The use of internal fixation in the subaxial cervical

Transcription

1 J Neurosurg Spine 19: , 2013 AANS, related complications in the subaxial cervical spine with the use of lateral mass versus cervical pedicle screws A systematic review Hiroyuki Yoshihara, M.D., Ph.D., Peter G. Passias, M.D., and Thomas J. Errico, M.D. Department of Orthopaedic Surgery, New York University Hospital for Joint Diseases, New York, New York Object. Lateral mass screws (LMS) have been used extensively with a low complication rate in the subaxial spine. Recently, cervical pedicle screws (CPS) have been introduced, and are thought to provide more optimal stabilization of the subaxial spine in certain circumstances. However, because of the concern for neurovascular injury, the routine use of CPS in this location remains controversial. Despite this controversy, however, there are no articles directly comparing screw-related complications of each procedure in the subaxial cervical spine. The purpose of this study was to evaluate screw-related complications of LMS and CPS in the subaxial cervical spine. Methods. A PubMed/MEDLINE and Cochrane Collaboration Library search was executed, using the key words lateral mass screw and cervical pedicle screw. Clinical studies evaluating surgical procedures of the subaxial cervical spine in which either LMS or CPS were used and complications were reported were included. Studies in which the number of patients who had subaxial cervical spine surgery and the number of screws placed from C-3 to C-7 could not be specified were excluded. Data on screw-related complications of each study were recorded and compared. Results. Ten studies of LMS and 12 studies of CPS were included in the analysis. Vertebral artery injuries were slightly but statistically significantly higher with the use of CPS relative to LMS in the subaxial cervical spine. Although the use of LMS was associated with a higher rate of screw loosening, screw pullout, loss of reduction, pseudarthrosis, and revision surgery, this finding was not statistically significant. Conclusions. Based on the available literature, it appears that perioperative neurological and late biomechanical complication rates, including pseudarthrosis, are similarly low for both LMS and CPS techniques. In contrast, vertebral artery injuries, although statistically significantly more common when using CPS, are extremely rare with both techniques, which may justify their nonroutine use in select cases. Given the paucity of well-designed studies available, this recommendation may be a reflection of deficiencies in the available studies. Surgeons using either technique should have intimate knowledge of cervical anatomy and an adequate preoperative evaluation for each patient, with the final selection based on individual case requirements and anatomical limitations. ( Key Words complication lateral mass screw cervical pedicle screw neurovascular injury vertebral artery injury hardware failure technique The use of internal fixation in the subaxial cervical spine has become the standard of care for several surgical indications, including certain cases of trauma, instability, deformity, and neoplastic destruction. However, various techniques and instrumentation options are available with regard to posterior stabilization of the subaxial cervical spine. These include wiring, interlaminar clamps, lateral mass screws (LMS), and most recently cervical pedicle screws (CPS). 1,2,6,11,20,38,40,41 Posterior cervical fixation with LMS was first described by Abbreviations used in this paper: ASD = adjacent-segment disease; CPS = cervical pedicle screws; LMS = lateral mass screws; SCI = spinal cord injury; VA = vertebral artery. Roy-Camille et al. 40 in The development of polyaxial screws made acutely angulated trajectories possible to place longer screws toward the superior-lateral-ventral corner of the lateral mass. In addition, the improvement of LMS constructs, from posterior cervical plating to rod-based systems, has provided additional advantages, such as greater amenability to multiplanar contouring, increased flexibility for extending to the occiput or thoracic spine, and allowing more precise screw placement. Roy-Camille et al. 39 initially stated that the placement of transpedicular screws into the C3 6 pedicles would be of unacceptable risk for injury to the vertebral artery (VA), spinal cord, and nerve roots. However, Abumi and colleagues 1,4 first described the successful clinical 614 J Neurosurg: Spine / Volume 19 / November 2013

