OBJECTIVE In a retrospective study we evaluated the predictive value of the traditional risk factors (Fontan criteria) for mortality and morbidity after total cavopulmonary anastomosis (Fontan operation). METHODS We studied 51 consecutive patients who underwent Fontan operations from 1982 through 1994. Morbidity was assessed by a score taking equally into account rhythm disturbances, cardiovascular hospital admissions and heart failure. RESULTS The mean age at operation was 4.1 +/- 3.4 years. Diagnoses were tricuspid atresia in 20, complex single ventricle in 13 and various complex congenital heart defects in 18 patients. All Fontan criteria were fulfilled in 11 patients while in the remaining 40 up to 6 criteria were not met. The criteria least fulfilled were age in 31 patients and pulmonary artery anatomy in 11. The mean follow-up was 4.6 +/- 3.0 years. The overall 30-day survival was 84% and 1- and 10-year survival 76.5% respectively. All patients who fulfilled all the Fontan criteria survived. The overall mortality and early mortality were associated with a significantly higher number of unfulfilled Fontan criteria (p < 0.01). Age below 2 years at operation and pulmonary resistance > 4 Woods units were risk factors for overall mortality (p < 0.01). The number of unmet Fontan criteria was not significantly different in patients with and without postoperative late morbidity. CONCLUSIONS The Fontan criteria are reliable predictors for mortality but have little predictive value for morbidity. Surviving patients who do not fulfil all criteria suggest that some criteria may need to be modified. Our limited data support a lower age limit of 2 instead of 4 years.