Diagnosis and Treatment of Tethered Spinal Cord: A Systematic Review

With this review, we followed the Agency for Healthcare Research and Quality (AHRQ) Evidence-based Practice Center methods detailed in a publicly available protocol.^10,11  We refined the key questions with the help of interest holders in key informant interviews. A multidisciplinary technical expert panel supported the review throughout and provided input on outcome prioritization. A draft evidence report of the findings was published on the AHRQ Web site to elicit public comments. The results of this review are aimed to support a Congress of Neurologic Surgeons clinical practice guideline.

For primary research studies, we searched PubMed (biomedical literature), Embase (pharmacology emphasis), CINAHL (allied nursing), Web of Science (technical innovation), and SCOPUS (general research). We searched US and international research registries (clinicaltrials.gov, ICTRP) to capture a complete record of adverse events. We used existing reviews for reference mining, which were identified through the same databases used for primary research, plus the Cochrane Database of Systematic Reviews and PROSPERO. We searched for clinical practice guidelines using the ECRI repository, G-I-N, MagicApp, and ClinicalKey. We reviewed the literature suggested by peer reviewers, public posting of the draft evidence report, submissions through the Supplementary Evidence and Data for Systematic Reviews portal to AHRQ, or responses to a Federal Register notice. The search strategies are documented in the Supplemental Information.

We used detailed pre-established criteria to determine the eligibility of publications, as outlined in the Supplemental Information and summarized as follows:

• Diagnosis: Studies of patients assessing for tethered spinal cord, evaluating screening and diagnostic approaches, tools, and criteria, such as physical examination, urodynamic studies, MRI, myelography, computed tomography (CT), or ultrasound, and revealing diagnostic performance, adverse events of diagnostic procedures, or clinical impact of the diagnosis.

• Prophylactic surgery: Studies of patients with tethered spinal cords and no symptoms or those who are marginally symptomatic without functional deficits undergoing prophylactic surgery and reporting on patient health outcomes and other patient effects such as quality of life.

• Symptomatic treatment: Studies of patients with symptomatic tethered spinal cords, evaluating surgical or non-surgical treatment or management interventions, such as surgical detethering, or other surgery (eg, spine-shortening vertebral osteotomy, spinal cord transection), physical therapy, bladder therapy for bladder function, or bracing, and revealing patient health outcomes and other patient effects, such as quality of life.

• Repeat tethering: Studies of patients who experienced retethering after previous spinal detethering surgery, evaluating surgical interventions, and revealing patient health outcomes and other patient effects, such as quality of life.

The authors of diagnostic studies had to report on a reference standard, and the authors of treatment studies had to report on a concurrent control group that either received no or an alternative intervention to be eligible for inclusion. Studies without reference standards or control groups were retained as a resource for clinicians but did not contribute to the evidence statements because of study limitations (see Supplemental Information).

To reduce reviewer errors and bias, all citations were reviewed by a human reviewer supported by a machine learning algorithm. Citations that were deemed potentially relevant were obtained as full text. Each full-text article was independently reviewed by 2 independent literature reviewers; disagreements were resolved through consensus. Publications involving the same participants were consolidated into 1 study record.

We abstracted details on the study, participant, intervention, comparator, outcome, and results. We contacted all authors of eligible studies and asked for any additional data relevant to the review. We assessed bias on the basis of patient selection, index test, reference standard, and flow of timing for diagnostic studies.¹²  We assessed the selection, performance, attrition, detection, reporting, and other sources of bias for treatment studies.¹³ 

The following 4 key questions guided the synthesis: Key question 1: What is the accuracy of radiographic and other diagnostic criteria in diagnosing tethered spinal cord? Key question 2: What are the benefits and harms of prophylactic surgery for asymptomatic tethered spinal cord patients? Key question 3: What are the effectiveness, comparative effectiveness, and harms of surgical and non-surgical treatments for symptomatic tethered spinal cord? Key question 4: Among individuals who experience retethering after spinal detethering surgery, what are the benefits, harms, and long-term outcomes of another surgery, compared with no treatment?

