Sagittal alignment changes and postoperative complications after adult spinal deformity surgery in patients with Parkinson’s disease: A multi-institutional retrospective study

Background: Parkinson’s disease (PD) has been reported to increase the risk of postoperative complications in patients with adult spinal deformity (ASD). However, those reports are limited, and few have made direct comparisons with patients who do not have PD. Methods: We retrospectively reviewed all surgically treated patients with ASD and at least 2 years of follow-up. Among them, 27 had PD (PD(+) group). Clinical data were collected on early and late postoperative complications and revision surgery. Radiographic parameters were evaluated before and immediately after surgery and at nal follow-up, including sagittal vertical axis (SVA), thoracic kyphosis, lumbar lordosis, sacral slope, and pelvic tilt. From the same database, we also retrieved 206 controls without PD (PD(−) group) matched for age, sex, and body mass index. Results: For early complications, the PD(+) group showed a higher rate of delirium than the PD(−) group. Deep vein thrombosis and pulmonary embolism rates tended to be higher in the PD(+) group. For late complications, the rate of pseudarthrosis was signicantly higher in the PD(+) group. Rates of rod failure and revision surgery due to mechanical complications also tended to be higher, but not signicantly, in the PD(+) group (p = 0.17, p = 0.13, respectively). SVA at nal follow-up and loss of correction in SVA were signicantly higher in the PD(+) group. Conclusion: Extra attention should be paid to perioperative complications, especially delirium and thrombosis, in PD patients undergoing surgery for ASD. Furthermore, loss of correction and rate of revision surgery due to mechanical complications were higher in these patients.

Generally, PD patients are presumed to have a higher risk of surgical complications, especially mechanical complications due to postural instability, a higher risk of falls, and lower bone quality [10][11][12][13] However, to date, only a few small case series have investigated complications after surgery for ASD in PD patients [14][15][16][17]. In addition, risk factors for complications and revision surgery in PD patients have not yet been con rmed. Therefore, we conducted this multicenter study to investigate 234 surgically treated ASD patients with at least 2 years' follow-up. We compared surgical outcomes and radiographic parameters of PD patients with those of non-PD patients matched for age, sex, and body mass index (BMI). We also investigated risk factors for revision surgery in PD patients.

Methods
This retrospective observational study follows the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines. We reviewed 233 ASD patients treated surgically between January, 2009, through December, 2016, at our hospital and related institutions. Institutional review board approval was obtained at each site for the patient enrollment and data collection protocols. Inclusion criteria were age 21 years or older at the time of surgery, follow-up period of at least 2 years, and surgery including posterior instrumentation of 4 spinal levels with su cient radiographic data. A total of 27 patients with PD were identi ed. Data on mean PD duration and Hoehn and Yahr (HY) stage were collected. A control group of 206 non-PD patients was retrieved from the same database.
Etiologies included degenerative kyphosis/kyphoscoliosis, post lumbar surgery, and previous vertebral fracture. Operative time and intraoperative blood loss were recorded. Perioperative complications were de ned as complications that occurred during and within 1 month after surgery, including neurological disorder, implant failure, DVT, PE, cerebrovascular disease, respiratory disorder, cardiovascular disorder, delirium, surgical site infection, and spinal epidural hematoma. The number of revision surgeries due to perioperative complications was also noted. Mechanical complications were de ned as proximal junctional kyphosis (PJK), distal junctional kyphosis (DJK), pseudarthrosis, rod breakage, and vertebral fracture with or without revision surgery. PJK was de ned as kyphosis > 10° between the upperinstrumented vertebra (UIV) and two-level proximal vertebrae. Mechanical failure was de ned as mechanical complications requiring revision surgery.
Demographic data were collected including age, sex, BMI, medical comorbidities, and bone mineral density (BMD) of the femoral neck. Measurements on radiographs included SVA, thoracic kyphosis (T4-T10), L1-S1 lumbar lordosis (LL), sacral slope, pelvic incidence (PI), and pelvic tilt. These parameters were evaluated before and at 4 weeks after surgery in the standing position. X-ray images were also evaluated at the nal follow-up. A three-column osteotomy was de ned as a procedure using pedicle subtraction osteotomy or vertebral column resection. The radiographic parameter of − 10° < PI − LL < 10° was de ned as ideal alignment based on the SRS-Schwab ASD classi cation.
Statistical analysis was performed using IBM SPSS Statistics for Macintosh, Version 25.0 (Released 2017 IBM Corp, Armonk, NY). We divided all patients into two groups: patients with PD (PD(+) group) and patients without PD (PD(−) group. We used a paired t-test or chi-squared test to compare the PD(+) and the PD(−) groups. Also, t-tests were used to compare the means of continuous variables and chi-square tests to compare the proportions of categorical variables between the groups. A p-value < 0.05 was considered statistically signi cant. Missing values were imputed using the last observation carried forward method. Table 1 shows the patient characteristics of the PD(+) group and the PD(−) group. There was no signi cant difference between the two groups in age, sex, BMI, or BMD. The number of xed levels was signi cantly higher in the PD(+) group than in the PD(−) group (9.3 ± 2.6 vs 7.8± 1.9, p = 0.006) and the rates of threecolumn osteotomy tended to be higher in the PD(+) group (63.0% vs 48.5%, p = 0.14). For preoperative radiographic parameters, SVA was signi cantly higher in the PD(+) group than in the PD(−) group (196.0 ± 63.8 mm vs 135.4 ± 69.9 mm, p < 0.001). Other parameters were not signi cantly different between these groups. In terms of etiology, ASD caused by previous vertebral fracture was signi cantly higher in the PD(+) group than in the PD(−) group (previous vertebral fracture/degenerative/post lumbar surgery: 38.5%/53.8%/7.7% vs 10.1%/64.7%/25.1%, p = 0.005).  Table 2 shows surgical invasiveness and postoperative complications in each group. There were no signi cant differences in operative time and intraoperative blood loss between the two groups. For early complications, the PD(+) group showed a higher rate of delirium than the PD(−) group. In the PD(+) group, rates of DVT and PE tended to be higher (DVT: 14.8% vs 6.8%, p = 0.12; PE: 3.7% vs 0.5%, p = 0.081). There were no signi cant differences between the groups in the rates of other complications, including neurological de cits, implant failure, cerebrovascular disorder, respiratory disorder, cardiovascular disorder, and surgical site infection. There was no signi cant difference in the rate of revision surgery due to early complications. For late complications, the rate of pseudarthrosis was signi cantly higher in the PD(+) group (15.3% vs 3.9%; p = 0.013). Rates of rod failure and revision surgery due to mechanical complications tended to be higher, but not signi cantly, in the PD(+) group (rod failure: 25.9% vs 13.1%, p = 0.174; revision surgery: 33.3% vs 18.0%, p = 0.13).  Table 3 summarizes changes in radiographic parameters postoperatively and at nal follow-up.

