Faculty Bibliography

BACKGROUND CONTEXT: Hospital acquired conditions (HACs) were established in the Affordable Care Act, and are defined as reasonably preventable complications that are nonreimbursable. In high risk patient populations for HACs, such as frail surgical spine patients, preoperative evaluation should consider modifiable factors. PURPOSE: To identify if optimizing the modifiable factors in the frailty index reduce the risk of developing HACs in population of surgical spine patients. STUDY DESIGN/SETTING: Retrospective cohort study. PATIENT SAMPLE: This study included 196,523 elective spine surgery patients in the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP). OUTCOME MEASURES: Modifiable patient frailty factors: hypertension and diabetes mellitus; rates of HACs, superficial or deep surgical site infection (SSI), deep venous thromboembolism (VTE) and urinary tract infection (UTI), across frailty scores.
METHOD(S): Patients >18 years who underwent elective spine surgery were identified in ACS-NSQIP database from 2005-2016. HACs identified: SSI, VTE, and UTI. Patient frailty was assessed utilizing the modified NSQIP 5-factor frailty index. The modified frailty score (mFI-5) is assessed on a scale between 0 and 1 (no frailty [NF] <0.3, mild frailty [MF] 0.3-0.5, severe frailty [SF] >0.5). Descriptive analysis quantified rates of patient characteristics, operative details, and HAC prevalence. Stepwise linear regression models determined significant predictors of baseline frailty using controllable patient factors. HACs were compared between 'optimized' and 'non-optimized' frailty status in the cohort. Level of significance was set to P<0.05.
RESULT(S): A total of 196,523 patients (57+/-14.2 years, 30.4 kg/m², 48%F) met inclusion criteria. Overall, 5,720 (2.9%) of patients developed an HAC, the most common was an SSI (1.3%), followed by UTI (1.1%). When stratified by the mFI-5 frailty severity groups at baseline, 83.6% of patients were categorized NF, 15.1% MF and 1.3% SF. Within the frailty severity groups, prevalence of overall HACs increased significantly (NF: 2.64%, MF: 4.17%, 5.93%, p<0.001). Rates of all individual postoperative HACs assessed also increased with greater baseline frailty severity: SSI (NF: 1.14%, MF: 1.93%, SF: 2.39%, p<0.001), UTI (NF: 0.91%, MF: 1.66%, SF: 2.85%, p<0.001), VTE (NF: 0.68%, MF: 0.80%, SF: 1.16%, p=0.002). Stepwise linear regression models determined that diabetes mellitus (beta = 0.493) and hypertension (beta = 0.679) were the most significant predictors for increased baseline frailty by way of the mFI-5 NSQIP index (Final model: R2= 0.897). Of total patients, 47.2% had the optimal modifiable frailty factors (no history of diabetes or hypertension). The optimal frailty patients had significantly less overall incidence of SSI (2.03% vs 2.5%, p<0.001), UTI (0.65% vs 1.4%, p<0.001), DVT (0.56% vs 0.84%, p<0.001), and any overall HAC (2.18% vs 3.56%, p<0.001).
CONCLUSION(S): Stepwise linear regression models determined that hypertension and diabetes account for 89.7% of variance in baseline mFI-5 score. Patients with these optimal controllable factors had reduced incidence of all hospital acquired conditions. In order to optimize hospital resources and treatment outcomes, physicians and patients should be aware of the modifiable factors that contribute to a patient's frailty that can ultimately impact acquiring HACs. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: The Scoliosis Research Society established appropriateness criteria for surgery for degenerative lumbar scoliosis (DLS) in order to improve and unify clinical decision making. Though utilized and validated in various cohorts, an appropriateness criteria has yet to be developed in a cervical deformity (CD) population. PURPOSE: To modify the DLS Appropriateness criteria to be utilized in a CD cohort. STUDY DESIGN/SETTING: Retrospective review of a single-surgeon CD database PATIENT SAMPLE: A total of 100 CD patients. OUTCOME MEASURES: Appropriateness criteria.
