**3. Results**

A total of 72 patients were enrolled in this study. Two patients were lost to follow-up and one patient died two months after surgery due to pneumonitis. Fifteen patients were excluded due to insufficient radiological data. Finally, 54 patients were included in the analysis. Patients with a mean age of 76.3 years ± standard deviation 6.1 were followed-up for 25.3 months ± 12.6. Twelve patients (22%) had a history of thoracic or lumbar surgery. Regarding medication for osteoporosis, 32 patients (59%) were treated by teriparatide, 3 patients (6%) by romosozumab, 8 patients (15%) by bisphosphonate, 4 patients (7%) by denosumab, and 7 patients (13%) by only vitamin D. Mean operative time and estimated blood loss was 269.8 ± 79.8 min and 289.5 ± 289.5 mL, respectively. Regarding fixation range, 32 patients (59%) were one above and one below fixation. Adjacent vertebral fractures were observed in 11 patients (20%) after surgery.

Table 1 shows the radiological parameters before and after surgery. Local kyphosis, thoracic kyphosis, lumbar lordosis, SVA, TPA and PI-LL significantly improved at final follow-up compared with before surgery, although there was no improvement in PT and SS. Local kyphosis improved from −17.5 degrees to 4.1 degrees immediately after surgery but was −0.6 degrees at final follow-up with 22.4% of correction loss. SVA was improved by only 18.7 mm (from 111.8 mm to 93.1 mm).


**Table 1.** Comparison of local and global alignment pre- and postoperatively.

SD, standard deviation; TK, Thoracic kyphosis; LL, Lumbar lordosis; SVA, Sagittal vertical axis; PT, Pelvic tilt; SS, sacral slope; TPA, T1 Pelvic Angle; PI-LL, Pelvic incidence- Lumbar lordosis.

Nineteen of the 54 patients (35%) showed global malalignment (SVA > 95 mm) postoperatively. Table 2 shows a comparison of baseline data, radiological parameters before surgery, and surgical complications between SVA > 95 mm and ≤95 mm groups. Adjacent vertebral fracture after surgery was significantly more frequent in the SVA > 95 mm group than in the SVA ≤ 95 mm group (37% vs. 11%, *p* = 0.038). TPA before surgery tended to be higher in the SVA > 95 mm group. Table 3 shows a comparison of clinical outcomes between SVA > 95 mm and ≤95 mm groups. VAS of back pain at final follow-up was significantly higher in patients with SVA > 95 mm than those in whom SVA was ≤95 mm (42.4 vs. 22.6, *p* = 0.015). Regarding the MCID, the better improvement was also observed in patients with SVA ≤ 95 mm (83% vs. 58%, *p* = 0.046). Multiple logistic regression showed a significantly increased odds ratio (OR) of adjacent vertebral fracture presence and TPA increase (OR = 4.76, 95% CI 1.10–20.58 and OR = 1.07, 1.00–1.14, respectively) (Table 4).


**Table 2.** Comparison between SVA > 95 mm and ≤95 mm groups by univariate analysis.

SD, standard deviation; BMD, Bone marrow density; TK, Thoracic kyphosis; LL, Lumbar lordosis; SVA, Sagittal vertical axis; PT, Pelvic tilt; SS, sacral slope; TPA, T1 Pelvic Angle; PI-LL, Pelvic incidence- Lumbar lordosis.



SD, standard deviation; PS, Performance Status; JOA score, The Japanese Orthopaedic Association score; MCID, minimal clinically important difference. \* There were significant differences between preop and final scores of JOA score and VAS of back pain.


**Table 4.** Adjusted odds ratio for SVA > 95 mm at final follow-up.

TPA, T1 Pelvic Angle; OR, odds ratio. \* The odds ratio was adjusted for age, preoperative TPA and adjacent vertebral fracture.

#### **4. Discussion**

This is the first study to reveal details about changes in sagittal balance following the minimally invasive procedure of corpectomy and reconstruction using an expandable cage with rectangular foot plates (APSF). Although there was 22.4% correction loss, local kyphotic changes using this system was 21.7◦, which was better than the previous reports for APSF [17–19]. As well, Kanayama et al. reported that 80% of patients with OVF could be successfully treated using Kaneda instrumentation without the need for posterior reinforcement [20]. However, nearly 40% of correction loss was observed at the final followup. Suk et al. compared anterior-posterior surgery versus closing wedge osteotomy for kyphotic OVF and reported that the correction loss of anterior-posterior surgery was 27.3% with a mean blood loss of 2892 mL, whereas that of posterior closing wedge osteotomy was 10.8% with a mean blood loss of 1930 mL [21]. Posterior closing wedge osteotomy might offer better kyphosis correction. However, the procedure is technically demanding with more blood loss compared with the system in this study.

