Endoscopic Surgery for the Treatment of Osteodiscitis: A Systematic Review
Article information
Abstract
Spondylodiscitis is a serious spinal infection that can cause spinal instability, neurologic deficits, and paralysis. Although endoscopic procedures are becoming increasingly common, their role in spondylodiscitis management remains undefined. In this systematic review, PubMed, Embase, and MEDLINE were searched for studies published from January 1950 to March 2024 on percutaneous endoscopic debridement and drainage (PEDD) surgery for spondylodiscitis. The inclusion criteria included clinical trials or case series of spondylodiscitis patients treated with endoscopy. Primary outcomes included culture yield, complications, reoperations, treatment success rates, and pain scores. Our review included 40 studies that assessed 667 patients. Most studies presented low-level evidence, consisting of case reports (40.0%) and case series (60.0%). Endoscopic techniques demonstrated a positive culture yield of 67.3%, with an overall complication rate of 7.8% (95% confidence interval [CI], 5.35%–10.47%) and a reoperation rate of 13.1% (95% CI, 10.08%–16.31%). There were no complications related to cerebrospinal fluid leaks or meningitis. The medical treatment success rate was higher in patients who underwent concurrent instrumentation than in those who received endoscopic debridement alone (96.4% vs. 85.9%, p=0.003). Thirteen studies presented patient-reported pain outcomes, and all cohorts showed improvements. This review suggests that PEDD shows promise as a safe and effective adjunctive treatment for spondylodiscitis, offering low complication rates and successful treatment. PEDD may play a role within a multidisciplinary treatment approach. Future clinical trials are needed to determine whether early PEDD can improve the overall success of medical therapy and reduce the need for more extensive debridement and fusion surgery.
INTRODUCTION
Spondylodiscitis is a serious spinal infection of the intervertebral disc and affects adjacent vertebral bodies potentially leading to spinal instability, neurologic deficits, and increased morbidity. The infectious extent can also include abscesses in the epidural space, with the paraspinal or psoas muscles, and multilevel involvement. It is estimated that the annual incidence of spondylodiscitis ranges from 0.4 to 2.4 per 100,000 people in the Western world [1]. Generally, spondylodiscitis arises due to bacterial causative agents that directly inoculate the vertebrae or through hematogenous spread of systemic infection, with common organisms stemming from Staphylococcus, Escherichia coli, Streptococcus pneumoniae, or Salmonella species [2]. Medical management, which remains the mainstay of treatment, consists of broad-spectrum antibiotic regimens upon first diagnosis followed by 4- to 6 weeks of targeted antibiotics once culture results are known [3].
Surgical treatment in spondylodiscitis can occur in the acute period, late period, or even after clearance of infection depending on the surgical goal. Indications for primary surgical management can include neurologic deficits, large epidural abscesses with spinal cord compression, and spinal instability. Indications for delayed surgical management of spondylodiscitis include a failure of infection to respond to antimicrobial treatment, development of significant neurologic deficits, instability and/or progressive deformity of the spine [4]. Thereby, the goal of surgery includes debriding the infected tissue, decompressing neural elements, and addressing the spinal alignment and instability [5].
There are still many cases of failure of medical management that then go on to require surgical intervention. Several scoring systems have been devised to predict such patients who will fail medical management [6,7]. For example, the MSI-20 score was developed to estimate the mortality risk of patients with spinal infections and to help decide whether surgery was superior to medical management based on factors such as body mass index, American Society of Anesthesiologists physical status classification, septic state, age-adjusted Charlson Comorbidity Index, neurological deficits, and C-reactive protein (CRP) levels at diagnosis [8]. The SITE (Spinal Infection Treatment Evaluation) score is another tool that was introduced to standardize surgical decision-making on patients with de novo spinal infections by evaluating neurological deficits, infection location, and extent, radiological changes, and overall patient comorbidities [6]. Although developing models predictive of failure of medical management and the need for surgical intervention is warranted, therapy to change the course of spondylodiscitis is critical.
Spinal endoscopic techniques present several potential advantages that may increase the success of medical therapy. With endoscopic techniques, more biopsy tissue is obtained which could increase the diagnostic yield, reduce the need for repeat biopsies, and allow for more cases of targeted antimicrobial therapy. Concurrent irrigation and debridement at the time of biopsy may aid in infectious source control and increase the efficacy of antimicrobial therapy. It is important to note that increased efficacy of medical management would reduce the need for salvage surgical techniques. With lower surgical approach complications and morbidity, endoscopic techniques such as percutaneous endoscopic debridement and drainage (PEDD) can be an option for patients deemed too high of a risk for surgical intervention [9].
