Cervical spondylotic myelopathy (CSM) is a common cause of neurological dysfunction. The onset of CSM is typically marked by fine motor dysfunction and decreased hand dexterity, as well as worsening gait and balance. Upper and lower extremity sensorimotor dysfunction and sphincter disturbance most commonly occur in a slow, stepwise pattern with disease progression, although rapid neurological decline can occur in a minority of cases. In keeping with the degenerative etiology of CSM, it is intuitive that the incidence increases with advancing age, and accordingly, CSM represents the most common cause of spinal cord dysfunction in individuals older than 55 years.1
It is well established that surgical decompression of the cervical spinal cord is an effective treatment option for CSM that, at a minimum, not only halts the progression of symptoms, but can also promote meaningful functional recovery in a significant portion of treated individuals.2,3 Spondylotic spinal cord compression can arise from pathologies located either anterior or posterior to the spinal canal, and, accordingly, surgical decompression can be performed via either an anterior surgical approach or a posterior surgical approach. Anterior surgery typically takes the form of either an anterior cervical discectomy and fusion or corpectomy, and posterior surgery typically involves a laminoplasty or laminectomy and fusion. Although laminectomy without fusion was widely used for the treatment of CSM in the past, the increasing recognition of the low but not insignificant incidence of postlaminectomy kyphotic deformities has resulted in a dramatic reduction in stand-alone laminectomy in the setting of CSM.4
Presently, it remains unclear whether multilevel spondylotic compression is best treated via an anterior or posterior surgical route and whether one of these surgical approaches is superior in terms of patient outcomes and/or complication profile. Several reports using large administrative databases have attempted to elucidate the safety and effectiveness profiles of the anterior versus posterior approach when treating multilevel CSM. Unfortunately, in the study by Shamji et al,5 no conclusions could be drawn regarding the effectiveness differences of anterior versus posterior surgery because of the lack of pathoanatomical patient data in this large administrative database. More recently, Fehlings et al6 described the complications of surgical treatment in patients with CSM and found that the posterior surgical group had a higher incidence of postoperative wound infection but overall complication rates, C5 palsy, and incidence of dysphagia were similar. Given this ongoing uncertainty as to the optimal surgical treatment paradigm for multilevel spondylotic cervical cord compression, there remains a substantive need for a rational algorithm on how best to surgically treat CSM.
Accordingly, the primary objective of this report was to perform a systematic review comparing anterior multilevel cervical discectomy or corpectomy with posterior cervical laminoplasty or laminectomy and fusion with regard to the following clinical outcomes: neurological outcome, postoperative neck pain, neck range of motion, and sagittal alignment, as well as the postoperative complication profile. Ultimately, we sought to answer the following key questions:
- What is the comparative effectiveness of anterior versus posterior decompression procedures for multilevel CSM?
- What is the comparative safety of anterior versus posterior decompression procedures for multilevel CSM?
MATERIALS AND METHODS
Electronic Literature Search
We conducted a systematic search in MEDLINE and the Cochrane Collaboration Library for literature published through September 2012. The search results were limited to human studies published in the English language. Reference lists of key articles were also systematically checked to identify additional eligible articles. We sought to identify comparative studies (e.g., randomized controlled trials, cohort studies) comparing anterior surgical approaches (discectomy and/or corpectomy) with posterior surgical approaches (laminectomy, laminectomy/fusion, and laminoplasty). We were interested in the following effectiveness and complication profiles: neurological, functional, quality of life, sagittal alignment, cervical range of motion, pain, and complications (Table 1). Reasons for exclusion included patients with degenerative disc disease or degenerative joint disease without CSM, patients with ossification of the posterior longitudinal ligament (OPLL), and a majority of subjects in one or both treatment arms with single-level CSM, spine tumor, trauma, and infection. We excluded studies that did not report results separately by treatment group because this prevented appropriate comparisons. Furthermore, studies that stated in the “Methods” section that patients were differentially selected for treatments based on important baseline characteristics were excluded. These studies created comparisons with significant imbalances with respect to key factors that influence outcome (e.g., age, disease levels) and put specific treatments at a clear disadvantage that we could not control for. Case series, case reports, data not reported separately for each comparison group, or studies that consisted of a sample size of less than 10 for either comparison group were excluded. Animal, cadaver, and biomechanical studies were also excluded.
