Adult spinal deformity is a complex disease with

Neurosurg Focus 36 (5):E9, 2014 AANS, 2014 Long fusion from sacrum to thoracic spine for adult spinal deformity with sagittal imbalance: upper versus lower thoracic spine as site of upper instrumented vertebra Takahito Fujimori, M.D., 1,2 Shinichi Inoue, M.D., 1,3 Hai Le, B.A., 1 William W. Schairer, M.D., 1 Sigurd H. Berven, M.D., 1 Bobby K. Tay, M.D., 1 Vedat Deviren, M.D., 1 Shane Burch, M.D., 1 Motoki Iwasaki, M.D., 2 and Serena S. Hu, M.D. 1 1 Department of Orthopedic Surgery, University of California, San Franicsco, California; 2 Department of Orthopaedic Surgery, Osaka University Graduate School of Medicine, Osaka, Japan; and 3 Department of Orthopedic Surgery, Hyogo College of Medicine, Nishinomiya, Hyogo, Japan Object. Despite increasing numbers of patients with adult spinal deformity, it is unclear how to select the optimal upper instrumented vertebra (UIV) in long fusion surgery for these patients. The purpose of this study was to compare the use of vertebrae in the upper thoracic (UT) versus lower thoracic (LT) spine as the upper instrumented vertebra in long fusion surgery for adult spinal deformity. Methods. Patients who underwent fusion from the sacrum to the thoracic spine for adult spinal deformity with sagittal imbalance at a single medical center were studied. The patients with a sagittal vertical axis (SVA) 40 mm who had radiographs and completed the 12-item Short-Form Health Survey (SF-12) preoperatively and at final follow-up ( 2 years postoperatively) were included. Results. Eighty patients (mean age of 61.1 ± 10.9 years; 69 women and 11 men) met the inclusion criteria. There were 31 patients in the UT group and 49 patients in the LT group. The mean follow-up period was 3.6 ± 1.6 years. The physical component summary (PCS) score of the SF-12 significantly improved from the preoperative assessment to final follow-up in each group (UT, 34 to 41; LT, 29 to 37; p = 0.001). This improvement reached the minimum clinically important difference in both groups. There was no significant difference in PCS score improvement between the 2 groups (p = 0.8). The UT group had significantly greater preoperative lumbar lordosis (28 vs 18, p = 0.03) and greater thoracic kyphosis (36 vs 18, p = 0.001). After surgery, there was no significant difference in lumbar lordosis or thoracic kyphosis. The UT group had significantly greater postoperative cervicothoracic kyphosis (20 vs 11, p = 0.009). The UT group tended to maintain a smaller positive SVA (51 vs 73 mm, p = 0.08) and smaller T-1 spinopelvic inclination ( 2.6 vs 0.6, p = 0.06). The LT group tended to have more proximal junctional kyphosis (PJK), although the difference did not reach statistical significance. Radiographic PJK was 32% in the UT group and 41% in the LT group (p = 0.4). Surgical PJK was 6.4% in the UT group and 10% in the LT group (p = 0.6). Conclusions. Both the UT and LT groups demonstrated significant improvement in clinical and radiographic outcomes. A significant difference was not observed in improvement of clinical outcomes between the 2 groups. (http://thejns.org/doi/abs/10.3171/2014.3.focus13541) Key Words adult spinal deformity sagittal imbalance upper thoracic lower thoracic proximal junctional kyphosis long fusion Adult spinal deformity is a complex disease with various clinical and radiographic presentations. 22 Many studies have reported that positive sagittal imbalance is associated with worse clinical outcomes. 9,19 Abbreviations used in this paper: EBL = estimated blood loss; LIV = lower instrumented vertebra; LL = lumbar lordosis; LT = lower thoracic; MCID = minimum clinically important difference; MCS = mental component summary; modi = modified Oswestry Disability Index; PCS = physical component summary; PI = pelvic incidence; PJK = proximal junctional kyphosis; PT = pelvic tilt; SF-12 = 12-item Short-Form Health Survey; SRS = Scoliosis Research Society; SS = sacral slope; SVA = sagittal vertical axis; T-1 Spi = T1 spinopelvic inclination; UCSF = University of California, San Francisco; UIV = upper instrumented vertebra; UT = upper thoracic. Neurosurg Focus / Volume 36 / May 2014 Schwab et al. classified cases in which patients had a sagittal vertical axis (SVA) 40 mm with the SVA modifier + 20 because of the negative effect that a positive SVA had on patient outcomes. Long fusion from the sacrum to the thoracic spine is a common surgical treatment option to correct for sagittal imbalance. 5,6,13,15 17,21 When deciding on the upper instrumented vertebra (UIV) for fusion, surgeons often face the difficulty of having to choose between the upper thoracic spine (UT group) or lower thoracic spine (LT group). Unfortunately, there have been few studies comparing the clinical and radiographic outcomes between these 2 groups. 18 The purpose of this study was to compare the radio- 1

