We read with great interest the recent article by Brown et al.1  regarding the impact of Bispectral Index (BIS)–guided sedation on the incidence of postoperative delirium during spinal anesthesia for spine surgery compared with BIS-masked general anesthesia. We appreciated the originality and efforts of the authors to clarify a still controversial topic such as the connection between depth of the hypnotic component of anesthesia and postoperative delirium. Nevertheless, after careful reading of the trial and its conclusions, we would like to address some critical input to the authors.

First, in the “Materials and Methods” section, the authors report that they achieved spinal anesthesia using intrathecal bupivacaine or lidocaine. Spinal anesthesia with lidocaine carries the risk of transient neurologic symptoms.2  Transient neurologic symptoms constitute an acute pain syndrome that could exacerbate postoperative pain and thus increase the incidence of postoperative delirium. In addition, acute postoperative urinary retention after spinal anesthesia could increase pain and discomfort and contribute to postoperative delirium.3  Both complications could have influenced the reported results on postoperative delirium, but their incidence is not reported by the authors.

Second, we would have been happy to see more information on intraoperative hemodynamic stability, all the more since moderately high doses of intrathecal local anesthetic agents and prone position for surgery may have an impact on it. Despite the fact that the relationship between intraoperative hypotension and the incidence of postoperative delirium is still not clearly established, it would have been informative to report not only on the lower intraoperative mean arterial pressure but also, and more importantly, on the decrease from initial mean arterial pressure and time spent below a patient-adapted threshold value, i.e., the time of relative cerebral hypoperfusion.4,5  Indeed, cerebral perfusion pressure beyond the autoregulatory limit is an independent risk factor for the development of neurologic complications, including postoperative delirium. Rightly, the authors report the incidence of postoperative stroke in their trial, but several publications have emphasized the importance of subclinical cerebral vascular events and their potential role in generating postoperative delirium.6 

Third, and in accordance with the results of this study, the BIS is probably not the right tool to guide anesthesia depth with the aim of avoiding postoperative delirium. Drug-induced alterations of brain function are complex, and BIS catches only a very small part of them.7  However, this does not mean that we should not seek a better understanding of the changes that are really relevant with regard to postoperative delirium and that should be prevented. The electroencephalogram is certainly the most accessible and noninvasive tool to be used in this respect, and several teams are currently performing an in-depth analysis of intraoperative electroencephalogram data to find out the most relevant markers. This must occur in a more general framework that takes account of the multifactorial nature of postoperative delirium, in which factors such as neuroinflammation, quality of organ perfusion, drug interactions, adequacy of antinociception, and patient comfort intervene.

Therefore, we should consider BIS as reflecting the tip of the iceberg only, while the immense mountain of ice remains hidden. Using BIS as the sole tool to prevent postoperative delirium is piloting a boat like the Titanic’s captain: the most important part of the problem could still be under the surface of the water.