Insights into posterior cervical biportal endoscopic spine surgery outcomes in radiculopathy: a systematic review and meta-analysis

Study design and inclusion criteria

This meta-analysis was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The study protocol was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO) before the systematic literature search. We focused on cohort studies reporting outcomes of PCBESS for radiculopathy. Eligible articles met all of the following criteria: (1) patients with radiologically diagnosed radiculopathy presenting with foraminal stenosis and/or disc herniation; (2) studies in which PCBESS was the primary intervention or was compared with other established approaches (e.g., traditional open surgery, posterior approaches, or conservative management); and (3) adult patients aged ≥18 years. Identified records were screened by titles and abstracts and subsequently assessed for eligibility. Exclusion criteria were studies of cervical spine conditions unrelated to radiculopathy, prior cervical spine surgery, individual case reports, reviews, conference abstracts, letters, and editorials. Studies reporting insufficient data for analysis were also excluded. A detailed list of inclusion and exclusion criteria and the full search strategy are provided in Tables 1 and 2.

Table 1

PICO framework and eligibility criteria

Table 2

Full search strategy

Literature search and selection

A systematic literature search was conducted across four electronic databases (ScienceDirect, PubMed, Google Scholar, and Cochrane) from database inception to August 31, 2024, with no restrictions on publication year. The search approach involved the use of Medical Subject Headings terms and relevant synonyms. Duplicate and ineligible records were removed using Mendeley Desktop ver. 1.19.6 (https://www.mendeley.com/). Titles and abstracts were subsequently screened against predefined criteria according to the PRISMA flowchart using Microsoft Excel 2011 (Microsoft Corp., Redmond, WA, USA). The reference lists of included studies were manually screened to identify additional studies. The literature search and study selection were conducted collaboratively by four investigators. Any disagreements were resolved through consensus, with a fifth independent reviewer adjudicating unresolved conflicts.

Quality assessment of included studies

The risk of bias for each study was assessed using the Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) tool (Cochrane, London, UK). This tool evaluates seven domains: confounding, participant selection, intervention classification, deviations from intended interventions, missing data, outcome measurement, and selective reporting. A study was deemed to have a low risk of bias if none of these domains raised concerns. Assessments were performed independently by at least two reviewers. Discrepancies were resolved by discussion; if consensus was not reached, a third reviewer provided final adjudication.

Data extraction

We included study location, sample size, mean age, number of male participants, body mass index, type of surgical interventions, pathology included (disc herniation, spinal stenosis, or both), affected disc levels, operation time, estimated blood loss, follow-up duration, and number and type of complications.

The research endpoints were predicated on perioperative, clinical, and spinal-specific domains. Perioperative outcomes included operative time, estimated intraoperative blood loss, length of hospital stay, reoperation rate, and surgical success rate. Clinical outcomes included complication rate, recurrence rate, Visual Analog Scale (VAS) scores, MacNab criteria, and functional measures such as the Neck Disability Index (NDI). Spinal-specific outcomes included pre- and postoperative levels of C-reactive protein (CRP) and creatine phosphokinase (CPK), segmental range of motion, and Cobb angle. To compare PCBESS and uniportal endoscopic techniques, we extracted data on surgical success rate, operative time, VAS scores, NDI, incidence of complications, and reintervention rate.

Data synthesis and analysis

Continuous variables were pooled as mean differences (MD) with corresponding 95% confidence intervals (CI); categorical variables were pooled as odds ratios (OR) or proportions with 95% CI. A random-effects model with restricted maximum likelihood estimation was used to account for between-study variability. Heterogeneity was quantified using the I² statistic and classified as low (<25%), moderate (25%–50%), or high (>50%). Forest plots were generated for each outcome to display pooled estimates and corresponding heterogeneity measures. For outcomes with high heterogeneity, subgroup analyses were performed (for example, by level of disc involvement, type of spinal disorder, or extent of decompression) to explore potential sources of heterogeneity. All statistical tests were two-tailed, and significance was set at p<0.05.

Certainty of evidence assessment

Certainty of evidence was evaluated using the GRADE framework (Grading of Recommendations, Assessment, Development, and Evaluations) across five domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias, classifying overall certainty as high, moderate, low, or very low [15]. Unlike the conventional approach that assigns observational studies a low initial rating, we applied the ROBINS-I–adapted approach wherein observational evidence was considered from a high initial certainty level when randomized trials were not feasible, with downgrading applied as indicated by the GRADE domain [16].

