Stratum-specific Likelihood Ratio Analysis Research Articles

Body Mass Index Greater Than 46 Associated With Increased Risk of 30 Day Complications Following Adult Tonsillectomy: A Retrospective Cohort Study.

Background: The study aimed to identify data-driven body mass index (BMI) thresholds that are associated with varying risk of 30 day complications following adult tonsillectomy. Methods: The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was utilized to conduct a retrospective cohort analysis of patients undergoing adult tonsillectomy from 2005 to 2019. Stratum-specific likelihood ratio (SSLR) analysis was conducted to determine data-driven BMI strata that maximized the likelihood of 30 day complications following adult tonsillectomy. Patient demographics and clinical comorbidities were compared using chi-squared analysis and student t tests, where appropriate, for each stratum. Multivariable regression analysis was conducted to confirm association between identified data-driven strata with 30 day complication rates. Results: In total, 44,161 patients undergoing adult tonsillectomy were included in this study. SSLR analysis identified 2 BMI categories: 18 to 45 and 46+. Relative to the 18 to 45 BMI cohort, the 46+ BMI cohort was more likely to have 30 day all-cause complications after surgery [odds ratio (OR): 1.62, P = .007]. Specifically, the 46+ BMI cohort had significantly higher odds for 30 day major medical complications (OR: 2.86, P = .001), pulmonary domain complications (OR: 1.86, P = .041), unplanned reintubation (OR: 2.65, P = .033), and deep vein thrombosis (OR: 6.54, P = .026). Conclusions: We identified a BMI threshold of 46+ that was associated with a significantly increased risk of 30 day all-cause complications following adult tonsillectomy. These BMI strata can guide preoperative planning and risk-stratifying models for predicting 30 day complications in tonsillectomy surgery.

Evaluating the Effect of Decreasing Preoperative Hemoglobin on Blood Transfusions, Major Complications, and Periprosthetic Joint Infection After Primary Total Knee Arthroplasty.

Preoperative anemia is associated with increased postoperative transfusion and complication rates after total knee arthroplasty (TKA). We aimed to create TKA-specific data-driven preoperative hemoglobin strata that quantify the likelihood of 90-day postoperative blood transfusion and evaluate whether these strata are associated with increased risk of 90-day major complications and 2-year prosthetic joint infection (PJI). Primary TKA patients from 2013 to 2022 were identified using a national database. Stratum-specific likelihood ratio (SSLR) analysis defined hemoglobin strata associated with the risk of 90-day blood transfusion. Each stratum was propensity score matched to the highest identified hemoglobin strata. Unmatched incidence rates and matched risk of 90-day major complications and 2-year PJI between strata were compared. SSLR identified four 90-day blood transfusion hemoglobin strata for men (strata [g/dL], likelihood ratio [<11.4, 8.06; 11.5 to 11.9, 4.34; 12.0 to 12.9, 1.70; 13.0 to 17.0, 0.54]) and women (<10.4, 8.22; 10.5 to 11.4, 2.84; 11.5 to 12.4, 1.38; 12.5 to 17.0, 0.50). Increased 2-year PJI risk was associated with three male strata (<11.4, 11.5 to 11.9, 12.0 to 12.9; all P < 0.001) and three female strata (<10.4, 10.5 to 11.4, 11.5 to 12.4; all P < 0.001). Increased 90-day major complication risk was associated with three male strata (<11.4, 11.5 to 11.9, 12.0 to 12.9; all P < 0.001) and three female strata (<10.4, 10.5 to 11.4, 11.5 to 12.4; all P < 0.001). Using SSLR analysis, we identified unique TKA-specific data-driven hemoglobin strata for both men and women that quantify the likelihood of 90-day blood transfusions and predict the risk of both 90-day major complications and 2-year PJI. These strata are a first in the TKA literature and can assist surgeons in stratifying patients' transfusion and complication risk based on their preoperative hemoglobin value. While optimizing patients in the preoperative setting, we recommend using these TKA-specific hemoglobin thresholds to help guide decision making on the need for presurgery anemia optimization and to help reduce the need for blood transfusion.

Reverse Total Shoulder Arthroplasty Within 6 Weeks of Proximal Humerus Fracture is Associated with the Lowest Risk of Revision

