10.20 Is Hospital Teaching Status Associated with Outcomes in Patients Admitted with EGS Conditions?

Posted on January 18, 2016

S. Jiao1,2, C. K. Zogg1, J. W. Scott1, L. L. Wolf1, A. Shah2, M. A. Chaudhary1, N. Changoor1, A. Salim3, E. B. Schneider1, A. H. Haider1 1Brigham And Women’s Hospital,Center For Surgery And Public Health, Harvard Medical School, Harvard T H Chan School Of Public Health,BOSTON, MA, USA 2Mayo Clinic In Arizona,Department Of Surgery,Phoenix, AZ, USA 3Brigham And Women’s Hospital,Division Of Trauma, Burns And Surgical Critical Care,BOSTON, MA, USA

Introduction:
Differences in outcomes between teaching (TH) versus non-teaching hospitals (NTH) have been reported among emergency general surgery (EGS) patients. Using a nationally representative sample, this study examined associations between hospital teaching status and surgical outcomes among EGS patients.

Methods:
Adult patients (≥18y) with a primary ICD-9 code for an EGS condition were identified in the 2007-2011 Nationwide Inpatient Sample. Analyses included only patients with complete data (except for race) who were non-electively admitted and underwent operative procedures. Cases were categorized into those managed at TH versus NTH and assessed for differences in patient- and hospital-level factors. Stratified analyses were conducted within the following diagnostic groups: colorectal pathology, intestinal obstruction, appendiceal disorder, hernias and vascular pathology. Outcomes included: major complications, in-hospital mortality, non-routine discharge, length of stay (LOS), and total hospital cost. Data were weighted to obtain national estimates, and clustering of patients within hospitals was accounted for. A combination of coarsened-exact matching and multivariable logistic/linear (family gamma, link log) regression was used to adjust for potential confounding.

Results:
A weighted total of 18,915,504 adult patients with complete data were admitted to inpatient care with EGS conditions between 2007-2011, 40.26% of whom were treated at THs. Overall, 6,395,624 (33.81%) underwent operations of which 69.02% were non-elective (64.82% in TH and 72.06% in NTH). The most common diagnoses within each of the 5 diagnostic groups listed were: diverticulitis, intestinal/peritoneal adhesion with obstruction, acute appendicitis, ventral hernia and arterial embolism, respectively. The odds of mortality was higher for all operatively managed EGS patients in TH (OR 1.07, 95% CI 1.01-1.13), which was also true for the most common diagnoses for colorectal pathologies (OR 1.40, 95% CI 1.13-1.73) and intestinal obstructions (OR 1.18, 95% CI 1.02-1.36), but not for the other 3 conditions. Interestingly, the odds of extended LOS was increased among patients with the most common diagnoses in colorectal pathology, intestinal obstruction and hernias if treated at TH, but was decreased for acute appendicitis. Treatment at TH was associated with increased odds of having total cost exceed the 75th percentile among patients with intestinal obstruction or ventral hernia.

Conclusion:
Differences in patient diagnosis-specific case-mix may account for much of the previously reported variability in EGS outcomes between TH and NTH. Future research should carefully consider the broad spectrum of diagnoses that constitute EGS conditions.

10.19 Disparities in the Receipt of Rehabilitation: A National Inspection of Acute Care Surgery Patients

Posted on January 18, 2016

M. A. Chaudhary1, A. Shah3, C. K. Zogg1, D. Metcalfe1, O. Olufajo1, E. J. Lilley1, A. Ranjit1, A. B. Chapital3, D. J. Johnson3, J. M. Havens2, A. Salim2, Z. Cooper1, A. H. Haider1 1Brigham And Women’s Hospital,Center For Surgery And Public Health, Harvard Medical School, Harvard T H Chan School Of Public Health,Boston, MA, USA 2Brigham And Women’s Hospital,Division Of Trauma, Burns And Surgical Critical Care,BOSTON, MA, USA 3Mayo Clinic In Arizona,Department Of Surgery,Phoenix, AZ, USA

Introduction:
Post-acute rehabilitation is increasingly recognized as a vital component of surgical care, necessary to restore patients’ pre-disease functional status. Unequal receipt of such care may prevent full recovery in underprivileged patients. Disparities in utilization of Acute Care Surgery (ACS) have been documented, but little is known about disparities in post-discharge rehabilitation services. This study sought to determine whether disparities exist in the receipt of post-discharge rehabilitation in ACS patients.

Methods:
The Nationwide Inpatient Sample (2007-2011) was queried for adult patients (≥18y) with an ACS primary diagnosis, including trauma and emergency general surgery (EGS) conditions defined by the American Association for the Surgery of Trauma (AAST). Two trauma subgroups – traumatic brain injury (TBI) and spinal cord injury (SCI) – were analyzed separately, as evidence suggests that they have specific rehabilitation needs. Patients that died in hospital were excluded and inpatients who survived were assessed for differential receipt of post-discharge inpatient rehabilitation. Multivariable logistic regression models were used to determine independent predictors of discharge to rehabilitation, adjusting for influence of patient- (age, race, gender, insurance status, income, disease severity, complications) and hospital- (volume, teaching status, location, bed size, geographical region) level covariates and accounting for clustering of patients within hospitals. Trauma, TBI and SCI models were further adjusted for Injury Severity Score (ISS).

