In recent years, the neurological manifestations of Ehlers-Danlos Syndrome (EDS) have become increasingly recognized due to their serious and potentially disabling implications. Though the causes of neurological symptoms have not been fully elucidated; Henderson et al. (2017) speculates that structural and functional defects of the brain and spine such as Chiari Malformation, syringomyelia, basilar impression, platybasia, and/or craniocervical instability may be partly to blame. Although many of these defects have been recognized for over half a century, our understanding is still very much evolving, and this has important implications on diagnosis and management. When abnormalities are severe and associated with obvious complications they can be managed surgically; however, the signs of “severe” defects and evidence of “obvious” complications are up for debate. This is largely because even pronounced defects may be asymptomatic, and because symptom presentation tends to be vague and non-specific, it is difficult to know with absolute certainty whether radiological findings are indeed responsible for a patient’s symptoms. Subsequently, while surgery can successfully manage symptomatic defects, indications for intervention are unclear, making the decision to intervene controversial. Surgical intervention is even more contentious in EDS, given the causes of neurological symptoms are unknown and likely multifactorial, and current standards of surgical care may not consider the inherent risks and special needs of this population. Thus, the efficacy of these procedures may differ for EDS patients, and treatment outcomes or long-term prognosis may not be comparable. This is especially true considering that EDS patients may also be more likely to experience intra and/or postoperative complications, delayed healing, or require revisions, making surgical intervention more risky. It is not difficult to understand why surgeons have reservations about operative management, especially considering the time and resources required to perform these interventions; however, few, if any, conservative options exist for the management of these defects, which may be highly symptomatic and severely disabling. Thus, even if surgery is only expected to offer modest relief, intervention may still be justified.
In light of the uncertainty surrounding the likelihood of clinical improvement and potential risk associated with these procedures, it is not certain whether the expected benefits of surgery truly outweigh the cost. Given the potential burden of this problem in the EDS population, and possible opportunities for management, there is great utility in clarifying indications for neurosurgical intervention. In September 2015 Critical Care Services Ontario assembled an expert panel to do precisely this, and at the end of an 8-week period the Panel produced a report that included an overview of the issues, criterion for neurosurgical referral and intervention, and additional considerations for neurosurgical pathways and the coordination of comprehensive care.
Aims
A critical appraisal of the Panel’s report was conducted to evaluate whether recommendations were sufficient to establish best practice guidelines. This appraisal relied on the AGREE II–an internationally recognized tool for the assessment of clinical guidelines. The principle aim of this appraisal was to highlight limitations of the EDS Expert Panel’s recommendations, identify persistent gaps in knowledge, and areas of uncertainty that need to be addressed in future research to make progress in the field.
Methods
The AGREE II include six domains including (1) scope and purpose, (2) stakeholder involvement, (3) rigor of development, (4) clarity of presentation, (5) applicability, and (6) editorial independence, and together, comprise a total of 23 questions. Each question is scored on a 7-point scale ranging from strongly agree to strongly disagree. Three independent reviewers were selected to appraise the Expert Panel’s Final Report, and scores were averaged for each reviewed across each of the domains. Agreement statistics were calculated and exceeded 0.75, suggesting a good degree of consensus among the reviewers. After averaging individual results, the Panel criteria received a score of 43.5%, and domains 1-3 were among the most problematic, receiving the lowest scores overall. It was concluded that while the recommendations were a vital first step, they were not sufficient to formulate best practice guidelines.
