Lumbosacral radicular syndrome, or sciatica, is responsible for sizeable personal or societal costs and is often caused by lumbar disc herniation (LDH), of which natural process is usually favorable.1 Surgery is recommended when patients with intractable pain are refractory to conservative treatment or experience progressive neurological deficits.2 The goal of lumbar discectomy is to remove pathological disc material to decompress the nerve root and relieve symptoms.3
In 1934, the first successful lumbar disc operation was described by Mixter and Barr.4 Since then, a variety of innovative and minimally microinvasive techniques have been developed. With the application of the surgical microscope and further enhancement of visualization, conventional microdiscectomy was introduced by Caspar5 and Yasargil,6 respectively. In 1997, Foley and Smith7,8 introduced minimally invasive transmuscular tubular discectomy, and performed minimally invasive discectomy using a combination of tubular retractors and trocar systems with endoscopy. In 2002, Greiner-Perth et al9 reported that combining tubular retractors and trocar systems with surgical microscopy for LDH overcomes disadvantages of the two-dimensionality of the endoscopic image.
The rationale behind trocar systems, replacing conventional subperiosteal muscle dissection by the muscle-splitting transmuscular approach of tubular microdiscectomy, is less tissue trauma caused by surgery and a subsequent faster rate of recovery.7 Although, in theory, TMD should benefit the patient, CMD currently remains the standard procedure for LDH.10,11 In 2008, the first RCT study12 comparing TMD with CMD was reported, but there were no significant differences except reduced postoperative consumption of analgesics in the TMD group. Subsequently, some RCT studies13–15 were successively introduced, though evidences supporting that TMD was superior in efficacy compared with CMD remained insufficient. As such, this study aimed to systematically review randomized controlled clinical studies relevant to determining the clinical effectiveness of TMD versus CMD for LDH.
MATERIALS AND METHODS
We comprehensively searched databases PubMed, EMBASE, and Cochrane Central Register of Controlled Trails (CENTRAL) for prospective RCTs, through using MeSH terms “microdiscectomy,” “tubular microdiscectomy,” “minimally invasive surgery,” and “spinal disease.” The retrieved results were last updated on March 15, 2018. References cited in the relevant literatures were also reviewed. To make an exhaustive search of all relevant literatures, two independent investigators (Y.Q.W. and Z.Y.L.) selected qualified studies based on search criteria and relevant content. When consensus could not be reached, a third reviewer (C.M.C.) was consulted to resolve the disagreement.
Study Selection Criteria
Inclusion criteria for original papers were: (I) age 18 to 80 years; (II) Magnetic Resonance Imaging confirmed (MRI-confirmed) disc herniation; (III) intolerable sciatica or rapidly progressive neurological deficits; (IV) single level virgin lumbar disc herniation and failure of 6 to 12 weeks of conservative treatment; (V) RCT of comparing TMD with CMD.
Exclusion criteria were: (I) conservatively treated lumbar disc herniation at adjacent levels; (II) spondylolytic or degenerative spondylolisthesis; (III) signs of spinal instability or other spinal abnormalities such as bone disease, spinal infection, malignancy; (IV) central spinal canal stenosis; (V) pregnancy; (VI) excessive obesity; (VII) incomplete follow-up data (lost >15%).
Two investigators (Y.Q.W. and Z.Y.L.) independently extracted data from predesigned forms. Any divarication was either resolved by discussion or by involving a third reviewer (C.M.C.) when necessary until a consensus for all items was achieved. For each study, the requisite data collected included study design, age, sample size, intervention details, and outcome variables with results. Primary outcomes included visual analogue scale (VAS), Oswestry disability index (ODI), and short form-36 (SF-36) scores and related complications (cerebrospinal fluid leakage, dural tear, reoperation, neurological deficits, etc.). Secondary outcomes included operation time, intraoperative blood loss, and duration of hospital stay. The VAS score16 was as follows: pain was assessed on a horizontal 100 mm scale varying from 0 mm (“no pain”) to 100 mm (the worst pain imaginable). The ODI scores17 use the ODI German version 2.0 standard for assessing physical and mental health. The SF-36 score18 ranges from 0 to 100, with higher scores indicating less severe symptoms.
