**3. Results**

#### *3.1. Demographic Data and Clinical Outcomes*

Patients (41 men, 12 women; follow-up rate, 100%) completed at least 2 years of followups (Table 2). Of these, 30 patients were categorized into the ACCF group and 23 into the hybrid group. Mean preoperative/postoperative JOA scores were 11.9/15.0 points and 11.1/14.6 points, respectively. The average number of fused segments was 3.3 levels in the ACCF group and 3.5 levels in the hybrid group. Mean estimated blood loss and operative time were 437 mL and 6.5 h in the ACCF group and 197 mL and 6.0 h in the hybrid group, respectively. Duration of intensive care unit (ICU) stay and timing of extubation were respectively 2.8 days and 0.5 days in the ACCF group and 3.3 days and 0.9 days in the hybrid group. Duration of hospitalization was 24.2 days and 29.3 days, respectively. There were no significant differences in demographic and clinical characteristics between the two groups.

The incidence of perioperative complications was similar between the groups (Table 3). In the ACCF group, there was persistent dysphagia in four cases categorized as moderate by their Bazaz score (13.3%), aspiration pneumonitis in three (10%), delirium in two (6.7%), segmental paralysis in two (6.7%), and deep vein thrombosis in one (3.3%). In one case, dysphagia did not resolve until 2 months after the operation. However, other complications were resolved during the hospital stay. In the hybrid group, there was one case each (4.3%) of dyspnea caused by internal hematoma, aspiration pneumonitis, delirium, and segmental paralysis.


**Table 2.** Demographic and clinical characteristics in the ACCF and hybrid groups.

Data are shown as the mean ± standard deviation. ACCF, anterior cervical corpectomy with fusion; JOA, Japanese Orthopaedic Association; ICU, intensive care unit.


**Table 3.** Perioperative complications in the ACCF and hybrid groups.

Data are shown as the mean ± standard deviation. ACCF, anterior cervical corpectomy with fusion; DVT, Deep vein thrombosis; \* Statistically significant, *p* < 0.05.

Revision surgery for implant failure was performed in five cases (16.7%) in the ACCF group but in none in the hybrid group (Table 3). However, additional corpectomy was required for one patient who developed segmental motor dysfunction in the hybrid group. The incidence of reoperation, especially due to strut dislodgement immediately postoperatively, was significantly higher in the ACCF group compared with the hybrid group (*p* = 0.04, Table 3).

#### *3.2. Radiographic Outcomes*

Chronological changes on radiographs were evaluated. Mean C2–7 angle increased immediately postoperatively and was maintained in both groups (Figure 2a). C-SVA was increased immediately postoperatively, but it had gradually decreased by 6 months postoperatively in both groups (Figure 2b). T1 slope did not change during the follow-up period (Figure 2c). In the ACCF group, mean FSA was increased immediately postoperatively and was maintained. However, there was a gradual decrease with a loss of 0.7 degrees 1 year after the operation. In the hybrid group, mean FSA was increased in segments treated

with either ACDF or ACCF after surgery and was unchanged at the 1-year follow-up (Figure 2d).

**Figure 2.** Radiographic measurement of (**a**) C2–7 angle; (**b**) C2–7 SVA; (**c**) T1 slope; (**d**) FSA; and (**e**) FSH in the ACCF and hybrid groups.

In the ACCF group, mean FSH was increased by 2 mm immediately postoperatively. However, there was a 2-mm decrease at 1 year postoperatively. In the hybrid group, mean FSH was unchanged postoperatively, even compared with preoperatively (Figure 2e). Strut subsidence was observed in eight cases (26.7%) in the ACCF group and three cases (13.0%) in the hybrid group (Table 4). Among patients with strut subsidence, secondary surgery was required for early implant dislodgement in four cases in the ACCF group, but no secondary surgeries were required in the hybrid group. The fusion rate was significantly higher in the hybrid group than in the ACCF group (100% vs. 80%).


**Table 4.** Radiologic outcomes in the ACCF and hybrid groups.

Data are shown as the mean ± standard deviation. ACCF, anterior cervical corpectomy with fusion; ACDF, anterior cervical discectomy with fusion; C-SVA, cervical-sagittal vertical axis; FSA, fused segmental angle; FSH, fused segmental height. \* Statistically significant, *p* < 0.05. # Compared between the ACCF group and all patients in the hybrid group.