Study selection and identification

The initial database search yielded 2,498 records. Prior to the screening phase, 2,333 records were excluded. Title and abstract screening led to the exclusion of an additional 112 articles. A total of 53 full-text articles were retrieved for detailed evaluation. Following a full-text eligibility assessment, 14 studies met all inclusion criteria and were included in the final quantitative and qualitative synthesis. The study selection process is illustrated in the PRISMA flow diagram (Fig. 1).

Fig. 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart.

Baseline characteristics and quality assessment of included studies

A total of 14 retrospective studies [10,13,14,17–27,] encompassing 712 patients with cervical radiculopathy were included in the meta-analysis. The underlying etiologies were predominantly foraminal stenosis and/or cervical disc herniation, treated with PCBESS (Table 3). Among the pooled cohort, 62.5 % of participants were male, and the mean age was 54.78±1.7 years. All included studies were conducted in Australia, China, or South Korea. Three studies evaluated multilevel disc involvement, whereas the remaining studies focused on single-level disease. Most studies described a combined foraminotomy and discectomy technique as the principal surgical procedure. The mean follow-up duration across studies was 15.6 months (range, 6–26.5 months).

Table 3

Characteristics of included studies

All included studies were non-randomized and assessed for methodological quality using the ROBINS-I tool (Fig. 2). Moderate risk of bias was frequently identified in the domains of confounding, missing data, outcome measurement, and selective reporting. One study exhibited a high risk of bias in the participant selection domain, leading to an overall serious risk classification for that study.

Fig. 2

ROBINS-I assessment of included studies.

Postoperative complications

Postoperative complications are summarized in Table 4. Seventeen patients experienced neurological complications, with temporary paralysis occurring in 47.1% of these cases. This notable proportion highlights the importance of meticulous surgical technique to minimize nerve irritation during decompression. Incomplete decompression was the next most frequent neurological event, reported in 35.3% of cases. Nineteen cases of mechanical, soft tissue, and miscellaneous complications were also identified. Among these, dural tears and epidural hematomas were the most common, occurring in 47.4% and 42.1% of cases, respectively. The overall pooled complication rate was 7% (95% CI, 4%–14%), indicating a relatively low incidence of adverse events (Fig. 3A). The certainty of evidence for this outcome was judged to be very low, primarily due to inclusion of studies with a high risk of bias, substantial heterogeneity, and wide CIs across the GRADE domains (Table 5).

Table 4

Reported postoperative complications (n=36)

Fig. 3

Forest plot of clinical outcomes. (A) Complications, (B) reoccurrance rate, (C) Visual Analog Scale (VAS) score, (D) Neck Disability Index (NDI) score, and (E) MacNab criteria. CI, confidence interval; REML, restricted maximum likelihood; C-BESS, cervical biportal endoscopic spine surgery; SD, standard deviation; MD, mean difference.

Table 5

Certainty of evidence across outcomes based on grading of recommendations assessment, development, and evaluation (GRADE)

Clinical outcomes

Clinical outcomes following PCBESS were favorable (Fig. 3). The pooled recurrence rate was low (effect size, 0.05; 95% CI, 0.02–0.09). Pain intensity, measured using the VAS, showed a significant reduction (MD, −4.77; 95% CI, −6.07 to −3.47) (Fig. 3C), and functional improvement was reflected by a marked decrease in NDI score (MD, −27.07; 95% CI, −32.83 to −21.30) (Fig. 3D). According to MacNab criteria (Fig. 3E), 53% of patients (95% CI, 34%–83%) achieved “excellent” and 31% (95% CI, 15%–64%) achieved “good” outcomes. The heterogeneity of these measures ranged from moderate to high. The certainty of evidence varied across outcomes (Table 5). Both VAS and NDI improvements were initially graded as very low due to the serious risk of bias, high heterogeneity, and wide CIs, but were upgraded to low certainty because of large effect sizes. The recurrence rate was supported by high-certainty evidence, as no serious limitations were identified. Certainty for the MacNab “fair” category was moderate owing to the wide CI.