For statistical analysis, odds ratio (OR) for dichotomous outcomes and standardized mean difference (SMD) were utilized for continuous outcomes with 95% confidence intervals (CI) for each outcome. The I2 test was used to assess statistical heterogeneity. This test demonstrated significant heterogeneity when I2 values exceeded 50%, and in these instances, the random effects model was used. Otherwise, a fixed effects model was used. The results are graphically summarized using forest plots. We assessed the robustness of our results by comparing different effect models for sensitivity analysis. The meta-analysis was performed by Review Manager (RevMan, Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). A P-value of 0.05 was set for determining statistical significance.
Two investigators (Y.Q.W. and Z.Y.L.) independently evaluated risk of bias using the 12 criteria recommended by the Cochrane Back Review Group.19 The items were scored as “low risk,” “high risk,” or “unclear.” If at least six of the criteria passed without serious potential flaws, studies were considered to have a “low risk of bias” overall. If not, the studies were defined as having “high risk of bias.” In addition, the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach20 was used to rate the quality of evidence for all pooled outcomes. According to the assessment of study design, risk of bias, consistency, directness and precision, the quality of outcomes was categorized as very low, low, moderate, or high.
Search Results and Study Characteristics
A total of 605 related studies from PubMed (n = 215), EMBASE (n = 210), and CENTRAL (n = 180) databases were initially reviewed. Of these, a total of four RCT researches12–15 with 610 individuals were included in the meta-analysis. The process of identifying related researches is presented in Figure 1. All patients (average age = 45 yr) had LDH confirmed by MRI with associated sciatica and were refractory to conservative treatment. Follow-up of all included studies occurred after at least 12 months. TMD was performed by the limited tubular retractor and CMD was performed by the Casper-type retractor. The extracted data of the characteristics of four included papers were recorded (Table 1).
Risk of Bias in Included Studies
We assessed all original studies as having an overall high risk of bias. The migration risk of each study is described in the migration risk table. The ratings across all included studies were summarized in Figure 2.
- (I) Allocation: Three studies were assessed as having low risk for random sequence generation.13–15 The study by Brock et al was assessed as having high risk of selection bias, due to inadequate generation of randomized sequence (assigned by admission date). Only one study performed allocation concealment and was assessed as having low risk of bias for selection bias.15 However, allocation concealment was often uncertain in assessment of other studies.12–14
- (II) Blinding: Two studies were blinded to both participants and outcome assessors.13,15 Two studies did not clearly report the use of types of blinding in assessment.12,14
- (III) Incomplete outcome data: All studies were assessed as having low risk of attrition bias.
- (IV) Selective reporting: All studies were assessed as having low risk of selective reporting bias except one study that reported incomplete SF-36 results, and this study was assessed as having a high risk of selective reporting bias.15
- (V) Other potential sources of bias: All studies were assessed as having a low risk of other potential sources of biases.
All included studies used VAS to measure pain intensity, but one set of VAS data in which data processing was different from other studies, had not been pooled. The results of pooled analyses indicated no significant difference of sumVAS scores 1 year postoperatively between TMD and CMD groups (SMD, 0.14, 95% CI, –0.40 to 0.12, P = 0.30, Figure 3). The research of Brock et al showed that 97% of patients showed “success” (VAS <4) in the TMD group, compared with 91.5% in the CMD group (P = 0.47).
Function and health were measured by an ODI questionnaire. Two studies were pooled. Meta-analysis discovered a significant difference in ODI scores about 1 year postoperatively between the two groups (SMD, –3.43, 95% CI, –4.64 to –2.21, P < 0.00001, Figure 4). The lower the ODI value, the better the postoperative recovery. Therefore, these findings suggested that TMD was more effective.
The studies of Arts et al and Ryang et al involved SF-36, but Arts et al only reported the results of bodily pain (BP) and physical function (PF). Therefore, only BP and PF values were analyzed. With respect to PF results at 12 months postsurgery, there were statistically differences between the two groups (SMD, –4.83, 95% CI, –8.94 to –0.72, P = 0.02, Figure 5). This indicated that patients who underwent TMD fared worse in PF. Meanwhile, there was no significant difference between the groups in BP outcomes (SMD, –3.42, 95% CI, –8.40 to 1.54, P = 0.18, Figure 6).