Reverse total shoulder arthroplasty (RTSA) has become an increasingly popular treatment strategy in the management of complex proximal humeral fractures (PHF). However, no definitive consensus has been reached regarding the optimal surgical timing of RTSA following PHF, particularly considering nonoperative management is often a viable option. Therefore, the aim of this study was (1) to identify optimal timing intervals that maximize the likelihood of revision following RTSA and (2) to determine differences in revision etiologies using the identified timing intervals. A retrospective cohort analysis of patients undergoing PHF-indicated RTSA from 2010 to 2021 was conducted using a national administrative claims database. Stratum specific likelihood ratio (SSLR) analysis was conducted to determine data-driven timing strata between PHF and RTSA that maximized the likelihood of revision surgery within 2-years of RTSA. To control for confounders, multivariable regression analysis was conducted to confirm the identified data-driven strata's association with 2-year revision rates as well as compare the likelihood of various indications for revision including mechanical loosening, dislocation, periprosthetic joint infection (PJI), and periprosthetic fracture (PPF). In total, 11,707 patients undergoing TSA following PHF were included in this study. SSLR analysis identified two timing categories: 0-6 weeks and 7-52 weeks from the time of PHF to TSA surgery. Relative to the 0-6 week cohort, the 7-52 week cohort was more likely to undergo revision surgery within 2-years (OR: 1.93, P < 0.001). Moreover, the 7-52 week cohort had significantly higher odds of revision indicated for dislocation (OR: 2.24, P < 0.001), mechanical loosening (OR: 1.71, P < 0.001), PJI (OR: 1.74, P < 0.001), and PPF (OR: 1.96, P < 0.001). Using SSLR, we were successful in identifying two data-driven timing strata between PHF and RTSA that maximized the likelihood of 2-year revision surgery. As it can be difficult to determine whether RTSA or nonoperative management is initially more appropriate, considering the results of this study, an early trial of 4 to 6 weeks of nonoperative management may be appropriate without altering the risks associated with RTSA.

509 Data-Driven Preoperative Hemoglobin Strata that Maximize the Likelihood of Blood Transfusion following Single-Level Lumbar Fusion Are Also Predictive of Major Complications and Deep Infections

INTRODUCTION: Optimizing patients undergoing single-level lumbar spine fusion helps reduce the use of blood transfusions which are associated with an increased risk of major complications and deep infection. Current hemoglobin thresholds for anemia severity were published in 1968 and are not surgery specific. METHODS: A retrospective cohort analysis was performed using a national database to identify patients undergoing primary single-level lumbar fusion with recorded hemoglobin values 30 days before surgery. Stratum-specific likelihood ratio (SSLR) analysis defined sex-based hemoglobin strata associated with the risk of 90-day postoperative blood transfusion. Each stratum was propensity-score matched to the highest identified hemoglobin stratum. Incidence rates and risk of 90-day major complications and 2-year deep infection between strata were observed. RESULTS: For 90-day blood transfusion, SSLR identified three female [Strata, Likelihood ratio (5.0-10.9, 2.41; 11.0-12.4, 1.35; 12.5-17.0, 0.78)] and three male hemoglobin strata (5.0-11.9, 2.95; 12.0-13.4, 1.46; 13.5-13.9, 0.71). Increased risk of 90-day major complications was associated with two female (11.0-12.4 [RR: 1.52; p &lt; 0.001], 5.0-10.9 [RR: 3.40; p &lt; 0.001]) and one male stratum (5.0-11.9 [RR: 2.02; p &lt; 0.001]). Increased risk of 2-year deep infection was associated with one female (5.0-10.9 [RR: 3.67; p &lt; 0.001]) and one male stratum (5.0-11.9 [RR: 2.11; p = 0.005]). CONCLUSIONS: SSLR analysis established single-level lumbar fusion-specific hemoglobin strata that maximize the likelihood of 90-day blood transfusions and predict the risk of 90-day major complication and 2-year deep infection. We recommend using these thresholds during preoperative optimization, noting the sequential increase in risk of complications starting with a hemoglobin of less than 13.5 in males and 12.5 in females.

The Creation of Data-Driven Preoperative Hemoglobin A1c and Same-Day Glucose Strata to Stratify Complication Risk Following Total Hip Arthroplasty

BackgroundThis study identifies data-driven strata for preoperative Hemoglobin A1c (HbA1c) and same-day glucose levels that maximize differences in the likelihood of complications following total hip arthroplasty (THA). MethodsPatients who underwent THA from 2013 to 2022 were identified using a national database. In total, 18,728 patients were identified with a mean age of 67 years (range, 18 to 80). Stratum specific likelihood ratio (SSLR) analysis determined separate strata for HbA1c and same-day glucose levels that minimized the likelihood of 90-day complications following THA. Each stratum was propensity-score matched based on age, sex, hypertension, heart failure, chronic obstructive pulmonary disease, and obesity to the lowest respective stratum. The risk ratio (RR) with respect to the lowest matched stratum was observed. ResultsOur SSLR analysis identified 3 data-driven HbA1c strata (4.5 to 5.9, 6.0 to 6.9, and 7.0+) and two same-day glucose strata (60 to 189 and 190+) that predicted 90-day major complications. For HbA1c, when compared to the lowest strata (4.5 to 5.9), the risk of 90-day major complications sequentially increased as the HbA1c strata increased: 6.0 to 6.9 (RR: 1.21; P = .041), 7+ (RR: 1.82; P < .001). For same-day glucose, when compared to the matched lowest strata (60 to 189), the risk of 90-day major complications was higher for the 190+ strata (RR: 1.5; P < .001). ConclusionsOur results support the use of multiple HbA1c strata that can be incorporated into preoperative risk-stratification models. Additionally, we identified a single cut-off level of 190 as a maximum target blood glucose level perioperatively.

Impact of Surgeon Case Volume on Outcomes After Reverse Total Shoulder Arthroplasty.