METHOD(S): CD patients>18yrs were included. Each patient was scored based upon the SRS-appropriateness criteria, comprised of clinical or radiographic characteristics and was modified for a cervical deformity cohort: (1) Severity of Symptoms [NDI], (2) Severity of Myelopathy[mJOA.], (3) Progression of Deformity, (4) Global Sagittal Malalignment [Schwab modifiers], (5) Severity of Risk Factors, (6) Degree of the T1S-CL curvature. For the category 'progression', radiographs were only available at one preoperative time point, and we were unable to include this in the total score. Based upon certain combinations of criteria, patients were stratified into Appropriate and Not Appropriate.
RESULT(S): A total of 100 patients were included (61yrs, 62% F, 29.5kg/m²). Assessment of the categories of Appropriateness of CD surgery: (1) Symptoms: 6% None to Mild, 94% Moderate to Severe; (2) Myelopathy: 16% None to Mild, 45% Moderate, 39% Severe; (3) All were grouped No Progression in the present study; (4) 19% demonstrated Global Malalignment (+/++Schwab), 81% did not; (5) Risk Factors: 27% had None to Mild, 63% Moderate, 10% Severe; (6) Curvature TS-CL demonstrated 88.6% of patients Severe (>20degree). Not Appropriate: None to Mild (NDI <28), with (1) None to Moderate myelopathy (mJOA >12) and Severe Risk Factors (2) None to Mild myelopathy (15-18) and No Progression or Imbalance (3) Moderate myelopathy (mJOA 12-15) and None to Moderate Risk Factors and no Progression, or imbalance and TSCL <25degree. Appropriate: Moderate to Severe Symptoms (NDI >=30), with (1) Severe myelopathy (mJOA<12) and Any Risk Factors (2) Moderate myelopathy and None to Moderate Risk Factors (3) Moderate myelopathy with Severe Risk Factors, Progression AND Imbalance (4) None to Mild myelopathy and None to Moderate Risk Factors and Progression OR imbalance (5) None to Mild myelopathy with None to Moderate Risk Factors and TSCL >25degree. In the present cohort, 94% were deemed Appropriate for surgery, 6% Not Appropriate. Less Appropriate patients demonstrated higher rates of postop dysphagia complications (17% vs 2%), met MCID for NDI less (0% vs 30.9%), and had more occurrences of DJK (16.7% vs 6.4%) by 2 years, p<0.05.
CONCLUSION(S): In light of the heterogeneity and uncertainty surrounding CD, this study developed CD appropriateness criteria, using established methodology, for surgeons to consider in the preoperative decision-making that correlate well with major postop occurrences. Application of the appropriateness criteria for CD may optimize patient selection and reduce the incidence of unwarranted surgery, although future validation is necessary. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Substantial adult spinal deformity (ASD) research has focused on sagittal plane radiological assessment and emphasized the importance of sagittal correction on patient-reported outcomes. It seems less attention has been given to evaluation of ASD in the coronal plane. However, recent ASD studies have demonstrated baseline global coronal malalignment (GCM; substantial displacement of the C7 coronal plumb line from midsacrum) in up to 35% of patients. Moreover, GCM can worsen postoperatively or may occur as an iatrogenic complication in previously balanced patients. Collectively, this may suggest that the associated clinical impact of GCM on ASD surgical outcomes has been underestimated. Currently, few reports focus on surgical outcomes in ASD patients with GCM. PURPOSE: Our objective was to assess treatment outcomes and complication rates associated with ASD surgery in the subset of patients with severe GCM. STUDY DESIGN/SETTING: Retrospective review of a prospectively collected multicenter consecutive case registry. PATIENT SAMPLE: Enrollment required: age >=18 years, scoliosis >=20degree, sagittal vertical axis (SVA) >=5cm, pelvic tilt >=25degree, and/or thoracic kyphosis >=60degree. OUTCOME MEASURES: Radiographic correction (both standard coronal and sagittal spinopelvic deformity measurements) and health-related quality of life (HRQL) outcomes, which included Oswestry Disability Index (ODI), Short Form-36 (SF-36) scores, Scoliosis Research Society-22 (SRS-22) scores, and back/leg pain numerical rating scale (NRS) scores.
