Physiological rationale for a bidirectional cavopulmonary shunt. A versatile complement to the Fontan principle.

The original Fontan procedure included a classic superior vena cava-to-right pulmonary artery (Glenn) shunt. Subsequent experience demonstrated that this anastomosis was not essential and was an unnecessary commitment of the larger right pulmonary circulation to the smaller blood volume of the superior vena caval return. With application of the Fontan principle to more complex cardiac malformations, there has been a reconsideration of possible benefits of a cavopulmonary shunt in selected patients. A modified shunt from the divided end of the superior vena cava to the side of the undivided right pulmonary artery utilized in 21 patients is described. This shunt is designed to allow bidirectional pulmonary arterial distribution of both superior vena caval inflow and right atrial outflow after completion of the Fontan procedure. Twelve patients had the bidirectional shunt performed prior to a Fontan operation; five of these had a subsequent atriopulmonary connection and seven await operation. Eight patients had construction of this shunt at the time of their Fontan procedure. One patient had a bidirectional shunt constructed following atriopulmonary anastomosis to help relieve right atrial outflow obstruction. Two patients with univentricular heart undergoing simultaneous Fontan procedure and a bidirectional shunt died while in the hospital. The remaining 19 patients have been followed up for 2 months to 9 years with one late sudden death at 9 years. There have been no bidirectional cavopulmonary shunt failures, stenoses, kinks, or recognized pulmonary arteriovenous malformations. Postoperatively, eight patients had assessment of pulmonary distribution of shunt blood flow by angiography. Seven of these patients were also evaluated by radionuclide angiography. Superior vena caval blood flow via the bidirectional cavopulmonary shunt tended to be greater to the right lung, but bilateral pulmonary flow was documented in all but one patient. After Fontan operation, six of seven patients tested also demonstrated bilateral distribution of atriopulmonary flow. We concluded from our experience that this modified shunt provides excellent relief of cyanosis, allows bidirectional pulmonary distribution of both superior vena caval return and also the right atrial blood flow after atriopulmonary connection, and may be done before, with, or after a Fontan procedure and is compatible with all currently recommended modifications. Perioperative hemodynamic adjustments to the Fontan procedure may be improved by reducing atrial volume, and this may also be of potential benefit in the long-term adaptation to Fontan physiology by minimizing atrial distention.