Despite a rapid increase in utilization of reverse total shoulder arthroplasty (rTSA), volume-outcome studies focusing on surgeon volume are lacking. Surgeon-specific volume-outcome studies may inform policymakers and provide insight into learning curves and measures of efficiency with greater case volume. This retrospective cohort study with longitudinal data included all rTSA cases as recorded in the Centers for Medicare & Medicaid Services Limited Data Set (2016 to 2018). The main effect was surgeon volume; this was categorized using two measures of surgeon volume: (1) rTSA case volume and (2) rTSA + TSA case volume. Volume cutoff values were calculated by applying a stratum-specific likelihood ratio analysis. Among 90,318 rTSA cases performed by 7,097 surgeons, we found a mean annual rTSA surgeon volume of 6 ± 10 and a mean rTSA + TSA volume of 9 ± 14. Regression models using surgeon-specific rTSA volume revealed that surgery from low (<29 cases) compared with medium (29 to 96 cases) rTSA-volume surgeons was associated with a significantly higher 90-day all-cause readmission (odds ratio [OR], 1.17; confidence interval [CI], 1.10 to 1.25; P < 0.0001), higher 90-day readmission rates because of an infection (OR, 1.46; CI, 1.16 to 1.83; P = 0.0013) or dislocation (OR, 1.43; CI, 1.19 to 1.72; P = 0.0001), increased 90-day postoperative cost (+11.3% CI, 4.2% to 19.0%; P = 0.0016), and a higher transfusion rate (OR, 2.06; CI, 1.70 to 2.50; P < 0.0001). Similar patterns existed when using categorizations based on rTSA + TSA case volume. Surgeon-specific volume-outcome relationships exist in this rTSA cohort, and we were able to identify thresholds that may identify low and medium/high volume surgeons. Observed volume-outcome relationships were independent of the definition of surgeon volume applied: either by focusing on the number of rTSAs performed per surgeon or anatomic TSAs performed. III.

Sarcopenia as a Risk Factor for Complications Following Pedicle Subtraction Osteotomy.

Retrospective cohort. The objective was to determine if sarcopenia is an independent risk factor for complications in adult spinal deformity (ASD) patients undergoing pedicle subtraction osteotomy (PSO) and define categories of complication risk by sarcopenia severity. Sarcopenia is linked to morbidity and mortality in several orthopedic procedures. Data concerning sarcopenia in ASD surgery is limited, particularly with respect to complex techniques performed such as PSO. With the high surgical burden of PSOs, appropriate patient selection is critical for minimizing complications. We identified 73 ASD patients with lumbar CT/MRI scans who underwent PSO with spinal fusion ≥5 levels at a tertiary care center from 2005 to 2014. Sarcopenia was assessed by the psoas-lumbar vertebral index (PLVI). Using stratum-specific likelihood ratio analysis, patients were separated into 3 sarcopenia groups by complication risk. The primary outcome measure was any 2-year complication. Secondary outcome measures included intraoperative blood loss and length of stay. The mean PLVI was 0.84±0.28, with 47% of patients having complications. Patients with a complication had a 27% lower PLVI on average than those without complications (0.76 vs. 0.91, P=0.021). Stratum-specific likelihood ratio analysis produced 3 complication categories: 32% complication rate for PLVI ≥ 0.81; 61% for PLVI 0.60-0.80; and 69% for PLVI < 0.60. Relative to patients with PLVI ≥ 0.81, those with PLVI 0.60-0.80 and PLVI < 0.60 had 3.2× and 4.3× greater odds of developing a complication (P<0.05). For individual complications, patients with PLVI < 1.0 had a significantly higher risk of proximal junctional kyphosis (34% vs. 0%, P=0.022), while patients with PLVI < 0.8 had a significantly higher risk of wound infection (12% vs. 0%, P=0.028) and dural tear (14% vs. 0%, P=0.019). There were no significant associations between sarcopenia, intraoperative blood loss, and length of stay. The increasing severity of sarcopenia is associated with a significantly and incrementally increased risk of complications following ASD surgery that require PSO. Level III.

A Novel Method for Stratification of 30-Day Major Complication Risk Using Body Mass Index Thresholds for Patients Undergoing Total Knee Arthroplasty: A National Cohort of 443,157 Patients

BackgroundMany organizations have used pre-established body mass index (BMI) cut-offs to guide surgical decision-making. As there have been many improvements in patient optimization, surgical technique, and perioperative care over time, it is important to reassess these thresholds and contextualize them to total knee arthroplasty (TKA). The purpose of this study was to calculate data-driven BMI thresholds that predict significant differences in risk of 30-day major complications following TKA. MethodsPatients who underwent primary TKA from 2010 to 2020 were identified in a national database. Stratum-specific likelihood ratio (SSLR) methodology was used to determine data-driven BMI thresholds at which the risk of 30-day major complications increased significantly. These BMI thresholds were tested using multivariable logistic regression analyses. A total of 443,157 patients were included, who had a mean age of 67 years (range, 18 to 89 years), mean BMI of 33 (range 19 to 59), and 11,766 (2.7%) patients had a 30-day major complication. ResultsSSLR analysis identified four BMI thresholds that were associated with significant differences in 30-day major complications: 19 to 33, 34 to 38, 39 to 50, and 51+. When compared to those who had a BMI between 19 and 33, the odds of sustaining a major complication sequentially and significantly increased by 1.1, 1.3, and 2.1 times (P < .05 for all) for the other thresholds. ConclusionThis study identified four data-driven BMI strata using SSLR analysis that were associated with significant differences in the risk of 30-day major complications following TKA. These strata can be used to guide shared decision-making in patients undergoing TKA.