T. Fujimori et al. graphic outcomes, clinical outcomes, and complication rates between using vertebrae from the UT spine or the LT spine as the UIV in fusion surgery for sagittal imbalance due to adult spinal deformity. Methods This was a retrospective study of 80 consecutive cases involving patients with adult spinal deformity treated at the University of California, San Francisco (UCSF) Medical Center between 2003 and 2009. Inclusion criteria of this study were the following: SVA 40 mm, fusion from the sacrum to the thoracic spine, pre- and postoperative radiographs, pre- and postoperative 12-item Short- Form Health Survey (SF-12) assessment, and a minimum follow-up period of 2 years. This study was approved by UCSF s institutional review board. These 80 patients were grouped by UIV. The UIV for the UT group ranged from T-1 to T-6 while for the LT group it ranged from T-7 to T-12. Surgical Strategy The indication for surgical treatment of adult spinal deformity was intolerable low back pain and/or radicular pain with deformity that was resistant to conservative therapy for more than 6 months. The UIV level was determined by the surgeon s preference based on the curve pattern. As a general rule, the apex of the thoracic kyphosis or kyphotic segment was avoided as the UIV. The UT spine was selected for those with coronal or sagittal thoracic curves, and the LT spine was selected for those with thoracolumbar curve. Combined anterior and posterior surgery was performed for patients with a hypolordotic lumbar spine. If the patient had previously undergone a successful procedure, a pedicle subtraction osteotomy was used to correct the deformity. In other cases, a Ponte osteotomy was performed for correction. Fusion to the sacrum was performed when the L5 S1 segment was symptomatic and/or sufficiently degenerated that it was not desirable to leave it as an unfused segment, or when inclusion was deemed necessary to achieve improved coronal and sagittal balance. Iliac screw fixation and interbody fusion was usually performed at the L5 S1 level to decrease the risk of pseudarthorosis. In the light of these factors, the final fusion level was determined by each surgeon s preference. Radiographic Parameters Radiographic parameters of interest included 1) cervical lordosis: C2 7 angle (positive means lordosis); 2) thoracic kyphosis: T5 12 angle (positive means kyphosis); 3) lumbar lordosis: T12 S1 angle (positive means lordosis); 4) cervicothoracic kyphosis: C7 T5 angle (positive means kyphosis); 5) C7 S1 sagittal vertical axis (SVA): distance between C-7 plumb line and posterosuperior sacrum; 6) C2 S1 SVA (C-2 SVA): distance between C-2 plumb line and posterosuperior sacrum; 7) pelvic parameters: sacral slope (SS), pelvic tilt (PT), and pelvic incidence (PI); 8) T-1 spinopelvic inclination (T-1 Spi): the angle between the vertical plumb line and the line drawn from the center of the T-1 vertebral body and the center of the bicoxofemoral axis; 19 and 9) T-1 slope: angle between the superior endplate of T-1 and the horizontal in the sagittal plane 14 (positive means T-1 tilts more forward in the sagittal plane) (Fig. 1). The sagittal modifier represented by PI- LL in the Scoliosis Research Society (SRS) Schwab adult spinal deformity classification system was calculated. The proximal junctional kyphosis (PJK) angle was defined as the sagittal angle subtended by the inferior endplate of the UIV and the superior endplate 2 levels above the UIV. All radiographic parameters were assessed in a digital viewer, Surgimap Spine (Nemaris, Inc.). Operative Data Parameters relevant to the operative procedure were evaluated. The UIV level, whether a hook or pedicle screw was used at the UIV, the use of osteotomy (Ponte or pedicle subtraction), and decompression were recorded. Operative time and estimated blood loss (EBL) were recorded; if surgical correction was a 2-staged operation, the operative time or EBL of each stage was summed. Clinical Outcomes Patients were evaluated preoperatively, at their first postoperative visit (mean 2 ± 1 month postoperatively), and at final follow-up. Clinical outcomes were determined by means of the SF-12, Scoliosis Research Society (SRS) score, and a modified version of the Oswestry Low Back Pain Disability Questionnaire (modified Oswestry Disability Index [modi]). 