2 -related complications in the subaxial cervical spine application of CPS for traumatic lesions in the subaxial cervical spine in 1994 and have since reported a low complication incidence of 2.2% for CPS fixation in The surgical indications for this technique have gradually expanded to the management of various kinds of nontraumatic cervical instability as well as the correction of cervical kyphosis. 2,3,5 However, because of the possible risk of neurovascular injury, 4 CPS in the subaxial cervical spine remain controversial for routine use. Although LMS instrumentation has been proven effective in stabilizing the cervical spine with various kinds of pathology, CPS may provide optimal stabilization for an unstable motion segment in cases in which posterior wiring or LMS cannot be applied. 53 Furthermore, CPS may obviate the requirement for anterior-posterior procedures in cases with a high risk of pseudarthrosis or construct failure. Despite these existing controversies there are no articles directly comparing screw-related complication rates between LMS and CPS techniques. Therefore, in this report we systematically review screw-related complications of LMS and CPS in the subaxial cervical spine based on the available published literature. Methods A literature search was performed using the National Center for Biotechnology Information databases using PubMed/MEDLINE and Cochrane Collaboration Library search engines covering the period from January 1950 until January A search using the key word lateral mass screw was performed and returned 540 articles. Studies in the English language were included. Abstracts from these articles were reviewed, and those that discussed clinical outcomes as well as those in which the average patient age was greater than 18 years were retained for more detailed review. Studies evaluating surgeries of the subaxial cervical spine performed using LMS, regardless of the indication for surgery, and reporting complications were included. Studies in which the number of patients who had subaxial cervical spine surgery and the number of screws placed from C-3 to C-7 could not be specified were excluded. In several studies only the data of C3 7 were extracted. If the same author published more than one article with overlap of the included patients, the article that included the greater number of patients and/or screws was included in the analysis. If the same author published more than one article without overlapping of the patients, both studies were included (Table 1). In the same way, a search using the key word cervical pedicle screw was performed and returned 419 articles. Inclusion and exclusion criteria were the same as for the lateral mass screw criteria. Screening included titles, subtitles, and abstracts. Two reviewers independently screened the titles and abstracts of all publications that were obtained by the initial search process. Investigators were contacted and asked to supplement additional data and clarifications when key information was missing from publications. We extracted data on study characteristics, patient enrollment, and the complications reported. Specifically, the following data were collected: type of study (prospective, J Neurosurg: Spine / Volume 19 / November 2013 retrospective); number of patients who had a subaxial cervical spine fusion in which LMS or CPS were used; number of screws inserted from C-3 to C-7; follow-up period; age; sex; screw trajectory; image methods used for screw placement (fluoroscopy or navigation); uni- or bicortical for LMS; and screw-related complications from C-3 to C-7 (perioperative and late screw-related complications). Perioperative screw-related complications that were recorded included intraoperative and early postoperative complications. Specific complications in this category that were included were nerve root injury, spinal cord injury (SCI), VA injury, fracture of lateral mass, facet violation, and malposition that required revision or removal. Late screw-related complications included screw loosening, screw pullout, screw breakage, plate/rod breakage, loss of reduction, pseudarthrosis, required revision surgery, and adjacent-segment disease (ASD) requiring surgery. Only complications that were specifically stated as having occurred or not having occurred in the articles were used in the analysis. Complications were not assumed to be absent just because they were not discussed, unless this was confirmed by direct communication with the corresponding authors (Table 1). Statistical analysis comparing the complication rates obtained from either technique was performed using a chi-square test with the Yates continuity correction or the Fisher exact test (when at least 1 cell had a count less than 5), and the level of significance was chosen at p < The statistical analysis was done using software (Statistica version 8.0, Statsoft, and R ). Results Our literature search identified 28 possible eligible studies for LMS. Eighteen studies were omitted due to the fact that they did not fulfill the inclusion and exclusion criteria. Therefore, a total of 10 studies were ultimately included in the analysis of LMS. 9,15 17,25,36,42,47 49 One study was prospective and 9 studies were retrospective. These studies included 766 patients in total, in whom 5328 screws were inserted from C-3 to C-7 (Table 2). Each particular complication of LMS was not reported in all studies (Table 3). The rate of each complication was calculated relative to the number of patients and screws involved. The perioperative complication rates were as follows: nerve root injury, 0.19% per screw and 1.36% per patient; SCI, 0% per screw and 0% per patient; VA injury, 0% per screw and 0% per patient; fractures of lateral mass, 1.62% per screw; facet violation, 0.62% per screw; and malposition requiring revision or removal, 0.38% per screw and 2.64% per patient. Late complication rates were as follows: screw loosening, 1.17% per screw; screw pullout, 1.1% per patient; plate/ rod breakage, 0.28% per patient; loss of reduction, 2.21% per patient; pseudarthrosis, 2.67% per patient; revision required, 2.81% per patient; and ASD requiring surgery, 0.74% per patient (Table 4). Our literature search identified 28 possible eligible studies for CPS. Sixteen studies were omitted due to the fact that they did not fulfill the inclusion and exclusion criteria. Therefore, a total of 12 studies were ultimately 615

