Blood-Brain Barrier Disruption After Cardiopulmonary Bypass: Diagnosis and Correlation to Cognition.

BACKGROUND: Cardiopulmonary bypass (CPB) elicits a systemic inflammatory response that may impair blood-brain barrier (BBB) integrity. BBB disruption can currently be detected by dynamic contrast enhancement magnetic resonance imaging (MRI), reflected by an increase in the permeability constant (Ktrans ). We aimed to determine (1) whether CPB induces BBB disruption, (2) duration until BBB disruption resolution, and (3) the obtainable correlation between BBB injury (location and intensity) and neurocognitive dysfunction.

METHODS: Seven patients undergoing CPB with coronary artery bypass grafting (CABG) were assigned to serial cerebral designated MRI evaluations, preoperatively and on postoperative day (POD) 1 and 5. Examinations were analyzed for BBB disruption and microemboli using dynamic contrast enhancement MRI and diffusion-weighted imaging methods, respectively. Neuropsychologic tests were performed 1 day preoperatively and on POD 5.

RESULTS: A significant local Ktrans increase (0.03 min-1 vs 0.07 min-1 , p = 0.033) compatible with BBB disruption was evident in 5 patients (71%) on POD 1. Resolution was observed by POD 5 (mean, 0.012 min-1 ). The location of the disruption was most prominent in the frontal lobes (400% vs 150% Ktrans levels upsurge, p = 0.05). MRI evidence of microembolization was demonstrated in only 1 patient (14%). The postoperative global cognitive score was reduced in all patients (98.2 ± 12 vs 95.1 ± 11, p = 0.032), predominantly in executive and attention (frontal lobe-related) functions (91.8 ± 13 vs 86.9 ± 12, p = 0.042). The intensity of the dynamic contrast enhancement MRI BBB impairment correlated with the magnitude of cognition reduction (r = 0.69, p = 0.04).

CONCLUSIONS: BBB disruption was evident in most patients, primarily in the frontal lobes. The location and intensity of the BBB disruption, rather than the microembolic load, correlated with postoperative neurocognitive dysfunction.