Faculty Bibliography

BACKGROUND AND OBJECTIVES/OBJECTIVE:The spectrum of patients requiring adult spinal deformity (ASD) surgery is highly variable in baseline (BL) risk such as age, frailty, and deformity severity. Although improvements have been realized in ASD surgery over the past decade, it is unknown whether these carry over to high-risk patients. We aim to determine temporal differences in outcomes at 2 years after ASD surgery in patients stratified by BL risk. METHODS:Patients ≥18 years with complete pre- (BL) and 2-year (2Y) postoperative data from 2009 to 2018 were categorized as having undergone surgery from 2009 to 2013 [early] or from 2014 to 2018 [late]. High-risk [HR] patients met ≥2 of the criteria: (1) ++ BL pelvic incidence and lumbar lordosis or SVA by Scoliosis Research Society (SRS)-Schwab criteria, (2) elderly [≥70 years], (3) severe BL frailty, (4) high Charlson comorbidity index, (5) undergoing 3-column osteotomy, and (6) fusion of >12 levels, or >7 levels for elderly patients. Demographics, clinical outcomes, radiographic alignment targets, and complication rates were assessed by time period for high-risk patients. RESULTS:Of the 725 patients included, 52% (n = 377) were identified as HR. 47% (n = 338) had surgery pre-2014 [early], and 53% (n = 387) underwent surgery in 2014 or later [late]. There was a higher proportion of HR patients in Late group (56% vs 48%). Analysis by early/late status showed no significant differences in achieving improved radiographic alignment by SRS-Schwab, age-adjusted alignment goals, or global alignment and proportion proportionality by 2Y (all P > .05). Late/HR patients had significantly less poor clinical outcomes per SRS and Oswestry Disability Index (both P < .01). Late/HR patients had fewer complications (63% vs 74%, P = .025), reoperations (17% vs 30%, P = .002), and surgical infections (0.9% vs 4.3%, P = .031). Late/HR patients had lower rates of early proximal junctional kyphosis (10% vs 17%, P = .041) and proximal junctional failure (11% vs 22%, P = .003). CONCLUSION/CONCLUSIONS:Despite operating on more high-risk patients between 2014 and 2018, surgeons effectively reduced rates of complications, mechanical failures, and reoperations, while simultaneously improving health-related quality of life.

STUDY DESIGN/METHODS:Retrospective cohort study of prospectively enrolled cervical deformity patients. OBJECTIVE:To investigate patients in whom fusion to the thoracic spine was warranted. SUMMARY OF BACKGROUND DATA/BACKGROUND:Thoracolumbar malalignment is often seen in patients presenting with cervical deformities. For cervical deformity (CD) patients, it is not always clear where in the thoracic spine to end the construct. METHODS:Patients were stratified into upper and lower thoracic groups: T1-T4 [Short Fusion], beyond T4 [Long Fusion]. Optimal outcome (OO) at 2 years postop was defined as: (1) did not have DJF and (2) had Virk et al good clinical outcome [≥2 of the following: NDI <20 or meeting MCID, mild myelopathy (mJOA ≥14), NRS-Neck ≤5 or improved by ≥2 points from baseline]. Univariate analysis compared variables between short and long fusion groups. Multivariate analysis analyzed associations between groups and factors related to treatment success and failures. RESULTS:One hundred forty-four cervical deformity patients were included (60.3±9.0 y, 60% F, 29.4±7.6 kg/m2, levels fused: 7.8±3.2). Eighty-two percent of patients had short fusions, whereas 18% had long fusions. 44.4% met the optimal outcome criteria, with no difference by fusion length (P=0.171). Factors associated with achieving OO in long fusions: baseline sacral slope ≤33.5 degrees (OR: 15.0), not undergoing high-grade osteotomy (OR: 12.3) and being Ames descriptor type C (OR: 13.5); all P<0.05. Factors associated with failure to achieve OO in short fusions: levels fused >6 (OR: 4.3), Ames descriptor type CT (OR: 11.5), Ames cSVA modifier grade 1 or 2 at BL (OR: 4.56), and Flatneck sagittal morphotype (OR: 4.5); all P<0.05. CONCLUSION/CONCLUSIONS:The choice of lower instrumented vertebra (LIV) in cervical deformity fusions may be influenced by multiple factors. In patients with global malalignment, severe frailty, flatneck deformity morphotype, and Ames type CT descriptor types, LIVs beyond T4 are associated with treatment success. LEVEL OF EVIDENCE/METHODS:Level III.