3 TABLE 1: Inclusion and exclusion criteria for the studies and inclusion criteria for complications* H. Yoshihara, P. G. Passias, and T. J. Errico inclusion studies evaluating ops of the subaxial cervical spine using LMS or CPS reporting complications regardless of the indication for op average pt age >18 yrs if the same author published >1 article w/ overlapping of the included pts, the article that included the greater no. of pts &/or screws was included if the same author published >1 article w/o overlapping of the pts, both studies were included English language exclusion studies in which the no. of pts who had subaxial cervical spine ops & the no. of screws placed from C-3 to C-7 could not be specified were excluded case reports inclusion criteria for complications only complications specifically stated in the articles as having or not having occurred were used in the analysis complications were not assumed to be absent just because they were not discussed, unless confirmed by direct communication w/ corresponding authors * Pt = patient. included in the analysis of CPS. 4,24,27,28,30,32 34,44,45,51,52 One study was prospective, 5 studies were retrospective, and for 6 the type of study was not described. These studies included 661 patients in whom 2668 screws were inserted from C-3 to C-7 (Table 2). Fluoroscopy was used in 9 studies, navigation was used in 2 studies, and both fluoroscopy and navigation were used in 1 study. As with the LMS analysis, each particular complication of CPS was not reported in all studies (Table 3). The rate of each complication was calculated relative to the number of patients and screws involved. Perioperative complication rates were as follows: nerve root injury, 0.31% per screw and 1.24% per patient; SCI, 0% per screw and 0% per patient; VA injury, 0.15% per screw and 0.61% per patient; and malposition requiring revision or removal, 0.29% per screw and 1.1% per patient. Late complications rates were as follows: screw loosening, 0.45% per screw and 1.73% per patient; screw pullout, 0.24% per patient; screw breakage, 1.76% per patient; plate/ rod breakage, 0% per patient; loss of reduction, 1.46% per patient; pseudarthrosis, 0.87% per patient; revision required, 1.03% per patient; and ASD requiring surgery, 1.19% per patient (Table 4). When directly comparing the complication rates between LMS and CPS techniques, there were no statistically significant differences in the development of complications for all of the categories analyzed, with the distinct exception of VA injuries, which occurred slightly more frequently among patients with CPS constructs (p < 0.05). Although the use of LMS was associated with a higher rate of screw loosening, screw pullout, loss of reduction, pseudarthrosis, and revision surgery, this finding was not statistically significant (Table 4). Discussion Perioperative Complications Numerous techniques for placing LMS have been described, including those reported by Jeanneret and colleagues (referred to as the Magerl technique), 21 Anderson et al., 8 An et al., 7 and Roy-Camille et al. 41 Subsequently, many authors have conducted anatomical studies to clarify the safety of each method of screw placement. 10,35,50 In a cadaveric study reported in 1995, Pait et al. 35 divided the lateral cervical mass (articular pillars of the cervical spine) into quadrants and concluded that the superior lateral quadrant was the safe quadrant for placing screws. Xu et al. 50 found in 1999 that the potential risk for nerve root violation is lower for the An technique than the Magerl and Anderson techniques. In 2005, using morphological CT scans, Barrey et al. 10 found that the Roy-Camille technique was the best option for C-3 and C-4, whereas the Magerl technique was a safer, although more technically demanding, procedure for C-5 and C-6. The risks of vascular and neural injury with the longer screws could be reduced by placing their tips in the ideal position; the superior-lateral-ventral corner of the lateral mass. 49 Heller and colleagues 17,19 predicted a maximum 3.6% incidence of nerve root injury using the Roy-Camille and Magerl trajectories, presumably because of the screw length used and more lateral trajectories, respectively. Although LMS can be placed at C-7, the lateral mass at C-7 is somewhat elongated in a rostral-caudal direction and is thinner, so CPS may be more appropriate at C-7. If LMS are to be used at C-7, great care must be taken so as not to insert too long a screw and injure the C-8 nerve root. Nerve root injury by LMS in our study was found to be very low, with a rate of 0.19% per screw. Although no vascular complication by LMS was reported in the current study, it has been documented to occur. Cho et al. 12 reported a case of VA injury by LMS, which was attributed to the use of a long screw. Facet violation can also occur. Some studies reported cases with facet violation, which rate is 0.62% per screw in our study. However, this risk may be minimized by using fluoroscopy. The use of CPS has been considered to pose a serious potential risk to the surrounding structures. 39 Many surgeons have advocated using CPS in C-2 and C-7 and LMS 616 J Neurosurg: Spine / Volume 19 / November 2013