The purpose of this systematic review was to identify and synthesize the results of current interventions that assessed the use of endoscopic techniques in the management of spondylodiscitis. We sought to evaluate if the current literature sufficiently supports any of the potential advantages of endoscopic techniques to change the disease course. If so, we sought to define a role for spinal endoscopy in the management of spondylodiscitis and identify any limitations of the technique that may prevent its wide adoption.
METHODS
1. Search Strategy
This systematic review has been reported according to the PRISMA-ScR (Preferred Reporting Items for Systematic Review and Meta-Analysis extension for Scoping Review Checklist) [10]. A protocol was formulated prior to starting the review to guide the search and extraction of data. The search terms (("endoscope"[All Fields] OR "endoscop*"[All Fields] OR "endoscope"[MeSH Terms] OR "endoscopes"[All Fields] OR "endoscopic*"[All Fields]) OR (("spinal"[All Fields] OR "spinal*"[All Fields]) AND "endoscop*"[All Fields])) AND ("osteodiscitis"[All Fields] OR ("epidural abscess"[MeSH Terms] OR ("epidural"[All Fields] AND "abscess"[All Fields])) OR ("vertebral"[All Fields] AND "osteomyelitis*"[All Fields]) OR "discitis"[All Fields] OR "osteodiscitis*"[All Fields] OR "diskitis"[All Fields] OR "spondylodiskitis*"[All Fields]) were ran through 3 electronic databases (Ovid MEDLINE, Ovid Embase, and PubMed) from January 1950 to March 2024. Citation chaining of existing systematic reviews/meta-analyses, clinical trial registries, and relevant primary studies was conducted to identify additional literature.
2. Inclusion and Exclusion Criteria
Inclusion criteria included (1) all patients presenting with spondylodiscitis, (2) patients who received endoscopic treatment for spondylodiscitis, (3) study-reported outcomes related to any of the following: the type and effects of treatment, duration of hospital stay, rates of reoperation/readministration of treatment, mortality rate (due to infection), need for further surgery, % of success from medical management, and any complications. Study designs including case reports, case series, cohort studies, and retrospective or prospective reviews were included. Only full-text articles were included.
Exclusion criteria included (1) review papers where the authors do not present new patients, (2) systematic reviews or meta-analyses, (3) texts not in English.
3. Study Selection & Data Extraction
The study selection was conducted independently by 2 reviewers, who screened titles and abstracts for potential inclusion (RR, SJ). If a study met the inclusion criteria based on the title and abstract, the full-text article was retrieved and reviewed for potential inclusion independently by 2 reviewers (RR, SJ).
Following study selection, 2 investigators (RR, SJ) independently collected the following variables: study design, patient population, mean age of the cohort, diagnosis type, location of infection on spine, anesthesia type, preoperative administration of antibiotics, complications, clinical and radiographic outcome measures reported, and patient-reported outcomes. Disagreements were resolved by discussion and consensus. All the data was extracted into an Excel document for organization and synthesis before tabulating.
4. Critical Appraisal of Evidence
The reviewers (RR, SJ) utilized the National Health and Medical Research Council (NHMRC) levels of evidence and grades of recommendations to critically evaluate the studies [11]. The NHMRC framework offers a systematic approach to classify research according to levels of evidence, taking into account whether the study design adequately addresses the research question. The grade of recommendation assigns an evidence value to a research study based on various factors, including quality, consistency, clinical impact, generalizability, and applicability. The reviewers conducted independent assessments of the studies, and any discrepancies were resolved through discussions. The level of evidence is subsequently presented.
5. Statistical Analysis
For every outcome, 2 numbers were extracted from all the studies: the total number of patients with reported the outcome and the total number of patients included in studies that reported on the complication. The first number was divided by the second number to estimate the incidence of each complication. Ninety-five percent confidence intervals (CIs) for the incidence of each complication were determined using exact binomial methods.
RESULTS
Our search initially yielded 569 unique studies, of which 64 remained after the title and abstract screening. After performing full-text screening, 40 studies met the inclusion criteria for our systematic review [12-51]. Figure 1 delineates the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) search strategy.
Study flow chart using PRISMA (Preferred Reporting Items for Systematic Review and Meta-analysis) 2020 guidelines.