7 All patients completed the SF- 12 preoperatively and at final follow-up. SRS and modi scores were available for 74% and 79% of the total patients, respectively. Complications Complications were classified as intraoperative, immediate postoperative, or post-discharge. Immediate postoperative was defined as the inpatient recovery period after the operation but before discharge. Major complications were defined as complications that prompted active medical intervention or return to the operating room. For example, deep wound infection, infection of vital organs, and revision surgery with instrumentation replacement were considered major complications. Proximal junctional kyphosis was defined by 2 criteria: 1) PJK angle 10 and 2) PJK angle 10 greater than the preoperative measurement. 11 PJK was further classified into symptomatic PJK and asymptomatic PJK (radiographic PJK). Symptomatic PJK that required revision surgery was defined as surgical PJK. Statistical Analysis In analyzing clinical outcomes, we performed both intragroup and intergroup analyses. The Shapiro-Wilk test was used to determine whether a data set was well modeled by a normal distribution or not. Normally distributed data were analyzed with a t-test, while data not normally distributed were analyzed with a Mann-Whitney test. The unpaired t-test was used to compare radiographic parameters between the UT and LT groups. The paired t-test 2 Neurosurg Focus / Volume 36 / May 2014

Level of upper instrumented vertebra in adult spinal deformity TABLE 1: Demographic and operative data Fig. 1. Radiographic parameters: cervical lordosis (CL); thoracic kyphosis (TK); lumbar lordosis (LL); cervicothoracic kyphosis (CTK, superior endplate of C-7 to inferior endplate of T-5); C7 S1 sagittal vertical axis (SVA, distance between C-7 plumb line and posterosuperior sacrum); C2 S1 SVA (C-2 SVA, distance between C-2 plumb line and posterosuperior sacrum); pelvic parameters sacral slope (SS), pelvic tilt (PT), and pelvic incidence (PI); T-1 spinopelvic inclination (T-1 Spi); and T-1 slope (angle between superior endplate of T-1 and the horizontal on sagittal plane). was used for intragroup comparison. The chi-square test was used to compare demographic data and complication rates. A p value less than 0.05 with a 2-tailed test was considered statistically significant. Statistical analysis was performed with SPSS Statistics (version 20; IBM, Inc.). Results Patient Demographic Characteristics The UT group included 31 patients with a mean age (± SD) of 60.3 ± 12.1 years and a mean follow-up time of 3.6 ± 1.6 years. The LT group included 49 patients with a mean age of 61.6 ± 10.2 years and a mean follow-up time of 3.7 ± 1.6 years. There were no significant differences between the 2 groups in terms of age, sex, and follow-up period (Table 1). However, there was a significant difference in the etiology of the deformity. The UT group included more patients with idiopathic scoliosis (UT, 42% vs LT, 14%), whereas the LT group included more patients with prior spine surgery (UT, 26% vs LT, 49%). Operative Data Operative data are reported in Table 1. The most common UIV in the UT group was T-3 (n = 13), followed by T-4 (n = 12). The most common UIV in the LT group was T-10 (n = 25), followed by T-11 (n = 11). For instrumentation at the UIV, in the UT group pedicle screws were used in 26 cases and a hook system in 5 cases, and in the LT group pedicle screws were used in 45 cases and a hook system in 4 cases. Iliac screws were used in 27 cases Neurosurg Focus / Volume 36 / May 2014 Variable UT Group (n = 31) LT Group (n = 49) p Value demographic data mean age (yrs) 60 ± 12 62 ± 10 0.6 sex 0.9 male 4 7 female 27 42 mean follow-up (yrs) 3.6 ± 1.6 3.7 ± 1.6 0.7 etiology of deformity 0.01* idiopathic 13 (42%) 7 (14%) degenerative 7 (23%) 17 (34%) prior surgery 8 (26%) 24 (49%) other 3 (10% 1 (2%) operative data UIV level T-1 2 T-2 1 T-3 13 T-4 12 T-5 3 T-7 2 T-8 3 T-9 2 T-10 25 T-11 11 T-12 6 UIV instrumentation 0.3 screw 26 45 hook 5 4 LIV level iliac/sacral 0.4 iliac 27 40 sacral 4 9 surgical approach 21/10 38/11 0.4 combined ant & pst 21 38 pst only 10 11 decompression 25 49 0.3 Ponte osteotomy 31 47 0.3 PSO 6 9 0.9 mean EBL (ml) 4922 ± 3056 3432 ± 2930 0.03* mean operative time (min) 623 ± 263 519 ± 170 0.1 * Values represent numbers of patients or cases unless otherwise indicated. Means are given with SDs. Ant = anterior; EBL = estimated blood loss; PSO = pedicle subtraction osteotomy; pst = posterior. (87%) in the UT group and in 40 cases (82%) in the LT group. Combined anterior and posterior surgery was performed in 21 cases (68%) in the UT group and in 38 cases (78%) in the LT group. Decompression was performed in 25 cases (81%) in the UT group and in all cases in the LT group. Pedicle subtraction osteotomy was performed in 6 cases (19%) in the UT group and in 9 cases (18%) in 3