From the included articles, the following data were extracted: study design, patient demographics, diagnosis and operated levels, follow-up duration and the rate of follow-up for each treatment group (if reported or calculable), and treatment intervention descriptions. We reported change in JOA (Japanese Orthopaedic Association) scores, postoperative axial pain, change in sagittal alignment, change in canal diameter, and complication rates.
Study Quality and Overall Strength of Body of Literature
Level of evidence ratings were assigned to each article independently by 2 reviewers (J.T.H., D.C.N.) using criteria set by The Journal of Bone & Joint Surgery, American Volume,7 for therapeutic studies and modified to delineate criteria associated with methodological quality described elsewhere.8 (See the Supplemental Digital Content, available at http://links.lww.com/BRS/A826, for individual study ratings.)
The overall body of evidence with respect to each key question was determined on the basis of precepts outlined by the GRADE (Grading of Recommendation Assessment, Development and Evaluation) Working Group9 and recommendations made by the AHRQ (Agency for Healthcare Research and Quality).10 Risk of bias was evaluated during the individual study evaluation described earlier in this section. This system, which derives a strength-of-evidence grade of “high,” “moderate,” “low,” or “insufficient” for each outcome or key question, is described in further detail in the Methodology article in this focus issue.8 A detailed description of how we arrived at the strength of evidence for each key question can be found in the Supplemental Digital Content (available at http://links.lww.com/BRS/A826).
We performed all analyses on an individual study level. For continuous outcome measures, we reported, or if necessary calculated, the mean change scores and standard deviations (SDs). We imputed the change score SDs using a formula recommended by the Cochrane Handbook for Systematic Reviews of Interventions.11 Mean baseline, mean follow-up, mean change scores, and their respective SDs allowed calculation of a standardized mean difference (SMD) that permitted comparison of change in scores between anterior and posterior surgical procedures within the different studies. SMDs and their 95% confidence intervals were displayed using forest plots generated using Review Manger 5.2.1. A random-effects model was assumed in an attempt to address heterogeneity. Cohen12 has reported an effect size of 0.2 to 0.3 as a “small” effect, around 0.5 as a “medium” effect, and 0.8 to infinity as a “large” effect. We used unpaired t tests to determine whether the outcome score changes were variable between different approaches. If the confidence interval did not cross the neutral line in the forest plot, then the effect can also be considered statistically significant. For dichotomous outcomes (e.g., axial pain rates or complications), we reported the raw data and rates and calculated risk differences and relative risks and their 95% confidence intervals. To test whether the differences in rates were statistically significant, we used the Fischer exact test. Risk differences are calculated by subtracting the 2 rates and relative risks by dividing one rate by another. Because of significant clinical (differences in demographics, interventions, and outcome assessment) and statistical (variability in intervention effects or follow-up rates) heterogeneity, we did not combine studies into a meta-analysis; however, we presented studies side by side in summary tables and figures to make a qualitative assessment of treatment effectiveness and complications. We performed statistical analyses using Stata 9.1.13
Clinical Recommendations and Consensus Statements
Clinical recommendations were made through a modified Delphi approach by applying the GRADE/AHRQ criteria that impart a deliberate separation between the strength of the evidence (i.e., high, moderate, low, or insufficient) from the strength of the recommendation. When appropriate, recommendations or statements “for” or “against” were given “strong” or “weak” designations based on the quality of the evidence, the balance of benefits/harms, and values and patient preferences. In some instances, costs may have been considered. A thorough description of this process can be found in the Methodology article in this focus issue.8
The search strategy yielded 135 potentially relevant citations. Of these, 115 were excluded on the basis of title and/or abstract. Twenty were selected for full-text review. An additional 12 were excluded on the basis of full-text review for the following reasons: differential treatment group selection based on heterogeneous characteristics such as disease levels and age, data not reported separately by treatment group, and majority of anterior surgery group containing single-level CSM (61%–77%)14–17 (Figure 1). A total of 8 studies18–25 were selected for inclusion and are summarized in this report. Among the 8 studies, 2 compared multilevel discectomy with laminoplasty, 1 study compared multilevel discectomy with laminectomy, 4 studies compared multilevel corpectomy with laminoplasty, and 1 study compared multilevel corpectomy with laminectomy with fusion in patients with multilevel CSM. No study was identified comparing multilevel corpectomy with laminectomy alone or multilevel discectomy with laminectomy with fusion.
