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

To the Editor: We have read the recently published “guidelines” for intraoperative neurophysiological monitoring (IONM) in spine surgery.1 We believe there are understandable, but insufficient, arguments to produce a document that professionally and medicolegally excuses those surgeons who will not accept IONM. We doubt anyone using and benefitting from well-done IONM will now give it up based on the content of these palpably biased guidelines. The authors’ review of evidence examining IONM’s role as a diagnostic, therapeutic, and cost/benefit procedure is neither complete nor balanced. Their tabulation of papers’ results (some reviewed more extensively in the text) is arbitrary and nonsystematized. Notably, they ignore 18 papers supporting IONM. The authors write, “…numerous study design, methodology, and statistical flaws weaken the strength of the resulting evidence…,” but without specifying the alleged flaws. Instead, they emphasize and detail mainly articles against the value of IONM. For example, they cite an older scoliosis correction case series where D wave recording was associated with frequent false positive reports, but D wave monitoring in this setting has not been recommended for many years.2,3 The authors’ biased, nonsystematized approach is highly problematic and misleading. Neurophysiology scholars must accept their share of the blame for the slender state of controlled evidence in IONM. But patients should not suffer catastrophic neurological injury because of this deficiency. In particular, the authors claim, “…significant clinical equipoise exists…” But their claim ignores the requirement that a broad spectrum of stakeholders review proposed controlled trials before putting patients at risk. The bioethics, opportunity cost, and equipoise assessments are not as straightforward as the authors suggest.4 Neither the bias of the authors nor the complexities of optimizing the evidence base justify abandonment of IONM. But that is precisely the implied outcome when the authors dismiss IONM as merely a “diagnostic adjunct” in the “perioperative period.” Surgeons will not order IONM unless they strongly associate well-done IONM with difficult real-time case management decisions intraoperatively. Considering potential harms when patients are denied access to IONM, the authors and their associated societies would have done well to extend their understanding of the evidence base beyond a static hierarchical pyramid. Leading evidence-based scholars have already moved beyond strictly hierarchical data assessments.5-7 Gordon Guyatt, Distinguished Professor of Clinical Epidemiology and Biostatistics and Medicine at McMaster University, has spearheaded the development of GRADE (Grading of Recommendations Assessment, Development and Evaluation), a flexible rather than static hierarchical evidence system.8 The GRADE framework includes practical considerations to downgrade flawed randomized control trials (RCTs) and upgrade observational or nonrandomized controlled studies with large effects. Organizations embracing GRADE include the World Health Organization, the American College of Physicians, the Cochrane Collaboration, and the British Medical Journal. According to Guyatt, “… when methodologically strong observational studies yield large or very large and consistent estimates of the magnitude of a treatment effect, we may be more confident about the results.”9 This means that the role of confounding variables falls away with larger effects. Guyatt sets large and very large effect risk ratios (RRs) at <0.5 and <0.2. The Sala historical control, case-matched, IONM vs no IONM intramedullary tumor resection trial achieved a RR of 0.27 at 3 mo (P < .0016), approaching a very large long-term treatment effect favoring IONM.10 Nevertheless, the guidelines1 authors rate this study as class III. In contrast, the Harel study employed the same historical control design but without case matching on prognostic factors. The IONM group was at a higher risk of neurological injury because 17% were instrumented (0% in the “no IONM” group). Follow-up McCormick scale ratings were missing in some no IONM cases. Harel et al11 found no outcome difference (IONM vs no IONM). The guidelines authors rated this study as class II. Did the authors use different weights and measures depending on which theory each paper was supporting? John Ney and colleagues12-14 published compelling evidence supporting the use of IONM based on sophisticated cost effectiveness analyses and careful analysis of national claims data (big data). Their work may be summarized: “Intraoperative monitoring is cost-saving for spinal surgeries in a theoretical model based on the current published literature.”13 Yet the guidelines authors ignore this evidence. Ironically, the single reference to Ney was meant to cite a 2012 editorial opinion critical of an IONM review, but the authors mistakenly cite a Ney and Van Der Goes review of spine IONM cost effectiveness. Instead of examining Ney's impressive body of work uniformly supporting IONM, the