T. Fujimori et al. the LT group. The estimated blood loss was significantly greater in the UT group (UT, 4922 ± 3056 ml vs LT, 3432 ± 2930 ml; p = 0.046). There was no significant difference in operative time (UT, 623 ± 263 minutes vs LT, 519 ± 170 minutes; p = 0.1). In the UT group, 6 of the 8 patients with prior spine surgery had prior fusion surgery. The prior UIV level was in the lumbar spine in 1 case, the lower thoracic spine in 2, or the upper thoracic spine in 3 cases. Three of the 6 patients (13%) had distal extension to the sacrum as well as proximal extension to the UT spine. In these 3 cases the prior lower instrumented vertebra (LIV) was L-3, L-4, and L-5. In the LT group, 18 of the 24 patients with prior spine surgery had prior fusion surgery. The prior UIV level was in the lumbar spine in 8 cases and the lower thoracic spine in 10. Seven of the 18 patients had distal extension to the sacrum as well as proximal extension to the LT spine. The prior LIV was L-4 in 4 cases and L-5 in 3 cases. Clinical Outcomes Intragroup Comparison. In the UT group, scores on the PCS, modi, and all SRS domains significantly improved from preoperative to final follow-up; however, there was no significant improvement in the SF-12 mental component summary (MCS) score. On the other hand, in the LT group, all clinical outcomes significantly improved from preoperative to final follow-up (Table 2). Intergroup Comparison. The preoperative SF-12 physical component summary (PCS) score was significantly higher in the UT group than in the LT group (34 vs 29; p = 0.03). There was no significant difference in preoperative MCS, modi, or any SRS domain except for the SRS mental health domain. At final follow-up, there was no significant difference in PCS, MCS, modi, or any SRS domain between the 2 groups (Table 3). Changes in Clinical Outcome Parameters Comparing preoperative values and values obtained at the final follow-up assessment, the change in PCS score was 6.7 in the UT group and 7.3 in the LT group (p = 0.8). The change in MCS score was 4.2 in the UT group and 4.4 in the LT group (p = 1.0). The change in modi was 13.3 in the UT group and 16.1 in the LT group (p = 0.6). With respect to the SRS domains, only the change in satisfaction with management domain was significantly greater in the UT group than in the LT group (p = 0.01). Radiographic Data Eleven patients underwent revision surgery with instrumentation replacement within 2 years (4 patients in the UT group and 7 in the LT group). The other 69 patients radiographic data were analyzed (Table 4) (Fig. 2). Significant differences were seen in the preoperative radiographic parameters between the UT and LT groups for lumbar lordosis (UT, 28 vs LT, 18 ; p = 0.03), thoracic kyphosis (UT, 36 vs LT, 18 ; p = 0.001), cervical lordosis (UT, 25 vs LT, 17 ; p = 0.02), and PI-LL (UT, 26 vs LT, 38 ; p = 0.02). There were no significant differences in other parameters. TABLE 2: Intragroup comparison of clinical outcomes Mean ± SD Group & Variable Preoperative Final p Value UT group (n = 31) SF-12 PCS 34 ± 8 41 ± 11 0.001* SF-12 MCS 39 ± 12 43 ± 12 0.1 modi 49 ± 16 35 ± 24 0.008* SRS function 2.6 ± 0.8 3.3 ± 0.7 0.002* SRS pain 2.5 ± 0.9 2.9 ± 0.8 0.02* SRS self-image 2.3 ± 0.6 3.6 ± 0.9 0.001* SRS mental health 2.9 ± 0.9 3.5 ± 1.0 0.003* SRS subtotal 2.6 ± 0.6 3.4 ± 0.7 0.001* SRS satisfaction 2.7 ± 0.9 4.2 ± 1.1 0.001* SRS total 2.6 ± 0.6 3.4 ± 0.7 0.001* LT group (n = 49) SF-12 PCS 29 ± 10 37 ± 11 0.001* SF-12 MCS 43 ± 14 48 ± 13 0.03* modi 50 ± 17 35 ± 18 0.001* SRS function 2.4 ± 0.6 3.3 ± 0.6 0.001* SRS pain 2.2 ± 0.8 3.1 ± 0.9 0.001* SRS self-image 2.7 ± 0.7 2.6 ± 0.7 0.001* SRS mental health 3.6 ± 0.9 4.0 ± 0.8 0.02* SRS subtotal 2.7 ± 0.6 3.5 ± 0.6 0.001* SRS satisfaction 3.3 ± 1.1 4.2 ± 0.9 0.001* SRS total 2.8 ± 0.6 3.5 ± 0.7 0.001* * Statistically significant (p < 0.05). After surgery, lumbar lordosis increased in both groups. Thoracic kyphosis decreased in the UT group but increased in the LT group. Finally, there was no significant difference in postoperative lumbar lordosis (UT, 46 vs LT, 42 ; p = 0.2) or in thoracic kyphosis (UT, 30 v. LT, 31 ; p = 0.6). Cervicothoracic kyphosis was signifi- TABLE 3: Intergroup comparison of improvement in clinical outcomes Variable UT Group (n = 31) Mean ± SD LT Group (n = 49) p Value SF-12 PCS 6.7 ± 10.6 7.3 ± 10.7 0.8 SF-12 MCS 4.2 ± 14.5 4.4 ± 13.8 1.0 modi 13.3 ± 20.7 16.1 ± 14.6 0.6 SRS function 1.1 ± 1.4 1.0 ± 1.2 0.7 SRS pain 0.8 ± 1.2 1.1 ± 1.3 0.4 SRS self-image 1.7 ± 1.3 1.1 ± 1.4 0.08 SRS mental health 1.1 ± 1.4 0.5 ± 1.4 0.1 SRS subtotal 1.2 ± 1.2 0.9 ± 1.2 0.2 SRS satisfaction 2.1 ± 1.5 1.1 ± 1.7 0.01* * Statistically significant (p < 0.05). Change greater than MCID. MCID was not available. 4 Neurosurg Focus / Volume 36 / May 2014