Delay of Timing of Anterior Cruciate Ligament Reconstruction Is Associated With Lower Risk of Arthrofibrosis Requiring Intervention

To conduct 2 separate stratum-specific likelihood ratio analyses in patients younger than 40 year of age (<40 years) and those aged 40 and older (40+ years) at time of anterior cruciate ligament (ACL) reconstruction to define data-driven strata between ACL tear and primary isolated ACL reconstruction in which the risk of arthrofibrosis, using manipulation under anesthesia and arthroscopic lysis of adhesions as surrogates, is significantly different. A retrospective cohort analysis was conducted using the PearlDiver Database. Patients who underwent ACL reconstruction were identified using the Current Procedure Terminology code 29888. Patients were stratified to those aged younger than 40 (<40) and those 40 and older (40+) at time of ACL reconstruction. The incidence of 2-year arthrofibrosis was calculated for weekly intervals from initial ACL injury to reconstruction. Stratum specific likelihood ratio analysis was conducted to determine data-driven intervals from initial ACL tear to reconstruction that optimize differences in 2-year arthrofibrosis. Following the identification of these intervals for both those <40 and 40+, multivariable analysis was conducted. For those <40, stratum-specific likelihood ratio analysis identified only 2 data-driven timing strata: 0-5 and 6-26 weeks. For those 40+, stratum-specific likelihood ratio analysis also only identified 2 data-driven strata: 0-9 and 10-26 weeks. A delay in ACL reconstruction from initial injury by at least 6 weeks in patients younger than 40 and at least 10 weeks in patients older than 40 years is associated with a 65% and 35% reduction of 2-year manipulation under anesthesia and arthroscopic lysis of adhesions, respectively. Our analysis showed a delay in ACLR of at least 6 weeks in patients younger than 40 years to be associated with a 65% reduction in the risk of surgical intervention for arthrofibrosis and a delay of at least 10 weeks in patients 40 years and older to be associated with only a 35% reduction in the risk of surgical intervention for arthrofibrosis. The authors propose this difference in reduction to be multifactorial and potentially associated with mechanism of injury, activity level, and preoperative factors such as amount of physical therapy, rather than solely timing. Level III, retrospective comparative prognostic study.

Interval Time of at Least 6 Weeks Between Bilateral Total Knee Arthroplasties is Associated With Decreased Postoperative Complications

BackgroundStaged, bilateral total knee arthroplasty (TKA) has an increased risk of complications if the second procedure is performed before physiologic recovery from the first. The aims of this study were to 1) determine whether there is a time-dependent relationship between TKA staging and rates of revisions and complications and 2) identify data-driven time intervals that reduce risk of revisions and complications. MethodsData were collected from a national insurance database from 2015 to 2018. Staged intervals were initially assessed using fixed 6-week intervals. Stratum-specific likelihood ratio analyses were subsequently conducted to observe data-driven staging thresholds. Bivariate and multivariable regression analyses were conducted to determine the associations between the time intervals and 2-year rates of revision surgery and 90-day major complications. We included 25,527 patients undergoing staged bilateral TKA. ResultsIn comparison to the shortest fixed time interval (1-6 weeks), as the staging interval increased the odds of 2-year all-cause revision and 90-day major complications significantly decreased (P < .05 for all). Stratum-specific likelihood ratio analysis identified 3 data-driven staging categories 1-5, 6-17, and 18-24 weeks that maximized the difference in both 2-year rates of revision and 90-day major complications. ConclusionOur data showed a time-dependent relationship between the timing of TKA stages and complications. If staging is considered, a delayed interval of at least 6 weeks between procedures may significantly reduce revision and major complications. Level of EvidenceLevel III Therapeutic Study.

204. The ALFA score predicts 30-day mortality after elective anterior lumbar interbody fusion

204. The ALFA score predicts 30-day mortality after elective anterior lumbar interbody fusion

Assessing the hospital volume-outcome relationship in total elbow arthroplasty.

