Letter: Guidelines for the Use of Electrophysiological Monitoring for Surgery of the Human Spinal Column and Spinal Cord

To the Editor: Hadley and colleagues’ recent publication in Neurosurgery1 is a bold step at addressing the evidence-based utility of intraoperative neuromonitoring (IONM) for surgery involving the spinal cord and spinal column.1 Their meta-analysis attempts to critically assess the literature on the benefits of IONM use based on 3 general categories: diagnostic, therapeutic, and cost effectiveness. They further categorize the published data according to the statistical levels of evidence (Level I, II, and III). While there is no doubt that evidence-based approaches need to be focused on the use of IONM, it is equally evident that this is not a straightforward task, primarily based upon the absence of randomized control trials and the ethical concerns central to conducting RCT for IONM. On a diagnostic level, the authors’ analysis found that multimodality IONM “is a reliable and valid diagnostic adjunct to assess spinal cord integrity.” However, for the therapeutic role of IONM, their recommendations are that IONM neither improves postoperative neurological outcome (level III) nor results in greater levels of gross tumor resection (level II). We think it is germane to this discussion to examine a few details of the 2 level II studies central to the author's recommendations. For example, Harel et al2 reported new neurological deficits in the IONM group to be 10% (4/41 patients) compared to 14% (10/70 patients) in a control group (P = .75). Examination of the data revealed that 10/41 (24%) patients in the IONM group received spinal instrumentation compared to 0% in the control group. Harel et al2 admit that this “probably poses increased neurological risk” to those patients. This study group discordance is unfortunate and in our view dramatically compromises the conclusions of both the Harel et al2 study as well as the weight this paper is given in the meta-analysis by Hadley et al.1 The second level II study reported by Hadley et al1 was published by Choi et al3 who evaluated the utility of IONM during intramedullary spinal cord tumor (IMSCT) resection. Although the authors clearly indicate that the Choi et al3 data “represents negative class II medical evidence,” they fail to mention that the trend for total excision was greater in the monitored group (76%) versus unmonitored group (53.8%; P = .049 univariate analysis). It is further evident that sample size and heterogeneity of the patient groups (tumor histology, gender, tumor size, etc) influenced the statistical outcomes as multivariate regression models were needed to adjust for the many patient factors resulting in the lack of significance (P = .119). Lastly, and most importantly, it should be noted that Choi et al3 did not use D wave monitoring in their study that has been shown to be a superior method to muscle motor-evoked potential for monitoring spinal cord during IMSCT surgery.4,5 Again, this omission was not addressed by the present authors. Finally, we must ask how IONM use possesses diagnostic efficacy but apparently lacks therapeutic influence? We would argue that this represents problems in communication between neuromonitorists and surgeons6 as well as the lack of universal interventional guidelines that can be used by surgical teams in the event of a legitimate IONM alert. These deficiencies may be mitigated with the use of interventional check-lists7 and placing the best trained and experienced IONM personal in the operating room. It has recently been shown that appropriate interventions following IONM alerts can significantly affect neurological outcomes when compared to IONM alerts that were not followed by an intervention.8 It may be time to encourage all neurophysiological-surgical teams to implement and use procedural checklists as part of our strategy during IONM alerts to improve patient outcomes. It is through a process of consistently applied interventional methods that the therapeutic benefits of IONM are more likely to synchronize with the established diagnostic abilities of these tests. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.