STUDY DESIGN/METHODS:Retrospective analysis of prospectively collected data. OBJECTIVE:This study evaluates the impact of knee osteoarthritis (OA) and knee arthroplasty on alignments and patient-reported outcomes measures (PROMS) of patients undergoing adult spinal deformity (ASD) corrective surgery. BACKGROUND:The relationship between knee OA and spinal alignment in patients with ASD is incompletely understood. It is also unknown how patients with knee arthroplasty and ASD compare to ASD patients with native knees. METHODS:Baseline full-body radiographs were used, and hip and knee OA were graded by two independent reviewers using the KL classification. Spinopelvic parameters and PROMs were compared across the different knee OA groups and compared between patients with knee replacement and native knees. RESULTS:199 patients with bilateral non severe OA (G1), 31 patients with unilateral severe knee OA (G2), and 60 patients with bilateral severe knee OA (G3). Patients with severe knee OA presented with worse spinopelvic parameters. However, after multivariable regression analysis controlling for age, frailty, PI, T1PA, knee OA was an independent predictor of knee flexion (G1:-0.02±7.3, G2: 7.8±9.4, G3: 4.5±8.7, P<0.001), and ankle dorsiflexion (G1: 2.3±4.0, G2: 6.6±4.5, G3: 5.1±4.1, P<0.001). There was no difference in PROMs (P>0.05). Secondary analysis included 96 patients: 48 patients (50%) with non-severe knee OA, and 48 patients (50%) with knee replacement. There was no difference in radiographic parameters or PROMs between the groups. CONCLUSION/CONCLUSIONS:In this study of complex ASD patients, patients with worse spinal deformity were more likely to have concomitant knee OA. Knee OA was shown to be a predictor of knee flexion and ankle dorsiflexion angles, but was not associated with worse PROMs in this study population. Patients with knee arthroplasty, however, had comparable spinal alignment and PROMs relative to those with mild OA.

BACKGROUND AND OBJECTIVES/OBJECTIVE:To investigate the cost-effectiveness and impact of prophylactic techniques on the development of proximal junctional kyphosis (PJK) and proximal junctional failure (PJF) in the context of postoperative alignment. METHODS:Adult spinal deformity patients with fusion to pelvis and 2-year data were included. Patients receiving PJK prophylaxis (hook, tether, cement, minimally-invasive surgery approach) were compared to those who did not. These cohorts were further stratified into "Matched" and "Unmatched" groups based on achievement of age-adjusted alignment criteria. Costs were calculated using the Diagnosis-Related Group costs accounting for PJK prophylaxis, postoperative complications, outpatient health care encounters, revisions, and medical-related readmissions. Quality-adjusted life years were calculated using Short Form-36 converted to Short-Form Six-Dimension (SF-6D) and used an annual 3% discount rate. Multivariate analysis controlling for age, sex, levels fused, and baseline deformity severity assessed outcomes of developing PJK/PJF if matched and/or with use of PJK prophylaxis. RESULTS:A total of 738 adult spinal deformity patients met inclusion criteria (age: 63.9 ± 9.9, body mass index: 28.5 ± 5.7, Charlson comorbidity index: 2.0 ± 1.7). Multivariate analysis revealed patients corrected to age-adjusted criteria postoperatively had lower rates of developing PJK or PJF (odds ratio [OR]: 0.4, [0.2-0.8]; P = .011) with the use of prophylaxis. Among those unmatched in T1 pelvic angle, pelvic incidence lumbar lordosis mismatch, and pelvic tilt, prophylaxis reduced the likelihood of developing PJK (OR: 0.5, [0.3-0.9]; P = .023) and PJF (OR: 0.1, [0.03-0.5]; P = .004). Analysis of covariance analysis revealed patients matched in age-adjusted alignment had better cost-utility at 2 years compared with those without prophylaxis ($361 539.25 vs $419 919.43; P < .001). Patients unmatched in age-adjusted criteria also generated better cost ($88 348.61 vs $101 318.07; P = .005) and cost-utility ($450 190.80 vs $564 108.86; P < .001) with use of prophylaxis. CONCLUSION/CONCLUSIONS:Despite additional surgical cost, the optimization of radiographic realignment in conjunction with prophylaxis of the proximal junction appeared to be a more cost-effective strategy, primarily because of the minimization of reoperations secondary to mechanical failure. Even among those not achieving optimal alignment, junctional prophylactic measures were shown to improve cost efficiency.