Total elbow arthroplasty (TEA) is an effective intervention for multiple elbow disorders including complex fracture in elderly patients, post-traumatic arthropathy, inflammatory arthropathy, and distal humeral nonunion. Given its known therapeutic value and low utilization rate, an investigation into the thresholds for TEA institutional volume-outcome relationships is warranted. The purpose of this study was to identify TEA volume thresholds that serve as predictors of institutional outcomes including complications, readmissions, revisions, cost of care, length of stay (LOS), and non-home discharge. We hypothesized that increased institutional volume would be associated with decreased 90-day adverse outcomes and resource utilization. The Nationwide Readmission Database was queried from 2010 to 2017 to identify all cases of TEA. Hospital volume was calculated using a unique hospital identifier and divided into quartiles. Outcomes such as complications, readmissions, revisions, cost of care, LOS, and non-home discharge were then analyzed by quartile. The same outcomes were assessed via stratum-specific likelihood ratio (SSLR) analysis to define volume strata among institutions. SSLR analysis defined statistically significant hospital volume categories for each 90-day outcome. The volume category with the lowest complication rate was ≥21 TEAs per year (5.6%). The volume categories with the lowest readmission rates were 1-3 TEAs per year (4.7%) and ≥18 TEAs per year (9.2%). Revision rates were lowest in the volume categories of 1-5 TEAs per year (0.1%) and ≥18 TEAs per year (0.1%). Hospitals with ≥21 TEAs per year had the lowest cost of care and the highest rate of extended LOS (>2 days). SSLR analysis showed that non-home discharges decreased in a stepwise manner as volume increased. The lowest non-home discharge rate was associated with the volume category of ≥22 TEAs per year (20.3%). This study defines TEA volume strata for institutional outcomes. The highest TEA volume strata were associated with the lowest rates of 90-day complications, revisions, and non-home discharges and the lowest cost of care. This trend is likely attributable to the benefits of high-volume institutional experience and standardized patient-care processes.

Allograft discard risk index for lung transplantation

The demand for donor lungs continues to outpace the supply, yet nearly 75% of donor lungs intended for lung transplantation are discarded. We reviewed all donation after brain death organ donors listed within the UNOS Deceased Donor Database between 2005 and 2020. Univariable and multivariable analyses were run on the training set (n=69,355) with the primary outcome defined as lung discard, and the results were used to create a discard risk index (DSRI). Discard data were assessed at DSRI value deciles using the validation set (n=34,670), and differences in 1-year mortality were assessed using stratum-specific likelihood ratio (SSLR) analysis. Donor factors most associated with higher DSRI values included age > 65, PaO2 < 300, hepatitis C virus, and cigarette use. Factors associated with lower DSRI values included donor age < 40 and PaO2 > 400. The DSRI was a reliable predictor of donor discard, with a C-statistic of 0.867 in the training set and 0.871 in the validation set. The DSRI was not a reliable predictor of 30-day, 1-year, 3-year, and 5-year survival following transplantation (C-statistic 0.519-0.530). SSLR analysis resulted in three 1-year mortality strata (SSLR 0.88 in the 1st DSRI value decile, 1.03 in the 2nd-5th, & 1.19 in the 6th-10th). The factors leading to lung allograft discard are not the same as those leading to worse recipient outcomes. This suggests that with proper allocation, many of the grafts that are now commonly discarded could be used in the future donor pool with limited impact on mortality.

P85. Delayed staging during same hospitalization increases complication risk following adult spinal deformity surgery

<h3>BACKGROUND CONTEXT</h3> Since circumferential adult spinal deformity (ASD) surgery can lead to high surgical burden for the patient, surgeons may elect to stage these procedures during the same hospitalization. As a result, there is a trend toward planning same-hospitalization staged ASD surgeries by performing an anterior lumbar interbody fusion (ALIF) prior to a subsequent staged posterior fusion. <h3>PURPOSE</h3> The aim of this study was to determine optimal timing for staging (early vs delayed) thoracolumbar ASD surgery within the same hospitalization. <h3>STUDY DESIGN/SETTING</h3> Retrospective review of a multicenter database. <h3>PATIENT SAMPLE</h3> A total of 158 surgical ASD patients undergoing anterior and staged posterior spinal fusion ≥5 levels. <h3>OUTCOME MEASURES</h3> Our outcome measures were 90-day complication rates, postoperative alignment, and 2-year ODI. <h3>METHODS</h3> Using a prospective, multicenter database, we identified 158 surgical ASD patients undergoing first an anterior surgery followed by a staged posterior spinal fusion ≥5 levels during the same hospitalization. Stratum-specific likelihood ratio (SSLR) analysis was performed to calculate a cutoff point beyond which 90-day complications were increased. The cutoff generated through SSLR were confirmed with multivariable logistic regression analysis controlling for age, gender, levels fused for each stage, preoperative alignment, three-column osteotomy, and Charlson Comorbidity Index. The outcome measures were 90-day complication rates, postoperative alignment, and 2-year ODI. Multivariable analyses were performed with logistic, Poisson and linear regressions where appropriate. <h3>RESULTS</h3> Utilization of staged procedures increased 4x from 2008-2019, and the mean staging interval was 3 days (range 1-8). On SSLR analysis, patients were divided into two staging categories based on complication risk: early (<6 days, range 1-5, N=139) versus delayed (≥ 6 day, range 6-8, N=19). On bivariate analysis, the delayed group had higher 90-day complication rates (68.4% vs 32.4%, p=0.002) and longer operative times (638 min vs 739 min, p=0.020) relative to the early group. Adjusting for covariates on multivariable analysis, patients with delayed staging had significantly greater odds of 90-day complications (OR=7.57, p=0.006), and longer total operative time (beta=119, p=0.035). With respect to specific complications, delayed staging carries increased odds of infection compared to early staging (OR=11.8 p=0.031). There were no significant differences in postoperative alignment or 2-year ODI between the groups (p>0.05 for all). <h3>CONCLUSIONS</h3> Compared to early staging (< 6 days) of ALIF and subsequent staged posterior fusion, delayed staging (≥ 6 days) is associated with greater odds of 90-day complications, particularly infection. With the growing trend toward staged ASD surgeries, efforts should be made to shorten the interval between ALIF and subsequent posterior fusion. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.