METHOD(S): Surgically treated ASD patients with severe GCM (coronal C7PL-midsacral offset magnitude >=1 SD above the mean) were identified. Baseline and follow-up radiographic and HRQL outcomes were analyzed. Percentages of patients achieving minimal clinically important difference (MCID) and substantial clinical benefit (SCB) were assessed. Demographic, frailty, surgical, and complications data were collected. The primary focus for analysis was on patients who reached minimum 2-year follow-up.
RESULT(S): Of 691 potentially eligible operative patients (mean GCM = 4+/-3 cm), 80 met criteria for severe GCM >= 7cm. Of these 80 patients, 62 (78%, mean age 64 years, 81% women) had minimum 2-year follow-up. Baseline frailty score was 3.9+/-1.5 (indicating patients were frail) and 48% had prior spine fusion. Surgical data included: posterior-only (58%), mean fusion 13 levels, sacroiliac fixation (97%), and three-column osteotomy (36%). Postoperative biplane alignment improved significantly (p<0.001): GCM 11 to 4 cm, lumbar Cobb 30degree to 11degree, C7-S1 SVA 13 to 4 cm, and PI-LL 31degree to 5degree. Overall HRQL improved significantly (p<=0.003): ODI 51 to 37, SF-36 PCS 29 to 37, SRS-22r 2.6 to 3.5, and back/leg pain NRS 7 to 4 and 5 to 3, respectively. Thresholds for MCID/SCB were met in 43-83% of patients. A total of 89 complications were reported (34 minor/55 major), and 45 (73%) patients had >=1 complication (most commonly rod fracture [19%] and PJK [18%]). There were 34 reoperations in 22 (36%) patients (most common indications rod fracture/PJK). The 18 patients who did not achieve 2-year follow-up had a mean of 0.63 years follow-up, and the types of complications encountered in these 18 patients were comparable to those encountered in patients with 2-year follow-up.
CONCLUSION(S): ASD surgery for patients with severe GCM is associated with significant improvements in biplane alignment and HRQL outcome measures despite high complication rates. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: As the field of corrective cervical deformity (CD) surgery progresses, surgeons continue to take on more challenging cases. In order to minimize suboptimal postoperative outcomes it is important to develop a tool that allows for proper preoperative risk stratification. PURPOSE: Develop individualized predictive models for identification of risk-factors that lead to the development of major complications, revisions, and unplanned reoperation. STUDY DESIGN/SETTING: Retrospective cohort study of a prospective adult cervical deformity database. PATIENT SAMPLE: A total of 105 CD patients. OUTCOME MEASURES: Major complications (prolonged hospitalization, invasive intervention, prolonged or permanent morbidity, death within 1 year[1Y]), unplanned revision or reoperation after CD surgery within 1Y, HRQOL (NDI, mJOA, EQ5D).
METHOD(S): CD patients (cervical kyphosis >10degree, coronal scoliosis >10degree, cSVA >4cm, TS-CL >10degree, or CBVA >25degree) >=18 years old with complete baseline (BL), 1Y HRQoL and radiographic data. Patients were stratified into two groups based on their outcome from surgery: revision (Rev) and major complication (MC). Descriptive analysis identified cohort demographics, radiographic parameters and surgical details. Univariate analysis of preoperative and surgical factors was conducted to determine associations with any of the two outcomes. Stepwise logistic regressions identified surgical, radiographic, and HRQL factors that were associated with Rev or MC. Decision tree analysis established cut-offs for predictive variables.