Three studies reported intraoperative and postoperative complications, involving 485 patients. We extracted related data, including dural tear, reoperation, cerebrospinal fluid leakage, and nerve deficits. In the meta-analysis, no significant differences in dural tear (OR, 1.28, 95% CI, 0.60–2.72, P = 0.52, Figure 7) or reoperation (OR, 1.01, 95% CI, 0.53–1.90, P = 0.98, Figure 8) were observed when comparing TMD with CMD. The unincorporated data of postoperative complications include: cerebrospinal fluid leakage, nerve deficits, etc. In the study of Arts et al, there was one case of cerebrospinal fluid leakage in the TMD group and two cases in CMD group and three cases of nerve deficits in the two groups. Brock et al reported five cases of postoperative recovery failure associated with sciatica in the TMD group and eight cases in the CMD group.
Concerning operation time, three studies with a total of 485 individuals were analyzed. A similar operation time was observed in the TMD group compared with the CMD group (SMD, –4.01, 95% CI, –23.06 to 15.04, P = 0.68, Figure 9). In addition, relative to the CMD group, Franke et al observed that operation time was significantly shortened in the TMD group at their index centre (CMD group [57.8 ± 20.2 min], TMD [33.3 ± 12.1 min]).
Intraoperative Blood Loss
Two studies calculated intraoperative blood loss. Arts et al reported 90% of patients in the TMD group with less than 50 mL intraoperative blood loss, and 85% in the CMD group (P < 0.08). In addition, Ryang et al also covered intraoperative blood loss in the TMD group (SMD, 26.2 ± 29.7 mL, range 0–100 mL) and the CMD group (SMD, 63.8 ± –86.8 mL, range 0–300 mL), but the difference between the two groups was not statistically significant (P < 0.31).
Duration of Hospital Stay
The duration of hospital stay was assessed in three studies, but these data were not pooled due to discrepancy in the definition and form of documentation among included studies. Arts et al calculated that hospital stay duration (including preoperative and postoperative days) was 3.3 ± 1.2 days (mean ± standard deviation, TMD group) and 3.3 ± 1.1 days (mean ± standard deviation, CMD group) (P < 0.82). With respect to the study by Franke et al, the average hospitalization time for patients at the Index centre was 4.9 days in the TMD group and 3.8 days in the CMD group (P-value not available). The mean hospital stay as presented by Ryang et al was not significantly different between the TMD (4 ± 2.3 d) and CMD groups (4.4 ± 2.8 d, P = 0.51). Therefore, there were no significant differences in the duration of hospital stay between TMD and CMD groups according to these studies.
We performed sensitivity analyses by comparing the meta-analysis results of different effect models. With regard to the pooled operation time results, there were significant differences in the TMD compared with the CMD group by using the fixed-effect model for analysis (SMD, 3.75, 95% CI, 0.26–7.24), but no significant difference was observed by using the random-effect model (SMD, –4.01, 95% CI, –23.06 to 15.04). With respect to other pooled outcomes, there were similar results by using both effect models, and we identified there was favorable robustness of results of the meta-analysis.
According to the previous reports, Rasouli et al,21 Wang et al,22 and Kamper et al23 retrospectively conducted meta-analysis comparing the clinical efficacy of minimally invasive discectomy (MID) and CMD. Rasouli et al21 identified that MID was inferior to CMD in terms of improving sciatica and reducing rehospitalization rate, while Wang et al22 and Kamper et al23 reported no significant differences in the efficacy of MID and CMD for LDH. However, MID may be performed with the assistance of total spine endoscopy, foraminal mirror technique, discoscope, and microscopy. Simultaneously, there were different operative indications, operating habits, and the extent of decompression among different minimally invasive discectomy techniques.24 For example, there were some advantages to endoscopy which included less trauma and shorter operative time, but disadvantages included one-handed performance, two-dimensional visual field, indication limitation and inadequate decompression.