Results:
A total of 5,228,453 patient records were included, weighted to represent 26,353,162 patients nationwide. Among ACS patients, 27.4% (n=1,460,934) were admitted for trauma and 72.6% (n=3,872,622) for EGS. Of these 133,439 (2.6%) were discharged to rehabilitation facilities. In the sub-groups, 7.7% of trauma, 0.6% of EGS, 6.5% of SCI and 2.3% of TBI patients were discharged to rehabilitation facilities. Black and Hispanic patients had 15-43% lower risk-adjusted odds of rehabilitation discharge relative to White patients, for both trauma and EGS diagnosis. Similarly, Medicaid beneficiaries (OR: 0.90 [0.87-0.93]) and uninsured patients (OR: 0.38 [CI: 0.37-0.40]) were less likely to be discharged to inpatient rehabilitation after trauma. ACS Patients treated at teaching (OR: 1.44 [CI: 1.42-1.46]) and urban (OR: 1.80 [CI: 1.75-1.83]) hospitals were more likely to be discharged to rehabilitation facilities. These findings persisted in the SCI sub-group analysis but became non-significant for the TBI group.

Conclusion:
Historically disadvantaged minorities, Medicaid beneficiaries, and those treated at rural and non-teaching centers had disparate receipt of inpatient rehabilitation, which may limit recovery. Further work should investigate the implications of these findings on post-operative and post-injury functional and quality-of-life outcomes.

10.18 Colonic Volvulus: An ACS-NSQIP Analysis

Posted on January 18, 2016

S. E. Koller2, E. A. Busch2, M. M. Philp2, H. Ross2, H. A. Pitt1,2 1Temple University,Health System,Philadelpha, PA, USA 2Temple University,School Of Medicine,Philadelpha, PA, USA

Introduction: Volvulus is uncommon and frequently occurs in older, frail patients who may be more likely to experience adverse surgical outcomes. Most analyses are single-institution studies performed over long time periods or come from administrative databases with poorly defined surgical outcomes. Thus, the aims of this study were to characterize the risk factors associated with poor outcomes in surgically managed volvulus patients using a large, contemporary database with well-defined 30-day outcomes.

Methods: The 2012 and 2013 American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) Participant Use Files were employed for this analysis. Standard ACS-NSQIP definitions were used for patient, procedure and outcome data. Eight preoperative variables were combined as a composite to determine patient frailty. Patients with five more of these variables were determined to be frail. Primary outcomes of interest were overall and serious morbidity as well as 30-day mortality. Secondary outcomes included anastomotic leak and return to the operating room (OR). STATA 13.1 was used for univariate analyses and multivariable logistic regression.

Results:Colorectal resection was performed for volvulus in 903 of 29,219 patients (3.1%). The mean age was 49.5 years, but 334 patients (37.0%) were 75 years or older. Three hundred seventy-eight were men (41.9%), 691 were Caucasian (76.5%) and 95 were African American (10.5%). Ninety-five were frail (10.5%) and fourteen had significant weight loss (1.6%). Fifty-eight percent of the operations were performed emergently. A stoma was performed in 164 patients (18.2%), most frequently with partial colectomy (27.2%). Outcomes for all patients as well as for those who were frail or had a stoma are presented in the Table.

Factors that were independently associated with overall morbidity included age 75 years or more (OR 1.62, 95% CI 1.16-2.26, p<0.01), male sex (OR 1.49, 95% CI 1.09-2.02, p<0.02), African American race (OR 1.62, 95% CI 1.01-2.58, p<0.05) and weight loss (OR 5.04, 95% CI 1.50-16.97%, p<0.01). BMI, emergent procedure and operative approach did not influence overall morbidity.

Conclusion:Outcomes of colonic resection for volvulus are worse in the elderly and frail as well as in African Americans and men. Performance of a stoma may reduce the risk of an anastomotic leak. The decision to perform surgery in these high-risk patients should be undertaken with caution.

10.17 WHipple-ABACUS, A Simple, Validated Risk Score for 30-Day Mortality After Pancreaticoduodenectomy

Posted on January 18, 2016

E. Gleeson1, M. F. Shaikh1, A. E. Poor1, P. A. Shewokis1, J. R. Clarke1, D. S. Lind1, W. C. Meyers1, W. B. Bowne1 1Drexel University College Of Medicine,Surgery,Philadelphia, Pa, USA

Introduction: Pancreaticoduodenectomy (PD) is a high-risk procedure. There is need for simple validated risk models to better identify 30-day mortality. The goal of this study was to identify independent, preoperative risk factors and to develop a simple risk score to predict 30-day mortality after PD.

Methods: We reviewed all patients who underwent PD from 2005-2012 in the ACS-NSQIP databases. Logistic regression was used to identify preoperative risks for 30-day mortality from a development cohort ([DEVEL] random 80% of the database). The WHipple-ABACUS score was created using weighted beta coefficients and predictive accuracy was assessed on the validation cohort ([VALID] the remaining 20%) using a receiver operator characteristic curve and calculating the area under the curve (AUC).

Results: The 30-day mortality rate was 2.7% for patients who underwent PD (n=14,993). The DEVEL identified 8 independent risk factors: hypertension with medication, history of cardiac surgery, age > 62, bleeding disorder, albumin <3.5g/dL, disseminated cancer, use of steroids and systemic inflammatory response syndrome (SIRS). The score created from weighted beta coefficients (see Figure) had an AUC=0.71 (95%CI 0.66 to 0.77) using the VALID. Using the WHipple-ABACUS score: WHipple-ABACUS = hypertension With medication + History of cardiac surgery + Age>62 + 2*Bleeding disorder + Albumin<3.5g/dL + 2*disseminated Cancer + 2*Use of steroids + 2*SIRS, mortality rates increase with increasing score (p<0.001).

Conclusion: While other risk scores exist for 30-day mortality after PD, we present a simple, validated score developed using exclusively preoperative predictors that surgeons should use to optimize co-morbidities and inform patients of risk with this procedure.