Results
Postoperative SVA was similar between the PD(+) and PD(−) groups, although preoperative SVA was much higher in the PD(+) group. The change in SVA between pre-and postoperative radiographs was signi cantly higher in the PD(+) group (−142.5 ± 82.0 mm vs −94.6 ± 69.7 mm, p = 0.014). In addition, SVA at the nal follow-up tended to be higher in the PD(+) group (p = 0.062). Loss of correction in SVA also tended to be higher in the PD(+) group (p = 0.11). No signi cant differences were found in other radiographic parameters (Fig. 1), including the proposed ideal alignment target of PI-LL < 10.  Table 4 shows the preoperative demographics, postoperative radiographic parameters, and surgical characteristics of the PD group comparing those who underwent revision surgery (revision subgroup) with those who did not (non-revision subgroup). No signi cant difference was found in the radiographic parameters between these two groups. For preoperative demographics, the duration of PD was signi cantly higher in the revision subgroup (87.0 ± 56.9 months vs 32.5 ± 48.0 months, p = 0.037). HY Page 10/17 stage was higher in the revision subgroup, although this was not signi cantly different (2.8 ± 1.2 vs 2.2 ± 1.3, p = 0.29).  19]. Furthermore, other surgical studies on the PD population have suggested high rates of postoperative medical complications such as pneumonia, delirium, and sepsis [5]. However, few case series have examined the relationship between PD and complications following corrective surgery for ASD. Furthermore, few studies have focused on postoperative loss of correction of radiographic parameters in PD patients. Thus, we retrospectively analyzed both postoperative complications and radiographic parameters in this multicenter database of ASD patients and compared PD patients with non-PD patients matched for age, sex, and BMI with at least 2-years' follow-up.
The results of this study demonstrated that the rate of delirium was signi cantly higher, and the rates of PE and DVT tended to be higher in the PD(+) group. In addition, Watanabe et al. previously reported that postoperative delirium was more common in patients with PD (23.1%) than in the controls (3.4%) [20]. This higher rate of delirium is possibly due to the use of antidopaminergic drugs and neurodegeneration in PD patients. Although delirium is a reversible condition, it is necessary to avoid additional complications due to falls or movement beyond the limits of restriction. For thrombotic events, PE is known to be a possible adverse reaction to antidopaminergic drugs, such as levodopa [21]. Yamane et al. reported a higher incidence (20%) of DVT in PD patients with a postural abnormality [22]. Since the correction surgery for ASD itself has a high risk for PE due to its long surgical time and large blood loss, it is assumed that the risk of PE of this surgery in PD patients is very high. In fact, in this study, the incidence of PE in PD patients was 7 times higher and that of DVT was about 2 times higher than in non-PD patients. Surgeons should thus consider PE in patients who develop chest pain and dyspnea after surgery.
In this study, we found signi cantly higher preoperative SVA and larger correction of SVA in the PD(+) group. We also found higher SVA at the nal follow-up and higher loss of correction in SVA, even though longer fusion was performed in the PD(+) group. This nding suggests that the deformities in the PD(+) group were mostly exible and largely corrected by the surgery. However, the corrected alignment could not be well maintained in this group, suggesting that a larger correction can lead to a higher rate of loss of correction. In addition, the progressive stooping posture of PD patients can pose a risk for loss of correction of global sagittal alignment, and this may be one of the causes for the poorer outcome in PD patients. Kawaguchi et al. reported that longer fusion until T4 yielded a good clinical outcome in a PD patient after corrective surgery for L1 to S1 was unsuccessful [23]. Watanabe et al. reported on 8 surgically treated ASD patients with PD who demonstrated poor clinical outcomes with a high non-union rate and adjacent segment disease [24]. Thus, it is crucial to consider possible prevention strategies, including longer fusion.
Previous studies reported that PD patients have a higher rate of osteoporotic vertebral fracture and lower BMD [1,20]. However, we found no signi cant difference in BMD in this study even though the rates of previous vertebral fractures were signi cantly higher in the PD group. We assumed that the postural abnormality in the PD(+) group added greater burden on the vertebral body than on that of a healthy person and resulted in vertebral fracture. Thus, caution should be exercised when considering the risk of PJK type fracture in PD patients. Yagi et al. reported that teriparatide treatment started immediately after surgery resulted in a decrease in rates of PJK-type bone fracture [25]. If BMD is low, then administration of osteoporotic drugs such as teriparatide could be a viable option to decrease the risk of complications, as suggested in the previous study.
This study showed that the revision rate due to mechanical complications was 33.3%, which was almost two times higher in the PD(+) group than in the PD(−) group. Similar to our study, Sheu et al. investigated 66 PD patients who underwent thoracolumbar or lumbar instrumented surgery due to degeneration or deformity; 29% of them needed revision surgery due to mechanical complications [13]. Bouyer et al.
reported a high revision rate of 42% in 48 ASD patients with PD, 89% of which were due to mechanical complications [16]. In terms of each complication, PJK has been reported to be signi cantly higher in PD patients. In contrast, a history of PD had no signi cant impact on the PJK rate in this study. Instead, rates of pseudarthrosis and rod failure were higher in the PD(+) group. This was probably because a longer fusion was performed in the PD(+) group, possibly as a result of the surgeon's decision considering the severe deformity with higher SVA. Long-segment xation may lower the risk of PJK but the stress on the rod and the mobile segment of the spine can be higher and may have caused rod fracture and pseudarthrosis.
Several studies have investigated risk factors for revision surgery in PD patients. Schroeder et al. reported that an HY stage > 2, diabetes mellitus, treatment for osteoporosis, and a combined anterior and posterior approach were risk factors for revision surgery in 94 lumbar spine surgeries [12]. According to Sheu et al., HY stage > 2, cancer history, osteoporosis, and three column osteotomy were risk factors for revision surgery [13]. In our study, PD duration was signi cantly longer in the revision subgroup, even though they were younger. HY stage can be affected by symptoms of sagittal malalignment and imbalance in ASD patients, and thus the stage itself may not re ect the exact severity of PD in ASD patients. However, the HY stage also tended to be higher in the revision subgroup. As the previous study indicated, poorly controlled PD negatively impacts the incidence of revision surgery. Thus, it is vital to consider duration and severity of PD when performing surgical treatment in ASD patients.
There are several limitations to this study. First, there was a selection bias in which the surgeon could change their choice of a surgical procedure because of the patient's PD status. Second, our study was retrospective in nature. Studies with more appropriate study designs and a larger sample size are needed.
Despite these limitations, our study demonstrated surgical outcomes in patients with PD compared with patients without PD using a large database with a su cient follow-up period.

