Response to Letter by Barili et al
We appreciate the comments by Dr Barili et al regarding our recent article on watershed strokes after cardiac surgery.1
They point out that we did not study patients who did not have strokes after cardiac surgery, and therefore cannot comment on risk of stroke, in general, in patients exposed to a decrease in mean arterial pressure. We agree and would like to emphasize that the increased risk of bilateral watershed strokes in patients exposed to a drop in mean arterial pressure of at least 10 mm Hg is compared with other stroke patients. We are in the process of conducting other studies to determine the role of this drop in mean arterial pressure in the development of stroke, in general.
In addition, the inclusion of unilateral watershed strokes in the “control” group was questioned. We do not necessarily agree that the mechanism is exactly the same in patients with unilateral and bilateral watershed strokes. The latter group is more likely to have been exposed to a global decrease in hypoperfusion, causing equivalent hypoperfusion to both sides of the brain, whereas patients with unilateral watershed infarcts are more likely to have carotid or intracranial disease leading to selective hypoperfusion of one particular hemisphere.2
Finally, we do acknowledge that the 2 patient groups were not homogeneous at univariate analysis, both with regards to procedure and cardiopulmonary bypass time. Multivariate adjustment for procedure made minimal to no difference in the point estimates or levels of significance of the other covariates and thus was not included in the final model. A relatively small number of patients in our sample underwent aortic procedures, and perhaps this accounts for our lack of further findings when procedure type was included in the multivariate model. We did, however, adjust for cardiopulmonary bypass time, given the discrepancy between mean duration in the 2 groups, and thus the potential confounding effect of differential bypass time should be reduced. We also adjusted for other factors that we felt to be clinically important and that were potential confounders in assessing the relationships between stroke type and outcome, as well as between change in mean arterial pressure and bilateral watershed strokes, as reported in the article. Obviously many of these problems are characteristic of a retrospective study of this nature.
We agree with Dr Barili and colleagues that we cannot definitively make conclusions about the role of hypoperfusion in the development of strokes after cardiac surgery. However, we feel that our results will lead to future studies that can answer this question more definitively. There is still a need for prospective, carefully designed studies to fully explore the relationship between blood pressure during cardiac surgery and postoperative neurological complications.