Perioperative results

Perioperative outcomes of PCBESS, including operative efficiency and recovery parameters, are presented in Fig. 4. The mean operative time was 84.89 minutes (95% CI, 57.08–112.70), and the mean intraoperative blood loss was 64.07 mL (95% CI, 47.69–80.45), consistent with minimally invasive surgical characteristics. The average hospital stay was 4.45 days (95% CI, 3.12–5.78). The pooled reoperation rate was low (3%; 95% CI, 1%–6%), while the surgical success rate was high (91%; 95% CI, 85%–98%; p=0.00). Considerable heterogeneity was observed across most perioperative parameters. The overall certainty of evidence for perioperative outcomes was predominantly low (Table 5). Operative time, estimated blood loss, and hospitalization duration were rated as very low certainty due to serious risk of bias, high heterogeneity, and wide CIs. The reoperation rate achieved moderate certainty after upgrading for a large effect size, while surgical success was upgraded from very low to low certainty for the same reason.

Fig. 4

Forest plot of perioperative outcomes. (A) Operative time, (B) estimated blood loss, (C) hospitalization time, (D) reoperation rate, and (E) success rate. CI, confidence interval; REML, restricted maximum likelihood.

Spinal-specific outcomes

Spinal outcomes following PCBESS demonstrated favorable changes in inflammatory markers and spinal alignment parameters. Postoperative levels of CPK and CRP showed mild postoperative increases, consistent with a controlled inflammatory response. The pooled MD in CRP was 0.28 (95% CI, 0.21–0.34; p=0.00) (Fig. 5A), and the pooled MD in CPK was 47.81 (95% CI, 6.13–89.09; p=0.02) (Fig. 5B). Segmental range of motion improved significantly after surgery (MD, 0.55; 95% CI, 0.27–0.84; p=0.00), indicating restoration of mobility at the operated level. However, changes in Cobb angle were not statistically significant (MD, 4.25; 95% CI, −2.69 to 11.18; p=0.23) (Fig. 5C). Heterogeneity among these parameters was minimal, suggesting consistent effects across studies. Based on GRADE assessment, the certainty of evidence for laboratory and radiological outcomes varied across measures (Table 5). CRP level and segmental range of motion were both rated as high-certainty evidence due to the absence of serious concerns in any domain. In contrast, CPK and Cobb angle outcomes were rated as low-certainty evidence because of inconsistency and imprecision.

Fig. 5

Forest plot of spinal variable outcomes. (A) C-reactive protein level, (B) creatine phosphokinase level, (C) segmental range of motion, and (D) Cobb’s angle. C-BESS, cervical biportal endoscopic spine surgery; SD, standard deviation; MD, mean difference; CI, confidence interval; REML, restricted maximum likelihood.

Comparison with uniportal endoscopy

PCBESS demonstrated outcomes comparable to those of uniportal endoscopic spine surgery across most evaluated parameters (Fig. 6). The ORs for complications (OR, 0.83; 95% CI, 0.32–2.14; p=0.7), nerve root palsy (OR, 0.40; 95% CI, 0.08–2.13; p=0.28), dural tear (OR, 1.63; 95% CI, 0.33–8.06; p=0.55), and reoperation (OR, 1.11; 95% CI, 0.15–8.14; p=0.92) revealed no significant differences between the two techniques, indicating similar safety profiles. Likewise, clinical outcomes showed no statistically significant differences between the two methods. MDs in VAS scores (MD, 0.18; 95% CI, −0.21 to 0.56; p=0.38) and NDI (MD, 0.10; 95% CI, −0.71 to 0.91; p=0.81) were not significant. Operative time was also comparable (MD, 2.41; 95% CI, −10.72 to 15.53; p>0.05). Although not statistically significant, PCBESS was associated with a lower incidence of nerve-related complications such as transient palsy (OR, 0.40; p=0.28) and a marginally higher, albeit non-significant, surgical success rate (OR, 1.45; 95% CI, 0.60–3.51; p=0.41). Heterogeneity across these parameters was predominantly low, suggesting consistent results among included studies. According to the GRADE assessment, most outcomes, including complication rate, nerve root palsy, dural tear, reoperation, success rate, and NDI score, were supported by moderate-certainty evidence (Table 5), reflecting wide CIs despite otherwise robust methodological consistency. Operative time and VAS score were rated as low-certainty evidence due to considerable heterogeneity and substantially wide CIs.

