The mortality rate of the Fontan operation for heart malformations with a single or dominant ventricle has been reduced by dividing the procedure into two stages. The hemi-Fontan procedure allows early reduction of the volume work of the single ventricle and remodeling of ventricular geometry before a completion Fontan operation. Despite the improvement of survival with this strategy (8% mortality for completion Fontan versus 16% mortality for primary Fontan operation), morbidity related to serous effusions remains substantial. Further technical modifications have been undertaken in an effort to reduce morbidity and further reduce mortality. From January 1990 through June 1993, 200 patients underwent completion Fontan procedures after previous hemi-Fontan operations. Mean age was 23 months, and 157 patients were less than 24 months of age. Diagnoses were hypoplastic left heart syndrome (127 patients), tricuspid atresia (19 patients), single left ventricle (17 patients), complex double-outlet right ventricle (16 patients), pulmonary atresia with intact ventricular septum (8 patients), and other (13 patients). Overall, early mortality rate was 8% (16 patients). In the last 112 patients, the procedure was modified technically by creating one or more fenestrations in the baffle used to separate systemic venous blood from pulmonary venous blood (36 patients), or by excluding one or more hepatic veins from the systemic venous pathway (76 patients). Early mortality for these 112 patients was reduced to 4.5% (5 patients). Substantial morbidity from serous effusions occurred at a rate of 45% (35 of 78 patients) among survivors who had received neither technical modification. Among survivors of a fenestrated Fontan procedure, the incidence of effusions was 39% (14 of 36 patients). Among survivors of a completion Fontan operation with partial exclusion of the hepatic veins, the incidence was only 14% (10 of 71 patients). We conclude that the mortality and morbidity rates of a completion Fontan procedure have been reduced further by either creation of fenestrations in the atrial baffle or association of one or more hepatic veins with the pulmonary venous atrium. The latter modification has significantly reduced the incidence of serous effusions.