We selected key outcomes with the help of the technical expert panel, considering clinical relevance and importance to patients to guide medical practice. We calculated effect sizes when possible and determined the feasibility of a meta-analysis for each intervention and outcome. We considered study limitations, consistency, directness, precision, and reporting bias to determine the following 4 strength of evidence (SoE) categories: high, moderate, low, or insufficient. We assessed the applicability and generalizability to current US clinical practice by evaluating patient, intervention, and setting factors.

With the searches, we identified 6285 citations (see Fig 1). Of these, we obtained 2005 as full-text studies. In total, 103 controlled studies met the inclusion criteria (see the Supplemental Information for the list of included studies). In addition, 355 case series provide further clinical information on the evaluated tests and interventions, but these did not contribute to the evidence statements because of study limitations. In total, 154 publications were retained as background and reference mined. The risk of bias and applicability concerns are documented in the Supplemental Information. In the following sections, we have summarized the main findings for the key questions.

We identified 59 diagnostic studies. The studies were conducted in 18 countries, the majority of which were US-based. The studies were generally small but ranged from 4 participants to a large-scale retrospective review of 3884 patient records.¹⁴  The studies included participants of different ages, one-third of which were exclusively pediatric patients. The main potential source of bias across the studies concerned patient selection. Most applicability concerns were associated with the patient population, either because the study included patients who were more complex (eg, all had severe neurologic symptoms) or the eligibility criteria were much narrower than those seen in usual practice.

The tests evaluated in the studies included the use of MRI, ultrasound, CT, myelography, evoked potentials, urodynamic studies, and intraoperative neurophysiological monitoring. The authors of MRI studies evaluated different sequencing modalities.¹⁵  The authors of most studies did not report concrete diagnostic accuracy measures, such as sensitivity or specificity. The authors of multiple older studies reported on the advantages of using MRI for the diagnosis of tethered spinal cords because of the achieved visual quality.

The reported sensitivity for MRI ranged from 99%¹⁶  to 62% (corresponding specificity 61%).¹⁷  The reported specificity ranged from 100% (corresponding sensitivity 93%)¹⁸  to 61% (corresponding sensitivity 62%).¹⁷  We downgraded SoE because of imprecision, given the wide range of reported results. The results for ultrasounds were also variable; the reported sensitivity ranged from 96% (corresponding specificity 96%)¹⁹  to 36% (corresponding specificity 84%;²⁰  low SoE). The reported specificity ranged from 96% (corresponding sensitivity 96%)¹⁹  to 84% (corresponding sensitivity 36%;²⁰  moderate SoE). Additional results are shown in Table 1.

We identified 8 studies that provided some information on the effects of prophylactic surgery. The studies were conducted in 6 countries, with 2 studies each from the United States and Japan. Sample sizes ranged from 41 to 354 patients, but the subgroups of asymptomatic patients were small. The studies included patients of different ages, and the oldest participant was 17. In this small set of studies, selection bias was suspected in several studies. The applicability issues included narrow eligibility criteria for patients, short follow-up times, and a suspected different level of care from current US standards. The study authors evaluated prophylactic surgical detethering approaches, such as in the context of scoliosis correction, or assessed other questions, such as the postoperative duration of the horizontal decubitus position. The studies compared the results of no surgery, delayed surgical detethering, or a different surgical intervention.

Table 2 contains a summary of the results for all key outcomes. The authors of 5 studies reported on neurologic status, and most studies reported that patients remained stable at follow-up after prophylactic surgery, but with few exceptions did not present equivalent data on the control group (eg, conservative observation).²²  One study reported benefits for urologic function after concomitant prophylactic detethering, compared with no detethering or detethering before scoliosis surgery.²³  The authors of 6 studies reported on post-operative complications. This included a study revealing an increased risk of surgical site infection with concomitant detethering compared with no detethering or detethering before scoliosis surgery (P = .01).²  Two studies revealed that one-third of patients experienced a surgical complication, and the remaining study authors reported on individual adverse events. SoE was low for all statements because the authors of no 2 studies reported on the same outcome measure, there were insufficient data to compute effect sizes, and data were based on observational studies with suspected selection bias. We identified no data on other key outcomes, including the need for repeat surgery.