The majority of studies were low-level evidence, mostly consisting of case reports (40.0%) and case series (60.0%). Specifically, 36 studies (90%) were graded as level IV, and 4 studies (10%) were graded as level III-2 based on the NHMRC framework [11]. Furthermore, most studies took place at academic hospitals throughout Southeast Asia (85.7%), with the remaining studies taking place in North America (11.9%), and South America (2.4%). There were 667 total participants included across all 40 studies after consideration of duplicates, the majority of whom were male (62.6%). The average age was found to be 60.9 years old (Table 1).
Characteristics of the included studies and demographic characteristics of the patients
1. Surgical and Postoperative Data
Operative characteristics and primary outcomes are delineated in Tables 2–7.
Diagnostic and operative characteristics of the included patients
Cultured microorganisms in the operating room
Reported postoperative complications following endoscopic treatment in 32 studies (449 total patients)
Reported revision procedures following endoscopic treatment in 29 studies (466 total patients)
Preoperative and postoperative patient-reported scores following endoscopic treatment
Infection treatment success rates
Patients with spinal instability and significant neurological deficits were excluded in most papers. 475 patients underwent PEDD alone and 121 patients underwent PEDD and concurrent fusion at the discretion of the treating team. Endoscopic procedures were initiated when surgery was indicated in all included studies.
Spondylodiscitis was the most reported pathology (81.1%) followed by the presence of epidural abscess (9.4%) and psoas/paraspinal abscess (9.4%). Most cases took place in the lumbar region (90.2%), with 9.5% of operations falling within the thoracic region and 0.3% in the cervical region. Local anesthesia was the predominant choice (66.6%), followed by general anesthesia (33.3%) in the studies that reported it. Furthermore, reported postoperative drains were left at the procedure site in 83.8% of patients, with 16.2% having no drains (Table 2).
Of the papers that reported endoscopic biopsy culture data, a composite positive culture yield was 67.3% (372 out of 553 patients). Preoperative antibiotic administration was specified in 336 out of 553 cases (60.7%), with 47.0% of these cases receiving preoperative antibiotics. 412 pathogens were reported in 404 cases (Table 3).
Complications reported in 32 of the 40 studies (n=449 patients) note an overall complication rate of 7.8% (95% CI, 5.35–10.47). Of the complications reported, postoperative paresthesia (3.3%; 95% CI, 0.89–3.34) was the most prominent, followed by postoperative radicular pain (2.0%; 95% CI, 0.89–3.34), screw loosening (0.4%; 95% CI, 0.00–1.11) and surgical site infection (2.0%; 95% CI, 0.89–3.34). Of note, all surgical site infections occurred during concurrent spinal fusion without specification of infection of the endoscopic incision versus the separate incision for spinal fusion. There were no reported complications related to cerebrospinal fluid (CSF) leaks or meningitis (Table 4).
Twenty-nine studies encompassing 466 total patients included surgical follow-up of at least 3 months to give an overall reoperation rate of 13.7% (95% CI, 10.68–16.74). Subsequent operations included 35 (7.1%; 95% CI, 4.79–9.33) debridement and fusions and 18 (3.6%; 95% CI, 1.95–5.31) repeat endoscopic procedures in the context of persistent infection or back symptoms. There were 15 (3.0%; 95% CI, 1.48–4.56) delayed fusions after eradication of infection for reasons such as persistent back pain, developing kyphosis, or concern for spinal instability (Table 5).
Seventeen studies provided data on patient-reported outcomes with regard to visual analogue scale (VAS), Oswestry Disability Index (ODI), or McNab scores [15-20,22,25,26,29,34,37,39,44,45,47]. Specifically, 13 studies that comprised 283 total patients provided preoperative and postoperative VAS scores, in which the median scores improved from 8.06±0.72 preoperatively to 2.43±0.58 postoperatively [15-17,19,20,22,25,26,34,39,44-45,47]. Furthermore, 7 studies that assessed 184 patients overall delineated improvements in ODI scores, with scores also improving from 63.7%±7.3% preoperatively to 17.9%±6.2% postoperatively [14,17,20,34,36,45,47]. Finally, 2 studies demonstrated advantageous outcomes in postoperative MacNab scores [18,29]. All 18 studies demonstrated improvements in the patient-reported outcomes scales, suggesting that on average, endoscopic surgery for spondylodiscitis may help improve postoperative pain (Table 6).