265. Sarcopenia as a risk factor for complications following pedicle subtraction osteotomy

<h3>BACKGROUND CONTEXT</h3> Sarcopenia, the loss of muscle mass and function, has been linked to morbidity and mortality in several orthopedic procedures. At present, there is a paucity of data concerning sarcopenia in adult spinal deformity (ASD) surgery and particularly with respect to the most complex techniques performed on this population such as pedicle subtraction osteotomy (PSO). Due to the high surgical burden placed on patients undergoing PSOs, appropriate patient selection is critical for minimizing complications. <h3>PURPOSE</h3> The aim of this study was to 1) determine if sarcopenia was an independent risk factor for complications in ASD patients undergoing PSO, and 2) define discrete categories of complication risk based on sarcopenia severity. <h3>STUDY DESIGN/SETTING</h3> Single-center retrospective review. <h3>PATIENT SAMPLE</h3> Seventy-three ASD patients who underwent PSO. <h3>OUTCOME MEASURES</h3> The primary outcome measure was the presence of any 2-year complication: cardiac, pulmonary, neurologic, dural tear, epidural hematoma, wound infection, proximal junctional kyphosis (PJK), nonunion and revision surgery. Secondary outcome measures included intraoperative blood loss (EBL) and length of stay (LOS). <h3>METHODS</h3> We identified 73 ASD patients with available lumbar CT/MRI scans who underwent PSO with spinal fusion ≥5 levels at a tertiary care center from 2005-2014. Sarcopenia was assessed based on the psoas-lumbar vertebral index (PLVI), a validated marker for central sarcopenia. The PLVI was calculated by dividing a patient's average iliopsoas cross-sectional area (CSA) by the CSA of the L4 vertebrae. Using stratum-specific likelihood ratio (SSLR) analysis, patients were separated into three distinct sarcopenia groups based on complication risk. Cutoffs generated through SSLR were confirmed using multivariable regression controlling for demographic and surgical factors. <h3>RESULTS</h3> The mean age of the cohort was 61±8 years, 75% female. The mean PLVI was 0.84±0.28, with 47% of patients having any complication. Patients with a complication had a 27% lower PLVI on average than those without complications (0.76 vs. 0.91, p=0.021). SSLR analysis produced 3 complication categories: 32% complication rate for PLVI ≥ 0.81; 61% for PLVI 0.60-0.80; and 69% for PLVI < 0.60. Relative to patients with PLVI ≥ 0.81, those with PLVI 0.60-0.80 and PLVI < 0.60 had 3.2x and 4.3x greater odds of developing a complication, respectively (p<0.05 for both). With respect to individual complications, patients with PLVI < 1.0 had significantly higher risk of PJK (34% vs. 0%, p=0.022), while patients with PLVI < 0.8 had significantly higher risk of wound infection (12% vs. 0%, p=0.028) and dural tear (14% vs. 0%, p=0.019). There were no significant associations between sarcopenia, EBL and LOS (p>0.05 for all). <h3>CONCLUSIONS</h3> Increasing severity of sarcopenia is associated with significantly and incrementally increased risk of complications following PSO for ASD surgery. We define PLVI cutoffs (0.60 and 0.81) below which complication risk is significantly increased. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.

P147. The at-HOME score: A novel scoring system predicting discharge disposition following ASD surgery

<h3>BACKGROUND CONTEXT</h3> Adult spinal deformity (ASD) surgery requires an extended postoperative recovery period and oftentimes non-routine discharge (DC). Predicting which patients will likely have non-home DC has both clinical and economic implications. <h3>PURPOSE</h3> The aim of this study was to develop and internally validate a novel and simple score-based risk stratification system (At-HOME Score) to preoperatively identify patients mostly likely to require non-home DC. <h3>STUDY DESIGN/SETTING</h3> Retrospective review of a single-center surgical registry. <h3>PATIENT SAMPLE</h3> A total of 195 adult thoracolumbar deformity patients. <h3>OUTCOME MEASURES</h3> Home vs non-home DC. <h3>METHODS</h3> In total, the study cohort included 195 adult thoracolumbar deformity patients who underwent spine fusion ≥ 5 levels from 2006-2018. We split this cohort 50% for training the model and 50% for the validating the model. Data collected included patient demographics, comorbidities, surgical parameters, ODI, SRS, Charlson Comorbidity Index (CCI), and 30-day postoperative complications. We performed multivariable logistic regression on the training set to identify preoperative risk factors for non-home DC. The risk factors were weighted based on nomogram analysis and aggregated into the 11-point At-HOME model. Using area under curve (AUC) analysis, we internally validated the At-HOME score by applying it to the validation set with comparison to CCI. Optimal cutoffs for the scoring system were determined using stratum-specific likelihood ratio (SSLR) analysis. <h3>RESULTS</h3> Mean age at index surgery was 60 ± 13 years, 77% females. 45% (88) had a non-home DC. Points in the At-HOME Score were assigned as follows: +1 for preoperative SRS Activity domain < 3 (OR=2.4), +1 for Hypothyroidism, (OR=2.8), +2 for preoperative ODI ≥ 42 (OR=2.7), +3 for preoperative Motor weakness (OR=2.8), +4 for Elderly age > 65 years (OR=5.6) (p<0.05 for all). The At-HOME Score had an AUC of 0.81 in both the training and validation set, which was significantly higher than the AUC of CCI (p<0.001). SSLR analysis produced 4 distinct categories based on risk of non-home DC: 4% for score 0, 31% for score 1-5, 65% for score 6-7, 86% for score 8+. Relative to a score of 0, scores of 1-5, 6-7, and 8+ had 11x, 46x, and 150x greater odds of non-home DC (p<0.05 for all). <h3>CONCLUSIONS</h3> By using a robust combination of baseline demographics, objective clinical criteria, and patient reported outcomes, the At-HOME score was highly predictive of non-routine DC following ASD surgery. This tool was shown to be more predictive than CCI. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.

267. Establishing safety thresholds for surgical invasiveness based on frailty status in ASD surgery

<h3>BACKGROUND CONTEXT</h3> Both frailty and surgical invasiveness are correlated with increased risk of complications following ASD surgery, yet there is no accepted risk-stratification system combining these factors. <h3>PURPOSE</h3> The aim of this study is to ascertain data-driven categories defining the risk for 90-day complications following ASD surgery based on frailty and surgical invasiveness. <h3>STUDY DESIGN/SETTING</h3> Retrospective review of a multicenter database. <h3>PATIENT SAMPLE</h3> A total of 1,272 surgical ASD patients undergoing spinal fusions ≥5 levels <h3>OUTCOME MEASURES</h3> Ninety-day complication risk, 90-day complication rate, invasiveness thresholds for increased complications. <h3>METHODS</h3> Using a prospective, multicenter database, we identified 1,272 surgical ASD patients undergoing posterior spinal fusion of at least five levels. Patients were separated into three frailty groups based on the ASD-FI frailty index. Within each frailty group, stratum-specific likelihood ratio (SSLR) analyses were performed to define frailty-based surgical invasiveness cutoffs associated with increased risk for 90-day complications. Cutoffs generated through SSLR were confirmed with multivariable logistic regression analysis controlling for age, alignment, and ODI. <h3>RESULTS</h3> Mean age was 60 ± 14 years, 74% females. Of 1,272 surgical ASD patients, 319 (35%) were nonfrail (NF), 667 (52%) frail (F), and 286 (23%) severely frail (SF). Mean SI was 93 ± 35. The mean complication rate was 35.4% for NF, 37% for F and 43.6% for SF. SSLR analysis of NF patients produced 2 complication categories: 33% complication rate for SI < 153 and 73% for SI ≥ 153. NF patients with SI ≥ 153 had 4.14x higher odds of complications than NF patients with SI <153 (p=0.04). SSLR analysis of F patients produced 3 complication categories: 26% complication rate for SI < 60, 42% for SI of 60 to 179, and 64% for SI ≥ 180. Relative to F patients with SI < 60, F patients with SI 60 to 179, and SI ≥ 180 had 2.4x and 6.7x higher odds of complications, respectively (p<0.01 for both). SSLR analysis of SF patients produced 2 complication categories: 35% complication rate for SI < 82 and 49% complication rate for SI ≥ 82. SF patients with SI ≥ 82 had 1.77x higher odds of complications than SF patients with SI < 82 (p=0.04). <h3>CONCLUSIONS</h3> Frailer groups were found to have lower SI cutoffs, indicating the less invasiveness surgeries can be performed on frail patients to minimize complication risk. This risk-stratification system is useful for counselling patients regarding their risk of complication. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.

Evidence-Based Hospital Procedural Volumes as Predictors of Outcomes After Revision Hip Arthroplasty

