Intermediate-term outcome following the fontan operation: A survival, functional and risk-factor analysis

Objective: To investigate survival and risk factors influencing intermediate outcome after the Fontan procedure. Methods: Retrospective analysis of 122 patients operated between April 1991 and September 2002. Poor outcome was defined as late death or poor functional status (intractable supraventricular arrhythmias/NYHA 3-4) necessitating revision surgery. Results: 64(52%) patients had an intermediate bi-directional cavo-pulmonary shunt (BCPS). 91 (76%) patients had a lateral tunnel total cavo-pulmonary connection, 21(17%) patients had an atrio-pulmonary connection and 10 (8%)patients had a Kawashima connection. There were 6 (5%)early deaths. Over a median follow up of 54 months (1-133), 12 (10%)patients have had surgical revision for poor functional status. There were 7 (6%) late deaths, 5 of which occurred after revision surgery. Univariate analysis identified older age at operation (>4 years) (P=0.04), higher postoperative pulmonary artery pressure at 24 h (P=0.012), arrhythmia postoperatively (P=0.03) or during follow-up (P=0.01) and the requirement for anticoagulation during follow-up (P=0.03) as significant predictors of poorer outcome. Patients who had an intermediate BCPS (P=0.002) or Norwood Stage 1 (P=0.05) had a better outcome. Multivariate analysis identified an intermediate Glenn shunt and lower postoperative pulmonary artery pressure as significant predictors of better outcome. Actuarial freedom from death or revision is 93% (±2), 88% (±3), 86% (±4) and 69% (±7) at 1, 5, 7 and 9 years respectively. Actuarial freedom from death or revision for the lateral tunnel group is 92% (±2), 89% (±3), 85% (±5) and 66% (±10) respectively. Conclusion: Atrio-pulmonary connection results in a higher incidence of arrhythmias and failure than the lateral tunnel Fontan. Even in patients with a lateral tunnel Fontan there is a continuing hazard phase in the intermediate term. Mid-term outcome appears to be favorably influenced by an intermediate BCPS.