We identified 38 studies addressing symptomatic tethered spinal cord treatment. The studies were conducted in 15 countries, with the largest group being US-based studies. Most studies were small, but sample sizes ranged from 6 to a large retrospective review of 6457 patients.²⁴  The studies included patients of different ages, but most studies exclusively used pediatric samples. Several studies were rated as having a high risk for bias because of suspected selection bias in which participants were not randomly assigned to treatments but were followed in observational cohorts. Multiple studies were flagged in areas for deviating from standard clinical practice in the United States. The selected studies also had issues such as addressing more complex patients than those who are typical in clinical practice.

The authors of most of the identified studies evaluated surgical detethering approaches. Of these, some evaluated combining cord release with other surgical interventions versus a sequential approach,^25–31  comparing surgical techniques such as laminectomy versus laminoplasty,³²  laminectomy versus the drainage approach,³³  duraplasty versus primary dural closure,³⁴  laminoplasty versus hemilaminectomy,³⁵  microsurgery versus open surgery,^36–38  untethering with or without a laser,³⁹  or complete versus incomplete detethering.⁴⁰ 

One study compared detethering surgery and spine-shortening osteotomy,⁴¹  and another compared detethering and spinal fusion in a large retrospective review.²⁴  Other studies compared surgery with the outcomes of patients who declined surgery or who were purposefully managed with standard medical treatment or observation.^24,42–47  Other studies compared different timings of the intervention after diagnosis, for example, early surgery versus surgery after conservative treatment first or surgery in the first year of life versus later.^22,48,49  The authors of some studies evaluated the incremental value of using intraoperative monitoring during surgery.^50–52 

A small set of studies addressed horizontal decubitus position after surgery, the ideal duration of or management with acetazolamide, or wound treatment approaches after detethering surgery.^53–59  The strongest evidence base was identified for open surgical detethering. The identified studies revealed improved neurologic status compared with control groups (low SoE).^32,41,76  However, open surgical detethering was associated with postoperative complications. Studies revealed cerebrospinal fluid leakage, in addition to other surgical complications^{22,31,32,34,36,41,42,47}  (moderate SoE). Table 3 includes a summary of the results for all key outcomes. The SoE was low for all other statements because only single studies contributed to each evidence statement, there were insufficient data to compute effect sizes, and data were based on observational studies with a high risk of confounding. Despite the substantial number of identified studies, the authors of no studies reported on a priori-selected, patient-centered outcomes.

We identified 3 studies addressing the treatment of retethering. The studies were conducted in the United States and Germany. The oldest participant was 26 years old (when specified).⁴⁰  Selection bias was identified as an issue in all retethering studies. The potential for reporting bias was also a concern in this small set of studies, which had short follow-up times and lacked numerical data for central outcomes. Applicability was limited in 1 because the patient sample composition did not reflect the full range of patients.

The studies compared vertebral column shortening versus revision detethering,⁶⁰  complete versus incomplete detethering during cord release,⁴⁰  and bilateral dural incision undercutting versus the conventional dural midline approach.⁶¹ 

Table 4 includes a summary of the outcomes of different interventions. The authors of 2 studies reported improvement in neurologic status after surgical detethering (low SoE) but reported no data on the control group or did not detect a statistically significant difference between groups.^40,61  One study revealed that 50% of patients had improved urinary function after spinal column shortening, whereas 75% of patients had worsened bowel and bladder function after revision detethering.⁶⁰  One study revealed improvement in bladder dysfunction in 26% of patients after repeat detethering (no data on the comparator group).⁶¹  One study revealed that 75% of patients reported improved pain after spinal column shortening.⁶⁰  The SoE was insufficient for multiple outcomes when no study was identified in which the authors reported on the outcome of interest.

With our review, we identified a large volume of research and addressed 4 critical questions relevant to patients with tethered spinal cords. This review provides a comprehensive dataset to support clinicians and researchers and documents remaining research gaps.