10.16 Economies of Scale in the Provision of Minimally Invasive Surgery

Posted on January 18, 2016

L. Kuo1, K. D. Simmons1, R. R. Kelz1 1Hospital Of The University Of Pennsylvania,Surgery,Philadelphia, PA, USA

Introduction: Minimally invasive surgery (MIS) is often associated with higher costs than open operations due to expensive equipment. Institutions that perform higher volumes of an operation are associated with improved clinical outcomes, such as shorter lengths of stay and fewer postoperative complications, over low-volume institutions, and these improved outcomes may translate into lower associated costs. We sought to examine if economies of scale existed in the provision of MIS surgery.

Methods: A unique inpatient database using discharge data from three high-volume surgical states (California, Florida and New York) over a five-year period was used. Four common general surgery operation groups, for which the MIS approach offers improved or equivalent outcomes to the open technique, were selected: cholecystectomy, appendectomy, bariatric, and antireflux. For each operation, the per-hospital number of MIS operations and the associated wage-adjusted cost was obtained. Hospitals were categorized as low-volume (lowest 20%), moderate, and high volume (highest 20%) for each of the five MIS operations. The median associated wage-adjusted cost was compared between low, moderate and high-volume categories for each operation using the Kruskal-Wallis. The Bonferroni correction was used to adjust for multiple comparisons.

Results: For cholecystectomy, appendectomy, and antireflux procedures, low-volume institutions had a significantly higher median cost than moderate- and high-volume institutions. For cholecystectomy and appendectomy, high-volume institutions had the lowest median cost. For antireflux operations, moderate-volume institutions had the lowest median cost, although the difference may not be clinically significant. For bariatric procedures, cost increased with volume. See Table 1.

Conclusion: Economies of scale exist between low- vs. moderate- and high-volume institutions for all operations studied except bariatric. The highest-volume institutions did not always have the lowest associated costs. Costs associated with bariatric operations increased with volume, which may be related to restrictions on where high-risk patients receive bariatric surgery.

10.15 Sarcopenia Predicts Outcomes in Patients Undergoing Pancreatectomy for Neoplasia

Posted on January 18, 2016

A. Sagnotta1, F. Carbonetti1, M. De Siena1, L. Mangogna1, C. Mattana1, P. Magistri1, G. Ramacciato1, G. Nigri1 1Sapienza University Of Rome,Department Of Medical And Surgical Sciences And Traslational Medicine,Rome, RM, Italy

Introduction: Sarcopenia is a surrogate marker of patient frailty that estimates the physiologic reserve of an individual patient. Few studies have examined the association between the presence of sarcopenia and outcomes following pancreatic surgery. These studies had traditionally defined sarcopenia measuring total psoas area (TPA) or volume (TPV). However, the entire skeletal muscle area (SMA) to the midpoint of L3 vertebral body is considered a more objective measure of sarcopenia. The aim of this study was to evaluate the prevalence of sarcopenia and to investigate its impact on long-term outcomes in patients who underwent pancreatectomy for tumor.

Methods: One hundred and six patients who underwent pancreatectomy for malignancy between 2003 and 2014 met the inclusion criteria. The skeletal muscle area (SMA) was measured on perioperative CT scans at the level of the third lumbar vertebra (L3). The L3 SMA (cm2) was normalized by the square of the height (m2) to obtain the skeletal muscle index (L3 SMI, cm2/m2). We established different gender specific cut-offs using the ROC curve method (43.35 cm2/m2in men and 34 cm2/m2 in women, respectively). The prevalence and impact of sarcopenia was assessed relatively to other clinic-pathological factors. Univariate and multivariate analyses evaluating prognostic factors of overall survival were performed, including preoperative, surgical and histopathological factors.

Results: Mean age was 67.4 years and 52.1% was female. Pancreatic cancer represents 69.9% of all cases. Pancreatoduodenectomy was performed in 81.5% and an R0 resection was obtained in 114 cases (78.1%). Mean SMI was lower in women (32.8 cm2/m2) versus men (41.1 cm2/m2, p<0.0001); 97 patients (66.4%) were sarcopenic, 66 (45.2%) were overweight/obese and 34 (23.3%) were both (p=0.003). Sarcopenia was significantly related to histology (p=0.031), body mass index (p<0.001), lower value of albumin (p=0.002), modified Glasgow Prognostic Score (mGPS, p=0.036) and prognostic nutritional index (PNI, p=0.003). Overall morbidity was 41.8% and 90-days mortality was 6.8%. Overall survival (OS) was significantly shorter in sarcopenic patients than in non-sarcopenic patients (3-year OS 60.3vs 35.6% and 5-year 47.5 vs 20.7%, log-rank p=0.008). In multivariate Cox regression analysis, sarcopenia remained independently associated with poor overall survival (HR 2.178; 95% CI 1.11-4.273; p=0.024).

Conclusion: Sarcopenia, as assessed by L3 SMI, was found in 66.4% of patients underwent pancreatectomy. It was correlated to worse nutritional status and could be a more objective measure of patient frailty even in overweight/obese patients. This condition independently impacts survival and may be a useful preoperative tool for clinical decision-making in patients undergoing curative pancreatic resection.

10.14 Partnerships between Surgeons and Industry

Posted on January 18, 2016

S. Chrabaszcz1, R. Rajeev1, B. Klooster1, T. Gamblin1, F. Johnston1, K. Turaga1 1Medical College Of Wisconsin,Division Of Surgical Oncology,Milwaukee, WI, USA

Introduction: Evolving healthcare demands have necessitated changing relationships between surgeons and industry. Honoraria, speakers’ fees, consulting/advisory fees, gifts and other non-ownership payments can significantly introduce a conflict of interest in healthcare. We hypothesized that there exists significant variation between the surgical subspecialties in the non-research payments from industry likely due to significant difference in adoption of modern technology between them.