RESULT(S): A total of 105 CD patients were included (61.6yrs, 67%F, BMI: 28.4kg/m2, CCI: 0.96+/-1.3). Surgical details: 7.6+/- 3.7 levels fused; mean EBL of 816 mL. By surgical approach, 45.71% had a posterior approach, 35.24% combined approach, and 19.05% anterior approach. Radiographically at baseline, patients presented with: PT: 18.9+/- 11.3; PI: 53.0+/-11.2; PI-LL: -0.45+/-17.2; SVA: -4.3+/-66.8, TS-CL: 37.9 +/-20.2; cSVA: 38.4+/-19.9. Postoperatively, 20 patients experienced a MC and 17 patients underwent a subsequent rev. instrumentation location (LIV: 1.1[1.0-1.3] and UIV: 1.5[1.1-2.1]) was significantly associated with undergoing a rev after index surgery (all p<0.05). The development of a postoperative MC was significantly associated with BL radiographic pelvic parameters (all <0.05). Predictive modeling incorporating preoperative and surgical factors identified development of a Rev to include: UIV>C3, LIV>T3, C2-T3 SVA<46.7degree, C2-C7 SVA>57.6degree, CTPA>7.8degree, and C2S<60.4 (AUC:0.80). For developing a MC, a model consisting of preoperative and surgical factors included BL EQ5D-VAS<30, TS-CL>59.2degree, C2-C7 SVA > 69.1degree, C2-T3 SVA < 18.6, Apex C2-C7 SVA >4.25, surgical invasiveness and posterior osteotomies (AUC:0.83).
CONCLUSION(S): Major adverse events were not uncommon following adult cervical deformity correction. Risk stratification models were developed to predict with high accuracy the occurrence of these common significant postoperative events. Revisions were predicted with an accuracy of 80% using a predominance of radiographic variables, while the occurrence of other major complications was also predicted with high reliability utilizing additional baseline HRQoL data and surgical factors. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Many different risk factors for PJK have been identified in those having surgery for ASD. The presence of CF prior to thoracolumbar fusion as a risk factor in this patient population has not been investigated. PURPOSE: To investigate if prior cervical fusion would be a risk factor for PJK when upper thoracic UIV was selected. STUDY DESIGN/SETTING: Retrospective review of prospective multicenter database. PATIENT SAMPLE: Inclusion: age >18 and one of the following inclusion criteria was queried: PI-LL >10, PT >20, Max Cobb>10, SVA > 5 cm. OUTCOME MEASURES: Development of PJK; HRQOL include NRS back and leg, ODI, SRS-22, EQ5D METHODS: Inclusion: age >18 and one of the following inclusion criteria was queried: PI-LL >10, PT >20, Max Cobb>10, SVA > 5 cm. 1023 patients were identified and grouped as CF (prior CF; n=101) and NoCF (n=922). Patients with 5 or fewer levels of CF and those without further cervical spine surgery at final follow-up were included (cASD= 86; ASD= 910). After initial analysis propensity matching was performed controlling for: age, BMI, pre-op alignment (PI, PI-LL, TPA, SVA), change in alignment (PI-LL), and UIV/LIV.
RESULT(S): Among 996 patients the CF group was older (64 v 60; p<0.000), higher BMI (29.7 v 28; p<0.014), with equal gender distribution. PJK at 1 year was 34.9% (n=30) in CF and 34.7% (n=316) in NoCF. All patients included had a minimum of four segments unfused between CF and TL fusion. No differences were found between groups when stratifying by UIV (UT=upper thoracic; LT=Lower thoracic; L=lumbar). Radiographically CF had worse preop deformity (PI-LL 21.3 v 17.2; TPA 27.3 v 23.6; SVA 90.6 v 68.5; p<0.05), with no difference in postop alignment between groups. CF was more likely to have LT UIV (61.6% v 46%; p=0.045) and fusion to the ilium (93% v 81%; p=0.005). After propensity matching 79 patients were included in each group. CF had 68 patients with ACDF, 5 patients with PSF, and 6 patients with A/PSF. There was no difference in the rate of PJK (39.2 v 34.2%), PJK magnitude (13.4 v 13.7) or change in PJK angle (10.1 v 10.4). PJK rate and magnitude are not different by UT or LT as UIV.