Because accurate and timely diagnosis is critical, clinicians aim to use tests associated with the highest sensitivity, specificity, and accuracy rates in clinical diagnostic tests. Our systematic review assessed the accuracy of diagnostic tests commonly used for patients with suspected tethered spinal cords. The identified studies evaluated MRI, ultrasound, CT, myelography, evoked potentials, urodynamic studies, and intraoperative neurophysiological monitoring in diagnosing tethered spinal cords. MRI had the strongest evidence base, with medium SoE for medium to good sensitivity and medium SoE for medium to good specificity. Ultrasound had low SoE for medium to good sensitivity and moderate SoE for medium to good specificity. On the other hand, both myelography and evoked potentials had low SoE for medium sensitivity and low SoE for medium specificity. There are several limitations to our analysis. First, the authors of most identified studies did not report diagnosis accuracy estimates such as sensitivity, specificity, or area under the curve. SoE was downgraded because of imprecision and inconsistency, given the lack of diagnostic effect estimates or range of reported results across studies and the absence of replication of results, thereby limiting the conclusions that can be drawn.

One of the most important considerations in tethered spinal cords is determining how best to manage asymptomatic patients, and a critical issue for clinicians and patients is how best to follow asymptomatic patients to detect any neurologic progression. Important considerations include the frequency of interval clinical assessments and the timing and type of diagnostic tests to triage patients. Historical clinical experiences suggest that many asymptomatic patients can develop irreversible neurologic injuries resulting in motor or sensory loss or bowel or bladder dysfunction. However, if prophylactic surgical treatment is pursued when patients are asymptomatic, patients and clinicians need to weigh surgically induced injury and adverse events. Most of the studies identified in this review included no asymptomatic patients or had a mixture of symptomatic and asymptomatic patients. Few studies were devoted to studying the benefits and harms of prophylactic surgery. Although predominantly suggesting benefits, sample sizes were small, and SoE was low.

In evaluating the effectiveness, comparative effectiveness, and harms of surgical and non-surgical treatments for patients with symptomatic tethered spinal cords, our review revealed that most studies revealed favorable outcomes associated with the surgical treatment of tethered spinal cords. Neurologic status was mostly improved after surgery, compared with no intervention. Early surgery is associated with more favorable neurologic outcomes in patients who are mildly symptomatic and diagnosed early on in their disease course. Conversely, surgery was associated with increased surgical risks, including infection, cerebrospinal fluid leaks, and wound complications, compared with non-surgical treatment. A significant limitation of many studies was the lack of a comparative analysis with a control group. In addition, sample sizes were small for many of the studies reviewed.

Even with improvement from prophylactic surgery or early detethering surgery, patients can develop recurrent tethered cord syndrome that leads to progressive neurologic and functional decline. In patients with recurrent tethering, a critical concern is weighing the benefits, harms, and long-term outcomes of another surgery compared with no treatment. Our systematic review revealed benefits, including in the areas of neurologic status, bowel and bladder functioning, and pain improvement, associated with revision detethering surgery. There was a lack of studies in which authors reported other important outcomes, including the number of patients with adverse events or the need for additional repeat surgeries.

With our systematic review, we have established a large and unique collection of relevant research on tethered spinal cords. It is based on a comprehensive literature search in multiple databases and research registries, incorporating input from the published literature and additional data provided by study authors. More than 450 studies contribute to this review and document the relevant research undertaken to support patients with suspected or confirmed tethered spinal cords.

In the review, we also identified the limitations of the research on tethered spinal cords. The current state of knowledge and the available data have many uncertainties. These are reflected in the wide variation in clinical practices in the community and clinical guidance is missing. To date, the Congress of Neurologic Surgeons has published guidelines for pediatric myelomeningocele;⁶²  guidelines by the National Institute for Health and Care Excellence on open prenatal repair for open neural tube defects in the fetus⁶³  and on fetoscopic prenatal repair for open neural tube defects in the fetus are also available.⁶⁴  In addition, a guideline published by the Spina Bifida Association,⁶⁵  the European Association of Urology/European Society for Pediatric Urology guidelines on the management of neurogenic bladder in children and adolescents,⁶⁶  guidelines by the International Children’s Continence Society on tethered cord syndrome in occult spinal dysraphism,⁶⁷  and an expert consensus document on best practices to minimize wound complications after complex tethered spinal cord surgery are available.⁶⁸  However, the existing guidelines are related to the diagnosis or high association with tethered cord but do not constitute clinical practice guidance for the diagnosis or treatment of tethered spinal cord. A planned Congress of Neurologic Surgeons guideline will address an important gap in clinical practice.¹⁰  Additionally, the current review indicates that although there are data to support the current practice to diagnose, triage, and treat patients with tethered cords, there is a significant gap and need for better clinical research, with comparative studies with well-defined and measurable outcomes.