4 -related complications in the subaxial cervical spine TABLE 2: Summary of studies included in the analysis* Authors & Year Study Type Pts s (C3 7) Follow-Up (range) Mean Age in Yrs (range) No. & Sex (M/F) Trajectory Method of Navigation Unicortical Bicortical LMS Heller et al., 1991 retrospective yrs (10 47 mos) 52.9 (14 82) ND An tech fluoroscopy tried for all Fehlings et al., 1994 retrospective mos (2 6 yrs) /7 Roy-Camille Graham et al., 1996 prospective >1 yr 52.4 (28 81) 17/4 Roy-Camille (C-7), Magerl (C3 6) 0 all Wellman et al., 1998 retrospective mos (1 63) /18 modified Magerl tried for all Sekhon, 2005 retrospective mos (1 50) /56 modified Anderson 948 screws 78 screws tech Pateder & Carbone, 2006 retrospective mos (24 48) 41 (18 85) 17/12 modified An tech fluoroscopy Wang & Levi, 2006 retrospective yrs minimum 41 (18 72) 8/10 Magerl tech fluoroscopy Wu et al., 2008 retrospective mos (4 35) 40.6 (18 82) 92/23 Chen tech occasionally fluoroscopy tried for all Katonis et al., 2011 retrospective mos (12 72) 68 (45 84) 120/ screws 366 screws Audat et al., 2011 retrospective mos (3 38) 46.8 (22 65) 33/17 modified Anderson or Sekhon techs fluoroscopy CPS Abumi et al., 2000 retrospective >2 yrs 51.8 (13 84) 106/74 fluoroscopy Neo et al., 2005 ND >6 mos /11 fluoroscopy Kast et al., 2006 prospective ND 60 (31 86) 17/9 navigation Takahashi et al., 2007 ND mos /10 navigation Yoshimoto et al., 2009 ND (6 113) 62.6 (21 87) 28/24 fluoroscopy Yukawa et al., 2009 ND >6 mos 44.1 (14 90) 125/19 fluoroscopy (oblique) Liu et al., 2009 retrospective (6 30) 51.2 (20 68) 9/16 fluoroscopy Miyamoto & Uno, 2009 ND ND 61.2 ± /10 fluoroscopy Kotil & Ozyuvaci, 2011 ND (30 37) 40.9 (34 65) 6/4 fluoroscopy Tofuku et al., 2012 retrospective (12 47) 66 (18 89) 28/4 fluoroscopy Lee et al., 2012 retrospective ND 58 (23 76) 36/14 fluoroscopy Nakashima et al., 2012 retrospective yrs (6 168 mos) /39 axis view, freehand, Iso- C3D navigation * ND = not described; tech = technique. Expressed as the mean ± SD. J Neurosurg: Spine / Volume 19 / November

5 H. Yoshihara, P. G. Passias, and T. J. Errico TABLE 3: Complications of the studies included in the analysis* Authors & Year Pts s (C3 7) Nerve Root Injury SCI Periop Complication Late Complication VA Injury Fx of Lateral Mass Facet Violation Malposition Reacquired; Postop Rev/ Removal Loosening Pullout Breakage Plate/Rod Breakage Loss of Reduction Pseudarthrosis Required Rev Op ASD Requiring Op LMS Heller et al., 1991 Fehlings et al., 1994 Graham et al., 1996 Wellman et al., screws, 4 pts 0 0 NR 1 screw 4 screws, 4 pts NR NR NR 8 screws, 5 pts screws, 0 0 NR NR 4 screws, 3 pts 7 screws 1 screw, 1 pt 2 screws 1 pt 2 pts 1/71 pts 4 pts 2 pts NR NR NR 3/42 pts NR NR 0 0 NR NR NR pt 1/35 pts 1 pt 1 pt Sekhon, NR 8 screws 0 NR 6 screws, Pateder & Carbone, 2006 Wang & Levi, 2006 Wu et al., 2008 Katonis et al., 2011 Audat et al., 2011 CPS Abumi et al., 2000 Neo et al., 2005 Kast et al., 2006 Takahashi et al., 2007 Yoshimoto et al., screws 1 pt 2 pts NR NR 1 pt NR NR NR NR NR NR NR 0 1 pt 0 1 pt NR 0 NR NR NR NR NR NR 0 NR NR NR NR NR NR 1 pt pt NR 0 NR screws, screws NR 3 screws, NR 0 0 NR 6 pts 9 pts NR 0 NR 4 screws 1 screw, 1 pt NR NR NR screws, 2 pts 0 1 screw, 1 pt NR NR 1 screw, 1 pt 0 0 NR NR 0 1 pt NR NR NR 0 0 NR NR NR NR NR 2 screws, 2 pts screws, 2 pts NR NR NR NR NR NR 0 NR NR 1 screw, 1 pt NR NR 0 NR NR NR NR 0 NR NR NR NR NR NR NR NR NR NR NR NR NR (continued) 618 J Neurosurg: Spine / Volume 19 / November 2013

6 -related complications in the subaxial cervical spine TABLE 3: Complications of the studies included in the analysis* (continued) Authors & Year Pts s (C3 7) Nerve Root Injury SCI Periop Complication Late Complication VA Injury Fx of Lateral Mass Facet Violation Malposition Reacquired; Postop Rev/ Removal Loosening Pullout Breakage Plate/Rod Breakage Loss of Reduction Pseudarthrosis Required Rev Op ASD Requiring Op CPS (continued) Yukawa et al., 2009 Liu et al., 2009 Miyamoto & Uno, 2009 Kotil & Ozyuvaci, 2011 Tofuku et al., 2012 Lee et al., 2012 Nakashima et al., screw, 1 pt 0 1 screw, 1 pt NR NR 0 NR 1 pt 1 pt NR 5 pts NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR 0 NR NR NR NR NR 0 NR NR NR NR NR NR NR NR NR 0 NR NR NR NR NR NR NR NR NR NR NR screws, 0 2 screws, 2 pts NR NR 3 screws, 3 pts 5 screws, 5 pts NR 3 screws, NR 2 pts 2 pts 1 pt 1 pt * Fx = fracture; NR = not reported; rev = revision. J Neurosurg: Spine / Volume 19 / November

