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

To the Editor: The author group of Guidelines for the Use of Electrophysiological Monitoring for Surgery of the Human Spinal Column and Spinal Cord1 read the Letter to the Editor2 with considerable interest. We are happy to respond, and should begin by addressing the more superficial criticisms that no doubt originate from a lack of understanding of the process of guideline development undertaken by organized neurosurgery and also from obvious unfamiliarity with the work/experience of the guideline's authors. But there appear to be even greater misunderstandings regarding the purpose and process of guideline development in general, and these need to be focused upon first, since this most likely, at least in part, provides the context for their point of view. Guidelines, practice parameters, or practice recommendations typically address points of clinical decision-making where there is equipoise or there are strongly opposing views, such as those, incidentally, shown by the authors of the Letter2 when interpreting the evidence. Our point of view (or “bias” according to the Letter authors), since we are not purveyors of intraoperative monitoring, and therefore have no vested interest in promoting its use, rather comes from the ranks of potential users. Our interest, like that of many (most) of our colleagues, is in performing due diligence in assessing a technology that has reached the status of “requirement” for some spinal surgery. Indeed, Inrtraoperative Monitoring (IOM) has been used as a legal standard of care so that its lack of use has led to punitive damages, in spite of the absence of case law regarding IOM use.3 This is based widely upon the assumption that what we hope will happen, ie, an alert leads to changes in intraoperative techniques that subsequently lead to restoration of signal and thus, an avoidance of injury or neurological deficit, will actually happen. We set out to examine whether this had been established in scientific study by our systematic review (always the basis of every guideline) of the literature, and without any point of view that we were trying to prove, contrary to the accusation by the authors of the Letter,2 and in contradistinction to any previous guidelines promulgated by purveyors of IOM.4 The fact that all of the authors1 represent a single institution is a mark of the commitment of the Department of Neurosurgery at the University of Alabama at Birmingham to practice, teach, and be productive in the area of evidence-based neurosurgery. The authors of the Letter2 are perhaps unfamiliar with the rigorous, and sometimes frustrating, peer review process required before endorsement by our specialty societies, which may lead to extensive revisions and in-depth questions regarding statements and approach. This guideline was no different than all guidelines endorsed by the Congress of Neurological Surgeons (CNS) and the American Association of Neurological Surgeons (AANS). The manuscript was first reviewed by the Guidelines Committee of the AANS/CNS Section on Disorders of the Spine and Peripheral Nerves, with several additions and changes to the document. Following successful review by that group of guidelines and content experts, the document was then reviewed, again in great detail, with more revisions and clarifications, by the AANS/CNS Joint Guidelines Committee, comprised of persons with further guidelines expertise prior to approval. Finally, the authors of the Guidelines for the Use of Electrophysiological Monitoring for Surgery of the Human Spinal Column and Spinal Cord1 have co-authored and written extensively on neurosurgical guidelines—over 125 guidelines documents for the senior author, more than 50 guidelines by the first author, more than 20 guidelines by the third author, with the second author having studied critical appraisal of the medical literature for 5 years, under the guidance of the other 3 authors, as an integral part of training in evidence-based neurosurgery. A recommendation was made to the authors to include this information in the IOM Guidelines text, with the authors declining to do so since we felt it would be seen to be self-serving. Other criticisms the authors of the Letter2 leveled at the Guidelines1 are aimed at interpretation of specific papers that formed a part of the evidence evaluated in the document, specifically, the papers by Choi,5 Sala,6 and Nuwer.7 We wish to respond to these criticisms below. It is important to reiterate that medical evidence-based guidelines are designed to collate, critically evaluate, and summarize the existing medical evidence on a particular topic. Therapeutic studies are designed to evaluate the therapeutic effect of a particular adjunct, usually compared to a control group of patients who did not receive this adjunct. The null hypothesis for a therapeutic study is typically stated as: no evidence of therapeutic benefit compared to the control group. As with diagnostic comparative studies, we assume the null position to be true (ie, there is no therapeutic benefit for the adjunct in question) until proven otherwise. If sufficient evidence demonstrates a therapeutic benefit, the null hypothesis is rejected. If the evidence fails to demonstrate a therapeutic benefit, we fail to reject the null hypothesis. In that case, therapeutic benefit remains unproven but has not been disproven. In their study, Choi and colleagues5 did not report an improvement in gross total resection rate or in neurological outcome in patients who underwent surgery with IOM compared to patients who underwent surgery without IOM. In short, they failed to demonstrate sufficient evidence to reject the null hypothesis, thus failing to demonstrate that IOM is an effective therapeutic adjunct. The IOM Guidelines authors do not believe that we misinterpreted the medical evidence presented in their study, nor do we believe that we misinformed readers of our publication. A failure to demonstrate sufficient evidence to reject the null hypothesis for whatever reason (eg, flawed methodology, small sample size, etc) is, nonetheless, a failure to demonstrate therapeutic benefit. Based on a comprehensive review of the existing medical evidence on the utilization of IOM as a therapeutic adjunct (including the Choi5 and Sala6 articles discussed herein), there is insufficient evidence to demonstrate a therapeutic benefit for IOM when used for this purpose. Furthermore, we reject the claim of selection bias in the evidentiary support surrounding the use of IOM as a therapeutic adjunct. Ostensibly, Sala6 and Choi5 conducted similar historical control studies evaluating the use of multimodality IOM during surgery for resection of intramedullary spinal cord tumors. While, strictly speaking, the methodology of these 2 articles is similar, the Sala6 article carries a significant methodological flaw that led to its reclassification as a class III medical evidence study by including nonconsecutive patients over a very lengthy time period in their control group (1988-2000). While the IOM Guidelines authors understand the benefits of matching control patients based on salient factors, doing so over such a lengthy time horizon has significant drawbacks. The field of neurosurgery underwent several fundamental changes from 1988 to 2000: the widespread use of the operating microscope, significant advancements in computational power for diagnostics and medical record keeping/database creation, and substantial changes in diagnostic methodology (most notably the consistent use of magnetic resonance imaging and the remarkable improvement in its quality). For these reasons, the control group in the Sala et al6 study was not assessed, evaluated, or managed in a contemporary or similar manner to the intervention group. Conversely, the historical control group provided by Choi and colleagues5 was comparatively recent and included consecutive patients who were evaluated using similar, contemporary protocols. While additional matching was not carried out, Choi and colleagues5 provided a better, contemporary comparison group and warranted a class II medical evidence rating. At the request of the Joint Guidelines Committee of the AANS/CNS, detailed in-text discussion of class III medical evidence studies was removed when class I or class II medical evidence existed for a particular topic. For this reason, the Sala et al6 study (class III medical evidence) was highlighted in the evidentiary table only, while the Choi et al5 study (class II medical evidence) was discussed in more detail within the text of the manuscript. Finally, in the Nuwer7 paper, we reclassified true positive to false positive because we were simply applying strict diagnostic test assessment methodology. An IOM alert followed by no postop deficit is a false positive while an alert in a patient who sustains a deficit is a true positive. We reject the idea that actions taken in response to an alert automatically equate to a deficit averted, which is the optimistic assumption that must be made to justify classifying those cases as true positives. The authors of the Guidelines for the Use of Electrophysiological Monitoring for Surgery of the Human Spinal Column and Spinal Cord1 therefore stand by the publication and its contents. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.