Results
Scope and purpose
The Panel did a good job of describing the sociopolitical context of the problem, but fail to elaborate on the principle source of conflict. Panelists report that there has been an increasing number of applications for out-of-country funding from EDS patients who are under the assumption that Ontario lacks the neurosurgical expertise to manage their condition. Panel speak to both side of the debate, explaining that while the expertise exists, physicians assert that”few…patients meet criteria to achieve clinical improvement.” In response, patients argue that while the province has expert neurosurgeons, “health care providers…have difficulty understanding indications for neurosurgical intervention.” The Panelists do not explain what about indications for intervention patient feel health care professionals do not understand, and why health care professionals feel so few patients meet criteria to achieve clinical improvement
[While physicians assert that few patients meet criteria to achieve clinical improvement, patients argue that health care providers have difficulty understanding indications for intervention. The Panelists do not reflect on the reasons why patients and physicians disagree, and therefore do not solve the problem ]
[…The lack of consensus stems from issues understanding the clinical significance of radiological finding, which is largely due to the lack of standardization related to the interpretation of imaging and use of craniometrics, as well as the lack of clinical outcome measures. The difficulty in capturing functional pathology of the craniovertebral junction also contributes to the problem; as well as the increasing specialization of medicine, which may miss pathology that arises between organ systems. In failing to address these core issues, the Panel’s recommendations were somewhat porous, inconcrete, and oftentimes unclear. The Panel’s description of the clinical problem did not differentiate between cause and effect and it was not clear what indications that the Panel actually sought to clarify, and what condition these indications applied. This led to numerous inconsistent and oversights throughout the report. Many defects relevant to understanding neurological symptoms in EDS were omitted (e.g. CSF leaks, intracranial hypertension, Tavlov cysts, spinal stenosis, brainstem compression, etc.); the report was missing detailed information about the clinical presentation, radiological findings, and indications for intervention for each defects, and the Panel did not mention the expected benefits and risks of the procedures that were indicated, nor did they estimate the impact their recommendations would have on the target population. The criteria also did not include differential diagnoses, or surrogate markers when primary indications are indeterminate. Options for conservative management were not included, and the Panel’s final included vague qualifiers which made it difficult to interpret. Lastly, recommendations were not based on a systematic review of the literature and the methods used to formulated recommendations were not transparent]
Reasons for conflict
The identification of defects on conventional imaging may rely on the use of craniometrics, however, these metrics have not been standardized for the assessment of these defects, and this has led to differences in clinical opinions. There are three main reasons why clinical impression may vary. Firstly, there are no gold standards for the diagnosis of these defects; therefore, it is up to the discretion of the physician to determine which tool is the most appropriate. Because each metric differs slightly in its approach to measurement, they may produce conflicting results, leading to differences in opinions. Second, while measurements are performed according to strict protocols, there is variance in the application of these protocols, and this can lead to discrepancies in measurement. Agreement statistics for most craniometrics are relatively good, however, Kappa scores range from 68% to 72%, implying that interpretation and clinical impressions can vary even when the same metric is used. The last, and among the most important problems, is the lack of consensus surrounding pathological thresholds, which leads to differences in conclusions even when the same metric is used and measurements are performed identically.
Debate surrounding pathological thresholds for craniometrics
Debate over pathological thresholds largely has to do with the lack of a consistent association with clinical outcomes. For example, while most agree that the pathological threshold for Chiari Malformation is ≥5mm below McRea’s line, accepted limits range from 3–5mm; however, herniations <1mm can still be associated with symptoms (i.e. Chiari 0 Malformation), while herniations <10mm may be asymptomatic (Elster & Chen, 1992; Milhorat et al., 1999; Strahle et al., 2011; Williams McVige & Leonardo, 1985).
While the lack of standardization explains some inconsistencies in the association between craniometrics and clinical outcomes, there are several additional reasons that may account for discrepancies. Among the most notable, is that craniometrics do not account for functional pathology, and may be indirect or overly simplistic representations of the mechanisms responsible for symptoms. While the pathophysiological mechanisms of these defects have not been fully elucidated; processes are theorized to be complex, and symptoms may arise as a result of an interaction between organ systems. For instance, symptoms associated with Chiari malformation may be secondary to changes in intracranial pressure (ICP) and blood flow, which results from the obstruction or disruption of the cerebrospinal fluid (CSF). The length of tonsillar descent is only one among many factors that influences a patient’s risk of developing complications such as intracranial hypertension and/or vertebrobasliar ischemia (Hofmann, Warmuth-Metz, Bendszus, & Solymosi, 2000). Other defects that crowd the CVJ and contribute to the stenosis of the foramen magnum such as odontoid retroflexion, reduction of the clivo-axial angle (CXA), and/or brainstem herniation, may potentiate the risk that any tonsillar herniation will disrupt normal physiological processes and result in symptoms. Comorbid conditions that impair or impact patients’ ability to compensate for the effects of anatomical abnormalities may greatly influence the risk that a defect will become symptomatic. For example, the brain may be able to compensate for changes in intracranial pressure by adjusting fluid volume in the body, altering blood pressure/volume, and modulating CSF uptake and production. Conditions that affect the patient’s ability to attenuate fluid dynamics in the body such as kidney disease, adrenal insufficiency, or underlying autonomic dysfunction, may contribute to the risk that these defects will impact normal physiology and becoming pathological. This may explain why some defects, while pronounced, may be symptomatic, while others that are mild to moderate are seriously symptomatic. The exploration and validation of novel tests and techniques to capture functional pathology of the CVJ such as CSF flow studies, diffusion tensor imaging, flexion-extension magnetic resonance imaging (MRI), and rotational computerized topography (CT), may provide surrogate markers that improve the clinical utility of craniometrics and should be explored in the future.