Improved reporting and concurrently controlled studies are urgently needed to advance the evidence base for this important clinical condition. Although we found numerous studies in which the authors reported on experience with a diagnostic modality, the reporting of diagnostic accuracy outcomes was lacking. Despite the importance of the outcomes, we identified no treatment studies in which the authors reported on key outcomes such as the number of participants with adverse events, and the available research for many other important outcomes was also insufficient to derive any evidence statements. Similarly, we identified hundreds of case series in which the authors reported on patients with tethered spinal cords. However, reporting within these studies was sparse, and limitations related to study design, in particular the lack of control or comparator information and the frequent lack of numerical outcome data for the complete sample, provided little information that can be adapted to clinical practice recommendations.

The authors of the existing research need to improve studies by including a control or comparison group. A tethered spinal cord is a progressive disease, and single-arm studies cannot distinguish the effects of the intervention from the natural course of the disease. The authors of diagnostic accuracy studies should clearly report on the performance of a test compared with a reference standard. For treatment studies, assignment to treatments not based on patient characteristics, such as randomized controlled trials, are needed to formulate stronger evidence statements. Ideally, the authors of randomized controlled trials would stratify tethered cord patients with similar degrees of neurologic impairment. Such studies might yield results that address the question of whether surgery or non-surgical treatments are better for tethered cord patients. Alternatively, large prospective observational studies that closely follow asymptomatic and symptomatic tethered cord patients with quantitative outcome measures and neurologic assessments would provide valuable information on the natural history and risk of progression for these patients.

Despite the documented use of and clinical experience with common diagnostic procedures in case series for tethered spinal cords, many open questions remain, particularly those related to the accuracy of these modalities. Our review revealed limited relevant data on diagnostic modalities beyond MRI and ultrasound. The authors of future research should more thoroughly investigate the different options available for patients to further refine diagnostic accuracy.

The authors of future research should report outcomes, including adverse events, quantitatively and systematically. For all key questions, experts advised on the key outcomes, and, despite the research volume, there was insufficient evidence for a number of outcomes. For tests used to diagnose a tethered spinal cord, we found no studies in which the authors reported information on over- or undertreatment or any clinical impact of using the diagnostic modality. For prophylactic surgery, we have no data on the effects on symptom scores, ambulation, quality of life, pain, the number of patients reporting adverse events, and the need for repeat surgery. Despite the large number of treatment studies evaluating interventions in symptomatic patients, we did not identify a study in which the authors reported on a standardized scale to assess relevant tethered spinal cord symptoms. Although we identified studies in which the authors reported on some postoperative complications, the authors of this small number of studies reported some conflicting results, and more research is needed for definitive evidence statements. The evidence base for addressing spinal cord retethering was insufficient for many expert-identified patient-centered outcomes that could guide clinical practice, despite the importance of the outcomes, indicating the need for more research. In general, greater emphasis is needed on reporting on patient-centered outcomes, such as ambulation or renal function preservation, which are key considerations for patients that could affect their decision to undergo surgical detethering.

We thank David Bauer, Azniz Bane, Daniel Donoho, Judy Thibadeau, Linda Bambrick, Eisha Christian, Wende Gibbs, Steven Hwang, Stephanie Kielb, Yogi Matharu, Justin Lantz, Robin Bowman, Lionel Banez, and Jennie Dalton for helpful comments. We thank Saad Ali, Haluk Altiok, Joshua Chern, Shlomi Constantini, Michael Fehlings, Toshiaki Hayashi, Fritz Karrer, Marc Levitt, Jonathan Norton, Vedantam Rajshekhar, Paul Steinbock, Goichiro Tamura, Adrian Elmi Terander, and Seow Wan Tew for assistance with the dataset. We thank Bob Bentz, Kevin Blythe, and Jeremy Miles for assistance with the analysis.

AHRQ

Agency for Healthcare Research and Quality

CT

computed tomography

SoE

strength of evidence