Methods: The Physician Payments Sunshine Act requires the health care industry to report payments made to physicians to the Centers for Medicare and Medicaid Services (CMS). This information is recorded in the Open Payments database (OPD) beginning with payments made on August 1, 2013. The OPD was reviewed to identify non-research payments made to physicians in surgical subspecialties between August 1, 2013 and December 31, 2013. Data on payments made to individual physicians and the number of surgeons receiving payments were compared across different surgical subspecialties. Focus group methodology was used to ascertain hierarchical ordering of adoption of technology amongst surgical subspecialties.

Results: In 2013, there were 131,819 general transactions from industry to 22,167 surgeons. Subspecialty classifications included General (and Trauma) Surgery (64.7%), Colorectal Surgery (4.3%), Critical Care Surgery (1.5%), Surgical Oncology (2%), Trauma only (1.3%), Vascular Surgery (11.2%), Thoracic (13%) and Transplant Surgery (1.1%). Median payment per transaction was $27 (IQR $14.4-$104.8) while the mean payment per transaction was $274 (± $7181.5). The mean number of payments per surgeon was 5.9. The specialty with the highest mean payment per transaction and mean payment per surgeon was critical care surgery ($2013 and $6872 respectively). Mean payment per surgeon was higher for subspecialists as compared to general surgeons ($2114 Colorectal Surgery, $6872 Critical Care Surgery, $1095 General Surgery, $1814 Pediatric Surgery, $1725 Surgical Oncology, $2383 Thoracic Surgery, $2566 Transplant Surgery, $545 Trauma Surgery, and $1800 Vascular Surgery). The maximum payment to one individual in a unique transaction was in critical care surgery ($2.3 million). Hierarchical ordering of specialties in magnitude of technological adoption was not possible using our current methods.

Conclusion: Critical care specialists and other surgical subspecialties are more likely to receive higher non-research payments from industry as compared to general surgeons. While we believe that this phenomenon is an effect of rapid adoption of evolving technology in ICUs and the operating rooms, we were unable to substantiate our hypothesis using current methods.

10.13 Effect of Perioperative Fluid Class, Volume, and Timing on Incidence of Postoperative Ileus

Posted on January 18, 2016

M. S. VandeHei1, C. M. Papageorge1, M. Murphy1, S. E. Tevis1, G. D. Kennedy1 1University Of Wisconsin,Department Of Surgery,Madison, WI, USA

Introduction: Postoperative ileus (POI) complicates approximately 20-30% of colon surgeries. It has been proposed that bowel edema at least partially contributes to the etiology of POI. Additionally, previous studies have shown decreased fluid administration can decrease POI incidence. We sought to investigate how the class, volume, and timing of fluid administration impact POI incidence.

Methods: We conducted a retrospective cohort study of 300 patients with rectal cancer who underwent partial colectomy or proctectomy at the University of Wisconsin Hospital and Clinics from 2008 to 2015. Chart review was performed to collect data on intra-operative fluid type and volume as well as fluid volume data for postoperative day one. The primary outcome was postoperative ileus, defined as presence of a nasogastric tube for nausea or vomiting or NPO status on postoperative day four or later. Volumes of crystalloid (0.9% saline and Ringer’s lactate) were divided into quartiles and volumes of colloid (5% albumin, hydroxyethyl starch, and blood products) were divided based on whether or not colloid was administered intraoperatively. Association with POI was assessed using Chi-square tests and binary logistic regression. Covariates were chosen from comorbidities, demographics, and operation specifics based on relevance to hypothesis and relative significance on univariate analysis. Secondly, we compared the ratio of the rates of fluid administration for the intraoperative period and the period through postoperative day 1.

Results: A total of 300 patients were included in our study, with an overall POI incidence of 30.0% (90 out of 300). On univariate analysis, we found that increasing intraoperative crystalloid administration was associated with increasing rates of POI (1st quartile: 16.3%; 2nd quartile: 31.5%; 3rd quartile: 34.2%; 4th quartile: 39.2%; P = 0.012). Total intra-operative colloid administration showed no association with development of POI (P = 0.065). Multivariate analysis confirmed a relationship between intra-operative crystalloid volume and risk of POI. Results are shown in table 1. Other significant predictors of POI included older age, and operative approach. Finally, we found no correlation between the incidence of POI and the ratio of fluid administration rates in the intraoperative and postoperative phases of care.

Conclusion: In patients undergoing colorectal surgery, lower intraoperative crystalloid administration was associated with lower rates of POI, while colloid did not appear to affect ileus rate. Therefore, limiting the volume of crystalloid administered may be effective at reducing the incidence of POI.

10.11 IHC as a Reliable Method for Detection of BRAF-V600E Mutation in Melanoma: A Meta-Analysis.

Posted on January 18, 2016

M. A. Anwar1, F. Murad1, E. Dawson1, Z. Y. Abd Elmageed1, K. Tsumagari1, E. Kandil1 1Tulane University School Of Medicine,General Surgery,New Orleans, LA, USA

Introduction:
BRAF-V600E mutation is associated with tumor aggressiveness and poor prognosis in melanoma patients. Identification of this mutation is clinically important as we now have FDA-approved targeted therapies, such as BRAF and MEK inhibitors, which have shown to retard the disease progression in these patients. Detection of BRAF-V600E by genetic analysis using PCR is the gold standard method for melanoma cases. However, immunohistochemistry (IHC) using VE1 antibody is rapidly immerging as a trustworthy method for determination of mutation status in patients’ specimens.

Objective:

To assess the reliability of IHC compared to genetic methods for successful identification of BRAF-V600E mutation in melanoma tissue specimens.