#### *3.3. Association of Change in FSH Immediately Postoperatively with Strut Subsidence in the ACCF Group*

To identify postoperative structural changes in the fused segments, associations between the incidence of graft dislodgement and parameters including FSA and FSH were evaluated in the ACCF group. Graft subsidence was more likely to occur in those with ΔFSH > 5 mm (Figure 3a). Of note, early strut dislodgement that required secondary surgery occurred in three of four cases with ΔFSH >10 mm. Stepwise logistic regression analysis demonstrated that only |ΔFSH| was a key risk factor for postoperative graft subsidence (odds ratio 1.328, 95% confidence interval 1.017–1.733; *p* = 0.04; Figure 3b).

**Figure 3.** (**a**) Plot showing changes in FSA and FSH (ΔFSA and ΔFSH) and the incidence of graft subsidence; (**b**) Plot showing associations between absolute values of ΔFSA and ΔFSH and graft migration. A large |ΔFSH| is more closely linked with graft migration than |ΔFSA|.

## **4. Discussion**

Various studies have discussed the superiority of certain surgical procedures for the treatment of patients with multilevel OPLL [5,6,8,9,21]. A meta-analysis revealed the anterior procedure provides more favorable results in terms of neurological recovery and postoperative cervical alignment [5]. Fujiyoshi et al. developed the K-line, which connects the midpoints of the spinal canal at C2 and C7 on neutral lateral radiographs, as a means of predicting poor clinical outcomes in patients with ossification of the posterior longitudinal ligament (OPLL) [22]. They classified OPLL patients into two groups, K-line (+) and K-line (−), and demonstrated that there was an insufficient posterior shift of the spinal cord and no neurologic improvement after posterior decompression surgery in patients in the K-line (−) group, in whom the anterior compression of OPLL exceeds the line. Notably, the classification is able to predict whether anterior compression of the spinal cord, which often impairs postoperative neural recovery [23], remains even after posterior decompression, and thus it is a very effective tool for deciding which surgical treatment—anterior or posterior—should be performed [8]. We also previously reported better outcomes after ACCF than after posterior procedures in patients who had severe OPLL with kyphotic alignment [9,21]. However, most comparative studies have shown that surgical complications were more frequent with anterior cervical surgery [6,21,24]. Of all complications after ACCF, airway obstruction and early graft migration are the most serious, often leading to emergency treatment and high reoperation rates [6,21,24]. For spine surgeons and neurosurgeons especially, it is important to recognize risk factors for predicting such perioperative complications.

This study also showed that the hybrid group had a relatively higher recovery rate in terms of JOA score (Table 1) and a lower incidence of complications (Table 2), although these differences did not reach significance. Additionally, no significant differences were found in terms of duration of ICU stay, the timing of postoperative extubation, hospital stay, or blood loss, although blood loss was certainly lower in the hybrid group. Fortunately, none of our patients needed a blood transfusion due to circulatory shock in this study. We speculate that corpectomy often violates the epidural venous plexus when decompressing the lateral portion of the spinal canal, thereby increasing the incidence of massive epidural bleeding in the ACCF procedure. Hospital stays were somewhat longer in the hybrid group than in the ACCF group, despite the lower complication rate in the hybrid group. Although there were more mechanical complications in the ACCF group, this finding suggests that the ACCF technique may be more invasive but may not affect the general condition of the patient as much as the hybrid procedure seems to.

Various studies have investigated the incidence of complications after anterior cervical surgery, which is reported to range from 11.3% to 64.3% [5,6,21,25–27]. Generally, it has been recognized that the hybrid procedure results in lower complication rates, ranging from 0% to 22.2%, and improved clinical outcomes compared with conventional ACCF, ranging from 6.2% to 43.6% [14,28,29]. In our study, the rate was 30.2% overall (ACCF, 40%; hybrid, 17.4%), relatively consistent with the results of previous studies. Of note, dysphagia occurred in four cases (13.3%) in the ACCF group, but did not occur in the hybrid group. A recent meta-analysis reviewed 38 studies and found an incidence of 16.8% after anterior cervical surgery [30]. Fortunately, the complication rate of dysphagia in this study was lower than previously reported. We speculate that because our institution is a high-volume center for cervical spine surgery, especially anterior cervical spine operation, the complication rate might be relatively low. Despite finding no significant difference between the two groups in our study, dysphagia caused by postoperative soft tissue swelling was likely lower in the hybrid group (0%) than in the ACCF group (13.3%). First, various surgical instruments, such as surgical osteotomes and retractors, were different between these two group because the patients categorized into the ACCF group underwent surgery in an earlier period than those in the hybrid group. Second, posterior cooling pads were placed on the anterior aspect of the neck immediately after surgery to prevent soft tissue swelling from 2012. In addition, we speculate that the superior laryngeal nerve and muscles related to swallowing might be less susceptible to mechanical damage in the hybrid group than in the ACCF group, likely because operative retractors can be loosened during decompression of the ossified lesions in the most compressed segmen<sup>t</sup> in the hybrid group, although they should be sufficiently expanded during decompression in the ACCF group. Because dysphagia can potentially develop even in patients treated with anterior hybrid fusion, it should be kept in mind postoperatively for both patients undergoing traditional ACCF and those undergoing hybrid fusion. Nevertheless, we believe that the hybrid technique could be a less invasive and more structurally stable strategy for treating severe OPLL compared with the conventional ACCF method.