Only 6 studies discussed postoperative radiographic observations, with the majority of these observations focusing on kyphosis-related findings [15,21,25,27,34,35]. Specifically, our review found that there was a large range of changes in postoperative kyphosis alignment, ranging from -13° to 19°. Chen et al. [15] evaluated the role of computed tomography (CT)-assisted endoscopic surgery in the treatment of infectious spondylodiscitis in immunocompromised patients. They presented an average preoperative kyphosis angle of 9.23±1.01 and a 3-month postoperative angle of 2.31±1.11, with only 7.7% of patients demonstrating a significant change between the 2 periods. Iwata et al. [27] treated fungal spinal infections with percutaneous posterolateral endoscopic surgery, where they reported progression of kyphosis by an average of 5° following a minimal 24-month postoperative period.
2. Treatment Success
A total of 35 papers comprising 505 patients reported treatment outcome data. A total of 444 patients (87.9%) had successful treatment of infection after undergoing endoscopic debridement. The medical treatment success was higher in patients that had concurrent instrumentation than endoscopic debridement alone (96.4% vs. 85.9%, p=0.003) (Table 7).
There were 54 cases where PEDD did not result in successful treatment of infection. Only 44 of the reported 54 treatment failures made note of any particular reason relevant to the failure of treatment. Over half of treatment failures had medical comorbidities such as human immunodeficiency virus, cancer, diabetes mellitus, and renal disease. Multilevel involvement (12 of 44), renal disease (7 of 44), preoperative bony destructions, and large abscess cavities were identified as predictors of failure of endoscopic debridement [41,47].
3. Changing Course of Disease
Two papers were identified that provided evidence to help address whether endoscopic treatment could alter the course of spondylodiscitis disease management. Yang et al. [49] performed a nonrandomized controlled trial of upfront PEDD versus CT-guided biopsy both followed by standard medical therapy. Fifty-two patients chose which group to join, of which 20 underwent PEDD and 32 underwent CT-guided biopsy. Patients in the PEDD group reported statistically significant findings of immediate back pain relief (90% vs. 0% CT-guided), higher positive culture rate (90% vs. 46.9%, p=0.002), and higher infection control (75% vs. 43.8%, p=0.027) compared to the CT-guided biopsy group [49]. In a direct comparison of percutaneous endoscopic interbody debridement and fusion (PEIDF) to open surgery, Wang et al. [44] showed earlier infection control with PEIDF in high-risk patients when medical therapy has failed. They reported significantly lower rates of sepsis (6.7% vs. 37.9%, p=0.030), and sooner return to normal CRP (26.93 days vs. 41.31 days, p=0.001) in PEIDF compared with open surgery [44].
DISCUSSION
Endoscopic surgery has emerged as a less invasive operation for spondylodiscitis, with previous studies reporting advantageous outcomes in culture rate, reoperation rates, and complications compared to other approaches [35,52]. However, there are only a limited systematic review that have defined the role of spinal endoscopy and assessed specific complications and outcomes that arise during endoscopic management of spondylodiscitis [53,54]. Abreu et al. [53] was the first review to emphasize the role of endoscopic surgery in cases refractory to medical management compared to open surgery. Giordan et al. [55] evaluated the effectiveness, safety, and outcomes of endoscopic surgery for thoracolumbar spondylodiscitis over the past 20 years, highlighting overall satisfactory clinical outcomes. Our systematic review sought to build on these previous reviews and provide a more extensive analysis by including more studies, offering detailed statistical assessments of specific complications and patient-reported outcomes, and exploring the efficacy of endoscopic techniques in high-risk patients. Furthermore, we sought to further identify the role, limitations, and potential for endoscopic techniques to alter the treatment course and medical outcome if possible.
Endoscopic debridement has a low complication profile with the most common side effects being transient paresthesia or radiculopathy. The proposed mechanism is irritation of the adjacent nerve by the endoscopic cannula. There were no instances reported of CSF leak or meningitis. The low morbidity and potential to be performed under local anesthesia presents endoscopic debridement as an option in high-risk patients not otherwise candidates for open surgery.
We found the culture yield with endoscopic biopsy (68.6%) to be within the range of reported culture rates from CT-guided biopsies [55,56]. Chen et al. [16] did a direct comparison of CT-guided biopsy to endoscopic biopsy and still only noted a 77% yield rate even when both techniques were used for each patient. Of note, the reported papers had a relatively high proportion of mycobacterium (11.2%) and fungal (4.8%) cases which may have suppressed the actual culture yield achievable in a Western patient setting [57,58].
Endoscopic debridement was consistently shown to improve patient-reported pain outcomes. This finding is consistent with literature on surgical management compared to medical management. In one retrospective cohort study, patients who received medical management and early surgical care had significantly lower mean ODI scores and back pain scores when compared to patients who were only treated medically [59].