The mean age of subjects was similar across studies, with age ranges of 48 to 59 years. All studies had a greater percentage of male patients (range, 58%–86%) than of female patients. The most common number of operated levels was 3 and relatively consistent across studies. Follow-up times differed between studies. Five studies had follow-up times between 2 and 4 years, and 2 studies had follow-up times between 7 and 11 years postoperatively. The main limitation of these studies was that the authors did not report follow-up rates or give the raw data to calculate follow-up rates. These differences in follow-up times and unknown follow-up rates should caution the final conclusions. Details regarding the demographics, operative characteristics, and results can be found in the detailed tables in the Supplemental Digital Content material (available at http://links.lww.com/BRS/A826).
We summarized the comparative effectiveness and safety of anterior versus posterior surgery in the following sections. Effectiveness outcomes consistent across studies and reported in a similar manner included change in JOA scores, postoperative axial pain scores, and change in canal diameter scores. The results by study on these outcomes are reported in summary tables and figures. Sagittal alignment was reported by several studies, but the methods were so heterogeneous that we were unable to summarize them in a comparable way. Neck range of motion was rarely reported and therefore not useful for summary. Unfortunately, none of the studies reported the NDI (Neck Disability Index) or a quality-of-life outcome such as the SF-36 (36-Item Short Form Health Survey). Consistent complications reported across all the studies included pseudoarthrosis, C5 palsy, infection, and dysphagia.
Comparative Effectiveness of Anterior Versus Posterior Decompression Procedures for Multilevel CSM
When evaluating the change in JOA scores among studies that reported both pre- and postoperative scores, 2 of the 4 studies18,24 reported greater improvement after anterior surgery than after posterior surgery. Among these studies, the difference in change scores was statistically significant in only one of them,24 with an SMD of 1.44, which represents a “large effect” (Figure 2). Among the 2 studies favoring posterior surgery,22,25 neither demonstrated a statistically significant difference between change scores with SMDs of 0.31 and 0.03, respectively. Both are considered small effects.
Axial Pain Scores
When evaluating the risk differences among studies that reported postoperative axial pain, 3 of the 4 favored anterior surgery20,22,23 with less postoperative axial pain and 1 reported no difference.19 Among the studies favoring anterior surgery, the risk differences were 3.7% (P = 0.34), 40.5% (P = 0.0004), and 28% (P = 0.03) (Table 2).
When evaluating the change in canal diameter measurements among studies that reported both pre- and postoperative measurements, all 3 studies22,24,25 reported a greater increase in canal diameter after posterior surgery than after anterior surgery. Among these studies, the difference in canal diameter increase was statistically significant in 2 of the 3 studies22,25 (P = 0.001 and P = 0.0003, respectively) and resulted in SDMs of 1.35 and 1.57. These both represent a “large effect” (Figure 3).
Comparative Safety of Anterior Versus Posterior Decompression Procedures for Multilevel CSM
The following complications among studies comparing anterior with posterior surgery for multilevel CSM were extracted from the 8 included studies: pseudoarthrosis, C5 palsy, infection, and dysphagia (Table 2). Three studies reported pseudoarthrosis rates19,22,23; however, unfortunately, the only studies comparing pseudoarthrosis rates used laminoplasty as the posterior comparator. Among these, pseudoarthrosis was observed only after anterior surgery with postoperative rates of 4%, 7.7%, and 17.6%, respectively, compared with 0% in the posterior group. Only one of these was statistically significant (P = 0.48, P = 0.31, and P = 0.004, respectively). Three studies reported dysphagia,19,21,22 with rates of 8%, 30.8%, and 7.1%, respectively, in the anterior groups compared with no events in the posterior groups. Two of the 3 studies were statistically significant (P = 0.13, P = 0.03, and P = 0.03, respectively). Three studies reported C5 palsy with anterior versus posterior rate comparisons of 0% versus 7.4% (P = 0.49),22 8.8% versus 10.2% (P = 0.83),23 and 9.8% versus 7.1% (P = 0.67),24 respectively. Infection rates were reported in both groups in only one study,21 with 2.3% in the anterior group and 6.5% in the posterior group (P = 0.33).