Level of upper instrumented vertebra in adult spinal deformity TABLE 4: Intergroup comparison of radiographic parameters Parameter UT Group (n = 27) Neurosurg Focus / Volume 36 / May 2014 LT Group (n = 42) p Value lumbar lordosis, T12 L5 ( ) preop 28 ± 18 18 ± 17 0.03* postop 46 ± 17 42 ± 11 0.2 final 42 ± 17 37 ± 16 0.2 thoracic kyphosis, T5 12 ( ) preop 36 ± 20 18 ± 16 0.001* postop 30 ± 15 31 ± 15 0.6 final 30 ± 17 31 ± 17 1.0 cervicothoracic kyphosis, C7 T5 ( ) preop 6 ± 7 9 ± 7 0.09 postop 17 ± 9 10 ± 9 0.002* final 20 ± 15 11 ± 13 0.001* cervical lordosis, C2 7 ( ) preop 25 ± 15 17 ± 11 0.02* postop 17 ± 14 14 ± 14 0.4 final 21 ± 16 15 ± 13 0.1 T-1 slope ( ) preop 34 ± 13 29 ± 12 0.07 postop 27 ± 12 24 ± 13 0.3 final 33 ± 15 27 ± 12 0.04* PJK angle ( ) preop 2 ± 5 4 ± 6 0.3 postop 11 ± 9 11 ± 9 0.8 final 11 ± 10 14 ± 10 0.3 SVA (mm) preop 104 ± 43 126 ± 75 0.3 postop 35 ± 49 47 ± 41 0.3 final 51 ± 56 73 ± 46 0.08 C-2 SVA (mm) preop 132 ± 47 146 ± 85 0.9 postop 61 ± 50 67 ± 46 0.6 final 82 ± 61 91 ± 47 0.5 T-1 Spi ( ) preop 2 ± 4 4 ± 9 0.6 postop 3 ± 4 2 ± 4 0.3 final 3 ± 5 1 ± 9 0.06 SS ( ) preop 23 ± 11 25 ± 14 0.7 postop 31 ± 11 31 ± 9 0.9 final 29 ± 11 30 ± 13 0.7 PT ( ) preop 30 ± 11 30 ± 10 0.9 postop 22 ± 10 26 ± 10 0.2 final 24 ± 12 29 ± 10 0.1 PI ( ) preop 54 ± 11 56 ± 11 0.4 (continued) TABLE 4: Intergroup comparison of radiographic parameters (continued) Parameter UT Group (n = 27) LT Group (n = 42) p Value PI-LL ( ) preop 26 ± 18 38 ± 17 0.02* postop 8 ± 14 15 ± 13 0.02* final 12 ± 16 19 ± 16 0.05 * Statistically significant (p < 0.05). cantly larger in the UT group than in the LT group (UT, 17 vs LT, 10 ; p = 0.002) (Fig. 3). PI-LL decreased in both groups (UT, 8 vs LT, 15 ; p = 0.02). PJK angle was increased by 9 in the UT group and 7 in the LT group (p = 0.4). At final follow-up, there were significant differences in cervicothoracic kyphosis (UT, 20 vs LT, 11 ; p = 0.001) and T-1 slope (UT, 33 vs LT, 27 ; p = 0.04). The UT group had a smaller positive SVA and a smaller T-1 Spi than the LT group (UT, 51 mm vs LT, 73 mm; p = 0.08; and UT, 2.6 vs LT, 0.6 ; p = 0.06) (Fig. 4). Preoperative PT was 30 in both groups (p = 0.9). Postoperatively, PT decreased to 22 in the UT group and 26 in the LT group. However, at the final follow-up, PT increased to 29 in the LT group; in the UT group, PT was 24 (p = 0.1). PI-LL was 12 in the UT group and 19 in the LT group (p = 0.05). Comparing sagittal curvatures immediately after surgery to those at final follow-up, there was a slight decrease in lumbar lordosis (UT, 4.5 vs LT, 4.7 ) (Fig. 5). Comparing the junctional measures immediately after surgery to those at final follow-up, the PJK angle had a greater increase in the LT group (2.8 ) than in the UT group (0.9 ), but this difference did not reach statistical significance (p = 0.4). A total increase in PJK angle from preoperative to final follow-up was almost the same in both groups (UT, 9.4 vs. LT, 9.7 ). Complications and Revision Surgery Complications are summarized in Table 5. Complication rates were not significantly different between the UT group and the LT group (overall complication UT, 63% vs LT, 73%, p = 0.2; major complication UT, 32% vs LT, 51%; p = 0.1). There was no significant difference in revision rate (42% vs 51%; p = 0.4). There was no mortality in either group. Causes for revision surgery in the UT group included implant failure (n = 6), pseudarthrosis (n = 5), deep wound infection (n = 3), and surgical PJK (n = 2). Causes for revision surgery in the LT group included pseudarthrosis (n = 9), deep wound infection (n = 6), surgical PJK (n = 5), nerve root irritation (n = 4), implant failure (n = 3), and epidural hematoma (n = 1). There was a significantly higher rate of dural tear in the LT group (n = 10; 20%) than in the UT group (n = 1; 3.2%) (p = 0.03). Of the 10 dural tears in the LT group, 5 occurred in patients with degenerative disease, 4 occurred in patients who had prior surgery, and 1 occurred in an idiopathic case. PJK was higher in the LT group (n = 20 [41%]) than 5

T. Fujimori et al. Fig. 2. Flow diagram of study participants. Eighty patients were enrolled. There were 31 patients in the UT group and 49 patients in the LT group. Four patients in the UT group and 7 patients in the LT group underwent revision surgery within 2 years. Fig. 3. Preoperative (left), postoperative (center), and final follow-up (right) standing lateral radiographs obtained in a woman who underwent fusion from the sacrum to T-4 at the age of 74 years. The final follow-up radiograph was obtained 5 years after surgery. The patient s PCS score improved from 33.5 to 44.6. Her SVA was 180 mm preoperatively and 25 mm at final follow-up. 6 Neurosurg Focus / Volume 36 / May 2014