We found a reoperation rate of 13.9% after endoscopic debridement. Of these patients, half of them were debridement and fusion surgeries performed in the context of ongoing spinal infection. These numbers echo the limitations of PEDD as a stand-alone treatment in high infection-burden cases with multilevel spinal involvement that can lead to instability. After eradication of the infection, only 2.9% of cases required a delayed fusion for developing kyphosis, persistent back pain, or concern for instability. Endoscopic techniques present an alternate strategy to reduce the overall surgical burden in the management of spondylodiscitis. A systematic review comprising 4,173 patients with spondylodiscitis found 33% of surgically treated patients had both an anterior and posterior approach, 26% of surgeries were staged, and the reoperation rate was 13% [60]. In an analysis of 9,983 patient with Staphylococcus aureus infections, Saad Berreta et al. [61] found 31% of spondylodiscitis patients ultimately required surgical management. Furthermore, 22% of S. aureus infections that were surgically managed required reoperation within 30 days of initial surgery, as well as a 37% 1-year reoperation rate [61].
We found several factors that identify the limitations of endoscopic debridement. Patients with immunosuppression or a heavy burden of infection were more likely to still require additional surgery despite endoscopic treatment. Specific factors identified include diabetes, renal disease, multilevel involvement of infection, large abscess cavity, and preoperative bony erosions [47]. All of these patients originally had advanced, multilevel infections that led to osteolytic destruction and increased spinal instability that required further debridement. Wang et al. [45] reported on the usage of a unilateral PEDD approach for lumbar spondylodiscitis, where they utilized an additional anterior debridement with instrumentation for 2 patients that had postoperative paresthesia and persistent severe back pain. Both of these patients had multilevel infections with paravertebral infections, suggesting that percutaneous endoscopic debridement may not be adequate for such cases [45].
We were not able to establish if the PEDD technique affects spinal alignment or leads to late-stage kyphosis. A limitation we found in the reporting of kyphotic measures is the inadequate detailing of external bracing. In some of the studies, patients had an increase in lordosis at follow-up. The PEDD technique does not create lordosis. It is suspected that preprocedural kyphosis was measured without a brace, whereas postoperative measurements were taken with an external brace applied which improved alignment. Additional studies with specific attention to spinal alignment after the PEDD technique with or without concurrent bracing therapy are needed.
In evaluating for a potentially new treatment algorithm for spondylodiscitis, there is a suggested early, adjunct role for endoscopic techniques. Pathogen identification for targeted antimicrobial therapy remains the mainstay of treatment. Biopsy can be performed by standard CT-guided means or with the use of the endoscope. While we did not find a superior biopsy culture yield with the endoscope, there are additional benefits including patient pain relief, sooner infection control, and lower likelihood for traditional open surgery thereby reducing the burden of surgical invasiveness and healthcare resource utilization. Larger studies are needed to further validate benefits such as lower sepsis and mortality after failure of medical treatment. When performed under local anesthesia, endoscopic techniques can be more liberally utilized in medically high-risk patients and even in high infection burden despite the limitations identified.
Our study has several limitations. First, there is limited literature on the assessment of endoscopic surgery for spondylodiscitis as the surgical techniques are not widely adopted. This results in our systematic review having lots of case reports and case series that provide low-quality data and a risk of bias. Additionally, the limited nature of studies specifically addressing pyogenic and tuberculous spondylodiscitis separately in the context of endoscopic treatment led us to combined both conditions in our analysis to provide a more comprehensive overview of endoscopic techniques, though future research with more granular data may enable distinct evaluations of these subgroups.
Secondly, most of the cases described in this review take place in Asia, making it difficult to generalize our findings in the greater context of spine surgery where patient demographics, comorbidities, and pathogens may differ. Additionally,
Finally, the lack of randomized controlled trials assessing this topic makes this systematic review and meta-analysis potentially subject to publication bias and increased statistical heterogeneity.
CONCLUSIONS
We compiled the available literature to on the safety, benefits, complication profile, and limitations of the PEDD technique. We presented a role for PEDD as an adjunct to current spondylodiscitis treatment algorithms and use in high-risk patients that will need to be further validated in clinical trials. The limited data suggesting PEDD can change the course of spondylodiscitis by reducing sepsis, mortality, and invasiveness of subsequent surgeries is encouraging and requires further investigation.
Notes
Conflicts of interest
The authors have nothing to disclose.
Funding/Support
This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Acknowledgments
This abstract was presented at the Spine Summit Conference in Las Vegas, NV (February 21-24, 2024).