PURPOSE/OBJECTIVE:To determine if iatrogenic posterior translation (UIV SPi) at the upper instrumented vertebrae (UIV) is associated with increased mechanical complications and secondarily to generate and validate a UIV SPi threshold for increased complications. METHODS:Two patient databases were utilized: one for generating a UIV SPi threshold and another for validation. Patients with a UIV between T8-L1 and a LIV at ilium were included. A receiver operating curve (ROC) curve analyses was performed to generate a threshold that predicted proximal junctional complications. This UIV SPi angle (-16.0°) was rounded to -15.0° for practical clinical use and validated in a separate cohort. Patients were stratified as above (most translated, MT) or below (least translated, LT) the threshold for comparative demographic and outcomes analyses. RESULTS:Generation of the threshold on 192 patients (122 LT, 70 MT) revealed that the MT group had higher absolute postoperative UIV SVA (MT=-56.1 ± 23.1 mm vs. LT=-10.4 ± 31.8 mm, p < 0.001), higher PT (25.7° vs. 19.3°, p < 0.001), and 2.8-5.8 times greater odds of postoperative proximal junctional complications at 2-years (p < 0.05). Validation on 135 patients (95 LT, 40 MT) revealed that the MT group had 11.7 times greater odds of radiographic PJK and had 4.5 times greater odds of all-cause reoperations (p < 0.05). CONCLUSION/CONCLUSIONS:Patients with UIV posterior translation, despite similar PI-LL and T1PA, exhibit a high PT and experience higher odds of proximal junctional complications. Our findings support limiting the UIV SPi to < 15° of posterior translation to mitigate postoperative mechanical complications. LEVEL OF EVIDENCE/METHODS:IV.

OBJECTIVE:Malalignment following cervical spine deformity (CSD) surgery can negatively impact outcomes and increase complications. Despite the growing ability to plan alignment, it remains unclear whether preoperative goals are achieved with surgery. The objective of this study was to assess how good surgeons are at achieving their preoperative goal alignment following CSD surgery. METHODS:Adult patients with CSD were prospectively enrolled into a multicenter registry. Surgeons documented alignment goals prior to surgery, including C2-7 sagittal vertical axis (SVA), C2-7 sagittal Cobb angle, T1 slope minus cervical lordosis (TS-CL), and C7-S1 SVA. Goals were compared with achieved alignment, and the offsets (achieved goal) were calculated. General linear models were created for offset magnitude for each alignment parameter, controlling for baseline deformity and surgical factors. RESULTS:The 88 enrolled patients had a mean age of 63.6 ± 13.0 years. The mean number of anterior and posterior instrumented levels was 3.5 ± 1.0 and 10.6 ± 4.5, respectively. Surgeons failed to achieve their preoperative alignment goals by an average of 17.2 (range 0.1-75.4) mm for C2-7 SVA, 10.3° (range 0.1°-45.5°) for C2-7 sagittal Cobb angle, 15.6° (range 0.0°-42.9°) for TS-CL, and 34.2 (range 0.3-113.7) mm for C7-S1 SVA. The sagittal alignment parameters with the highest rate of extreme outliers were TS-CL and C7-S1 SVA, with 32.2% exceeding 20° and 60.8% exceeding 20 mm from goal alignment, respectively. After controlling for baseline deformity and operative parameters, the only factor associated with achieving targeted alignment for C2-7 sagittal Cobb angle was greater baseline thoracic kyphosis (TK; B = -0.148, 95% CI -0.288 to -0.007, p = 0.040), and for TS-CL, the only associated factor was lower baseline TS-CL (B = 0.187, 95% CI 0.027-0.347, p = 0.022). Both lower TK and greater TS-CL may reflect increased baseline deformity through greater thoracic compensation and increased TS-CL mismatch, respectively. No significant associations were identified for C2-7 SVA and C7-S1 SVA. CONCLUSIONS:Surgeons failed to achieve their preoperative alignment goals by an average of 17.2 mm for C2-7 SVA, 10.3° for C2-7 sagittal Cobb angle, 15.6° for TS-CL, and 34.2 mm for C7-S1 SVA. The few factors identified that were associated with offset between goal and achieved alignment suggest that achievement of goal alignment was most challenging for more severe deformities. Further advancements are needed to enable more consistent translation of preoperative alignment goals into the operating room for adult CSD correction. Clinical trial registration no.: NCT01588054 (ClinicalTrials.gov).

BACKGROUND:Early-term complications may not predict long-term success after adult cervical deformity (ACD) correction. OBJECTIVE:Evaluate whether optimal realignment results in similar rates of perioperative complications but achieves longer-term cost-utility. STUDY DESIGN/METHODS:Retrospective cohort study. METHODS:ACD patients with 2-year data included. Outcomes: distal junctional failure (DJF), good clinical outcome (GCO):[Meeting 2 of 3: (1) NDI>20 or meeting MCID, (2) mJOA≥14, (3)NRS-Neck improved≥2]. Ideal Outcome defined as GCO without DJF or reoperation. Patient groups were stratified by correction to 'Optimal radiographic outcome', defined by cSVA 9 (<40 mm) AND TS-CL (<15 deg) upon correction. Cost calculated by CMS.com definitions, and cost-per-QALY was calculated by converting NDI to SF-6D. Multivariable analysis controlling for age, baseline T1-slope, cSVA, disability, and frailty, was used to assess complication rates, clinical outcomes, and cost-utility based on meeting optimal radiographic outcome. RESULTS:One hundred forty-six patients included: 52 optimal radiographic realignment (O) and 94 not optimal (NO). NO group presented with higher cSVA and T1-slope. Adjusted analysis showed O group suffered similar 90-day complications (P>0.8), but less DJK, DJF (0% vs. 18%; P<0.001) and reoperations (18% vs. 35%; P=0.02). Patients meeting optimal radiographic criteria more often met Ideal outcome [odds ratio: 2.2, (1.1-4.8); P=0.03]. Despite no differences in overall cost, O group saw greater clinical improvement, translating to a better cost-utility [mean difference: $91,000, ($49,000-$132,000); P<0.001]. CONCLUSION/CONCLUSIONS:Despite similar perioperative courses, patients optimally realigned experienced less junctional failure, leading to better cost-utility compared with those sub-optimally realigned. Perioperative complication risk should not necessarily preclude optimal surgical intervention, and policy efforts might better focus on long-term outcome measures in adult cervical deformity surgery. LEVEL OF EVIDENCE/METHODS:Level III.

STUDY DESIGN/METHODS:Retrospective analysis of prospectively collected data. OBJECTIVE:Evaluate the impact of prior cervical constructs on upper instrumented vertebrae (UIV) selection and postoperative outcomes among patients undergoing thoracolumbar deformity correction. BACKGROUND:Surgical planning for adult spinal deformity (ASD) patients involves consideration of spinal alignment and existing fusion constructs. METHODS:ASD patients with (ANTERIOR or POSTERIOR) and without (NONE) prior cervical fusion who underwent thoracolumbar fusion were included. Demographics, radiographic alignment, patient-reported outcome measures (PROMs), and complications were compared. Univariate and multivariate analyses were performed on POSTERIOR patients to identify parameters predictive of UIV choice and to evaluate postoperative outcomes impacted by UIV selection. RESULTS:Among 542 patients, with 446 NONE, 72 ANTERIOR, and 24 POSTERIOR patients, mean age was 64.4 years and 432 (80%) were female. Cervical fusion patients had worse preoperative cervical and lumbosacral deformity, and PROMs (P<0.05). In the POSTERIOR cohort, preoperative LIV was frequently below the cervicothoracic junction (54%) and uncommonly (13%) connected to the thoracolumbar UIV. Multivariate analyses revealed that higher preoperative cervical SVA (coeff=-0.22, 95%CI=-0.43--0.01, P=0.038) and C2SPi (coeff=-0.72, 95%CI=-1.36--0.07, P=0.031), and lower preoperative thoracic kyphosis (coeff=0.14, 95%CI=0.01-0.28, P=0.040) and thoracolumbar lordosis (coeff=0.22, 95%CI=0.10-0.33, P=0.001) were predictive of cranial UIV. Two-year postoperatively, cervical patients continued to have worse cervical deformity and PROMs (P<0.05) but had comparable postoperative complications. Choice of thoracolumbar UIV below or above T6, as well as the number of unfused levels between constructs, did not affect patient outcomes. CONCLUSIONS:Among patients who underwent thoracolumbar deformity correction, prior cervical fusion was associated with more severe spinopelvic deformity and PROMs preoperatively. The choice of thoracolumbar UIV was strongly predicted by their baseline cervical and thoracolumbar alignment. Despite their poor preoperative condition, these patients still experienced significant improvements in their thoracolumbar alignment and PROMs after surgery, irrespective of UIV selection. LEVEL OF EVIDENCE/METHODS:IV.

PURPOSE/OBJECTIVE:To evaluate the variability in intraoperative fluid management during adult spinal deformity (ASD) surgery, and analyze the association with complications, intensive care unit (ICU) requirement, and length of hospital stay (LOS). METHODS:Multicenter comparative cohort study. Patients ≥ 18 years old and with ASD were included. Intraoperative intravenous (IV) fluid data were collected including: crystalloids, colloids, crystalloid/colloid ratio (C/C), total IV fluid (tIVF, ml), normalized total IV fluid (nIVF, ml/kg/h), input/output ratio (IOR), input-output difference (IOD), and normalized input-output difference (nIOD, ml/kg/h). Data from different centers were compared for variability analysis, and fluid parameters were analyzed for possible associations with the outcomes. RESULTS:Seven hundred ninety-eight patients with a median age of 65.2 were included. Among different surgical centers, tIVF, nIVF, and C/C showed significant variation (p < 0.001 for each) with differences of 4.8-fold, 3.7-fold, and 4.9-fold, respectively. Two hundred ninety-two (36.6%) patients experienced at least one in-hospital complication, and ninety-two (11.5%) were IV fluid related. Univariate analysis showed significant relations for: LOS and tIVF (ρ = 0.221, p < 0.001), IOD (ρ = 0.115, p = 0.001) and IOR (ρ = -0.138, p < 0.001); IV fluid-related complications and tIVF (p = 0.049); ICU stay and tIVF, nIVF, IOD and nIOD (p < 0.001 each); extended ICU stay and tIVF (p < 0.001), nIVF (p = 0.010) and IOD (p < 0.001). Multivariate analysis controlling for confounders showed significant relations for: LOS and tIVF (p < 0.001) and nIVF (p = 0.003); ICU stay and IOR (p = 0.002), extended ICU stay and tIVF (p = 0.004). CONCLUSION/CONCLUSIONS:Significant variability and lack of standardization in intraoperative IV fluid management exists between different surgical centers. Excessive fluid administration was found to be correlated with negative outcomes. LEVEL OF EVIDENCE/METHODS:III.

BACKGROUND:Recombinant human bone morphogenetic protein-2 (rhBMP-2) has not shown superior benefit overall in cost-effectiveness during adult spinal deformity (ASD) surgery. STUDY DESIGN/SETTING/METHODS:Retrospective PURPOSE: Generate a risk score for pseudarthrosis to inform the utilization of rhBMP-2, balancing costs against quality of life and complications. METHODS:ASD patients with 3-year data were included. Quality of life gained was calculated from ODI to SF-6D and translated to quality-adjusted life years (QALYs). Cost was calculated using the PearlDiver database and CMS definitions for complications and comorbidities. Established weights were generated for predictive variables via logistic regression to yield a predictive risk score for pseudarthrosis that accounted for frailty, diabetes, depression, ASA grade, thoracolumbar kyphosis and three-column osteotomy use. Risk score categories, established via conditional inference tree (CIT)-derived thresholds were tested for cost-utility of rhBMP-2 usage, controlling for age, prior fusion, and baseline deformity and disability. RESULTS:64% of ASD patients received rhBMP-2 (308/481). There were 17 (3.5%) patients that developed pseudarthrosis. rhBMP-2 use overall did not lower pseudarthrosis rates (OR: 0.5, [0.2-1.3]). Pseudarthrosis rates for each risk category were: No Risk (NoR) 0%; Low-Risk (LowR) 1.6%; Moderate Risk (ModR) 9.3%; High-Risk (HighR) 24.3%. Patients receiving rhBMP-2 had similar QALYs overall to those that did not (0.163 vs. 0.171, p = .65). rhBMP-2 usage had worse cost-utility in the LowR cohort (p < .001). In ModR patients, rhBMP-2 usage had equivocal cost-utility ($53,398 vs. $61,581, p = .232). In the HighR cohort, the cost-utility was reduced via rhBMP-2 usage ($98,328 vs. $211,091, p < .001). CONCLUSION/CONCLUSIONS:Our study shows rhBMP-2 demonstrates effective cost-utility for individuals at high risk for developing pseudarthrosis. The generated score can aid spine surgeons in the assessment of risk and enhance justification for the strategic use of rhBMP-2 in the appropriate clinical contexts. LEVEL OF EVIDENCE/METHODS:III.