7 H. Yoshihara, P. G. Passias, and T. J. Errico TABLE 4: Summary of complications in the studies included in the analysis and comparison between LMS and CPS* Type of Complication LMS CPS p Value periop root injury 0.19 (10/5130) 0.31 (8/2598) (10/737) 1.24 (8/643) 0.96 SCI 0 (0/687) 0 (0/569) VA injury 0 (0/5328) 0.15 (4/2668) (0/766) 0.61 (4/661) Fx of lateral mass 1.62 (27/1662) NA facet violation 0.62 (13/2085) NA malposition requiring postop rev/removal 0.38 (12/3144) 0.29 (5/1711) (11/417) 1.1 (5/455) 0.15 late screw loosening 1.17 (15/1287) 0.45 (5/1110) 0.09 unknown 1.73 (5/289) screw pullout 1.1 (8/722) 0.24 (1/418) 0.17 screw breakage unknown 1.76 (6/340) plate/rod breakage 0.28 (2/722) 0 (0/94) 1.00 loss of reduction 2.21 (10/452) 1.46 (7/478) 0.54 pseudarthrosis 2.67 (11/412) 0.87 (3/343) 0.10 required rev op 2.81 (17/605) 1.03 (1/97) 0.49 ASD requiring op 0.74 (4/539) 1.19 (1/84) 0.51 * Data are presented as percentages (event/no.). Complications shown in bold have percentages based on the number of screws. All other percentages are based on the number of patients. NA = not applicable; unknown = unknown, rate could not be calculated because the number of patients was not given in some articles; = not done. Calculated using the Fisher exact test; the other p values were determined using the chi-square test with the Yates continuous correction. in the C3 6 vertebrae to reduce the risk of neurovascular injury associated with CPS in the smaller C3 6 vertebrae. Normal values for pedicle widths have been established in the subaxial cervical spine. Pedicle widths of the C-3 vertebra average 4.9 mm in males and 4.5 mm in females, with the minimum reported width being 3 mm. 13 The C-4 averages 4.7 mm in males and 4.6 mm in females, with a minimum reported width of 3.1 mm. 13 The pedicle widths of C-5 and C-6 are slightly higher. The significance of this is reflected in the literature. Kast et al. 24 reported that 91% of the CPS in C-6 were correctly placed, compared with only 48% of screws in C-4; the most obvious risk factor for screw misplacement in their series was the level of surgery. In terms of clinical outcomes in cases in which CPS fixation was used, most series have focused on the incidence of pedicle perforation. The reported incidence of pedicle perforation in the previous studies ranged from 1.1% to 29.8%. 24,34 Some authors have described safer CPS placement when a computer-assisted navigation system is used. In 2003, Kotani et al. 26 reported that the screw misplacement rate was significantly lower in a computerassisted group (1.2%) than in a freehand group (6.7%). Similarly, Richter et al. 37 reported that the rate of pedicle perforations was 3% in the computer-assisted group and 8.6% in the conventional freehand group when using cannulated screws and separate stab incisions. Although the incidence of lateral pedicle perforation was relatively high in previous studies, the actual cases in which a VA injury was sustained were found to be rare. Reported cases of VA injuries were 1 of 180 cases reported by Abumi et al., 4 1 of 144 cases by Yukawa et al., 52 and 2 of 84 cases by Nakashima et al. 33 In our study, the overall rate of VA injury in procedures performed using CPS was 0.15% per screw and 0.61% per patient, which as predicted was statistically significantly higher than those using LMS. There was no case reported in which SCI occurred in a procedure using CPS, whereas the rate of nerve root injury was 0.31% per screw and 1.24% per patient, which was not found to be significantly different from that of LMS. Late Complications Instrumentation with LMS has been proven to be biomechanically stable 14,18,31,43,46 and effective in stabilizing the cervical spine with various kinds of pathological conditions. 29,36 Heller et al. 18 have examined the pullout strength of unicortical versus bicortical purchase, and the bicortical screw pullout force was 20% higher than unicortical screw pullout force, although both were deemed acceptable. Seybold et al. 43 reported in a cadaveric study of unicortical and bilateral LMS placement that, although the mean pullout force was higher for the bicortical screws, there was no statistically significant difference between unicortical and bicortical screws. Jones et al J Neurosurg: Spine / Volume 19 / November 2013

8 -related complications in the subaxial cervical spine showed a lower load-to-failure resistance for LMS than for CPS; however, there was no difference in pullout strength, bone density, screw length, or vertebral level. In 2002, Merola et al. 31 reported that aiming at the superior-lateral corner of the lateral mass itself offers the maximum amount of bone for screw purchase. Clinically, Pateder and Carbone 36 reported that LMS can be used effectively for different patterns of cervical trauma because they maintain excellent alignment and are associated with minimal complications. In our study, the rate of LMS loosening was 1.17% per screw and that of LMS pullout was 1.1% per patient. The rate of loss of reduction after using LMS was 2.21% per patient and the pseudarthrosis rate was 2.67%. Cervical pedicle screws offer extremely rigid fixation. The pedicles are the strongest structural