Level of upper instrumented vertebra in adult spinal deformity Fig. 4. Preoperative (left), postoperative (center), and final follow-up (right) standing lateral radiographs obtained in a woman who underwent fusion from the sacrum to T-10 at the age of 64 years. The final follow-up radiograph was obtained 4.6 years after surgery. The patient s PCS score improved from 39 to 55. Her SVA was 78 mm preoperatively and 69.9 mm at final follow-up. in the UT group (n = 10 [32%]) (p = 0.4). In the UT group, 4 patients (13%) had symptomatic PJK not requiring surgery, compared with 8 (16%) in the LT group (p = 0.6). Five patients (10%) in the LT group had surgical PJK compared with 2 (6.4%) in the UT group (p = 0.6). Thus, the percentage of PJK cases requiring surgery was 25% in the LT group and 20% in the UT group. There was no significant difference in incidence of PJK or surgical PJK between the two groups. Discussion Compared with treatment of adolescent idiopathic scoliosis, the surgical treatment of adult spinal deformity is more challenging due to increased age, less flexible curvatures, sagittal deformity, and comorbid medical diseases. 3,10,18,23 Additionally, complications are not uncommon. Cho et al. reported that the complication rate for degenerative lumbar scoliosis was 68%. 2 At final follow-up, both the UT and LT groups Neurosurg Focus / Volume 36 / May 2014 showed significant improvements in PCS score, modi, and all SRS domains. The MCS score also significantly improved in the LT group but not in the UT group. The change in PCS score was 6.7 in the UT group and 7.3 in the LT group. The change in modi was 13.3 in the UT group and 16.1 in the LT group. The improvement in PCS score and modi exceeded the minimum clinically important difference (MCID) in both groups. Copay et al. reported that the MCID of lumbar surgery was 4.9 for PCS score and 12.8 for modi. 4 In terms of SRS, Carreon et al. reported that the MCID of adolescent scoliosis was 0.2 for the pain domain, 0.08 for the function domain, and 0.98 for the self-image domain. 1 In our study, improvements in all of these domains met MCID. Preoperative thoracic kyphosis was significantly greater in the UT group than in the LT group, suggesting that surgeons tended to fuse to the UT spine if thoracic kyphosis was significant. After surgery, thoracic kyphosis was almost the same magnitude (30 ) in both groups, and this was maintained at final follow-up. On the other hand, 7

T. Fujimori et al. Fig. 5. Line charts of radiographic parameters. The x-axis shows time course; the y-axis represents degrees or millimeters. The error bars indicate standard deviations. approximately 5 of correction loss occurred in the lumbar spine in both groups. This correction loss suggests that anterior support might be desirable in the lumbar spine in many patients with long fusion, because loss of correction in the distal spine could cause more sagittal imbalance than loss of correction in the proximal spine. Pelvic tilt (PT) is an important parameter that shows the compensatory mechanism of pelvic retroversion. In the LT group, PT decreased after surgery from 30 to 26 ; however, it had increased to 29, nearly the preoperative value, at the final follow-up. This suggested that compensatory function was still maximized in the LT group at the final follow-up. Global sagittal balance, as represented by final SVA and T-1 Spi, was also likely to be maintained in the UT group. The mean increase in PJK angle from the preoperative assessment to final follow-up was approximately 10 in both groups. Ninety percent of this change occurred immediately after surgery in the UT group, compared with 71% in the LT group. Hostin et al. reported a 5.6% incidence of acute PJK. 12 An advantage of fusion to the UT spine is that the UT segments surrounded by the rib cage and the scapulae are the most stabilized segments in the thoracic spine. 8 However, our data suggest that the stress caused by the deformity correction was distributed over fewer motion segments (only in the cervicothoracic spine) in the UT group than in the LT group (in the thoracic spine and cervicothoracic spine). This might be one of the reasons for the occurrence of surgical PJK in the UT group. On the other hand, because the LT spine is biomechanically more vulnerable than the UT spine, PJK was likely to occur in the LT group. One major consideration for spine surgeons and patients is how to reduce complication and reoperation rates. Our radiographic data suggest that the lumbar spine might be able to tolerate more rigid fixation than the thoracic spine. Therefore, an option might be to use a stiffer rod in the lumbar spine and a more elastic rod in the thoracic spine. Usage of a hook system rather than a pedicle screw at the UIV might be another way to reduce PJK. There were some limitations to our study. As this study was not randomized, the background between the 2 groups was different. The UT group included 3 times more patients with idiopathic etiology than the LT group, and the LT group included 2 times more patients with prior surgery. Given that the LT group had a significantly lower mean preoperative PCS score, surgeons might select fewer fusion segments in patients whose health con- 8 Neurosurg Focus / Volume 36 / May 2014