CONCLUSION(S): The presence of CF is not associated with the development of PJK or PJK magnitude in this propensity matched cohort. All CF patients had a minimum of four segments unfused between the CF and the TL fusion which may act as a buffer and be protective of PJK. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Adult spinal deformity (ASD) correction involves complex, invasive procedures. However, it is unknown how metabolic syndrome affects the cost efficiency of ASD surgery. PURPOSE: Investigate the differences in ASD surgery cost for metabolic syndrome patients. STUDY DESIGN/SETTING: Retrospective review of a single center ASD database. PATIENT SAMPLE: A total of 557 ASD patients. OUTCOME MEASURES: Complications, revisions and costs.
METHOD(S): ASD patients (scoliosis >=20degree, SVA>=5cm, PT >=25degree, or TK >=60degree) >=18, undergoing >=4 level fusions were included. Descriptive analysis assessed mean baseline demographic, radiographic and surgical data were assessed. Patients diagnosed with: BMI >30, diabetes mellitus, and HTN were classified metabolic syndrome patients (MetS). Independent samples T-tests assessed baseline differences in SVA, PT, PI-LL, and levels fused between MetS and non-MetS patients. Total surgery costs for MetS and non-MetS ASD patients were compared. Costs were calculated using the PearlDiver database, which reflects both private insurance and Medicare reimbursement claims. Complications and comorbidities (CC) and major complications and comorbidities (MCC) were assessed according to CMS.gov manual definitions.
RESULT(S): A total of 557 patients met inclusion criteria. Baseline demographics and surgical details: age 60.8 +/- 13.2, 67.5% female, BMI 28.3 +/- 6.67, 6.2% anterior approach, 81.0% posterior approach, 12.8% combined approach, 10.4 +/- 3.9 levels fused, op time 423.5 +/- 166.4 minutes, EBL 2061.4 +/- 1631.5 cc, LOS 7.6 +/- 5.4 days. 11 MetS patients were included. There was no significant difference in BL SVA (54.2degree vs 70.3degree), PT (26.7degree vs 23.7degree), PI-LL (16.4degree vs 21.2degree), or levels fused between groups (10.5 vs 10.6) (all p>0.05). Twenty-seven percent of MetS patients experienced a CC or MCC, with 18% undergoing revisions within 90 days (vs 13.2% and 3.2% for non-MetS, p<0.05). The average costs of surgery for MetS patients was $68,004.06 vs $48,726.40 for non-MetS patients (p<0.05).
CONCLUSION(S): Metabolic syndrome is a growing concern among health care providers. In metabolic syndrome patients undergoing corrective adult spinal deformity surgery, the total costs of surgery was 28.4% higher for metabolic syndrome patients ($68,004.06 vs $48,726.40). Optimizing modifiable factors like diabetes or BMI in preoperative patients may have an impact on cost effectiveness. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Amidst a current US opioid epidemic, it is important to understand factors that contribute to long-term or sustained opioid use after elective spine surgery. Recent reports have reported up to 25% of opioid-naive patients are taking opioids at 2-year follow-up after adult spinal deformity (ASD) surgery. This suggests ASD surgery, or at least the postoperative course, could be a significant factor contributing to long-term or sustained opioid use. ln Asian countries, opioids are rarely prescribed for postoperative pain. Potentially, there is a significant cultural difference that can play a role in patient expectations in the postoperative period leading to long-term opioid use. PURPOSE: To compare long-term or sustained (2-years) opioid use after ASD surgery between American and Japanese Cohorts STUDY DESIGN/SETTING: Propensity-matched comparison of American and Japanese ASD patients PATIENT SAMPLE: Patients from two (United States vs Japanese) prospective, multicenter ASD databases were compared OUTCOME MEASURES: Baseline and 2-year opioid use and SRS-22 domains were compared METHODS: We propensity matched 127 JPN to 619 US ASD patients based on age, sex, BMI, 3-column osteotomy, pelvic fixation, number of levels fused and preoperative SVA. Pre-op and 2-year postoperative opioid use was determined using responses to SRS22R Q11 and grouped into 3 categories (none, weekly, or daily). We have previously validated the use of SRS22R Q11 as an accurate measure for opioid use by comparing to a prospective study specifically designed to study opioid use and postoperative outcomes in ASD.