Study

Design:

Systematic review and meta-analysis of English language studies comparing IHC with genetic analysis for the detection of BRAF-V600E mutation in melanoma patients was performed.

Methods:

A literature search of PubMed, Web of Science, and Embase was performed for studies comparing IHC to genetic analysis for the detection of BRAF in melanoma patients published through May 28, 2015. Pooled sensitivity, specificity, diagnostic odds ratio, positive and negative likelihood ratios were calculated using a bivariate model. Logit estimates of sensitivity and specificity with their respective variances were used to plot a hierarchical ROC curve and area under the curve. Heterogeneity was assessed using the Q and I-squared statistics.

Results:
Initial literature search resulted in 287 articles. After two independent reviews and consensus-based discussion to resolve disparities; 21 studies involving a total of 1,753 cases met the eligibility criteria and were included in the analysis. The pooled sensitivity of IHC was 0.97; 95% CI (0.94-0.98), specificity 0.99; 95% CI (0.97-1.00), positive likelihood ratio 158.3; 95% CI (33.8-740.8), negative likelihood ratio 0.03; 95% CI (0.02-0.06), and diagnostic odds ratio 4672 (1104-19780). A high heterogeneity was observed between these studies (Q value of 28.77 & I2 value of 93; 95% CI (87-99) which may be explained by studies using different cutoff values for interpretation of IHC. High accuracy of IHC was depicted by AUC in the ROC curve which was 0.99; 95 % CI (0.98-1.00).

Conclusion:
Meta-analysis demonstrates that IHC is highly sensitive and specific for the detection of BRAF-V600E in melanoma cases. IHC is likely to be useful method in BRAF mutation detection because it is highly comparable to the genetic methods. Any negative or low staining cases may be selected to undergo genetic analysis based on other clinical and histopathological features.

10.12 Comparative ERCP Outcomes of Gastroenterologists and Surgeons:Analysis of National Inpatient Sample

Posted on January 18, 2016

J. A. Cooper1, S. Desai1, S. Scaife1, C. Gonczy1, J. Mellinger1 1Southern Illinois University School Of Medicine,General Surgery,Springfield, IL, USA

Introduction:
Endoscopic retrograde cholangiopancreatography (ERCP) is performed by both gastroenterologists and surgeons. There has been recent controversy regarding training paradigms for gastrointestinal endoscopy. No prior studies have evaluated comparative outcomes as a function of training background. This study utilized the National Inpatient Sample (NIS) to assess ERCP outcomes as a function of specialty background.

Methods:

NIS data was queried from 2007-9. Gastroenterologists and surgeons were identified by procedural profiles and unique physician identifiers. Comorbidity was assessed via Charlson Score. Outcomes including cost, length of stay (LOS), and mortality were analyzed, with and without propensity score matching (PSM). Comparison for statistical significance was accomplished via t-test.

Results:
A total of 198,661 ERCP’s were identified, of which 158,318 (79.7%) were performed by surgeons. Surgeons exhibited longer LOS (8.7 vs. 7.2 days), overall cost ($24,739 vs. $16,960), and mortality (3.9% vs. 1.2%, odds ratio 3.3), with p<0.001 for all measures. 71.6% of surgical patients, vs. 19.6% of gastroenterologic, underwent subsequent inpatient laparoscopic cholecystectomy or laparotomy. Outcome differences persisted when PSM included performance of subsequent laparoscopic cholecystectomy. Evaluation of minimum performance standards revealed up to 5-fold increased mortality for providers who performed less than 5 ERCP’s/year, irrespective of specialty background.

Conclusion:

Gastroenterologists demonstrate favorable gross outcomes compared to surgeons performing ERCP. Differences may correlate in part with more frequent subsequent surgical management of comorbid conditions by surgical providers. Lower volume providers achieve inferior outcomes regardless of specialty background. Analyses of this type may help inform discussions on optimal training and proficiency paradigms for therapeutic endoscopy.

10.10 A Calculated Risk: Performing Laparoscopic Cholecystectomy on Patients with End Stage Renal Disease

Posted on January 18, 2016

P. J. Chung1, M. J. Lee2, M. C. Smith1, A. E. Alfonso1, G. Sugiyama1 1SUNY Downstate Medical Center,Department Of Surgery,Brooklyn, NY, USA 2SUNY Downstate Medical Center,College Of Medicine,Brooklyn, NY, USA

Introduction: End stage renal disease (ESRD) is a multifactorial disease linked to socioeconomic status and associated with worse surgical outcomes. We explore intraoperative and postoperative outcomes in patients with cholecystitis undergoing laparoscopic cholecystectomy (LC) while adjusting for clinical, socioeconomic, and demographic variables.

Methods: The Nationwide Inpatient Sample from 2005 to 2012 was used to identify adult patients that underwent LC for cholecystitis using International Classification of Diseases 9th Revision codes. Patients who underwent surgery for gallbladder disease without cholecystitis or had acute renal failure were excluded. Outcomes of interest were mortality, common bile duct injury, conversion to open, intraoperative complications, postoperative complications (mechanical, respiratory, digestive system, cardiovascular, nervous system, postoperative infections), length of stay (LOS), and total charge. Univariate analysis was performed using t-test for continuous variables and chi-square test for categorical variables. Multivariable models were created that adjusted for age, gender, race, type of insurance, income demographics, year of admission, elective status, number of Elixhauser comorbidities, and presence of end stage renal disease (ESRD).

