Congenital heart disease affects 1% of US births, with many babies requiring major cardiothoracic surgery under cardiopulmonary bypass (CPB), exposing the more critical patients to neurodevelopmental impairment. Optimal surgical parameters to minimize neuronal injury are unknown. We used proton magnetic resonance spectroscopy (¹H MRS) and blood ammonia assays in a neonatal pig model of CPB to compare 2 approaches, complete circulatory arrest (CA) versus antegrade cerebral perfusion.

Two-week old piglets (N = 17) were put on a CPB pump and placed in a 3-T magnetic resonance imaging to study brain metabolism during CPB. Dynamic single-voxel ¹H MRS brain data were acquired while animals underwent 1 of 4 CPB protocols: ∼50 minutes CA at 18 °C and 28 °C or antegrade cerebral perfusion at 18°C and 28 °C, followed by a ∼1-hour recovery period. On the basis of ¹H MRS findings suggesting the presence of brain ammonia upon reperfusion, a second cohort of piglets (N = 22) underwent the same CPB conditions without MRS to allow regular venous blood sampling with ammonia assays.

All animals showed a transitory temperature-dependent increase in blood ammonia (P < .001) immediately after restart of whole-body perfusion. In contrast, metabolic processing of brain ammonia, as detected by an increased ¹H MRS glutamine/glutamate ratio, was also temperature dependent (P = .002) but only significantly observed in the CA studies (P = .009).

Serial ¹H-MRS and blood ammonia assays in this preclinical CPB model identified a previously unreported build-up of ammonia, hypothesized to arise from gut bacterial production, after reperfusion, that may contribute to brain injury in these pediatric surgeries.