Adequacy of Brain Oxygen Saturation for Goal-Directed Therapy

In Response: Drs. Nemergut and Mauermann (1) offer some insightful comments and we agree that cerebral oximetry may represent an important new perioperative monitor. While they are correct that the results of the trial did not disprove the null hypothesis for the primary endpoint, it should be noted that the secondary hypotheses of the study protocol were established a priori to additionally examine pooled major adverse outcomes, selected individual complications, and cost drivers such as lengths of stay between groups. Therefore, these analyses, as referenced in the text, were not the result of retrospective “data mining,” but rather were established prospectively as part of the original protocol and consequently support our conclusion that monitoring cerebral rSO2 is associated with a significantly lower incidence of major organ dysfunction (2). To our knowledge, this is the first time a monitoring modality has been shown in a randomized, controlled trial to aid in the reduction of major adverse events and intensive care unit (ICU) length of stay. Their characterization of the mean decrease in ICU length of stay of 0.62 day as “clinically questionable” should also be reconsidered. While this may appear numerically small, the cumulative ICU stay in the control group was 185 days vs 125 days in the intervention group, a reduction of 60 days or 32%. Not only is this an important direct cost savings, but in many institutions ICU bed occupancy is the rate limiting step in daily cardiac surgical caseload. In such settings, patients being discharged earlier versus later, even during the same day, will directly impact surgical throughput. Additionally, the large differences between groups in the standard deviation of the means for duration of ICU stay indicate that more patients in the control group had prolonged ICU stays which was, in fact, the case. Again, such “outliers” consume a disproportionate amount of hospital resources—this reduction in ICU length of stay is thus a very clinically relevant finding. The observation regarding the association between lower rSO2 with longer hospitalization, “but only in the cohort of patients with a hospital admission longer than 10 days” is not meant to imply that low intraoperative rSO2 is a primary determinant of long term hospitalization, but rather to demonstrate that there is some association. Clearly, many other factors including comorbidities, concomitant disease, and postoperative events can all profoundly influence postoperative convalescence. However, in concert with the observations that patients experiencing MOMM had lower baseline rSO2, lower nadir rSO2 and significantly more profound cerebral desaturations, this does imply that rSO2 is an important index of cerebral, and, as we posit here, of systemic “well-being.” This observation is also not inconsistent with the findings of the study by Fenton et al., in which they demonstrated that low baseline rSO2 predicts perioperative mortality in children with congenital heart disease (3). Drs Nielsen and Børglum (4) question whether the cardiac index (CI) was similar between groups and we assure them that on average it was. However, at particular times during cardiopulmonary bypass (CPB), e.g., dislocation of heart for performance of distal anastamoses, CI may be decreased and in the absence of early change in other indicators of adequacy of perfusion such as decreased mixed venous oxygen saturation, steps to increase CI were not undertaken. Similarly, Fio2 during CPB is generally maintained between 0.3 and 0.4, which in some patients at some times may be inadequate as demonstrated here. The fact that various interventions were required to reverse desaturation within and between patients reinforces the importance of monitoring and individualizing treatment strategies. Whether the general conduct of CPB can be altered to routinely increase Fio2 and CI in the absence of such specific indicators as cerebral desaturation cannot be addressed here, but some steps in that direction have already been taken (5). We concur with Dr. Dworschak (6), that in a majority of instances, decreases in rSO2 were bilateral. However, we did observe unilateral desaturations of greater than 10% in 11 patients (6%) and, as referenced in the Introduction (2), given the high incidence of occult cerebrovascular disease in this population the question remains that if only one sensor is employed, which side to monitor (7)? We also believe that bilateral monitoring can assist in diagnosing both baseline (8) and developing (9) asymmetry related to anatomic, pathophysiologic, or emergent issues. For clarification the intervention algorithm was prioritized to focus initially on optimizing those parameters, e.g., CI, MAP etc., which were out of range then secondarily to further increase perfusion pressure and/or Fio2 with a goal of the optimization of rSO2. We are puzzled, however, by the comment regarding the preemptive nature of the interventions and their potential for methodological flaw. The whole point of the study was that monitoring rSO2 would decrease the duration of cerebral desaturations and improve outcomes so rather than methodological flaw, preemptive intervention was the actual focus of the study. Traditionally, improvements in patient care and outcomes are incremental and often the risk of the therapy is questioned with regard to its impact. However, cerebral oximetry is a noninvasive monitoring tool that appears to contribute to a meaningful reduction in major adverse events and ICU length of stay at no additional risk to the patient. We would hope that the results of this randomized, controlled trial, in combination with the positive findings associated with its use as demonstrated by others, will stimulate greater consideration and investigation of its impact on patient care. John M. Murkin, MD, FRCPC Daniel Bainbridge, MD, FRCPC Department of Anesthesiology and Perioperative Medicine University Hospital-LHSC University of Western Ontario, London Ontario, Canada [email protected] Richard Novick, MD, FRCSC Division of Cardiac Surgery University Hospital-LHSC University of Western Ontario, London Ontario, Canada