It is known that multiple ACDF provides several advantages over multiple corpectomies [11]. ACDF is less invasive and is technically easier than ACCF. There is also less graft dislodgement, better correction of kyphotic deformity, and less need for postoperative immobilization because the screws are inserted into the preserved vertebral bodies [11]. However, there is some degree of difficulty involved in decompressing multilevel continuous lesions in the spinal canal and foramen using ACDF [8,20]. In patients with multilevel OPLL, corpectomy is often required to decompress the lesion entirely [20,21]. Vaccaro et al. reported that early graft dislodgement was seen in 9% of patients treated with 1-vertebra body corpectomy, and in as high as 50% of those receiving 2-vertebral body corpectomy [31]. Similarly, Okawa et al. reported that implant migration was observed in 30% of patients undergoing ACCF with average 3.8-level decompression [32]. The present study is the first to focus exclusively on severe OPLL affecting three or more levels and demonstrated that the incidence of reoperation for graft dislodgement was significantly higher in the ACCF group (16.7%) compared with the hybrid group (0%). In addition, mean FSH was decreased by 2 mm at 1 year after surgery in the ACCF group, whereas

it was unchanged in the hybrid group (Table 3 and Figure 2e). Interestingly, mean FSH at 1 year returned to the preoperative value in the ACCF group. These findings sugges<sup>t</sup> that the longer sized-graft bone could be applied to patients with ≥3 levels OPLL. As we previously reported, a postoperative increase in FSH can affect graft stability and lead to early implant migration [13]. The long lever arm created by long strut graft and fixation with four screws in the ACCF technique often stresses the screws and sometimes results in graft dislodgement [20]. A biomechanical study demonstrated that the hybrid technique was more effective for strengthening cervical stability and reconstructing sagittal alignment compared with the ACCF technique [33]. We also found that the fusion rate was significantly higher in the hybrid group at 1 year after the surgery. Taken together, our findings and previous findings indicate that the placement of two additional anchors in the hybrid technique can reduce stress on the proximal and distal screws and prevent loosening of the screws and graft dislodgement.

In this study, C2–7 lordotic angle and C-SVA were improved in both the ACCF and hybrid groups. Various studies have shown improvement and maintenance of cervical sagittal alignment after anterior cervical surgery [8,10,20,21]. This suggests that complete anterior decompression of ossified lesions may allow patients to assume a posture with more extension of the cervical spine. However, repeated micromotion can loosen the screws and could lead to severe reconstruction failure even while wearing a neck collar postoperatively. Particularly in cervical OPLL in patients with ankylosing spondylitis, reconstruction surgery with rigid fixation should be applied for the affected levels adjacent to the lesion. Interestingly, the prevalence of diffuse idiopathic skeletal hyperostosis (DISH) w shown to be approximately 45% in cervical OPLL patients [34,35]. Other observational studies have also demonstrated that multiple ossified lesions in the cervical spine often coexist with ossification of the spinal ligaments in the thoracolumbar spine [34,36–38]. In our study, of the four patients who underwent secondary surgery due to early implant failure in the ACCF group, three had DISH. Therefore, reconstruction surgery with rigid fixation and adequate structural alignments, such as hybrid fusion or anterior pedicle screw fixation [39], should be performed for cervical OPLL in patients with ankylosing thoracic spondylitis.

This study has several limitations. First, this was a retrospective study with a relatively small number of subjects and heterogeneity in terms of the implants used. Second, the degree of preservation of the cephalic and caudal endplates was not consistent in our series, although it was preserved as much as possible. Third, we could not assess factors associated with the location of the graft or screw angle and length. Fourth, these two techniques were performed by several independent surgeons. Fifth, the period was different between the two methods. Nevertheless, despite these limitations, our study highlights the fact that FSH should not be increased extremely after graft placement and plating should be carried out to prevent postoperative graft dislodgement in ACCF. Additionally, the hybrid fusion procedure is recommended for multilevel cervical OPLL in patients with ankylosing spondylitis to achieve sufficient stability of fused segments.