The overall strength of evidence that one approach (anterior vs. posterior) is more effective than the other when treating patients with CSM is “insufficient” (the evidence does not permit a conclusion) with regard to improvements of JOA/modified JOA (mJOA) scores; “low” with regard to postoperative pain favoring anterior surgery; and “moderate” (moderate confidence that the evidence reflects the true effect) with regard to increasing canal diameter favoring posterior surgery. The overall strength that one approach (anterior vs. posterior) is more safe than the other when treating patients with CSM is “insufficient” with regard to pseudoarthrosis, the development of postoperative C5 palsy, and infection. There is “low” evidence that posterior surgery has fewer occurrences of dysphagia than anterior surgery (Table 3).
When developing a surgical strategy to address CSM, there are many variables that one must consider. In this systematic review, an attempt was made to evaluate both the comparative effectiveness and safety of the anterior versus posterior surgical approach for the treatment of multilevel cervical degenerative spinal cord compression, with the goal of defining superiority of one surgical approach. Unfortunately, given the paucity of well-designed studies and the heterogeneity of outcome and complications reporting, defining a superior approach is not possible. With that said, our analysis suggests that there is clinical equipoise with regard to efficacy and safety between the anterior and posterior approaches for the treatment of multilevel CSM. Thus, given the multitude of variables that one must consider when deciding on an anterior or posterior surgical approach, including ventral versus dorsal compression, the number of levels of spinal cord compression, the presence or absence of radiculopathy, and sagittal alignment of the cervical spine, our review supports using a customized approach to each individual patient, given the similarity with regard to outcome and safety.
When evaluating effectiveness across these studies, several variables were consistently available for comparison; specifically neurological outcome, as measured by JOA/mJOA scores, axial neck pain, and radiographical measurement of postoperative spinal canal diameter. Unfortunately, none of the studies used health-related quality-of-life scores, such as the SF-36, or functional measures, such as the NDI, when reporting their outcomes. Of the 5 studies for which pre- and postoperative JOA scores were available for comparison, only one study determined a statistically significant improvement after anterior surgery.24 Interestingly, in this study, the preoperative JOA scores were not equivalent between groups, with the anterior group having significantly lower scores than the posterior group (8.1 vs. 9.3). Postoperatively, patients in both groups had similar JOA scores (13.3 vs. 12.8). Of the 3 remaining studies that compared JOA/mJOA scores, 2 favored posterior surgery and 1 favored anterior surgery, but none of these were statistically significant. As such, the current literature is insufficient for making a statement with regard to the superiority of the anterior versus posterior surgical approach for neurological outcome.
Similar confounding factors exist regarding postoperative axial neck pain comparisons. A total of 4 studies, all comparing anterior procedures with laminoplasty, reported on comparisons of postoperative neck pain, with 2 of the studies favoring anterior surgery.20,23 It is important to point out that none of the studies reported pre- to postoperative differences because no preoperative pain assessments were reported in any of these studies. Although this is consistent with clinical experience and is likely secondary to musculotendinous disruption inherent in the posterior approach, it is difficult to make conclusions based solely on postoperative data due to potential confounding by baseline pain.
Because of the lack of concrete outcome measure differences between anterior and posterior surgical procedures, it is difficult to ascertain the effects of the radiographical measurement—canal diameter—that was published. Although it seems a larger canal diameter can be created more effectively with posterior surgery, it does not seem to correlate with improvements in JOA/mJOA scores. Therefore, determining the clinical relevancy of this particular radiographical assessment is not possible. Perhaps, a measurement such as pre- to postoperative cord compression would be more clinically relevant; however, unfortunately this measurement was not included in any of the studies meeting our criteria.
When evaluating safety across these studies, similar complication profiles were reported. Unfortunately, significant heterogeneity with regard to reported complications exists, making conclusions about the safety of the anterior versus posterior surgical approach difficult. In addition, the number of events was so low that statistical significance was difficult to achieve due to low study power and comparators were not clinically relevant in all circumstances. For instance, pseudoarthrosis was compared with laminoplasty as the posterior comparator. Arthrodesis is not a goal in laminoplasty and therefore not comparable with anterior cervical discectomy and fusion or corpectomy. It does seem that posterior surgery may be associated with lower rates of postoperative dysphagia, although not statistically significant. Anterior surgery may be associated with lower infection rates, but the reported infection rates in this review are much higher than those reported in the literature overall.6 This may be due to the small numbers reported, but it may also influence the outcomes of this particular group of patients, given the morbidity associated with postoperative wound infections. The rates for C5 palsy were similar.