Level of upper instrumented vertebra in adult spinal deformity TABLE 5: Complications data Complication UT Group (n = 31) ditions were poor. These differences of the background and preferences of each surgeon might have effects on the different outcomes between the groups. Because of these limitations, our ability to generalize from our findings was restricted. Our data do not allow us to conclude whether fusion to the UT spine or LT spine is better. However, the fact that both groups had significant improvements in both clinical and radiographic parameters demonstrated the effectiveness of the long fusion surgery for adult spinal deformity. Further studies that include more patients more closely matched for comparison and with a longer follow-up period are desirable. Neurosurg Focus / Volume 36 / May 2014 LT Group (n = 49) p Value any complication 18 (63%) 36 (73%) 0.2 major complication 10 (32%) 24 (51%) 0.1 any revision surgery 13 (42%) 25 (51%) 0.4 intraop dural tear 1 (3.2%) 10 (20%) 0.03* nerve root irritation 0 5 0.07 lymphatic leak 1 0 0.5 skin or cutaneous nerve damage 1 2 0.8 immediate postop epidural hematoma 0 1 0.4 arrhythmia 4 3 0.3 meningitis 0 1 0.4 electrolyte abnormality 4 6 0.9 delirium 0 2 0.2 pulmonary edema 5 2 0.06 deep vein thrombosis 1 0 0.2 urinary tract infection 1 0 0.2 post-discharge deep wound infection 3 (9.7%) 6 (12%) 0.7 superficial wound infection 2 (6.4%) 0 (0%) 0.07 pseudarthrosis 5 (16%) 9 (18%) 0.8 PJK 10 (32%) 20 (41%) 0.4 asymptomatic (radiographic) 6 (19%) 12 (24%) 0.6 nonsurgical symptomatic PJK 4 (13%) 8 (16%) 0.7 surgical PJK 2 (6.4%) 5 (10%) 0.6 implant failure 6 (19%) 3 (6%) 0.07 * Statistically significant (p < 0.05). Any complication from the intraoperative period to final follow-up. Complications needing active medical intervention or return to the operating room. Four patients required revision surgery. Rod breakage, screw prominence, screw pullout, and disconnection of rod and screw. Conclusions Patients in both the UT group and the LT group demonstrated significant improvements in health-related outcome measures after long fusion surgery for adult spinal deformity. However, surgeons and patients need to consider the potential complications of the surgery. With respect to maintaining radiographic global balance and limiting the risk of PJK, the UT group had advantages over the LT group. However, the rate of implant failure was high in the UT group. Fusion to the UT spine is specifically indicated in patients with thoracic hyperkyphosis, patients at high risk for acute PJK, and patients with untreated idiopathic scoliosis with major thoracic curves. In most other cases, the LT spine would be the primary choice for location of the UIV, and if failure occurred, revision surgery to the UT spine would be a realistic salvage strategy. Further innovative fusion techniques or devices that can distribute the correction stress may be necessary to maintain the improvement over the long term. Acknowledgment We thank Linda Racine, CCRP, for helping with data collection. Disclosure This study was supported by a grant-in-aid from the Nakatomi Foundation. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript. Author contributions to the study and manuscript preparation include the following. Conception and design: Fujimori, Le, Tay, Burch, Hu. Acquisition of data: Fujimori, Inoue, Le, Schairer, Berven, Tay, Deviren, Burch, Hu. Analysis and interpretation of data: Fujimori, Inoue, Schairer, Iwasaki, Hu. Drafting the article: Fujimori, Inoue, Schairer, Deviren, Hu. Critically revising the article: Fujimori, Inoue, Schairer, Berven, Tay, Iwasaki, Hu. Reviewed submitted version of manuscript: Fujimori, Berven, Tay, Deviren, Burch, Iwasaki, Hu. Approved the final version of the manuscript on behalf of all authors: Fujimori. Statistical analysis: Fujimori, Schairer, Hu. Administrative/technical/material support: Hu. Study supervision: Schairer, Berven, Hu. References 1. Carreon LY, Sanders JO, Diab M, Sucato DJ, Sturm PF, Glassman SD: The minimum clinically important difference in Scoliosis Research Society-22 Appearance, Activity, And Pain domains after surgical correction of adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 35:2079 2083, 2010 2. Cho KJ, Suk SI, Park SR, Kim JH, Kim SS, Choi WK, et al: Complications in posterior fusion and instrumentation