Objectives: We previously reported that postoperative hemodynamics and developmental outcomes were better among infants randomized to a higher hematocrit value during hypothermic cardiopulmonary bypass. However, worse outcomes were concentrated in patients with hematocrit values of 20% or below, and the benefits of hematocrit values higher than 25% were uncertain.

Methods: We compared perioperative hemodynamics and, at 1 year, developmental outcome and brain magnetic resonance imaging in a single-center, randomized trial of hemodilution to a hematocrit value of 25% versus 35% during hypothermic radiopulmonary bypass for reparative heart surgery in infants undergoing 2-ventricle repairs without aortic arch obstruction.

Results: Among 124 subjects, 56 were assigned to the lower-hematocrit strategy (24.8% +/- 3.1%, mean +/- SD) and 68 to the higher-hematocrit strategy (32.6% +/- 3.5%). Infants randomized to the 25% strategy, compared with the 35% strategy, had a more positive intraoperative fluid balance (P = .007) and lower regional cerebral oxygen saturation at 10 minutes after cooling (P = .04) and onset of low flow (P = .03). Infants with dextro-transposition of the great arteries in the 25% group had significantly longer hospital stay. Other postoperative outcomes, blood product usage, and adverse events were similar in the treatment groups. At age 1 year (n = 106), the treatment groups had similar scores on the Psychomotor and Mental Development Indexes of the Bayley Scales; both groups scored significantly worse than population norms.

Conclusions: Hemodilution to hematocrit levels of 35% compared with those of 25% had no major benefits or risks overall among infants undergoing 2-ventricle repair. Developmental outcomes at age 1 year in both randomized groups were below those in the normative population.