authors describe Ney and van der Goes as “prominent electrophysiologists” to suit their own bias, and then present the following misleading and unjustified conclusion: “At present, there is insufficient evidence to support IOM during spinal surgery as a cost-effective measure. Due to the lack of evidence supporting IOM as an effective therapeutic adjunct, the expense of IOM and its interpretation during spinal column/spinal cord surgery may not justify its use in attempting to prevent iatrogenic spinal cord injury.”1 The authors point to the lack of class I evidence on the therapeutic value of IONM to conclude that it cannot be considered a “standard of care.” We agree that “standard of care” can be misleading because it is often used in medicolegal expert testimony regarding the care of a particular patient. But as much as this paper may reflect the authors’ concern for the medicolegal implication of IONM guidelines, it cannot justify a radical interpretation of evidence-based medicine supporting the principle that everything which is not supported by RCTs does not represent “meaningful medical evidence”. A more balanced framework would assess IONM for benefits vs the harms if withheld. In fact, the Institute of Medicine recommends this very approach: ‘‘Clinical practice guidelines fundamentally rest on appraisal of the quality of relevant evidence, comparison of the benefits and harms of particular clinical recommendations (italic added)…’’15 If the authors question the role of IONM due to the lack of class I and II evidence, shouldn’t they also question most of neurosurgical practice that also lacks such evidence? In fact, most of neurosurgical practice is not based on class I or II evidence. But, even in the absence of RCTs, many neurosurgical interventions are established as “standard of care,” and there will continue to be plausible interventions for which the evidence is not there yet.16 Considering this problem, Peter McCulloch recommends carefully acknowledging (but continuously re-evaluating) fully adopted surgical practice while demanding better (controlled) evidence to support new practices before they become fully adopted.17 One of us has already suggested this approach in IONM.18 These guidelines, as endorsed by both the AANS and the CNS, are promulgated by 4 neurosurgeons from a single institution: 2 senior adult spine neurosurgeons, 1 pediatric spine neurosurgeon, and 1 senior resident (5 yr training in critical appraisal of the medical literature). Although this might warrant “collective experience in clinical epidemiology and management of spinal column/spinal cord pathologies”, it is unclear to what extent it warrants “collective experience… in neurosurgical guidelines development and authorship.”1 Notably, none of the authors declares any specific expertise in IONM, which is, after all, the subject of these guidelines. Missing this expertise, the risk of inappropriate interpretation of IONM data such as specific techniques or warning criteria is very substantial. Unfortunately, the authors simply did not follow even the rudiments of proper guideline development. The Institute of Medicine provides an easily accessible treatise on Clinical Practice Guideline (CPG) development.15 Herein find pertinent quoted examples where the authors failed CPG principles: “The guideline development group should be multidisciplinary and balanced, comprising a variety of methodological experts…; CPG developers should use systematic reviews …; For each recommendation, the following should be provided… including: A clear description of potential benefits and harms…; External reviewers should comprise a full spectrum of relevant stakeholders, including scientific and clinical experts, organizations….”15 Clearly, the many failings of these guidelines could have been avoided simply by requesting the relevant opinions of IONM academic experts and societies such as the International Society of Intraoperative Neurophysiology, the American Clinical Neurophysiological Society, or the American Society of Neurophysiologic Monitoring. We are well aware of the high level of medicolegal litigation in the United States and the fact that Neurosurgery is the official journal of one of the two most prestigious North American societies. Nevertheless, the Editor and the Editorial Board should consider the large diffusion, impact, and influence within the global neurosurgical community of papers published in their internationally acclaimed journal. In this perspective, the endorsement of these guidelines by two of the most outstanding neurosurgical scientific societies inside and outside North America (CNS and AANS) is a matter of great concern. In summary, these guidelines are unacceptable because the article meets few or no CPG standards. It is a biased and incomplete narrative literature review and opinion piece. It fails to address the benefits of IONM vs the harms of withholding IONM. Readers should not accept its recommendations at face value. No patient should be deprived of IONM on account of these guidelines. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.