Because of the lack of well-designed randomized controlled trials, it remains unclear whether there is an optimal approach for surgical treatment of patients with CSM. In light of this, individualizing the approach decision to the details of each patient (location of compression, the number of levels involved, overall cervical alignment, the presence or absence of neck pain, etc.) seems most appropriate. What is known is that patients with moderate and severe myelopathy report improved outcomes and improved neurological recovery after surgical decompression irrespective of approach. Given the differing approaches to differing pathoanatomy, several illustrative cases are presented: 1 case depicts that an anterior approach is preferred (Figure 4); 1 case for which a posterior approach is preferred (Figure 5); and for 2 cases that are equivocal, either an anterior approach or a posterior approach could be performed (Figures 6 and 7).
OPLL was excluded in this systematic review; however, well-established treatment algorithms have been described for the surgical treatment of this disease. Most surgeons prefer to treat OPLL with a posterior approach, given the risk of cerebrospinal leakage, but if more than 60% of the canal is compromised by OPLL, then neurological outcomes have been shown to improve with anterior surgery.26 A recent systematic review was attempted by Xu et al27 comparing anterior with posterior approaches in patients with multilevel OPLL. Given the lack of comparable outcomes reported between the anterior and posterior approaches, as well as the inconsistency across the studies, no conclusions could be drawn.
This systematic review strengthens the strategy that most surgeons use when treating patients with CSM, one that requires understanding of the disease process and the complications associated with surgical treatment and one that is individualized on the basis of the pathoanatomy. Clearly, consensus on outcome measures and reported complication profiles would facilitate comparisons across differing treatment centers and surgical techniques. Such studies should use more rigorous methods to include a prospective design to diminish heterogeneity in patient populations, follow-up times, and patient-reported measures that take into account both neck disability and extremity disability.
Evidence-Based Clinical Recommendations.
Recommendation. We recommend a individualized approach when treating patients with CSM accounting for pathoanatomical variations (ventral vs. dorsal, focal vs. diffuse, sagittal, dynamic instability), as there appears to be similar outcomes between the anterior and posterior approaches in regards to effectiveness and safety.
Overall Strength of Evidence. Low
Strength of Recommendation. Strong
- The comparative effectiveness is similar between the anterior and posterior surgical approach when treating patients with multilevel CSM.
- The comparative safety is similar between the anterior and posterior surgical approach when treating patients with multilevel CSM.
- Using an individualized approach when treating multilevel CSM is recommended because differing pathologies require differing surgical approaches in the cervical spine.
Supplemental digital content is available for this article. Direct URL citation appears in the printed text and is provided in the HTML and PDF versions of this article on the journal's Web site (www.spinejournal.com).
1. Emery SE. Cervical spondylotic myelopathy: diagnosis and treatment. J Am Acad Orthop Surg 2001;9:376–88.
2. Furlan JC, Kalsi-Ryan S, Kailaya-Vasan A, et al. Functional and clinical outcomes following surgical treatment in patients with cervical spondylotic myelopathy: a prospective study of 81 cases. J Neurosurg Spine 2011;14:348–55.
3. Sampath P, Bendebba M, Davis JD, et al. Outcome of patients treated for cervical myelopathy. A prospective, multicenter study with independent clinical review. Spine 2000;25:670–6.
4. Kaptain GJ, Simmons NE, Replogle RE, et al. Incidence and outcome of kyphotic deformity following laminectomy for cervical spondylotic myelopathy. J Neurosurg 2000;93:199–204.
5. Shamji MF, Cook C, Tackett S, et al. Impact of preoperative neurological status on perioperative morbidity associated with anterior and posterior cervical fusion. J Neurosurg Spine 2008;9:10–6.
6. Fehlings MG, Smith JS, Kopjar B, et al. Perioperative and delayed complications associated with the surgical treatment of cervical spondylotic myelopathy based on 302 patients from the AOSpine North America Cervical Spondylotic Myelopathy Study. J Neurosurg Spine 2012;16:425–32.