Conclusion
In patients with PD, postoperative delirium was common and the rate of PE and DVT tended to be higher.
Late complications included rod fracture, pseudarthrosis, and revision surgery, which were higher in the PD(+) group. The rate of revision surgery due to mechanical complications was about two times that of the PD(−) group. In the PD(+) group, SVA was signi cantly larger before surgery and at nal follow-up, as was loss of correction in SVA, suggesting that PD-speci c postural abnormalities are involved in sagittal parameter deterioration. PD duration was signi cantly higher in the PD patients who underwent revision surgery than in those who did not.

List Of Abbreviations
ASD: adult spinal deformity; BMI: body mass index; BMD: bone mineral density; DVT: deep vein thrombosis; DJK: distal junctional kyphosis; HY: Hoehn and Yahr; LL: L1-S1 lumbar lordosis; PD: Parkinson's disease; PI: pelvic incidence; PJK: proximal junctional kyphosis; PE: pulmonary embolism; T4-T10: thoracic kyphosis; UIV: upper-instrumented vertebra Declarations Ethics approval and consent to participate The study was approved by the ethics committee of all institutions involved. Informed consent was waivered by the above ethics committee as the present retrospective cohort study involved already existing data and records at the time of investigation, and did not retain personal identi ers of the gathered information.

Consent for Publication
Not applicable Availability of data and materials The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
Not applicable Funding Not applicable Authors' contributions AK analysed the data and wrote the original draft. TY conceived of, reviewed, and edited the paper. KS, TH, MY, HI, YM, MT, IT, and KK acquired the data. KO performed the investigation. YA validated the paper. YA, SS, and AO supervised the researched. All authors contributed to the writing of the nal manuscript. All authors approved the manuscript to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.