Although it is reported that anatomical and biomechanical restoration of vertebra is an advantage of anterior surgery resulting from the placement of anterior struts, our results indicated that restoration of sagittal alignment was not achieved by anterior surgery with 1–2 level posterior fixation in OVF patients. The parameters of SVA and TPA were used to evaluate sagittal spinal balance in this study. SVA increases with aging, and it is affected by movement of the hip and knee joint, such as "sway back" TPA, which combines information of SVA and PT and is a reliable indicator to address sagittal balance, including pelvic inclination [22]. TPA in this series was 33.2◦ preoperatively and 30.1◦ postoperatively. Thus, the improvement in TPA might not be significant. Ryan et al. demonstrated that TPA > 20◦ was the severe deformity threshold [23]. The main reason for this observation in our study was postoperative development of adjacent vertebral fracture. Low BMD, older age, an upper instrumented vertebra (UIV) level at the thoracolumbar spine, and a high preoperative SVA have been reported as risk factors for proximal junctional failure following surgical treatment for adult spinal deformity [24]. In the current series, BMD, medicine for osteoporosis, and level of surgery was not different between SVA > 95 mm and SVA ≤ 95 mm groups, probably because all the patients had comparatively severe osteoporosis. Posterior tethers and vertebral augmentation might be effective in preventing the failure of instrumentation, especially in patients with a high risk for proximal junctional kyphosis [25].

The relationship between PI and LL (PI-LL) is also considered an important parameter to evaluate sagittal spinal balance. Schwab et al. reported that SVA of 47 mm or more, PI-LL > 11◦ or more, and PT < 22◦ predicted severe disability (ODI > 40) [26]. Yamato et al. [31] described that the ideal LL angle can be determined using the equation 'LL = 0.45 × PI + 31.8'. Inami et al. [27] reported that the optimum value of PI-LL is inconsistent, in that it depends on the individual PI. [28]. In this study, although PI-LL improved significantly (from 35.1◦ to 24.2◦), the final PI-LL did not reach the ideal value. In addition, the preoperative decrease in SS did not change postoperatively, indicating absence of improvement of pelvic retroversion. If lumbar lordosis is restored by surgery, the retroverted sacrum must be improved to maintain spino-pelvic harmony. Otherwise, reciprocal changes in the thoracic spine might develop to maintain sagittal balance [29,30]. Our results showed an increase in TK from 26.8◦to 32.8◦, which concurred with the theory mentioned above. This reciprocal change might be one of the reasons SVA did not change significantly in the OVF patients in our study. Improvement of the retroverted sacrum requires extension

of the hip joint, with the erector spinae and gluteus muscles playing an important role in this action. In aged OVF patients, weakness of these muscles is responsible for the pelvic retroversion [8,31]. The average age of patients in this study was 76.3 years; hence, although we did not measure muscle volumes in these patients, they might have had age-related muscle wasting and weakness. A retroverted pelvis can be managed surgically by osteotomy of the lower lumbar vertebra or long fixation involving the pelvis. However, these are extremely invasive surgeries and it is not clear whether such invasive correction surgery is necessary for aged OVF patients.

SVA changed from 111.8 mm to 93.1 mm, which, although a statistically significant change, might be an insufficient improvement to correct malalignment. Based on the classification of Scoliosis Research Society [16], SVA (>95 mm) was reported as a risk factor with the deterioration of QOL measures [32]. In the current study, the number of patients who acquired one or more level improvement of PS was 15/19 (78.9%) in SVA > 95 mm and 33/35 (94.3%) in SVA ≤ 95 mm groups, which although better in the SVA ≤ 95 mm group, was not significantly different. Postoperative VAS was better in the SVA ≤ 95 mm than the SVA > 95 mm group. As also reported by Hu et al. [5]. SVA correlated with back pain in this study, which significantly improved after surgery. However, age and preoperative comorbidities influence the complication rate in deformity surgery [33]. Thus, we thought that the strategy for aged OVF patients should differ from those in ASD patients to relieve pain and improve mobility. Our results also showed the significant improvement of JOA score and VAS even in the SVA > 95 mm group compared with those before surgery. It is not always necessary to restore sagittal imbalance in aged OVF patients to the same level as in young people, although the clinical results are worse in patients with SVA > 95 mm.

There are some limitations to this study. First, the number of patients was small because some patients were excluded due to lack of data from standing whole spine X-ray films before surgery because of intractable back pain. Second, due to the lack of apparatus, we did not take whole spine X-rays including the lower extremity. Hence, we could not evaluate knee and hip joint flexion, which might have been used to compensate for sagittal imbalance [34]. Despite these limitations, this is the first report describing the correlation between anterior spinal surgery and changes in sagittal alignment, which might contribute to preoperative planning in OVF patients. For further study, the prediction methods for postoperative sagittal balance are necessary, since this might contribute to decision-making in the surgical planning for OVF patients.