The lack of clinical outcomes measures
Another important problem is the lack of robust clinical outcomes measures for neurological symptoms, especially pain and changes in sensation and perception. While some neurological deficits such as reflex abnormalities and motor dysfunction can be observed in clinic, many symptoms of tremendous importance to understanding these defects are difficult to capture from an objective standpoint, and this can make robust scientific investigations difficult. New psychogenic metrics and nerve conduction studies such as somatosensory evoked potentials (SSEP), nerve block/ablation, and nerve conduction velocity studies (CVS) are being explored, but still require testing and validation. A recent systematic review looking at outcome methods used in clinical studies of Chiari malformation report “[t]he methods used to evaluate clinical outcomes…are inconsistent and frequently not comparable, complicating efforts to analyze results across studies. Development, validation, and incorporation of a small number of disease-specific patient-based instruments will improve the quality of research and care…” (Greenberg et al., 2019). The lack of valid clicometrics is a major limitation in understanding the clinical significance of these defects, and is an important subject for future research
The increasing specialization of medicine
Lastly, conventional practice and the increasing specialization of medicine may be contributing to uncertainty surrounding these defects. In isolation, individual defects may appear mild to moderate and clinically insignificant. However, the pathogenicity of these defects may be potentiated by other defects and morbidities, therefore, looking at systems in isolation may act as a source of confirmation bias that skews clinical perspective to favour defect severity over complexity. While mild to moderate defects are not necessarily implicit indications for surgical intervention, they may still be clinically significant, and ought to be distinguished and classified accordingly. This will not only aid in the study and understanding of these defects, but their diagnosis and management.
The Panel do not speak to any of these issues and therefore do not resolve the sources of uncertainty that fuel this conflict. This led to substantial oversights regarding the Panel’s understanding of the problem, and gaps/inconsistencies in their recommendations. These issues do not suggest the Panel’s work is not without value, however, the report may be more useful in edifying persistent gaps in knowledge, than providing clear guidance on the neurosurgical management of the defect in question.
The Panel’s diagnostic criteria and recommendations for intervention
Upon initial review of the report, it seemed like Panel members had difficulty deciphering between cause and effect, and delineating disease mechanisms. For example, the clinical problem was introduced broadly as “neurological manifestations of EDS,” and was refined to “brainstem and spinal cord symptoms…” This was characterized by symptoms including headache and widespread pain; several conditions and syndromes (i.e. urogenital dysfunction and cerviomedullary syndrome [CMS]); and specific defects of the spine and CVJ (Incl. tethered cord, craniocervical instability, and kyphosis of the clivo-axial angle [CXA]). Based on this description, it was not clear what the clinical problem was and what indications the Panel sought to clarify. The focus of the Panel’s report remained inconsistent throughout.
The only procedure the Panel recommend was a craniocervical fusion; however, criteria for neurosurgical referral read “evidence of craniocervical instability OR brainstem and cord dysfunction,” and there were a number of conditions listed under imaging evidence that may be a cause of brainstem/cord dysfunction but are not necessarily suggestive of instability, nor implicit indications for a fusion (i.e. Chiari Malformation, basilar impression, Syringomyelia, etc.). In fact, several findings including C1-2 instability, subaxial cord compression, assimilation anomalies, and subaxial instability may be counter-indications for a fusion, especially in the context of EDS in which fixation may lead to compensatory instability. The Panel do not distinguish between findings in this list that are suggestive of instability versus other causes of brainstem/cord dysfunction, and despite the inclusion and relevance of defects such as Chiari and Syringomyelia, the Panel do not discuss their clinical significance or management.