for degenerative lumbar scoliosis. Spine (Phila Pa 1976) 32:2232 2237, 2007 3. Cho W, Mason JR, Smith JS, Shimer AL, Wilson AS, Shaffrey CI, et al: Failure of lumbopelvic fixation after long construct fusions in patients with adult spinal deformity: clinical and radiographic risk factors. Clinical article. J Neurosurg Spine 19:445 453, 2013 4. Copay AG, Glassman SD, Subach BR, Berven S, Schuler TC, Carreon LY: Minimum clinically important difference in lumbar spine surgery patients: a choice of methods using the Oswestry Disability Index, Medical Outcomes Study questionnaire Short Form 36, and pain scales. Spine J 8:968 974, 2008 5. Devlin VJ, Boachie-Adjei O, Bradford DS, Ogilvie JW, Transfeldt EE: Treatment of adult spinal deformity with fusion to the sacrum using CD instrumentation. J Spinal Disord 4:1 14, 1991 6. Emami A, Deviren V, Berven S, Smith JA, Hu SS, Bradford 9

T. Fujimori et al. DS: Outcome and complications of long fusions to the sacrum in adult spine deformity: Luque-Galveston, combined iliac and sacral screws, and sacral fixation. Spine (Phila Pa 1976) 27: 776 786, 2002 7. Fritz JM, Irrgang JJ: A comparison of a modified Oswestry Low Back Pain Disability Questionnaire and the Quebec Back Pain Disability Scale. Phys Ther 81:776 788, 2001 8. Fujimori T, Iwasaki M, Nagamoto Y, Ishii T, Kashii M, Murase T, et al: Kinematics of the thoracic spine in trunk rotation: in vivo 3-dimensional analysis. Spine (Phila Pa 1976) 37:E1318 E1328, 2012 9. Glassman SD, Bridwell K, Dimar JR, Horton W, Berven S, Schwab F: The impact of positive sagittal balance in adult spinal deformity. Spine (Phila Pa 1976) 30:2024 2029, 2005 10. Glassman SD, Hamill CL, Bridwell KH, Schwab FJ, Dimar JR, Lowe TG: The impact of perioperative complications on clinical outcome in adult deformity surgery. Spine (Phila Pa 1976) 32:2764 2770, 2007 11. Glattes RC, Bridwell KH, Lenke LG, Kim YJ, Rinella A, Edwards C II: Proximal junctional kyphosis in adult spinal deformity following long instrumented posterior spinal fusion: incidence, outcomes, and risk factor analysis. Spine (Phila Pa 1976) 30:1643 1649, 2005 12. Hostin R, McCarthy I, O Brien M, Bess S, Line B, Boachie- Adjei O, et al: Incidence, mode, and location of acute proximal junctional failures following surgical treatment for adult spinal deformity. Spine (Phila Pa 1976) [epub ahead of print], 2013 13. Kim YJ, Bridwell KH, Lenke LG, Rhim S, Cheh G: Pseudarthrosis in long adult spinal deformity instrumentation and fusion to the sacrum: prevalence and risk factor analysis of 144 cases. Spine (Phila Pa 1976) 31:2329 2336, 2006 14. Knott PT, Mardjetko SM, Techy F: The use of the T1 sagittal angle in predicting overall sagittal balance of the spine. Spine J 10:994 998, 2010 15. Kostuik JP, Hall BB: Spinal fusions to the sacrum in adults with scoliosis. Spine (Phila Pa 1976) 8:489 500, 1983 16. Kuhns CA, Bridwell KH, Lenke LG, Amor C, Lehman RA, Buchowski JM, et al: Thoracolumbar deformity arthrodesis stopping at L5: fate of the L5-S1 disc, minimum 5-year followup. Spine (Phila Pa 1976) 32:2771 2776, 2007 17. Maeda T, Buchowski JM, Kim YJ, Mishiro T, Bridwell KH: Long adult spinal deformity fusion to the sacrum using rh- BMP-2 versus autogenous iliac crest bone graft. Spine (Phila Pa 1976) 34:2205 2212, 2009 18. O Shaughnessy BA, Bridwell KH, Lenke LG, Cho W, Baldus C, Chang MS, et al: Does a long-fusion T3-sacrum portend a worse outcome than a short-fusion T10-sacrum in primary surgery for adult scoliosis? Spine (Phila Pa 1976) 37:884 890, 2012 19. Schwab F, Lafage V, Patel A, Farcy JP: Sagittal plane considerations and the pelvis in the adult patient. Spine (Phila Pa 1976) 34:1828 1833, 2009 20. Schwab F, Ungar B, Blondel B, Buchowski J, Coe J, Deinlein D, et al: Scoliosis Research Society-Schwab adult spinal deformity classification: a validation study. Spine (Phila Pa 1976) 37:1077 1082, 2012 21. Schwab FJ, Lafage V, Farcy JP, Bridwell KH, Glassman S, Shainline MR: Predicting outcome and complications in the surgical treatment of adult scoliosis. Spine (Phila Pa 1976) 33:2243 2247, 2008 22. Silva FE, Lenke LG: Adult degenerative scoliosis: evaluation and management. Neurosurg Focus 28(3):E1, 2010 23. Yadla S, Maltenfort MG, Ratliff JK, Harrop JS: Adult scoliosis surgery outcomes: a systematic review. Neurosurg Focus 28(3):E3, 2010 Manuscript submitted December 22, 2013. Accepted March 20, 2014. Please include this information when citing this paper: DOI: 10.3171/2014.3.FOCUS13541. Address correspondence to: Takahito Fujimori, M.D., M.Sc., De partment of Orthopaedic Surgery, Osaka University Graduate School of Medicine, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan. email: takahito-f@hotmail.co.jp. 10 Neurosurg Focus / Volume 36 / May 2014