elements of the cervical vertebrae, as is the case in the thoracic and lumbar spine. The results of a comparative biomechanical study of the cervical spine by Jones et al. 23 provided evidence that CPS had a statistically significantly greater load-to-failure resistance compared with LMS. Johnston et al. 22 also reported that CPS demonstrated a significantly lower rate of loosening at the bone-screw interface, and a higher strength after fatigue testing compared with LMS. Therefore, CPS may be appropriate for conditions that require greater mechanical strength, such as for the correction of kyphotic deformity. Surprisingly, however, Yukawa et al. 52 reported 5 cases with loss of reduction and Nakashima et al. 33 reported 2 cases among patients in whom CPS fixation was used. In our study, the rate of CPS loosening was 0.45% per screw and that of CPS pullout was 0.24% per patient. The rate of loss of reduction using CPS was 1.46% per patient and the pseudarthrosis rate was 0.87% per patient. Interestingly, although the use of LMS was associated with a higher rate of screw loosening, screw pullout, loss of reduction, pseudarthrosis, and revision surgery, this finding was not statistically significant. J Neurosurg: Spine / Volume 19 / November 2013 Summary Although our study represents the first series comparing the 2 techniques, there are clear limitations. Whereas CPS appear to be used more commonly outside of North America, in North America LMS are more commonly used, as indicated by the origin of studies included. We cannot comment on the underlying reasons for this, although this may theoretically influence the homogeneity of our comparison groups. To ensure the most accurate possible comparison between the 2 techniques in terms of complications, we limited our series to studies in which the amount of screws placed per level could be determined and also in which the complications could be clearly attributed to the technique. Although this narrowed the number of studies included, we believe that a systematic review with these more stringent inclusion criteria would more accurately reflect the incidence of screw-related complications. Recently, Coe presented a systematic review of LMS at the Cervical Spine Research Society s meeting (Coe JD, presentation to the Cervical Spine Research Society annual meeting, December 2012). Coe s study was a systematic literature review including all studies available in which LMS fixation was used, and found similar complication rates to those our study revealed, demonstrating that the inclusion criteria for our systematic review provides an accurate assessment of the overall literature, while still allowing us to report on complications per screw inserted. The findings of this review should also be interpreted with caution, because the available literature on CPS primarily reflects the work of experienced surgeons following completion of a steep learning curve. As Yoshimoto et al. 51 reported, the learning curve of CPS and the perforation rate in their study was 12% in the initial 19 cases and was decreased to 7% in the next 18 cases, and to 1.1% in the last 15 cases; thus less experienced surgeons should use such techniques with extreme caution and perhaps with the assistance of more experienced surgeons. Injury to bilateral VAs might cause serious problems such as cerebral infarction or death. When occlusion of the VA on one side is seen preoperatively, the insertion of CPS in the remaining side should be avoided. 52 In such cases, CPS fixation on the occluded side and LMS fixation on the remaining side are recommended. 52 In summary, the results of our study reveal that according to the existing literature, the reported screw-related complication rates of both LMS and CPS are similarly very low, although statistically significantly higher rates of VA injuries were seen in cases treated using CPS. The best tool for providing safe and accurate screw placement is the spine surgeon s intimate knowledge of cervical anatomy. 52 In addition, preoperative evaluation is critical regardless of technique. Thus, based on the available literature, it appears that as long as the surgeon has sufficient expertise and adequate preoperative evaluation has been performed, spine surgeons may select either procedure. Higher-quality studies are clearly needed with direct comparison between the 2 techniques to better answer this question. Conclusions The following conclusions can be drawn. 1) The complication rate of LMS and CPS in the subaxial cervical spine was reviewed based on the previous literature. These rates can be used as a database for comparison in future studies. 2) Vertebral artery injuries were statistically significantly higher with the use of CPS relative to LMS in the subaxial cervical spine. Although the use of LMS was associated with a higher rate of screw loosening, screw pullout, loss of reduction, pseudarthrosis, and revision surgery, this finding was not statistically significant. 3) Surgeons need to have an intimate knowledge of cervical anatomy and adequate preoperative evaluation for each patient, with selection of either procedure in appropriate cases a reasonable option. Acknowledgment The authors thank Dr. Issei Kurahashi for assistance with the statistical analysis. Disclosure Dr. Errico has financial conflicts of interest as follows: research 621

9 H. Yoshihara, P. G. Passias, and T. J. Errico support from Paradigm Spine ($104,000) and fellowship support from the Orthopaedic Research and Education Foundation ($75,000) and the AO ($75,000). The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: Yoshihara. Acquisition of data: Yoshihara. Analysis and interpretation of data: Yoshihara. Drafting the article: Yoshihara. Critically revising the article: Yoshihara, Passias. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Yoshihara. Statistical analysis: Yoshihara. Administrative/technical/material support: Yoshihara. Study supervision: Yoshihara, Errico. References 1. Abumi K, Itoh H, Taneichi H, Kaneda K: Transpedicular screw fixation for traumatic lesions of the middle and lower cervical spine: description of the techniques and preliminary report. J Spinal Disord 7:19 28, Abumi K, Kaneda K: Pedicle screw fixation for nontraumatic lesions of the cervical spine. Spine (Phila Pa 1976) 22: , Abumi K, Kaneda K, Shono Y, Fujiya M: One-stage posterior decompression and reconstruction of the cervical spine by using pedicle screw fixation systems. J Neurosurg 90 (1 Suppl):19 26, Abumi K, Shono Y, Ito M, Taneichi H, Kotani Y, Kaneda K: Complications of pedicle screw fixation in reconstructive surgery of the cervical spine. Spine (Phila Pa 1976) 25: , Abumi K, Shono Y, Taneichi H, Ito M, Kaneda K: Correction of cervical kyphosis using pedicle screw fixation systems. Spine (Phila Pa 1976) 24: , Aldrich EF, Weber PB, Crow WN: Halifax interlaminar clamp for posterior cervical fusion: a long-term follow-up review. J Neurosurg 78: , An HS, Gordin R, Renner K: Anatomic considerations for platescrew fixation of the cervical spine. Spine (Phila Pa 1976) 16 (10 Suppl):S548 S551, Anderson PA, Henley MB, Grady MS, Montesano PX, Winn HR: Posterior cervical arthrodesis with AO reconstruction plates and bone graft. Spine (Phila Pa 1976) 16 (3 Suppl): S72 S79, Audat ZA, Barbarawi MM, Obeidat MM: Posterior cervical decompressive laminectomy and lateral mass screw fixation. Neurosciences (Riyadh) 16: , Barrey C, Mertens P, Jund J, Cotton F, Perrin G: Quantitative anatomic evaluation of cervical lateral mass fixation with a comparison of the Roy-Camille and the Magerl screw techniques. Spine (Phila Pa 1976) 30:E140 E147, Benzel EC, Kesterson L: Posterior cervical interspinous compression wiring and fusion for mid to low cervical spinal injuries. J Neurosurg 70: , Cho KH, Shin YS, Yoon SH, Kim SH, Ahn YH, Cho KG: Poor surgical technique in cervical plating leading to vertebral artery injury and brain stem infarction case report. Surg Neurol 64: , Ebraheim NA, Xu R, Knight T, Yeasting RA: Morphometric evaluation of lower cervical pedicle and its projection. Spine (Phila Pa 1976) 22:1 6, Errico T, Uhl R, Cooper P, Casar R, McHenry T: Pullout strength comparison of two methods of orienting screw insertion in the lateral masses of the bovine cervical spine. J Spinal Disord 5: , Fehlings MG, Cooper PR, Errico TJ: Posterior plates in the management of cervical instability: long-term results in 44 patients. J Neurosurg 81: , Graham AW, Swank ML, Kinard RE, Lowery GL, Dials BE: Posterior cervical arthrodesis and stabilization with a lateral mass plate. Clinical and computed tomographic evaluation of lateral mass screw placement and associated complications. Spine (Phila Pa 1976) 21: , Heller JG, Carlson GD, Abitbol JJ, Garfin SR: Anatomic comparison of the Roy-Camille and Magerl techniques for screw placement in the lower cervical spine. Spine 16 (Phila Pa 1976) (10 Suppl):S552 S557, Heller JG, Estes BT, Zaouali M, Diop A: Biomechanical study of screws in the lateral masses: variables affecting pull-out resistance. J Bone Joint Surg Am 78: , Heller JG, Silcox DH III, Sutterlin CE III: Complications of posterior cervical plating. Spine (Phila Pa 1976) 20: , Holness RO, Huestis WS, Howes WJ, Langille RA: Posterior stabilization with an interlaminar clamp in cervical injuries: technical note and review of the long term experience with the method. Neurosurgery 14: , Jeanneret B, Magerl F, Ward EH, Ward JC: Posterior stabilization of the cervical spine with hook plates. Spine 16 (Phila Pa 1976) (3 Suppl):S56 S63, Johnston TL, Karaikovic EE, Lautenschlager EP, Marcu D: Cervical pedicle screws vs. lateral mass screws: uniplanar fatigue analysis and residual pullout strengths. Spine J 6: , Jones EL, Heller JG, Silcox DH, Hutton WC: Cervical pedicle screws versus lateral mass screws. Anatomic feasibility and biomechanical comparison. Spine (Phila Pa 1976) 22: , Kast E, Mohr K, Richter HP, Börm W: Complications of transpedicular screw fixation in the cervical spine. Eur Spine J 15: , Katonis P, Papadakis SA, Galanakos S, Paskou D, Bano A, Sapkas G, et al: Lateral mass screw complications: analysis of 1662 screws. J Spinal Disord Tech 24: , Kotani Y, Abumi K, Ito M, Minami A: Improved accuracy of computer-assisted cervical pedicle screw insertion. J Neurosurg 99 (3 Suppl): , Kotil K, Ozyuvaci E: Multilevel decompressive laminectomy and transpedicular instrumented fusion for cervical spondylotic radiculopathy and myelopathy: a minimum follow-up of 3 years. J Craniovertebr Junction Spine 2:27 31, Lee SH, Kim KT, Abumi K, Suk KS, Lee JH, Park KJ: Cervical pedicle screw placement using the key slot technique : the feasibility and learning curve. J Spinal Disord Tech 25: , Liu H, Ploumis A, Schwender JD, Garvey TA: Posterior cervical lateral mass screw fixation and fusion to treat pseudarthrosis of anterior cervical fusion. J Spinal Disord Tech 25: , Liu Y, Hu JH, Yu KY: Pedicle screw fixation for cervical spine instability: clinical efficacy and safety analysis. Chin Med J (Engl) 122: , Merola AA, Castro BA, Alongi PR, Mathur S, Brkaric M, Vigna F, et al: Anatomic consideration for standard and modified techniques of cervical lateral mass screw placement. Spine J 2: , Miyamoto H, Uno K: Cervical pedicle screw insertion using a computed tomography cutout technique. Technical note. J Neurosurg Spine 11: , Nakashima H, Yukawa Y, Imagama S, Kanemura T, Kamiya M, Yanase M, et al: Complications of cervical pedicle screw fixation for nontraumatic lesions: a multicenter study of 84 patients. Clinical article. J Neurosurg Spine 16: , Neo M, Sakamoto T, Fujibayashi S, Nakamura T: The clinical risk of vertebral artery injury from cervical pedicle screws inserted in degenerative vertebrae. Spine (Phila Pa 1976) 30: , J Neurosurg: Spine / Volume 19 / November 2013

10 -related complications in the subaxial cervical spine 35. Pait TG, McAllister PV, Kaufman HH: Quadrant anatomy of the articular pillars (lateral cervical mass) of the cervical spine. J Neurosurg 82: , Pateder DB, Carbone JJ: Lateral mass screw fixation for cervical spine trauma: associated complications and efficacy in maintaining alignment. Spine J 6:40 43, Richter M, Cakir B, Schmidt R: Cervical pedicle screws: conventional versus computer-assisted placement of cannulated screws. Spine (Phila Pa 1976) 30: , Rogers WA: Treatment of fracture-dislocation of the cervical spine. J Bone Joint Surg Am 24: , Roy-Camille R: Rationale and techniques of internal fixation in trauma of the cervical spine, in Errico T, Bauer RD, Waugh TR (eds): Spinal Trauma, ed 2. Philadelphia: JB Lippincott, 1991, pp Roy-Camille R, Gaillant G, Bertreaux D: Early management of spinal injuries, In McKibben B (ed): Recent Advances in Orthopedics. Edinburgh: Churchill-Livingstone, 1979, pp Roy-Camille R, Saillant G, Laville C, Benazet JP: Treatment of lower cervical spinal injuries C3 to C7. Spine (Phila Pa 1976) 17 (10 Suppl):S442 S446, Sekhon LH: Posterior cervical lateral mass screw fixation: analysis of 1026 consecutive screws in 143 patients. J Spinal Disord Tech 18: , Seybold EA, Baker JA, Criscitiello AA, Ordway NR, Park CK, Connolly PJ: Characteristics of unicortical and bicortical lateral mass screws in the cervical spine. Spine (Phila Pa 1976) 24: , Takahashi J, Shono Y, Nakamura I, Hirabayashi H, Kamimura M, Ebara S, et al: Computer-assisted screw insertion for cervical disorders in rheumatoid arthritis. Eur Spine J 16: , Tofuku K, Koga H, Komiya S: Cervical pedicle screw insertion using a gutter entry point at the transitional area between the lateral mass and lamina. Eur Spine J 21: , Ulrich C, Arand M, Nothwang J: Internal fixation on the lower cervical spine biomechanics and clinical practice of procedures and implants. Eur Spine J 10:88 100, Wang MY, Levi AD: Minimally invasive lateral mass screw fixation in the cervical spine: initial clinical experience with long-term follow-up. Neurosurgery 58: , Wellman BJ, Follett KA, Traynelis VC: Complications of posterior articular mass plate fixation of the subaxial cervical spine in 43 consecutive patients. Spine (Phila Pa 1976) 23: , Wu JC, Huang WC, Chen YC, Shih YH, Cheng H: Stabilization of subaxial cervical spines by lateral mass screw fixation with modified Magerl s technique. Surg Neurol 70 Suppl 1: S1:25 S1:33, Xu R, Haman SP, Ebraheim NA, Yeasting RA: The anatomic relation of lateral mass screws to the spinal nerves. A comparison of the Magerl, Anderson, and An techniques. Spine (Phila Pa 1976) 24: , Yoshimoto H, Sato S, Hyakumachi T, Yanagibashi Y, Kanno T, Masuda T: Clinical accuracy of cervical pedicle screw insertion using lateral fluoroscopy: a radiographic analysis of the learning curve. Eur Spine J 18: , Yukawa Y, Kato F, Ito K, Horie Y, Hida T, Nakashima H, et al: Placement and complications of cervical pedicle screws in 144 cervical trauma patients using pedicle axis view techniques by fluoroscope. Eur Spine J 18: , Yukawa Y, Kato F, Yoshihara H, Yanase M, Ito K: Cervical pedicle screw fixation in 100 cases of unstable cervical injuries: pedicle axis views obtained using fluoroscopy. J Neurosurg Spine 5: , 2006 Manuscript submitted February 17, Accepted August 5, Please include this information when citing this paper: published online September 13, 2013; DOI: / SPINE Address correspondence to: Hiroyuki Yoshihara, M.D., Ph.D., Department of Orthopaedic Surgery, NYU Hospital for Joint Diseases, 301 East 17th St., New York, NY hiroyoshi hara55@yahoo.co.jp. J Neurosurg: Spine / Volume 19 / November