Bidirectional cavo-pulmonary anastomosis with an additional flow source to the lungs: clinical experience in 21 cases.

The Fontan operation is considered a palliative procedure for the treatment of complex congenital heart diseases with a morbidity and mortality that are not negligible, especially in high-risk patients. In these cases, a bidirectional cavopulmonary anastomosis with additional source of blood flow to the lungs can represent a valuable surgical option. Between November 1992 and December 1995, 21 patients with univentricular heart physiology were submitted to bidirectional cavopulmonary anastomosis with additional blood flow the additional flow consisted of an aortopulmonary shunt in 13 cases, pulmonary artery banding in seven cases, and pulmonary artery stenosis in one case. Mean age at operation was 35.5 months (range 4 months to 12 years) and mean body weight 10.9 (range 4-24.4) kg. Hospital mortality was zero. Morbidity (pleuropericardial effusions) was significant in one case. There were three late deaths (14%) caused by worsening atrioventricular valve regurgitation: two of these occurred in patients with right isomerism. Late haemodynamic evaluation documented a significant increase in cardiac output (P < 0.01) and oxygen saturation (P < 0.05) during pharmacological stress testing. Angiographic characterization of blood flow distribution to the lungs showed better results when the additional blood flow was represented by an anterograde source (pulmonary artery banding or native pulmonary artery stenosis). In conclusion, bidrectional cavopulmonary anastomosis with additional blood flow is a good alternative to classic or fenestrated Fontan technique in high-risk patients. Secondly, the diagnosis of isomerism could be a contraindication to this type of surgery, because the persistent volume overload could worsen atrioventricular valve regurgitation. Thirdly, the pharmacological stress testing shows a favourable behaviour of cardiac output and arterial oxygen saturation. Fourth, the type of additional blood flow seems to condition both pulsatility and distribution of pulmonary blood flow.