Fig. 6

Forest plot of posterior cervical biportal endoscopic spine surgery (C-BESS) vs. uniportal endoscopic. (A) Complications, (B) nerve root palsy, (C) dural tears, (D) reoperation, (E) operative time, (F) success rate, (G) Neck Disability Index (NDI) score, and (H) Visual Analog Scale (VAS) score. OR, odds ratio; CI, confidence interval; REML, restricted maximum likelihood; SD, standard deviation; MD, mean difference.

Subgroup analysis

As shown in Table 6, subgroup analysis revealed that postoperative complications following PCBESS were more frequent in multi-level procedures compared to single-level procedures (18% vs. 4%, p=0.01). Regarding the extent of decompression, combined foraminotomy and discectomy achieved a significantly greater reduction in VAS pain scores compared with foraminotomy alone (p<0.001). As summarized in Table 7, multi-level PCBESS was associated with higher intraoperative blood loss compared with single-level procedures (93.60 mL vs. 58.61 mL, p<0.001). Hospitalization duration was also longer among patients with spinal stenosis than those with disc herniation (7.90 days vs. 3.19 days, p<0.001). Performing discectomy in all patients and applying inclinatory foraminotomy, compared with foraminotomy with or without discectomy, resulted in prolonged operative time (p=0.03). Postoperative biomarkers and spinal mobility parameters were not significantly affected by the number of disc levels treated (CPK: p=0.11; CRP: p=0.65; segmental range of motion: p=0.56) (Table 8). Collectively, these findings underscore the influence of surgical strategy on postoperative outcomes, reinforcing the importance of individualized, anatomy-based planning to optimize efficacy, safety, and patient-centered outcomes in the management of cervical radiculopathy.

Table 6

Subgroup analyses results of clinical outcomes (single-arm intervention)

Table 7

Subgroup analyses results of perioperative outcomes (single-arm intervention)

Table 8

Subgroup analyses results of spinal-specified outcomes (single-arm intervention)

Incidence of complications

The present subgroup analysis did not identify a significant association between the extent of disc involvement and complication rates, suggesting that disc pathology alone may not directly influence surgical risk. Instead, overall procedural complexity, particularly in multi-level decompressions, appears to be the predominant determinant. Nevertheless, previous studies have reported that complex or extensive disc herniations can increase intraoperative difficulty and complication risk [19,34,35]. Patients with such lesions may therefore benefit from meticulous preoperative evaluation and individualized surgical planning [36]. High-resolution imaging modalities, such as magnetic resonance imaging, are essential for evaluating the severity and configuration of disc pathology and for guiding operative strategy [37]. A tailored approach enables surgeons to anticipate potential complications, refine their technique, and minimize the risk of adverse events, thus improving patient safety and outcomes.

The markedly higher complication rate observed in multi–level PCBESS procedures (18% vs. 4%) likely reflects the increased technical complexity of decompressing multiple spinal segments. Prior literature has similarly demonstrated that multi–level lumbar and cervical fusion surgeries are associated with longer operative times and greater blood loss, variables that correlate with higher complication rates [38–40]. Similarly, a comparative analysis of single–versus multi–level lateral lumbar interbody fusion reported significantly prolonged operative time (145.1 minutes vs. 98.9 minutes) and greater intraoperative blood loss in the multi–level cohort [39]. In the context of endoscopic multilevel decompression, a higher incidence of dural tears, epidural hematomas, and incomplete decompression has been documented, indicating that expanded surgical exposure and manipulation may predispose to neural irritation or suboptimal decompression [33,38,41,42]. These findings underline the importance of meticulous surgical planning and, when appropriate, the consideration of staged or alternative approaches to mitigate risk in complex multi-level cases.

Alleviation of pain and restoration of function

The present findings demonstrate that the surgical approach substantially influences pain relief, as measured by the VAS, and functional recovery, as reflected by the NDI score (both p=0.00). These findings underscore the significance of surgical precision and proficiency in achieving optimal results. In minimally invasive surgery, where the goal is to minimize tissue damage while maintaining procedural effectiveness, the surgeon’s proficiency in manipulating endoscopic instruments and performing accurate decompressions is paramount [38,43]. This highlights the need for extensive training programs that foster endoscopic competence and familiarity with sophisticated surgical instruments.

Variations in postoperative discomfort and functional improvement across surgical approaches may stem from differences in the degree of decompression achieved, the management of neural structures, or the ability to address complex anatomical constraints. Surgeons should continually refine their techniques and integrate feedback from clinical outcomes to optimize performance and tailor interventions to each patient’s unique anatomy and pathology [38].