The Panel also did not address all defects subject to current debate. For instance, the Panel mention that tethered cord release is among procedures being performed out-of-country, and therefore, among conditions requiring clarification, however, indications for tethered cord syndrome were not included in the Panel’s criteria. Many other defects that were omitted despite being relevant to understanding neurological symptoms in EDS including, but not limited to, CSF leaks, intracranial hypotension, intracranial hypertension, Tavlov cysts, spinal stenosis, brainstem herniation, brainstem compression, and/or cranial settling.
The report was missing a lot of specific information about the clinical presentation of each defect/syndrome, radiological findings for each defect that serve as reliable indications for intervention, how reliable they are, and what the expected benefits and risks of the various procedures are. This is the very crux of the problem, and while there is still a great deal of uncertainty surrounding the clinical utility of various outcome measures (i.e. clinicometrics, craniometrics, psychometrics). The criteria also did not include differential diagnoses, or surrogate markers useful in rendering a verdict when the diagnosis is indeterminate. Options for conservative management were not included, and the Panel did not estimate what implications their recommendations would have on the target population.
Lastly, the Panel’s final recommendations included vague qualifiers such as “significant” and “possible,” which should be avoided as it adds to the subjective nature of interpretation, and may lead to continued disagreement and confusion.
Rigor of development
Another important limitation of the Panel’s recommendations were not based on a systematic review of the literature and the methods used to formulated recommendations were not transparent. Admittedly, the Panel mention that there are few high quality studies specific to EDS, thus, recommendations regarding surgical candidacy were based on expert consensus. While a systematic review on the neurosurgical management of EDS may not have yielded a sufficiently large number of studies; the review process still should have followed systematic and clear methods. Because recommendations relied on expert consensus, strategies such the Delphi Techniques could have been used and clearly described so decision-making remained methodical and reportable.
Stakeholder involvement and Applicability
The Panel did a good job involving various Stakeholders and representing the various medical specialties, and the report did consider the various organizational barriers EDS patients in pursuit of care. The Panel then conclude with the following remark “information will require further dissemination…CCSO [is] to develop a brief background on Ehlers-Danlos Syndrome and share referral criteria with Ontario’s Medical Community.” criteria included in their report require dissemination within the medical community. Criteria still reflect a substantial about of uncertainty, and despite not including concrete, evidence-based indications that address the full scope of the clinical problem, the Panel assert criteria are ready to be implemented in practice. This is concerning because recommendations do not truly clarify indications for intervention, and therefore may just perpetuate uncertainty and disagreement, especially under the assumption that the current recommendations are complete.
Conclusion
Building upon the EDS Expert Panel’s work and addressing gaps in knowledge
In order to ensure the accuracy of risk reporting and patient prognosis, research should also focus on improving the diagnosis and classification of EDS, and specifically, developing criteria that distinguish between hEDS, other forms of EDS or HCTD, and the various joint hypermobility spectrum disorders. Measures for neurological symptoms, as well as other patient-centred clinical outcome (i.e. quality of life, disease-specific surveys, psychogenic metrics, and functional outcome measures) are also needed, and require disease-specific validation. Exploring novel diagnostic modalities capable of capturing functional and pathophysiological processes of the CVJ and CNS is needed, and requires correlation with craniometrics and validation against patient-important clinical outcome measures. Interdisciplinary perspectives that consider the combined effects of comorbid conditions and the interactions between organ systems are also needed. Next, the development of robust clinical practice guidelines necessarily depends on a systematic review of the literature. To capture a sufficiently large number of studies, eligibility criteria can be expanded to include other HCTD or the broader population diagnosed with one or more of the defects of interest. Special considerations related to risk factors inherent to EDS can be theorized until sufficient outcome data has been generated. To ensure stakeholder inclusiveness, data gleaned from the literature should consider endpoints deemed important to experts in the field as well as patients. Patient perspectives can be gleaned using a population-based survey; where expert opinion can be collected by circulating review protocols in the medical community. Data that does not exist should be generated by prospective investigations/trials in specialized clinics with access to the technology and medical expertise required to diagnose and manage the defects of interest.
[Understanding the clinical significance of these defects and their involvement in the neurological manifestations of EDS hinges on future research that understands the causes of symptoms in EDS patients, explores functional imaging of the CVJ and CNS, and standardizes the use of clinical and craniometrics. Future guidelines should be based on a systematic review of the literature.]