RESULT(S): From each cohort, 34 cases were successfully matched, with no difference in baseline parameters, including baseline opioid use (p=0.095), between the cohorts. The relatively low numbers in each cohort arise from difficulty matching a lower BMI and lower baseline opioid use in the JPN cohort. At 2-years postop, 22 US (65%) vs 31 JPN (91%) reported no opioid use, 1 US (3%) vs 0 JPN (0%) reported weekly use, and 11 US (32%) vs 3 JPN (9%) reported daily use (p=0.009). There was no difference in 2-year SRS-22 Self-image and Mental Health between the two groups while US patients had better Satisfaction (4.29 vs 3.84, p=0.032) but lower Function scores (3.52 vs 3.90, p=0.029).
CONCLUSION(S): In propensity-matched ASD cohorts, 35% of US patients were still using opioids 2 years after surgery compared to 9% of JPN patients, with the most taking opioids daily. Further studies are needed to identify sources of this variability. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Adult spinal deformity (ASD) corrective surgery is often associated with high rates of adverse outcomes during the recovery course. With the growth of predictive analytics in the medical field, it is important to utilize the developing machine-learning resources to consider the risks associated with major operations. PURPOSE: To assess the validity of the ACS-NSQIP risk index for postop complications and mortality in a prospective multicenter ASD database. STUDY DESIGN/SETTING: Retrospective cohort study of a prospective multicenter ASD database PATIENT SAMPLE: A total of 1,281 ASD patients. OUTCOME MEASURES: Predictive performance of the ACS-NSQIP calculator in an ASD database.
METHOD(S): Patients>18yrs undergoing surgery for ASD (scoliosis>=20degree, SVA>=5cm, PT>=25degree or TK>=60degree). Calculated perioperative complication risk averages via the ACS-NSQIP surgical calculator with CPT codes were compared with 30-day complication rates in the ASD database. Outcomes assessed (as defined by NSQIP): serious complication, any complication, pneumonia, cardiac complication, SSI, UTI, VTE, renal failure, return to OR, death, sepsis and length of hospital stay. Predictive performance of the calculator was analyzed by Brier score (sum of squared differences between the binary outcome and the predicted risk). It ranges from 0 to a maximum Brier score [(mean observed outcome)*(1-[mean observed outcome])]. Values closer to 0 are suggestive of better predictive performance (>0.05 considered poor). Length of stay was assessed with a Bland-Altman plot, observed LOS on the x-axis and the difference between the observed and predicted LOS on the y-axis. Performance of the calculator for serious/any complications for surgeries indicated as 'Risk significantly higher than estimate' was assessed by comparing means of patients who were in the top quartile for their ISSG-invasiveness scores.
RESULT(S): A total of 1,281 ASD patients (60.2 yrs, 73.5% F, 28 kg/m2) were included. A total of 49.4% of procedures involved decompression and 100% involved fusion, with a mean number of levels fused of 10.98. The means for individual patient characteristics that coincide with the variables entered into the online risk calculator interface are as follows: functional status through baseline ODI scores (Independent [0-40]: 35.6%, Partially Dependent [41-80]: 61.8%, Totally Dependent [81+]: 2.6%), 0% emergent cases, ASA Class (I: 5.6%, II: 46.8%, III: 41.7%, IV: 1.4%, V: 0%), 11.9% disseminated cancer, 9.1% diabetes mellitus, 36.5% use of hypertensive medications, 10.8% CHF, 5.9% current smoker, 5.2% COPD, and 4.4% acute renal failure. Predictive of any 30-day postoperative complications ranged from 2.8-18.5% across CPT codes, where the actual rate in the cohort was 9.0%, and demonstrated good predictive performance via Brier Score (0.00064516, Max: 0.00819), as well as pneumonia, SSI, UTI, VTE, renal failure, death and sepsis (Brier Max: 0-0.01458096). Serious and cardiac complications, as well as return to OR were poorly predicted via the NSQIP risk index (Brier Max > 0.05). Mean difference between observed and predicted LOS was 4.276 days with a 95% confidence interval of 9.484 - -0.932. When indicated for significantly high risk, the calculator poorly predicted overall and serious complications (Brier Max >0.1).