7. Wright JG, Swiontkowski MF, Heckman JD. Introducing levels of evidence to the journal. J Bone Joint Surg Am 2003;85-A:1–3.
8. Norvell DC, Dettori JR, Fehlings MG, et al. Methodology for the systematic reviews on an evidence-based approach for the management of chronic low back pain. Spine 2011;36:S10–8.
9. Atkins D, Best D, Briss PA, et al. Grading quality of evidence and strength of recommendations. BMJ 2004;328:1490.
10. West S, King V, Carey TS, et al. Systems to Rate the Strength of Scientific Evidence (prepared by the Research Triangle Institute–University of North Carolina Evidence-based Practice Center, Contract No. 290-97-0011). Rockville, MD: Agency for Healthcare Research and Quality; 2002. Evidence Report/Technology Assessment No. 47.
11. Deeks JJ, Higgins JPT, Altman DG. Chapter 9: Analysing data and undertaking meta-analyses. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0. Updated March 2011. The Cochrane Collaboration, 2011. Available at: www.cochrane-handbook.org
. Accessed September 14, 2013.
12. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. 2011.
13. Stata Statistical Software [computer program]. Version 9.1. College Station, TX: Stata Corporation LP; 2005.
14. Cabraja M, Abbushi A, Koeppen D, et al. Comparison between anterior and posterior decompression with instrumentation for cervical spondylotic myelopathy: sagittal alignment and clinical outcome. Neurosurg Focus 2010;28:E15.
15. Ebersold MJ, Pare MC, Quast LM. Surgical treatment for cervical spondylitic myelopathy. J Neurosurg 1995;82:745–51.
16. Kawakami M, Tamaki T, Iwasaki H, et al. A comparative study of surgical approaches for cervical compressive myelopathy. Clin Orthop Relat Res 2000;381:129–36.
17. Wada E, Suzuki S, Kanazawa A, et al. Subtotal corpectomy versus laminoplasty
for multilevel cervical spondylotic myelopathy: a long-term follow-up study over 10 years. Spine (Phila Pa 1976) 2001;26:1443–7; discussion 1448.
18. Benzel EC, Lancon J, Kesterson L, et al. Cervical laminectomy and dentate ligament section for cervical spondylotic myelopathy. J Spinal Disord 1991;4:286–95.
19. Edwards CC II, Heller JG, Murakami H. Corpectomy versus laminoplasty
for multilevel cervical myelopathy: an independent matched-cohort analysis. Spine 2002;27:1168–75.
20. Hosono N, Yonenobu K, Ono K. Neck and shoulder pain after laminoplasty
. A noticeable complication. Spine 1996;21:1969–73.
21. Kristof RA, Kiefer T, Thudium M, et al. Comparison of ventral corpectomy and plate-screw-instrumented fusion with dorsal laminectomy and rod-screw-instrumented fusion for treatment of at least two vertebral-level spondylotic cervical myelopathy. Eur Spine J 2009;18:1951–6.
22. Liu T, Yang HL, Xu YZ, et al. ACDF with the PCB cage-plate system versus laminoplasty
for multilevel cervical spondylotic myelopathy. J Spinal Disord Tech 2011;24:213–20.
23. Shibuya S, Komatsubara S, Oka S, et al. Differences between subtotal corpectomy and laminoplasty
for cervical spondylotic myelopathy. Spinal Cord 2010;48:214–20.
24. Yonenobu K, Hosono N, Iwasaki M, et al. Laminoplasty versus
subtotal corpectomy. A comparative study of results in multisegmental cervical spondylotic myelopathy. Spine 1992;17:1281–4.
25. Yoshida M, Tamaki T, Kawakami M, et al. Indication and clinical results of laminoplasty
for cervical myelopathy caused by disc herniation with developmental canal stenosis. Spine 1998;23:2391–7.
26. Iwasaki M, Okuda S, Miyauchi A, et al. Surgical strategy for cervical myelopathy due to ossification of the posterior longitudinal ligament, part 2: advantages of anterior decompression and fusion over laminoplasty
. Spine 2007;32:654–60.
27. Xu J, Zhang K, Ma X, et al. Systematic review of cohort studies comparing surgical treatment for multilevel ossification of posterior longitudinal ligament: anterior vs
posterior approach. Orthopedics 2011;34:e397–402.