Although a variety of minimally invasive techniques have been developed, CMD remains the standard procedure LDH patients.10,11 Recently Li et al25 published a meta-analysis comparing the clinical efficacy of TMD and CMD, and found no significant difference in the clinical efficacy of TMD and CMD for LDH in outcomes with respect to blood loss, operation time, hospitalization duration, complication rate and functional scores. In that study, however, the extraction and analysis of data was imprecise, and there was high risk of bias due to including and pooling non-randomized controlled trials. The quality of evidence was not assessed for the included study and pooled results, and sensitivity analysis was not performed. Therefore, evidence for comparing the efficacy of TMD with CMD was not adequate.
The present study is a further systematic review, and primarily found that there was similar short-term effectiveness of TMD compared with CMD for treating LDH. The strengths of our review included a sensitive search strategy, only RCTs, and the use of systematic review methodology as endorsed by the Cochrane Collaboration. This includes independent screening of identified studies, performing sensitivity analyses for pooled outcomes, assessing risk of bias in included studies and assessing the quality of the evidence according to the GRADE approach (Tables 2 and 3).
The ODI results 1 year following operation indicated a 3.43% lower score in the TMD group compared with the CMD group. Although CMD is still the most common procedure for LDH, the pooled ODI result in this study is more favorable to TMD. It is important to note the level of evidence quality was downgraded from moderate to low due to the sample size not meeting the OIS standard.26 The state of health and quality of life was evaluated by using SF-36. The analysis result of PF, one of the SF-36 items, was favorable to CMD group at postooperative 12 months. However, the level of evidence quality was low. In addition, the mean PF result in the study of Arts et al also supported the benefit in the CMD group, while there was no statistically significant difference between the two groups in the study by Ryang et al. With regard to intraoperative blood loss, which was assessed to identify surgical safety, Arts et al12 reported the percentage of participants with intraoperative blood loss less than 50 mL in the two groups and Ryang et al compared the mean intraoperative blood loss in the groups. According to the results, there were no statistically significant differences between the TMD and CMD groups.
Sciatica was evaluated by using VAS in the present study. According to the pooled results, no statistically significant difference was observed between the TMD and CMD groups regarding postoperative improvement in VAS score. In addition, Franke et al mentioned that at 6 months after surgery, the TMD group exhibited reduced pain compared with the CMD group, though no significant difference was found in other included studies. The outcomes of operation time, related complications, and duration of hospital stays were also calculated to assess the safety and efficacy of the two operations. The mean operation time of the TMD group was 4.01 minutes less than for the CMD group. The differences were small and did not meet standard thresholds for clinically meaningful differences. The heterogeneity in this outcome is considerable, and may be explained by differences in the experience of the surgeons. There were no meaningful differences between the two groups in the outcomes of dural tear, reoperation, nerve deficits, and duration of hospital stays.
There are still several limitations in this study. First, our study had a limited sample size. Although all included studies were RCT studies, the scale of the studies was small a total sample number was low. For lack of enough available data, a separate analysis hadn’t been conducted on change in VAS, ODI, and SF-36 scores from baseline at 1 year. In addition, each study had differences in the sample populations. SF-36 and ODI were merely mentioned by two studies, highlighting assessment variation. A high bias risk in original research was also prevalent. For example, there was a high risk of bias with regards to the generation of allocation sequences in the study by Brock et al. The studies also demonstrated short follow-up time periods postsurgery. In our review, the follow-up time was limited to 1 year, without long-term evaluation of clinical effectiveness. Therefore, more precise and prospective RCT, with sufficient samples and long-term follow-up, comparing TMD to CMD, are yet to be performed. Such RCTs should pay attention to more rigorous methodology and some outcome indicators, including more accurate estimation of intraoperative blood loss, damage of soft tissue, consumption of analgesic and surgical costs.
In our review, we found that TMD and CMD had similar clinical effectiveness in the treatment of LDH as assessed at 12 months postsurgery. TMD may be superior in terms of postoperative ODI scores, but inferior in terms of postoperative SF-36 scores. There were no significant differences in VAS scores, intraoperative blood loss, dural tear, reoperation, operation time, nerve deficits, and length of hospital stay. Given these potential advantages, additional research is needed to confirm TMD as a beneficial alternative to CMD.
- Clinical effectiveness was similar between TMD and CMD for treating LDH.
- TMD may be superior in terms of postoperative ODI scores and intraoperative blood loss, but inferior concerning postoperative SF-36 scores.