CONCLUSION(S): While the ACS-NSQIP risk index had acceptable predictive performance in regards to the occurrence of overall post-operative complications, notable exceptions were detected. Specifically, deficiencies in assessing serious complications, cardiac complication and return to OR were seen, and performance was noted to diminish with procedures of greater invasiveness. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Following cervical deformity (CD) corrective surgery, durability remains a challenge and distal junctional kyphosis (DJK) is an important risk for a surgeon to consider. DJK is the result from fixation failure, adjacent level fracture or spondylolisthesis. The timing of DJK onset has yet to be investigated. PURPOSE: To determine the timing of DJK development following CD corrective surgery. STUDY DESIGN/SETTING: Retrospective review. PATIENT SAMPLE: A total of 139 CD patients. OUTCOME MEASURES: Early/Late DJK Development, Severe DJK, Symptomatic DJK.
METHOD(S): Included: surgical CD patients (C2-C7 Cobb>10degree, CL>10degree, cSVA>4cm, or CBVA>25degree) without preop DJK. DJK angle was defined as the change in kyphosis between lower-most instrumented vertebra (LIV) and LIV-2 from preop to postop (<-10). X2 analysis and post-hoc testing assessed annual and overall incidence of early (3M follow-u) and late (DJK development at 6M, 1Y, 2Y) DJK development among operative patients. Differences between early and late DJK (development after 3 months) were assessed. Pearson correlations determined significant variables associated with development of early versus late DJK. Severe DJK was defined as DJK angle <-20degree. DJK was considered symptomatic if associated with reoperation or the previously published thresholds of NDI>24 or mJOA<14. The natural progression of the disease was predicted with the use of a Kaplan-Meier survivorship analysis.
RESULT(S): A total of 139 patients without preop DJK were included (16 re-op) (61.8yrs, 62.3%F, 29.1kg/m2). Surgical characteristics: 20.1% anterior approach, 42.4% posterior, 36.7% combined (levels fused: 7.6). Incidence of DJK from 2013-2018: 23%. Early: 9.4%, late: 10.1% (6M delayed: 3.6%, 1Y delayed: 6.5%, 2Y delayed: 2.2%). No differences were observed between op and reop for development of early or late DJK at all f/u (p>0.050). Presence of upgoing plantar response at baseline neuro exam, C2-C7 angle, T1-C2 angle, CBVA and combined approach correlated with development of early DJK. For late DJK: history of tumor and pulmonary disease, pelvic incidence, T10-L2 angle, and L4 pelvic angle. Apex of the secondary driver was significantly lower in the Late group (mean early: T2/3; late: T10; p=0.023). 3M radiographically, Early DJK patients had greater TK (-57.9degree vs -40degree, p=0.024, while L4PA remained larger in the late DJK group (14.5degree, p=0.009). Between early and late groups, 41.7% of early DJK patients met criteria for severe DJK, while 0% of Late DJK patient were severe(p=0.010); symptomatic DJK between the two groups was not significant (p=0.941). Kaplan-Meier survivorship analysis determined patients within the cohort to have a 91.1% cumulative probability of maintaining non-DJK status by 3-month follow-up, 85.9% at 6-month, 80% at 1-year, and 77.0% by 2-year follow-up.
