), male gender (p = 0.03), hepatomegaly (p = 0.03), De Vega or Carpentier annuloplasty (repair group only), (p = 0.01), and older age at operation (p = 0.06). Ninety-eight percent of the patients were followed up. There were 185 late deaths (41%). The actuarial survival rate was 20% at 180 months. Risk factors for late death included male gender (p = 0.03), hepatomegaly (p = 0.04), and lack of postoperative warfarin therapy (p less than 0.001). Actuarial freedom from reoperation was 25.5% at 180 months. There was no difference in reoperation rates (p = 0.10) or survival (p = 0.42) whether the tricuspid valve had been repaired or replaced. We conclude that the requirement for surgical treatment of tricuspid valve insufficiency in patients with multivalvular disease constitutes a high risk group for cardiac surgery. Preoperative variables may predict the result of tricuspid valve replacement. Tricuspid valve replacement may be performed with the expectation of a low risk of valve-related events.