Atrial cardiomyoplasty in a Fontan circulation

Abstract

Introduction: The Fontan circulation is a direct connection between the systemic veins and the pulmonary artery (PA). Consequently, the pulmonary flow is passive due to the gradient between the right and left atrial pressure. In patients with increased pulmonary vascular resistance, the surgical procedure of atrio-pulmonary connection is therefore prone to failure. The goal of this experiment was to increase the pulmonary flow in an experimental model of a Fontan circulation by performing a right atrial cardiomyoplasty (ACMP).

Methods: In 19 Foxhounds the left m. latissimus dorsi (LD) was mobilised and transferred as a pedicle into the chest. After sternotomy a 'Fontan circulation' was created under cardiopulmonary bypass (CPB) by connecting the right atrium (RA) with the PA by a valveless conduit. The tricuspid valve was closed with a patch. In 11 dogs (group 1) a valve was implanted in the inferior vena cava (IVC) and pulmonary inflow impedance was increased by partial occlusion of the conduit to a gradient of 10 mmHg between RA and PA. In the other eight dogs (group 2) no valve was implanted, but flowmeters were placed in the IVC and the superior vena cava (SVC). In all dogs the RA was enlarged by a fascia lata patch before the LD was wrapped over the RA and stimulated synchronously to the R-wave with burst impulses.

Results: After coming off CPB, relatively high central venous pressures (22.5+/-5.8 mmHg) were necessary to maintain haemodynamic stability. With LD-stimulation in a 1:3 mode in group 1, RA pressure (P) increased from 23.1+/-7.7 to 45+/-10.5 mmHg (P<0.001), pulmonary atrial pressure (PAP) from 15.5+/-4.3 to 25.5+/-7.6 mmHg (P<0.001) and central venous pressure increased to 33.1+/-11.3 mmHg (P<0.05). Stroke volume increase from 11.4+/-4.7 to 17.2+/-4.3 ml and peak conduit-flow from 1286.3+/-880.3 to 2329+/-1173 ml/min (all P<0.001). In a 1:1 stimulation mode a pulsatile pressure/flow profile was obtained in the PA-conduit. Furthermore, at higher frequencies of about 120 beats/min muscle relaxation was still fast enough as not to interfere with the RA filling. In group 2 caval flow without stimulation occurred mainly during diastole. However, with LD-stimulation, a strong backflow into IVC and SVC was observed resulting in a less pronounced pressure/flow increase in the PA.

Conclusions: Our experimental model demonstrates the possibility of a 'ventricularisation' of the RA by using the force of the LD. However, the haemodynamic benefit of ACMP was achieved only, when a valve was implanted in the ICV.