CONCLUSION(S): Patients undergoing CD corrective surgery have incidence of early and incidence of late DJK. While the majority of DJK development occurs within the first 6 months, late DJK occurs and differs in presentation and etiologic factors. Early DJK occurrence is more likely to be severe radiographically, associated with neurological decline, and is more strongly related to biomechanical factors such as use of a combined approach at the time of surgery. Contrarily, although equally likely to be symptomatic as reflected in loss of clinical gains from surgery, late DJK is more likely mild radiographically and associated with suboptimal cervical realignment and lack of addressing secondary drivers, likely related to negative compensatory mechanisms. Customized prophylactic approaches for both occurrences is mandated. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Recombinant human bone morphogenetic protein 2 (BMP) is increasingly utilized in minimally invasive (MIS) transforaminal lumbar interbody fusions (TLIF) in order to increase rate of fusion by promoting bone growth through the induction of osteoblast differentiation, awhile reducing morbidity related to iliac crest autograft. Despite these benefits, BMP use is still controversial due to its pro-inflammatory mechanism of action and potential to cause radiculitis. PURPOSE: To assess whether BMP is a risk factor for postoperative radiculitis in TLIF. STUDY DESIGN/SETTING: Single-center retrospective cohort study. PATIENT SAMPLE: A total of 397 TLIFs from June 2012 to December 2018. OUTCOME MEASURES: Perioperative clinical characteristics, post-operative risk of radiculitis and complication, and future reoperation rates.
METHOD(S): Patients >= 18 years-old undergoing elective single-level TLIFs from 2012 to 2018 were included. Outcome measures included perioperative clinical characteristics, postoperative risk of radiculitis and complication, and future reoperation rates. Radiculitis was defined the delayed onset of radicular symptoms postoperatively in patients whom had initial resolution of radicular symptoms immediately postoperatively, in the absence of persistent neurological compression on postoperative imaging. Statistical analyses included independent t-tests and chi-square analysis. Propensity score matching was utilized to control for demographic differences between the groups. Independent predictors for post-operative radiculitis were assessed by multivariate logistic regression. Significance set at p<0.05.
RESULT(S): A total of 397 cases were included in the study (59.33 +/- 13.49 mean age, 28.98 +/- 6.29 mean BMI, 52.90% female, 2.29 +/- 1.92 average Charleston comorbidity Index). There were 223 open procedures and 174 MIS. For the entire cohort, 238 cases utilized BMP and 159 did not, with 102 MIS pairing with BMP use. The MIS TLIFs had a higher percentage of BMP use than open TLIFs (58.6% vs 25.7%, p<0.001), lower estimated blood loss (212.28 +/- 193.79 mL vs 410.91 +/- 337.98 mL, p<0.001) higher fluoroscopy dosage (52.43 +/- 48.61mGy vs 16.77 +/- 27.84mGy, p<0.001), and a lower length of stay (3.20 +/- 2.55 days vs 4.11 +/- 2.52 days, p<0.001). There were no other differences in perioperative clinical characteristics. There was a significantly higher rate of postoperative radiculitis in the MIS TLIFs compared to open (12.6% vs 6.8%, p=0.046) and use of BMP compared to no BMP (13.2% vs 6.7%, p=0.029). There was a 15.7% radiculitis rate when MIS was paired with BMP use. There were no other notable differences in complication rates or rates of reoperation. Individually, MIS had a 12.6% radiculitis rate (p=0.046) and BMP use had a 13.2% rate (p=0.029). Propensity score match controlled for the significant difference in CCI between the MIS and open groups (N=168 each). Multivariate regression indicated that MIS (p=0.314) and BMP (p=0.109) were not independent predictors individually when controlling for age, gender, and BMI. When technique was paired with biologic use the regression revealed MIS + BMP is a risk factor of post-operative radiculitis (2.265(4.753-1.079), p=0.031).
CONCLUSION(S): While BMP and MIS technique were not independent risk factors for postoperative radiculitis, there is an increased risk of radiculitis when using BMP in MIS TLIF. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright