In Reply

We thank Drs. Fellahi and Portran for bringing attention to these important issues and for putting into context some of the methods leading to the results obtained in our feasibility study. The main purpose of the feasibility trial was to verify that the high rate of success of reversal of cerebral desaturations obtained in our single-center study¹  (88% of patients) could be repeated in centers that would eventually participate in a multicenter randomized controlled trial on the benefit of reversal of cerebral desaturations on patient outcomes in cardiac surgery. While discussing the methodology to adopt for the feasibility trial, we came to realize that after years of using this technology, it has become routine to start interventions to bring cerebral saturation back to normal values well before the full 20% decrease from baseline is reached. We, therefore, wanted to ask the question, do early interventions on cerebral desaturations help to avoid the progression of desaturations below the 20% threshold from baseline? If most of the variations within 10% of baseline are noise or extracranial contaminations, interventions should not impact on further progression to 20% baseline values. Our results show that early interventions prevented further decreases from 10 to 20% of baseline.²  Only 48% of patients in the intervention group had desaturations continue decreasing below 20% baseline, while almost all patients without interventions (82%) went on to desaturations below 20% baseline. We still considered 20% baseline the threshold for significant desaturations and calculated the cerebral desaturation load (area under the curve of cerebral desaturation over time) from this level. Early interventions are probably the reason why the reversal rate per patient was even higher (97%) than in our previous study. If extracranial contamination for the devices is within 10% of baseline value, why would interventions at this level matter for further desaturations? To answer this question, we have to look at the methodology used in the feasibility trial, desaturations had to last at least 15 s or more before interventions were started, and this is a long enough time to exclude most variations due to noise or extracranial contamination. Furthermore, in Davie and Grocott’s study,³  in order to estimate extracranial contamination, a pneumatic cuff had to be inflated around the head to unmask skin blood flow, which is considered contamination. Without cuff inflation, the skin contamination is just part of the signal, and we have no reason to believe that skin blood flow was specifically affected independently of cerebral blood flow in our study. Therefore, in a multicenter, prospective randomized control trail on the impact of reversing decreases in cerebral saturation on outcomes, we would very likely use the 10% threshold for intervention and 20% threshold for significant desaturation as we did in the feasibility study.

We agree thoroughly with Drs. Fellahi and Portran that there are many situations in which blood pressure, flow states, and oxygenation do not match, sepsis being a very good example, and therefore, the sequence of algorithmic choices for the interventions to reverse cerebral desaturations could vary. Presenting these options in a straight sequence for the purpose of a trial is one the limitations of working with an algorithmic approach. In clinical practice, the choice of an intervention to reverse a cerebral desaturation does not usually flow linearly; additional information, including transesophageal echography, hemoglobin level, or expired carbon dioxide levels, helps greatly in focusing on the appropriate intervention. The goal is not to follow the algorithm rigorously but to reverse cerebral desaturations in our patients. Still, most interventions suggested in the algorithm have been used with success to reverse cerebral desaturations.¹ 

The final point of the letter referring to the lack of improvement in adverse events in the group of patients with successful reversal of cerebral desaturations touches upon one of the most interesting findings of our feasibility study. We made the choice to study a patient population at high risk of complications (EUROSCORE II more than 10), because in many institutions, these are the very patients in which the use of this expensive technology is justified. In Murkin et al.’s study,⁴  the patients recruited were scheduled for primary elective coronary artery bypass, a population of patients with few comorbidities, not comparable to ours. Therefore, it is not surprising that the results from Murkin et al.’s study were not confirmed in our feasibility study. Nevertheless, we would have expected (or liked) to see more of a trend toward less adverse events in our intervention group. But that was not the case. Even more telling was the results from our study showing that patients without cerebral desaturations had an equivalent rate of adverse events as patients with cerebral desaturations (table 6).²  As related in our discussion: “we also have to consider the possibility that our patient population may have been so high risk that interventions to reverse decreases in rso₂ (regional cerebral oxygen saturation) have very little impact on outcomes.”²  We, therefore, now believe that for the purpose of a large, multicenter, randomized controlled trial, the inclusion of low- and high-risk patients appears to be the most advantageous strategy to use. It would provide data on the entire range of patients in whom cerebral saturation devices are used in common practice and may indicate which category of patients will benefit most from the device in terms of outcomes. This is a complex and high-cost study to conduct that will need careful planning. One should remember, however, that even though trials showing statistical significance in terms of outcomes are important, during complex high-risk surgeries where numerous difficult technical procedures must be performed, having a device that will inform you of the impact of an obstructed cannula on brain oxygenation and blood flow is also very beneficial for the patient.