PCBESS demonstrated substantial improvements in pain and function, corroborating its therapeutic efficacy for radiculopathy. When compared with uniportal endoscopic techniques, PCBESS yielded comparable outcomes, with no significant differences in mean VAS or NDI scores. These findings indicate that both methods are effective for symptom relief. However, the biportal configuration of PCBESS offers improved visualization and a wider operative field, potentially facilitating more thorough decompression in complex anatomical regions. Although clinical outcome measures appear comparable, the biportal approach may confer technical advantages in terms of surgeon ergonomics and intraoperative flexibility.

Perioperative results

The influence of disease type on intraoperative blood loss highlights the interplay between patient-specific variables and surgical performance in determining perioperative outcomes. Conditions with greater soft-tissue or vascular involvement may predispose patients to increased blood loss, necessitating meticulous hemostatic control and surgical precision. Minimizing blood loss is a key advantage of minimally invasive spine surgeries such as PCBESS, as it contributes to reduced morbidity, faster recovery, and lower rates of postoperative complications, a finding consistently supported by prior literature on minimally invasive spine surgery [44]. Hospitalization duration was significantly longer in patients with spinal stenosis than in those with disc herniation, underscoring the effect of underlying pathology on recovery, whereas operative duration did not differ significantly across subgroups. The consistency of these parameters across varying patient profiles and surgical techniques underscores the procedural reliability of PCBESS and its adaptability across diverse clinical contexts.

In terms of perioperative parameters, PCBESS demonstrated mean operative time, blood loss, and hospital stay durations comparable to those reported for other minimally invasive spine procedures. When compared with the uniportal endoscopic approach, operative time and hospital stay did not significantly differ. However, the modular dual-port configuration of PCBESS may facilitate more efficient handling in multi-level or anatomically constrained cases, reducing surgeon fatigue and potentially shortening the learning curve. Although both techniques achieved favorable functional outcomes, PCBESS showed slightly superior segmental range of motion, likely reflecting its capacity for more complete decompression. These subtle advantages may justify PCBESS as a first-line option, particularly in technically demanding cases.

Spine-specific outcomes

Analysis of spinal-specific outcomes revealed that the degree of disc herniation did not significantly influence biochemical markers, including CPK and CRP levels, or segmental range of motion. These findings suggest that, although anatomical parameters are important for surgical planning, postoperative biochemical responses and spinal mobility are likely influenced by additional factors such as surgical technique, individual patient physiology, and rehabilitation protocols [31,45]. This emphasizes the multifactorial nature of spine surgery outcomes and highlights the importance of a holistic, multidisciplinary approach to patient management that accounts for both surgical and non-surgical elements.

Patient satisfaction and surgical revision frequencies

Patient satisfaction, as assessed by the MacNab criteria, was not significantly influenced by the extent of disc involvement. This finding suggests that patient-reported outcomes may depend more on subjective experiences, expectations, and the interaction between the patient and the healthcare team than on anatomical parameters alone [46–48]. Effective preoperative counseling and the establishment of realistic expectations are, therefore, essential for ensuring patient satisfaction and perceived surgical success [46].

Reoperation rates demonstrated borderline significance in relation to the level of disc herniation (p=0.08), indicating that certain disc pathologies may predispose patients to additional surgical interventions. This highlights the importance of thorough preoperative assessments and meticulous surgical planning to anticipate and mitigate potential complications. Ongoing postoperative monitoring and long-term follow-up are likewise crucial for the early detection and management of recurrence or residual symptoms, enhancing long-term outcomes and patient satisfaction [49,50].

Despite these insights, this study has several limitations. The retrospective nature of the included studies introduces the possibility of selection and reporting biases. Substantial heterogeneity among studies likely reflects variations in patient demographics, surgical approaches, and outcome assessment tools, limiting the generalizability of the findings. Moreover, the lack of large, randomized controlled trials restricts the ability to draw definitive conclusions. Future research should focus on prospective, multicenter designs with standardized outcome measures to enhance comparability and strengthen the evidence base. Overall, this study reinforces the efficacy and safety of PCBESS for cervical radiculopathy, while emphasizing the influence of patient-specific factors and surgical methodology on outcomes.