Results: Of 512,073 patients that underwent LC, 5,020 had ESRD. On univariate analysis, the ESRD cohort had a higher incidence of mortality (3.8% vs 0.37%, p <0.0001) and complications: intraoperative (3.0% vs 1.4%, p<0.0001), mechanical wound (1.7% vs 0.4%, p <0.0001), respiratory (3.8% vs 1.8%, p<0.0001), cardiovascular (1.6% vs 0.7%, p<0.0001), nervous system (0.1% vs 0.04%, p=0.0426), and postoperative infections (8.4% vs 5.3%, p<0.0001). ESRD patients had higher median LOS (7.0 days vs 3.0 days, p<0.0001) and total charge ($55,920 vs $27,371, p<0.0001). Multivariate analysis showed ESRD as an independent risk factor for mortality (OR 3.94, 95% CI 3.30-4.69, p<0.0001), mechanical wound complications (OR 1.90, 95% CI 1.48-2.45, p<0.0001), and intraoperative complications (OR 1.46, 95% CI 1.21-1.76, p<0.0001). Patients with ESRD were found to have decreased risk for postoperative infections (OR 0.84, 95% CI 0.75-0.94, p=0.0023). Negative binomial regression analysis showed that ESRD patients had LOS 50.4% longer than non-ESRD patients (95% CI 47.5%-53.3%, p<0.0001). Linear regression analysis showed that, after adjustment, ESRD patients had total charge 6.82% higher than non-ESRD patients (95% CI 5.02%-8.66%, p<0.0001).

Conclusion: In this large retrospective analysis, we find that after adjusting for clinical, socioeconomic, and demographic variables, ESRD is an independent risk factor for increased mortality, intraoperative complications, mechanical wound complications, increased LOS and cost for patients undergoing LC. Prospective studies exploring risk optimization strategies for patients with ESRD undergoing LC are warranted.

10.09 Clinical & Patient Reported Outcomes in Nonoperative Ventral Hernia Management:A Prospective Cohort

Posted on January 18, 2016

J. L. Holihan1, Z. M. Alawadi1, J. R. Flores-Gonzelez1, T. C. Ko1, L. S. Kao1, M. K. Liang1 1University Of Texas Health Science Center At Houston,Surgery,Houston, TX, USA

Introduction: Little is known about clinical or patient-reported outcomes (PRO) with non-operative management of patients with a ventral hernia. The aim of this prospective study is to determine the outcomes of patients undergoing initial non-operative treatment of their ventral hernia.

Methods: A prospective, longitudinal study of patients undergoing non-operative management of their ventral hernia at a safety-net hospital was initiated in 2014. Non-operative management was elected either by the surgeon based upon poor predicted risk-benefit ratio or by the patient. Primary outcome was rate of surgical repair of the ventral hernias. Secondary outcomes included rate of emergency repair, elective repair, and emergency room visits. Validated surveys for patient satisfaction, cosmetic satisfaction, abdominal pain, and patient function (modified Activities Assessment Scale, AAS) were utilized. Baseline and 6-month follow up surveys were compared using a Wilcoxon signed rank test with p<0.05 considered significant.

Results: Of 114 patients who underwent non-operative management of their ventral hernia, 85 (74.6%) were followed for a median (interquartile range) of 7.8 (5.6-9.0) months. Reasons for non-operative management were obesity (81, 71.1%), smoking (22, 19.2%), other co-morbidities (23, 20.2%), patient choice (10, 8.8%), and surgical complexity (3, 2.6%). Six (7.1%) patients underwent ventral hernia repair following enrollment: 4 (4.7%) were repaired electively and 2 (2.4%) were repaired emergently. Among those undergoing elective repair, 3 patients met their goals required for elective surgery (weight loss and/or smoking cessation) and one underwent surgery at another institution. Ten (11.8%) patients had at least 1 emergency room visit associated with their hernia. Based on the modified AAS survey, patients who were managed non-operatively had increased feelings of sadness because of their hernia during the time they were followed (p=0.045) and worsening overall functional status (p=0.049). Patients who were converted to operative management had improved satisfaction with their abdominal wall (p=0.027), satisfaction with the appearance of their abdominal wall (p=0.027), and improvement in how much their abdominal wall affected them at work (p=0.034). There were no differences in abdominal wall pain (table).

Conclusion: Patients who undergo non-operative management of their ventral hernias experience significant declines in patient-reported outcomes. While the short-term risk of emergency surgery among patients with ventral hernias managed non-operatively is low, the long-term risk is unknown. Patients with ventral hernias should be optimized so that they can undergo elective repair.

10.08 Medical Malpractice in Hernia Surgery: Methods to Save Yourself a Trip to the Courts

Posted on January 18, 2016

A. J. Choudhry1, N. N. Haddad1, E. B. Habermann2, D. S. Morris1, S. F. Heller1, B. D. Kim1, D. H. Jenkins1, M. D. Zielinski1 1Mayo Clinic,General Surgery,Rochester, MN, USA 2Mayo Clinic,Robert D. And Patricia E. Kern Center For The Science Of Health Care Delivery, Surgical Outcomes Program,Rochester, MN, USA

Objective: Deemed a high risk specialty, general surgery litigation rates are rising with some of the highest malpractice premiums in the medical field. Hernia surgery is among the most common surgical procedures performed by general surgeons estimated to be 750,000 operations yearly. Our objective was to analyze malpractice litigation related to hernia surgery and discuss ways to prevent future litigation for physicians.

Methods: Using Westlaw Next (Thomson Reuters, New York, NY), a comprehensive law database with publically available records collected from many jurisdictions, we searched all jury verdicts and settlements for the terms: ‘medical malpractice’, ‘hernia’ and, ‘herniorrhaphy.’ Dates of cases ranged from 1985 to 2015. Exclusion criteria included cases where hernia surgery was not the primary reason for litigation. Information was collected from all cases on patient demographics, physician speciality, procedural characteristics, alleged reason for malpractice claim and outcome of the trial. Data is presented as means ± standard deviation with a p value of <0.05 is considered statistically significant.

