Prospective randomized neurocognitive and S-100 study of hypothermic circulatory arrest, retrograde brain perfusion,...

To determine the optimal method of brain protection during deep hypothermic circulatory arrest (DHCA) for arch repair.

Of 139 potential aortic arch repairs (denominator), we randomized 30 patients to either DHCA alone (n = 10), DHCA plus retrograde brain perfusion (RBP) (n = 10), or antegrade perfusion (ANTE) (n = 10); a further 5 coronary bypass (CAB) patients were controls. Fifty-one neurocognitive subscores were obtained for each patient at each of four intervals: preoperatively, 3 to 6 days postoperatively, 2 to 3 weeks postoperatively, and 6 months postoperatively. Intraoperative and postoperative S-100 blood levels and electroencephalograms were also obtained.

For the denominator, the 30-day and hospital survival rate was 97.8% (136 of 139) and the stroke rate 2.8% (4 of 139). For the randomized patients, the survival rate was 100% and no patient suffered a stroke or seizure. Circulatory arrest (CA) times were not different (DHCA: RBP:ANTE) for 11 total arch repairs (including 6 elephant trunk; mean, 41.4 minutes; standard deviation, 15). Hemiarch repairs (n = 17) were quickest with DHCA (mean 10.0 minutes; standard deviation, 3.6; p = 0.011) and longest with ANTE (mean 23.8 minutes; standard deviation, 10.28; p = 0.004). Of the patients, 96% had clinical neurocognitive impairment at 3 to 6 days, but by 2 to 3 weeks only 9% had a residual new deficit (1 DHCA, 1 RBP, 1 ANTE), and by 6 months these 3 patients had recovered. Comparison of postoperative mean scores showed the DHCA group did better than RBP patients in 5 of 7 significantly different (p < 0.05) scores and versus 9 of 9 ANTE patients. There were no S-100 level differences between CA groups, but levels were significantly higher versus the CAB controls, particularly at the end of bypass (p < 0.0001); however, these may have been influenced by other variables such as greater pump time, cardiotomy use, and postoperative autotransfusion. Circulatory arrest (p = 0.01) and pump time (p = 0.057) correlated with peak S-100 levels.

The results of hypothermic arrest have improved; however, there is no neurocognitive advantage with RBP or ANTE. Nevertheless, retrograde brain perfusion may, in a larger study, potentially reduce the risk of strokes related to embolic material. S-100 levels may be artificial. In patients with severe atheroma or high risk for embolic strokes, we use a combination of retrograde and antegrade perfusion on a selective basis.