BackgroundThe aim of this study is to define the evidence-based institutional volume-outcome relationship in revision hip arthroplasty. We hypothesized that high-volume centers would be associated with superior outcomes, and that stratum-specific likelihood ratio (SSLR) analysis would delineate concrete volume thresholds for optimizing outcomes. MethodsThe Nationwide Readmission Database was queried from 2011 to 2016 for patients undergoing revision hip arthroplasty. SSLR analysis was used to determine hospital volume cutoffs specific for outcomes of interest. Volume categories were confirmed with multivariate regression. ResultsSSLR analysis produced distinct hospital volume cutoffs for all outcomes. Each subsequent volume threshold diminished patients’ risk for adverse outcomes. Tertiles were identified for 90-day infection (≤6, 7-51, ≥52 cases per year). Quartiles were found for 90-day readmission (≤5, 6-15, 16-79, ≥80), 90-day prosthesis-related complication (≤5, 6-16, 17-65, ≥66), 90-day dislocation (≤5, 6-19, 20-79, ≥80), and non-home discharge (≤5, 6-15, 16-40, and ≥41). Quintiles were generated for extended length of stay >2 days (≤2, 3-10, 11-20, 21-30, ≥31). Heptiles were produced for medical complications within 90 days (≤2, 3-8, 9-16, 17-51, 52-89, ≥90). ConclusionThis is the first known study to define evidence-based thresholds for the impact of hospital volume on revision joint arthroplasty. This supports the notion that institutional volume functions as a surrogate for protocolized interdisciplinary coordination of care and surgical experience, and that high-volume centers offer enhanced outcomes for complex cases. Additional studies should investigate the potential role for incentivization of such institutions, as they offer optimal outcomes for revision hip arthroplasty.

Data-Driven Volume Thresholds Predictive of Improved Outcomes after Orthotopic Heart Transplantation

It is well established that "high volume" centers have better outcomes for complex surgical care including transplantation. However, few studies have used data-driven methodology to establish clinically meaningful volume thresholds predictive of improved outcomes for centers performing orthotopic heart transplantation (OHT). All OHT cases performed between 2005 and 2015 (with follow-up through 2016) within the United Network for Organ Sharing (UNOS) database were reviewed for potential eligibility. All patients ≥18 years of age undergoing isolated OHT were included. Stratum-specific likelihood ratio analysis (SSLR) was used to establish volume thresholds predictive of the primary outcome of mortality at 1-year. Preoperative characteristics and postoperative outcomes were compared between volume strata using χ2 tests (categorical) or ANOVA (continuous). A total of 21,800 recipients (mean age 52.6 ± 12.8 years; 25.2% female) were included in the final cohort. Three distinct center volume strata were identified from the analysis: low (<10 OHT/year), intermediate (10-18 OHT/year), and high volume (≥19 OHT/year). More than half of recipients (57.4%) were transplanted at high volume centers. Compared to low volume centers, recipients at high volume centers had higher survival at 30-days and 1-year (96.0% vs 94.2% and 90.0% vs 86.6%, respectively; both P<.001). These recipients also had a lower rate of post-transplant dialysis (10.0% vs 12.8%; P<.001) despite higher rates of intra-aortic balloon pump and inotropic medication usage (table), along with shorter length of stay (20 days vs 23 days; P=.002). The results of our study demonstrate that higher volume centers perform better than their lower volume counterparts in terms of early mortality after OHT. Our study adds to the existing literature by providing specific, data-driven volume thresholds predictive of improved early outcomes in the contemporary transplant and circulatory support landscape.

Defining the Volume-Outcome Relationship in Reverse Shoulder Arthroplasty: A Nationwide Analysis.

As the utilization of reverse total shoulder arthroplasty (RSA) grows, it is increasingly important to examine the relationship between hospital volume and RSA outcomes. We hypothesized that hospitals that perform a higher volume of RSAs would have improved outcomes. We also performed stratum-specific likelihood ratio (SSLR) analysis with the aim of delineating concrete definitions of hospital volume for RSA. The Nationwide Readmissions Database was queried for patients who had undergone elective RSA from 2011 to 2015. Annual hospital volume and 90-day outcome data were collected, including readmission, revision, complications, hospital length of stay (LOS), supramedian cost, and discharge disposition. SSLR analysis was performed to determine hospital volume cutoffs associated with increased risks for adverse events. Cutoffs generated through SSLR analysis were confirmed via binomial logistic regression. The proportion of patients receiving care at high-volume centers increased from 2011 to 2015. SSLR analysis produced hospital volume cutoffs for each outcome, with higher-volume centers showing improved outcomes. The volume cutoffs associated with the best rates of 90-day outcomes ranged from 54 to 70 RSAs/year, whereas cost and resource utilization cutoffs were higher, with the best outcomes in hospitals performing >100 RSAs/year. SSLR analysis of 90-day readmission produced 3 hospital volume categories (1 to 16, 17 to 69, and ≥70 RSAs/year), each significantly different from each other. These were similar to the strata for 90-day revision (1 to 16, 17 to 53, and ≥54 RSAs/year) and 90-day complications (1 to 9, 10 to 68, and ≥69 RSAs/year). SSLR analysis produced 6 hospital volume categories for cost of care over the median value (1 to 5, 6 to 25, 26 to 47, 48 to 71, 72 to 105, and ≥106 RSAs/year), 5 categories for an extended LOS (1 to 10, 11 to 25, 26 to 59, 60 to 105, and ≥106 RSAs/year), and 4 categories for non-home discharge (1 to 31, 32 to 71, 72 to 105, and ≥106 RSAs/year). We have defined hospital surgical volumes that maximize outcomes after RSA, likely related to surgical experience, ancillary staff familiarity, and protocolized pathways. This information may be used in future policy decisions to consolidate complex procedures, such as RSA, at high-volume destinations, or to encourage lower-volume institutions to strategize an approach to function as a higher-volume center. Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.