Results: The search criteria yielded 568 case briefs; 260 cases met inclusion criteria and were included in the study. Sixty-five percent (170) of the patients involved in cases were male; mean patient age was 50 ± 15. The two most common hernia procedures which led to litigation were inguinal and ventral hernia repairs (>90%). The state of New York had the highest number of medical malpractice cases (47) followed closely by California (46). In 38 cases (15%) the patient claimed informed consent was breeched by the surgeon. The most common alleged reason for litigation was damage to surrounding structures. Overall 59% (152) of cases were decided in favor of the defendant (physician). Median jury verdict and settlement sum was $700,000.

Conclusion: Surgical complications and patient death played a significant role in outcomes of malpractice litigation. Although a majority of the cases were decided in the favor of the defendant, jury verdicts and settlements decided in favor of the plaintiff were costly. By understanding the reasons why surgeons go to trial; the risk of future lawsuits may lessen if proper measures are taken to prevent such outcomes.

10.07 Inadvertent Parathyroidectomy During Thyroidectomy: Incidence, Risk Factors and Long-Term Outcome

Posted on January 18, 2016

H. Y. Zhou1,2, J. C. He1,2, C. R. McHenry2,3 1University Hospitals Case Medical Center,Surgery,Cleveland, OH, USA 2Case Western Reserve University School Of Medicine,Cleveland, OH, USA 3MetroHealth Medical Center,Surgery,Cleveland, OH, USA

Introduction:
Preservation of the parathyroid glands either in situ or by autotransplantation is important in preventing permanent hypoparathyroidism. The parathyroid glands are 5mm or less in size, often intracapsular or intrathyroidal and obscured by central compartment lymph nodes, making preservation a challenge. The purpose of this study was to determine the incidence of inadvertent parathyroidectomy (IP) during thyroidectomy, its risk factors and whether it contributes to the development of hypoparathyroidism.

Methods:
A retrospective review of all thyroid operations performed by a single surgeon from January 1, 2010 through July 31, 2014 were reviewed to determine the rate of IP and permanent hypoparathyroidism. Patient medical records were assessed for demographics, extent of thyroidectomy, central compartment lymph node dissection, thyroid gland weight, retrosternal goiter, parathyroid autotransplantation, reoperation, pathology, postoperative calcium levels and number of parathyroid glands inadvertently removed.

Results:
Three hundred eighty-six patients underwent thyroidectomy, including 243 total and 17 completion thyroidectomies and 126 lobectomies. Mean age was 51.8 ± 14.1 years and 327 (85%) patients were female. There were 25 (6%) reoperations, 40 (10%) patients had central compartment lymph node dissection and 128 (33%) underwent parathyroid gland autotransplantation. IP occurred in 78 (20%) patients, 16 of whom had intrathyroidal glands. Permanent hypoparathyroidism occurred in 7 (2.7%) of 260 patients following total or completion thyroidectomy, 4 (5%) with IP compared to 3 (1%) without IP (p = 0.033). Logistic regression analysis revealed that female gender (odds ratio = 2.768, p = 0.040), central compartment lymph node dissection (odds ratio = 9.584, p = 0.001), and thyroid gland weight (odds ratio = 0.994, p = 0.022) were independent factors associated with IP.

Conclusion:
IP, which occurred in 20% of patients undergoing thyroidectomy, is a remediable factor that was associated with a higher rate of hypoparathyroidism. Central compartment lymph node dissection is an independent risk factor for IP.

10.06 Robotic vs. Lap Colectomy Results In Increased Operative Time Without Improved Perioperative Outcomes

Posted on January 18, 2016

B. Ezekian1, Z. Sun1, M. Adam1, J. Kim1, M. C. Turner1, B. F. Gilmore1, C. T. Ong1, C. R. Mantyh1, J. Migaly1 1Duke University Medical Center,Department Of Surgery,Durham, NC, USA

Introduction: Interest in robotic technologies is burgeoning within the field of colorectal surgery. However, benefits of utilizing a robotic approach compared to the standard laparoscopic approach have not been well defined, especially in resections not involving pelvic anastomoses.

Methods: Patients who underwent either a robotic or laparoscopic colectomy from the 2012-2013 National Surgical Quality Improvement Program (NSQIP) database were selected. Short-term perioperative outcomes were compared between 1:1 propensity-matched groups. A subset analysis was performed among patients who underwent segmental resections only.

Results: Among the 15,976 patients included, 498 (3.1%) colectomies were performed robotically. After matching for demographic, clinical, and treatment characteristics, there were no differences between robotic and laparoscopic resections with regard to wound complications or other complications such as urinary tract infection, cardiopulmonary or thromboembolic events, renal insufficiency, anastomotic leaks, transfusion requirements, unplanned readmissions, or 30-day mortality (all p >0.05). However, operative time was significantly higher for the robotic colectomies (196 vs. 166 minutes, p <0.001). Among segmental resections only, operative time was still significantly longer for robotic colectomies (190 vs. 153 minutes, p <0.001) without differences in postoperative outcomes (all p >0.05).

Conclusion: In this early population-based experience, robotic colectomies appear to result in similar perioperative outcomes when compared to laparoscopic colectomies, but are associated with longer operative times. Given the current focus on value-based healthcare, utilizing robotic technologies in straightforward resections may not justify the cost.

10.05 Physician Practice Patterns in Decision Making for Older Patients with Symptomatic Gallstone Disease

Posted on January 18, 2016

N. P. Tamirisa1, T. S. Riall1, F. M. Dimou1 1University Of Texas Medical Branch,Galveston, TX, USA

Introduction: Recent data demonstrate that elective cholecystectomy is performed in fewer than 25% of older patients with symptomatic gallstones; moreover, receipt of cholecystectomy is independent of patients’ risk of developing acute biliary complications. Our goal was to explore physicians’ clinical decision making in older patients.

