Review of Cortical Bone Trajectory: Evidence of a New Technique.

This article summarizes recent evidence on the cortical bone trajectory (CBT) obtained from published anatomical, biomechanical, and clinical studies. CBT was proposed by Santoni in 2009 as a new trajectory that can improve the fixation of pedicle screws in response to screw loosening in osteoporotic patients. Recently, research interest has been growing with increasing numbers of published series and frequent reports of new applications. We performed an online database search using the terms "cortical bone trajectory," "pedicle screw," "CBT spine," "CBT fixation," "MISS CBT," and "traditional trajectory." The search included the PubMed, Ovid MEDLINE, Cochrane, and Google Scholar databases, resulting in an analysis of 42 articles in total. These covered three aspects of CBT research: anatomical studies, biomechanical parameters, and clinical cases or series. Compared to the traditional trajectory, CBT improves pullout strength, provides greater stiffness in cephalocaudal and mediolateral loading, and shows superior resistance to flexion/extension; however, it is inferior in lateral bending and axial rotation. CBT seems to provide better immediate implant stability. In clinical studies, CBT has shown better perioperative results for blood loss, length of stay in hospital, and surgery time; similar or better clinical postoperative scores; and similar comorbidity, without any major fixation system complications due to instrumentation failure or screw misplacement. In addition, advantages such as less lateral exposure allow it to be used as a minimally invasive technique. However, most of the clinical studies were retrospective case series or case-control studies; prospective evidence on this technique is scarce, making a definitive comparison with the traditional trajectory difficult. Nevertheless, we can conclude that CBT is a safe technique that offers good clinical results with similar biomechanical and perioperative parameters to those of the traditional trajectory. In addition, new applications can improve its results and make it useful for additional pathologies.

Introduction

Pedicle screw fixation is a standard procedure in spine surgery and the main technique for maintaining spinal stability and biomechanical features in spinal disease. Pedicle screws have been successfully used with good results for multiple pathologies, such as spinal deformities, degenerative diseases, or fractures. However, some complications have been associated with this procedure, with system failure being one of the most important. Incidence of screw loosening has been estimated to range from 1% to 15% in non-osteoporotic patients and exceed 60% in patients with osteoporotic bones [1]; however, the actual incidence of this problem has not yet been confirmed [2]. Obtaining a more solid internal fixation is therefore an important issue for spinal surgeons, and various approaches, including changing the screw trajectory, have been proposed with the aim of improving longevity and avoiding such complications [34].

In 2009, Santoni et al. [3] proposed a new insertional pathway in place of the traditional trajectory (TT), referred to as the cortical bone trajectory (CBT). The traditional insertional pathway runs through the pedicle axis with a lateral-to-medial trajectory starting at the junction between the transverse process and the lateral wall of the facet and ending at the vertebral body. In contrast, CBT involves a medial-to-lateral direction and a caudocephalad path with the objective of maximizing thread contact with higher-density bone. The aims of this track are to improve the adhesion of the screws in osteoporotic vertebrae and to prevent instrumentation failure [3].

This paper summarizes the biomechanical and clinical studies of CBT published to date and evaluates its possible advantages for daily clinical practice.

Materials and Methods

A literature search was performed in the PubMed, Ovid Medline, Cochrane, and Google Scholar databases using the search terms “cortical bone trajectory,” “CBT spine,” “CBT fixation,” “MISS CBT,” “traditional trajectory,” and “pedicle screw.” The search included papers published up to May 2016. All retrieved abstracts were analyzed to determine whether they presented relevant information about anatomical, biomechanical, or radiological results in either clinical or cadaveric studies. For those that did, complications, clinical results, and technical and surgical features were recovered from the associated papers for subsequent review. In addition, the reference lists of the identified publications were checked to determine further relevant articles that should be incorporated in the analysis.

Results

Finally, 42 articles were included in the analysis. These covered three aspects of CBT: anatomical studies, biomechanical parameters, and clinical cases or series. Six of the articles considered anatomical landmarks of CBT with regard to previous citations or possible complications in CBT procedures. Biomechanical parameters were discussed in 13 articles, which considered physical parameters that could show how CBT resulted in improved biomechanical properties compared to TT. Finally, 20 articles and two communications to meetings were analyzed for clinical series and cases. These 42 articles comprised all eligible reports published up to the date the review was finished. No previous review studies were included in the present narrative review.

1. Biomechanical and anatomical studies

In 2007, Sterba et al. [5] published a cadaveric study showing that straight screw insertion provided a more stable pedicle screw construct than the pathway parallel to the sagittal plane, because the angled trajectory resulted in lower average total fatigue damage. Previously, Roy-Camille et al. [6] had proposed that a vertical trajectory through the pedicle would increase thread contact with the cortical bone at the end point. In this context, Santoni et al. [3] proposed CBT as a way to avoid screw system failure in osteoporotic patients and demonstrated that this trajectory increased resistance in a uniaxial pullout test when compared to traditional screws. Some previous anatomic studies about the limitations of pedicle screw insertion should be considered. Li et al. [7] reviewed 41 computed tomography scans that focused on the height, width, and isthmus inclination of the pedicle, concluding that width was the most important factor for screw placement. They also found that the cortical thickness of the superior and medial walls was generally greater than 2 mm. This could explain why pedicles are more likely to break laterally. This demonstrated that the CBT technique could result in enhanced screw purchase and greater interface strength independent of the trabecular bone mineral density (BMD), because the screws have four points of fixation: between the dorsal cortex and the site of insertion, on the posteromedial and anterolateral pedicle walls, and on the marginal region of the vertebral body wall (Fig. 1) [48]. Thus, although CBT screws are shorter in length and smaller in diameter than those used with TT, they are in contact with bone surface of higher density. In fact, Mai et al. [9] conducted an observational study that measured BMD (in Hounsfield units) at the theoretical end point in CBT and TT, demonstrating greater density for CBT, especially in the osteoporotic cohort (Table 1).

Fig. 1

Comparison of the traditional trajectory (A, B) and cortical bone trajectory (C, D). (A, B) Axial and sagittal views of the traditional trajectory following the pedicle axis in a lateral-to-medial trajectory parallel to the superior and inferior endplates. (C, D) Axial and sagittal views of the cortical bone trajectory with a medial-to-lateral disposition and a 25°–30° cranial direction along the inferior border of the pedicle.

Table 1

Summary of anatomical studies for CBT fixation

CBT, cortical bone trajectory; LE, level of evidence; CT, computed tomography; TT, traditional trajectory; HU, Hounsfield unit; 3D, three dimensional.

Since the introduction of CBT, studies have been conducted to determine its biomechanical characteristics. Perez-Orribo et al. [14] published a cadaveric study using screws that were oriented more sagittally and cranially; they did not observe differences in stability compared to traditional pedicle screws, although this orientation gave worse results under axial rotation. Matsukawa et al. [15] showed that CBT provided greater pullout strength, stronger stiffness during cephalocaudal and mediolateral loading, and superior resistance to flexion/extension compared with TT, but that it was inferior with regard to lateral bending and axial rotation. However, different results were obtained when testing spondylolytic vertebrae; CBT showed worse results for all these parameters, making it unsuitable for fixation in these cases [16]. However, as discussed later, CBT has been applied to spondylolytic cases with good outcomes in other series and case-control studies. Range of motion has been compared between CBT and TT in animal [17] and human [1418] cadaveric studies, finding no differences for flexion/extension, lateral bending, or axial rotation. Calvert et al. [19] used CBT with rescue screws after instrumentation failure; this provided adequate stiffness in flexion/extension and axial rotation. Insertional torque has been previously correlated with pedicle screw pullout strength [20] and Matsukawa et al. [21] investigated how it changed during implant positioning, as an indirect measure of implant stability, founding that it is 1.7 times higher with CBT than with TT. These authors determined, in a multiple regression analysis, that BMD of the femoral neck, screw length within the lamina, and cephalad angle were significant independent factors affecting the torque [2122]. However, in contradiction to previous articles, Akpolat et al. [23] published in 2016 the results of a cadaveric study which concluded that implantation via TT required more cycles for screw loosening and showed better resistance to pullout. Finally, Sansur et al. [24] compared CBT and TT in destabilized cadaveric lumbar spines with osteoporosis, showing that CBT gave better results in the lower spine because of an increase in cancellous bone limit fixation with TT (Table 2) [3141516171819212223242526].

Table 2

Summary of biomechanical studies for CBT fixation

CBT, cortical bone trajectory; LE, level of evidence; qCT, quantitative computer tomography; TT, traditional trajectory; CT, computed tomography; ROM, range of motion.

2. Clinical trials

Many biomechanical studies have been published since Santoni et al. [3] first proposed CBT, but clinical evidence is still lacking, with no systematic reviews that can provide clear indications for its use. In 2004, Steel et al. [27] were the first to propose mediolateral fixation at a single thoracic level. They described 18 patients with thoracolumbar fractures who underwent operations without major complications. All achieved stable fusion after 6 weeks, without deformity, system failure, or neurological deficit, except for one patient with non-union after 12 months. The authors concluded that this trajectory was safe and effective for fixation and stabilization, and that it was less invasive than anterior or lateral approaches. In 2013, Ueno et al. [28] published a case report of a patient with degenerative scoliosis and osteoporosis who was operated on with a double-trajectory technique (CBT combined with TT at each level), with good morphometric results, no complications, and improvement in daily-life activities 14 months after surgery. Biomechanical studies demonstrated that this technique resulted in greater strength than the TT or CBT constructs for flexion, extension, lateral bending, and axial rotation [26]. Gonchar et al. [29] presented two series at a meeting of the Society for Minimally Invasive Spine Surgery (data not published), one of which compared 100 CBT versus 63 TT patients with pathologies such as degenerative diseases, osteoporosis, trauma, or deformities. CBT showed similar clinical results to TT for visual analogue scale (VAS), Oswestry disability index (ODI), Japanese Orthopedic Association Score (JOA), and surgical time, but superior results for blood loss, screw loosening, and pseudarthrosis. The authors reported two cases of screw breakage in CBT that were not repeated with the use of screws of a larger diameter (5.5 mm instead of 4.75 mm). Furthermore, in a prospective study of 60 patients with spondylolisthesis (30 CBT and 30 TT with minimally invasive spine surgery), the CBT patients showed a significantly lower rate of screw loosening and loss of correction, and CBT was shown to be less invasive than TT, as measured by creatine phosphokinase (CPK) values [29].

Iwatsuki et al. [8] proposed an isthmus-guided approach for CBT (IGCBT) to avoid the complications associated with this pathway and compared it with CBT. No complications occurred and only one screw was misplaced with the IGCBT technique compared to four misplaced screws with the traditional CBT. Ohkawa et al. [30] subsequently used this procedure and compared CPK levels between the original CBT and IGCBT, showing improved results with the latter. They reported 4% screw misplacement in 12 patients (22%), without major complications.

Mizuno et al. [31] investigated patients with spondylolisthesis and CBT fixation with posterior lumbar interbody fusion (PLIF) or transforaminal lumbar interbody fusion. This revealed one intraoperative complication due to a cortical bone fracture, but this was not associated with neurological deficit. At 20 months' follow-up, there were no signs of screw loosening, although four screws (8.3%) had perforated the pedicle wall without clinically adverse effects. Okudaira et al. compared PLIF versus mid-line lumbar fusion with CBT and showed significant benefits with CBT in surgical time and blood loss, with similar clinical outcomes. There were no complications associated with CBT, whereas in the PLIF group, there was one wound infection with neural injury. The authors concluded that CBT was less invasive, requiring less exposure, and that recovery was faster (data not published).

Rodriguez et al. [32] proposed a double-fixation system in patients with adjacent-segment lumbar disease using CBT without removing the previous screws. No complications and good clinical outcomes were recorded at 6 months' follow-up, with clinical improvement and radiographic fusion. Like Ueno et al. [28] and Rodriguez et al. [32], Takata et al. [33] also presented a hybrid technique in which TT was used for caudal segments and CBT for cranial vertebrae. They showed that this hybrid technique was less invasive than TT because it reduced muscular retraction in the upper level.

Pacione et al. [34] published a case report of an 83-year-old woman with osteoporotic L4 compression treated with kyphoplasty, L4 decompression, and CBT fixation on L3 and L5 that needed to be rescued 3 months later with kyphoplasty after L3 compression with the standard technique, because CBT enabled the usual pedicle pathway without complications. Glennie et al. [35] were the first to report complications with CBT after a 1-year follow-up. Eight patients underwent operations with CBT; at 1-year follow-up, two had undergone revision surgery, five showed screw loosening, and four failed to maintain reduction on radiographic control. The authors concluded that CBT should be evaluated at medium- or long-term clinical follow-up to validate outcome measures and check for complications. Other series have reported similar complication incidences: Patel et al. [36] reported complications in five out of 22 patients (13.6%) and Cheng et al. [37] reported two pars and pedicle fractures and two instances of early screw loosening in 22 patients. A subsequent cadaveric study by these authors revealed the proportions of pars and pedicle fractures to be 2.7% and 16.2%, deviations which resulted in gross loosening. These results were similar to those of the study by Ninomiya et al. [38], which evaluated screw loosening with and without the presence of a clear zone around the screws in patients with CBT fixation. Clear zones were observed around six screws (5.5%) in five patients (26.3%), showing better results than those of previous studies. This group also compared radiological results of slippage correction and lordosis change 1 year after surgery, with significant improvement compared with preoperative values, unlike with TT [39]. Snyder et al. [40] reported a study with 79 patients that had CBT screws with or without other fixation techniques; they observed only 10 complications in seven patients: two hardware failures, two cases of pseudoarthrosis, two pulmonary embolisms, two deep vein thromboses, a wound infection, and an epidural hematoma. Surgical intervention was required for both pseudoarthroses, the wound infection, and the epidural hematoma. Mori et al. [41] reported one-year follow-up data for CBT procedures, showing good morphometric and clinical results after surgery without major complications; however, the follow-up revealed apparent non-union in 9.4% (three) of the patients, which was a higher rate than that for TT [4243]. Dabbous et al. [44] also found good results with CBT procedures for intraoperative time and blood loss, good recovery in terms of ODI score and walking distance without medication, and a reduction in analgesic medication, with 44% of patients ceasing medication. They found no major complications; minor complications were a durotomy, a pedicle fracture, and a cage migration, with only the former related to the technique due to the space limitations imposed by minimally invasive surgery (MIS). Hung et al. [45] measured fat infiltration after CBT or TT fixation by postoperative magnetic resonance imaging; they found no significant differences in clinical outcome or perioperative parameters, although minor postoperative blood loss, operative time, and hospital stay were reduced with CBT. Fat infiltration was higher with TT (Table 3) [2728293031323334353637383940414445464748].

Table 3

Summary of clinical series published with CBT fixation

CBT, cortical bone trajectory; LE, level of evidence; TT, traditional trajectory; MISS, minimally invasive spine surgeries; PLIF, posterior lumbar interbody fusion; TLIF, transforaminal lumbar interbody fusion; IGCBT, isthmus-guided cortical bone trajectory; CPK, creatine phosphokinase; MIDLF, midline lumbar fusion; ODI, Oswestry disability index; SF-36, 36-item short form health survey; ALIF, anterior lumbar interbody fusion; LLIF, lateral lumbar interbody fusion; PE, pulmonary embolism; DVT deep vein thrombosis; JOA, Japanese Orthopedic Association Score; pCBT, percutaneous cortical bone trajectory; pTT, percutaneous traditional trajectory; VAS, visual analogue scale.

New articles about CBT have been published in the last few months. Orita et al. [46] described percutaneous CBT with better results measured with a VAS at 6 months compared to percutaneous TT, with a shorter time of fluoroscopy and skin incision. Ashayeri et al. [47] used a hybrid technique for congenital multilevel spinal non-segmentation, concluding that CBT could improve pedicle screw fixation when bone quality was suboptimal or pedicle anatomy was distorted. Sakaura et al. [48] compared CBT with PLIF versus TT with PLIF and concluded that the JOA were significantly better in the CBT group and that there were fewer cases of symptomatic adjacent-segment disease (three [3.2%] with CBT vs. nine [11%] with TT, p<0.05), with no differences in solid spinal fusion.

Discussion

In 1986, Roy-Camille et al. [6] proposed a vertical trajectory that did not follow the pedicle axis and contacted a greater proportion of cortical bone at its end point. However, until Sterba et al. [5] reported on its improved biomechanical properties, the vertical straight pathway was not used as a conventional technique. Anatomical studies have demonstrated that CBT enhanced screw purchase and interface strength through increasing thread contact with cortical bone [4789]. Although CBT uses shorter and smaller-diameter screws compared to TT, it has shown better results in biomechanical studies for pullout strength, insertional torque, greater stiffness during cephalocaudal and mediolateral loading, and a superior resistance to flexion/extension; conversely, it is inferior with regard to lateral bending and axial rotation [31415182127]. The benefits were especially important in patients with osteoporosis or in lower lumbar vertebrae where cancellous bone is more important [324]. Biomechanical studies have also demonstrated greater immediate stability with CBT screws due to the higher insertional torque [21]. However, other studies have shown worse results with CBT in terms of resistance to cycling loading [23] and decreased pullout strength, flexion, extension, lateral bending, and axial rotation compared to the TT construct in spondylolytic vertebrae [16]. However, clinical evidence does not support this negative conclusion from Matsukawa et al. [16], and instead shows good outcomes without major complications [829303132333538394144454648]. Biomechanical studies provide poor evidence in support of clinical outcomes, providing only indirect information which suggests that CBT is associated with less screw loosening and improved fixation after surgery, with the vast majority of these studies seeming to agree with its biomechanical advantages.

Prior to Santoni et al.'s proposal for CBT [3], clinical evidence for mediolateral fixation was described by Steel et al. [27] for thoracolumbar fractures, with good results. Since 2009, 22 clinical series or case reports have been reported and interest in CBT is growing. The majority of these showed favorable clinical results, with good clinical outcomes and better results for blood loss, surgical time, time to hospital discharge, and morphometric corrections [2829303132333440414243]. Clinical outcomes for CBT were similar to those for TT, with better perioperative parameters, which can lead to fewer complications [40]. CBT is also less invasive, as measured by serum or clinical parameters, such as incision length [3046]. However, although these data show a theoretical benefit of CBT fixation, over the last few years, some authors have published clinical series reporting high complication rates. We differentiated complications related to screw loosening or misplacement from those related to clinical results. Results related to screw loosening are not clear in clinical series, because clear zone, as a sign of screw loosening, has been observed to disappear in two-thirds of the cases after a 3-year follow-up. In this context, the results of Ninomiya et al. [38], who did not observe significant screw loosening at 6 months' follow-up, contrasted with those from the series reported by Glennie et al. [35], with 62% screw loosening; however, this high level was not reflected in other studies, where the screw loosening rate was estimated at around 0%–16.2% [36374041]. Construct failure appears to be around 2.5% [40], and pedicle fracture, which is a frequent complication with CBT, has an incidence of around 4% (0%–8.3%) [3036374144]. Screw misplacement is more frequent with CBT than with TT, as shown in systematic reviews of neuronavigation systems, in which TT showed misplacement at around 10% in the worst cases [49], whereas pedicle fracture could reach 6.6% with TT [50]. The incidence of misplacement with CBT has been reported to be around 4%–12.5% [830], which does not represent an important difference from TT. Clinical complications were not frequent compared to those with other techniques described (0%–8.1%) [27283336404144]. Note that, in most of the studies, the follow-up period was short (median, 12 months); this could result in the underestimation of posterior complications, especially those due to screw loosening and non-union, as well as postoperative pars and pedicle fractures. In contrast, clinical outcomes with CBT were at least as good as those with TT, and new studies have reported better results for CBT [29444648].

CBT has been shown to have applications for various pathologies, such as the double-fixation system [28], hybrid techniques for avoiding major exposure [33] or for congenital spinal deformation [47], treatment of adjacent segment disease [32], in combination with other techniques such as kyphoplasty [34], or the recently proposed percutaneous CBT [46]. However, these have been reported only as case studies and the indication has not been clearly elucidated. Sakaura et al. [48] reported a minor rate of adjacent segment disease in their series with CBT, but in the other cases, the main indications were to obtain greater rigidity of fixation using CBT in combination with TT [2832], but also CBT can be used to avoid TT in case it was previously used for kyphoplasty or for removing pedicle screws previously placed in that position [3234], or to achieve adequate vertebral fixation with minor exposure [3346]. It should also be emphasized that the double-fixation system demonstrated better strength in flexion, extension, lateral bending, and axial rotation compared to CBT or TT alone [26].

However, most of the studies published to date have been retrospective case series or case-control studies. Prospective evidence on the CBT technique is scarce, making it difficult to reach a definitive conclusion about its superiority over TT.

In summary, it is our opinion that the majority of the literature confirms that CBT is a safe technique that can improve results in some situations compared to traditional techniques. The vast majority of authors found that, because CBT needs less lateral exposure as there is no need to reach the transverse process, it could improve perioperative parameters such as blood loss, surgical time, or hospital stay [293132334044], and result in lower levels of CPK and postoperative fat infiltration [293145]; it could therefore benefit patients where MIS techniques are more suitable, such as for an obese population [4]. However, these potential benefits should be attributed not only to the lesser exposure but also to an appropriate selection of patients [44]. We also think that CBT should be considered with caution as a new technique. Although Dabbous et al. [44] reported that results were favorable even during the learning curve, a period of previous training is required before this technique is offered to patients, particularly because the incidence of pars of pedicle fracture and misplacement is not negligible. CBT presents good biomechanical parameters compared to TT, opening the field to new applications for this technique; however, the technique was initially devised for osteoporotic patients, but fewer clinical studies have been conducted with this group of patients [213436]. Furthermore, although in this article we have tried to summarize all the information about CBT published to date, we emphasize that the majority of the studies offer weak evidence because most involved small cohorts, case series, clinical cases, or comparisons with historical control groups. Thus, further investigation of CBT must involve randomized controlled trials or homogeneous systematic studies rather than low-evidence studies. Finally, most of the previous studies compared CBT with other techniques, and most of the cases that have compared it with TT also presented the PLIF technique. The study by Sakaura et al. [48] is one of the latest to show good results. However, to date, few studies have compared CBT with TT alone.

Conclusions

CBT is increasingly used, and is a new and interesting minimal invasive technique that is demonstrating good results with acceptable morbidity. Various new applications have been proposed. However, more clinical studies are required to clarify several aspects of this technique.