Biologically Variable Bypass Reduces Enzymuria After Deep Hypothermic Circulatory Arrest

Abstract

Renal injury is common after open-heart surgery. Cardiopulmonary bypass contributes to the problem. We compared conventional nonpulsatile perfusion (NP) to biologically variable perfusion (BVP), which uses a computer controller to restore physiological beat-to-beat variability to roller pump flow. We hypothesized BVP would decrease renal injury after deep hypothermic circulatory arrest. Pigs were randomly assigned to either BVP (n = 9) or NP (n = 9), cooled, arrested at 18 degrees C (1 hour), reperfused, and rewarmed and maintained normothermic (3 hours). Additional pigs had NP for a similar time as above, but without circulatory arrest (n = 3), or were sham-treated without bypass (n = 3). Hemodynamics, acid-base status, temperature, and urine volumes were measured. Urinary enzyme markers of tubular injury were compared post-hoc for gamma glutamyl transpeptidase, alkaline phosphatase, and glutathione S-transferase and by urine proteomics using mass spectrometry. Urine output at 1 hour after arrest was 250 +/- 129 mL with BVP versus 114 +/- 66 mL with NP (p < 0.02). All three renal enzyme markers were higher with NP after arrest compared with BVP. In animals on bypass without arrest or those sham-treated, no elevations were seen in renal enzymes. Urine proteomics revealed abnormal proteins, persisting longer with NP. Biologically variable perfusion decreased cooling to 21.0 +/- 9.0 minutes versus 31.7 +/- 7.5 minutes (p < 0.002), and decreased rewarming to 22.1 +/- 3.9 minutes versus 31.2 +/- 5.1 minutes (p < 0.002). Biologically variable perfusion improved urine output, decreased enzymuria, and attenuated mass spectrometry urine protein signal with more rapid temperature changes. This strategy could potentially shorten bypass duration and may decrease renal tubular injury with deep hypothermic circulatory arrest.