- No significant differences were observed between TMD and CMD regarding VAS scores, dural tear, reoperation, operation time, nerve deficits, and length of hospital stay.
1. Konstantinou K, Dunn K. Sciatica: review of epidemiological studies and prevalence estimates. Spine (Phila Pa 1976)
2. Jacobs W, van Tulder M, Arts M, et al. Surgery versus conservative management of sciatica due to a lumbar herniated disc: a systematic review. Eur Spine J
3. Angevine P, McCormick P. Outcomes research and lumbar discectomy. Neurosurg Focus
4. Mixter W, Barr JS. Rupture of the intervertebral disc with involvement of the spinal canal. New Eng J Med
5. Caspar W. A new surgical procedure for lumbar disk herniation causing less tissue damage through a microsurgical approach. Adv Neurosurg
6. Yasargil MG. Microsurgical operation of herniated lumbar disc. Adv Neurosurg
7. Foley KTSM. Microendoscopic discectomy. Tech Neurosurg
8. Smith MM, Foley KT. Microendoscopic discectomy (MED): the first 100 cases. Neurosurgery
9. Greiner-Perth R, Böhm H, El Saghir H. Microscopically assisted percutaneous nucleotomy, an alternative minimally invasive procedure for the operative treatment of lumbar disc herniation
: preliminary results. Neurosurg Rev
10. Gibson JN, Waddell G. Surgical interventions for lumbar disc prolapse: updated Cochrane Review. Spine (Phila Pa 1976)
11. Gibson JN, Waddell G. Surgical interventions for lumbar disc prolapse. Cochrane Database Syst Rev
12. Brock M, Kunkel P, Papavero L. Lumbar microdiscectomy: subperiosteal versus transmuscular approach and influence on the early postoperative analgesic consumption. Eur Spine J
13. Ryang YM, Oertel MF, Mayfrank L, et al. Standard open microdiscectomy versus minimal access trocar microdiscectomy: results of a prospective randomized study. Neurosurgery
2008; 62:174–181. discussion 81-2.
14. Franke J, Greiner-Perth R, Boehm H, et al. Comparison of a minimally invasive procedure versus standard microscopic discotomy: a prospective randomised controlled clinical trial. Eur Spine J
15. Arts MP, Brand R, van den Akker ME, et al. Tubular diskectomy vs conventional microdiskectomy for sciatica: a randomized controlled trial. JAMA
16. Price D, McGrath P, Rafii A, et al. The validation of visual analogue scales as ratio scale measures for chronic and experimental pain. Pain
17. Fairbank JC, Pynsent PB. The Oswestry Disability Index
. Spine (Phila Pa 1976)
18. Ware J, Sherbourne C. The MOS 36-item short-form health survey (SF-36). I: Conceptual framework and item selection. Med Care
19. Furlan AD, Malmivaara A, Chou R, et al. Updated method guideline for systematic reviews in the cochrane back and neck group. Spine
20. Atkins D, Best D, Briss P, et al. Grading quality of evidence and strength of recommendations. BMJ
21. Rasouli M, Rahimi-Movaghar V, Shokraneh F, et al. Minimally invasive discectomy versus microdiscectomy/open discectomy for symptomatic lumbar disc herniation
. Cochrane Database Syst Rev
22. Wang X, Sun R, Ji Q, et al. A meta-analysis of interlaminar minimally invasive discectomy compared to conventional microdiscectomy for lumbar disk herniation. Clin Neurol Neurosurg
23. Kamper S, Ostelo R, Rubinstein S, et al. Minimally invasive surgery
for lumbar disc herniation
: a systematic review and meta-analysis. Eur Spine J
24. Clark A, Safaee M, Khan N, et al. Tubular microdiscectomy
: techniques, complication avoidance, and review of the literature. Neurosurg Focus
25. Li X, Chang H, Meng X. Tubular microscopes discectomy versus conventional microdiscectomy for treating lumbar disk herniation: systematic review and meta-analysis. Medicine (Baltimore)
26. Guyatt G, Oxman A, Kunz R, et al. GRADE guidelines 6. Rating the quality of evidence–imprecision. J Clin Epidemiol