Methods: Semi-structured interviews with 10 clinical vignettes were conducted with 9 primary care physicians (PCP), 7 gastroenterologists, and 6 surgeons. In the vignettes patients were >65 years; the severity of their gallbladder disease and associated comorbidities varied. Using a previously validated PREOP-Gallstones risk prediction model, predicted rates of gallstone-related hospitalization were calculated for each vignette. For each vignette, physicians were asked: 1) To estimate the patient’s 2-year risk of developing gallstone-related complications, and 2) If they would recommend or refer the patient for cholecystectomy. Model-predicted rates were compared to physician-predicted rates.

Results:The results from 5 vignettes are summarized in Table 1. Across specialties, physician-predicted risk of 2-year acute gallstone-related hospitalization was inconsistent with model-predicted risk. Physician-predicted risk generally increased with increased gallstone disease severity but both the direction and magnitude of the inconsistency varied with the clinical scenario and physician specialty. On average, surgeons were more likely to recommend cholecystectomy. In many cases surgeons behaved paradoxically, recommending cholecystectomy less often for patients at higher risk; they recommended cholecystectomy in 83% of patients with biliary colic (predicted a 21% 2-year risk of complications) and only 67% of patients with gallstone pancreatitis and comorbidities (predicted a 61% risk). Among PCPs and gastroenterologists, referral for cholecystectomy generally increased with perceived risk, but even when they predicted a high risk they did not uniformly refer for surgical evaluation. In most cases where surgeon recommended cholecystectomy rates were high, PCPs and gastroenterologists did not recommend referral.

Conclusion:Across specialties, physicians were inaccurate at predicting risk as compared to a recently validated risk-prediction model. Patients perceived to be at the highest risk are not routinely referred for surgical evaluation. Moreover, there appears to be a disconnect, with non-surgeons frequently not referring patients in whom surgeons would recommend cholecystectomy. Physician education and incorporation of risk prediction models into clinical practice can better align treatment with risk and improve outcomes in older patients with symptomatic gallstones.

10.04 Cholecystectomy Outcomes In Obese Patients: An ACS NSQIP Analysis

Posted on January 18, 2016

C. J. Neylan1, D. T. Dempsey1, R. R. Kelz1, G. C. Karakousis1, K. Lee1, N. N. Williams1, A. Furukawa1, M. G. Peters1, K. R. Dumon1 1University Of Pennsylvania,Department Of Surgery,Philadelphia, PA, USA

Introduction: Obesity is a risk factor for cholelithiasis and subsequent cholecystitis. Cholecystectomy remains the mainstay of treatment for acute cholecystitis. Little is known about the risk of obesity on outcomes following cholecystectomy for acute cholecystitis. The purpose of this study was to examine the impact of obesity on the outcomes of cholecystectomy for acute cholecystitis.

Methods: Patients who underwent a cholecystectomy (laparoscopic, open, or converted) for acute cholecystitis from 2007-2013 were identified from the American College of Surgeons NSQIP database. Patients were classified into normal (BMI < 30) and obese (BMI ≥ 30) groups. Obese patients were subcategorized into the following groups: 30 ≤ BMI < 40, 40 ≤ BMI < 50, and 50 ≤ BMI. The outcomes of interest were operative time, mortality, morbidity, and post-operative length of stay. Independent multivariable logistic and linear regression models were used to examine the association between obesity and each of the outcomes of interest. A procedure was considered converted if there was a primary CPT code for a laparoscopic procedure and a secondary CPT code for an open procedure.

Results: Of 22, 808 patients included in the study, 46% were obese. After adjustment for potential confounders, obesity (BMI ≥ 30) was found to be significantly associated with prolonged operative time (defined as a procedure-specific operative time in the 90th percentile or above) (OR = 1.43, p < 0.0001) relative to normal BMI. Obesity was not significantly associated with overall morbidity, mortality, or post-operative length of stay. Subgroup analysis revealed that each obese subgroup was significantly associated with a prolonged operative time, relative to the normal BMI group. The subgroup with a BMI between 40 and 49.9 had a significantly higher mortality rate relative to the normal BMI group (OR 3.16, p = 0.003), and the subgroup with a BMI between 30 and 39.9 had significantly more thromboembolic complications (pulmonary embolism or DVT) relative to the normal BMI group (OR 1.86, p = 0.02). Among open cholecystectomies, obesity was significantly associated with prolonged operative time and increased mortality compared to a normal BMI. Among laparoscopic cholecystectomies, obesity was significantly associated with prolonged operative time but not increased mortality, relative to normal BMI. The percent of patients who received laparoscopic vs. open cholecystectomies did not vary with BMI.

Conclusion: The data suggest that cholecystectomy for acute cholecystitis is safe even in high BMI patients, and that the operation can usually be completed laparoscopically. However, certain groups of obese patients are at increased risk for postoperative morbidity or mortality.

10.03 Early Versus Late Readmission After Pancreatectomy

Posted on January 18, 2016

I. Folkert1, S. Damrauer1, G. Karakousis1, M. K. Lee, IV1 1Hospital Of The University Of Pennsylvania,Department Of Surgery,Philadelphia, PA, USA

Introduction: Readmission rates have been highlighted as a target for cost-control in healthcare. Pancreatectomy is associated with high readmission rates, and multiple studies have identified factors associated with 30-day readmission. This study sought to delineate differences in patients readmitted early after discharge versus those readmitted remotely.

Methods: Patients who underwent open pancreatectomy were identified in the 2008-2010 California State Inpatient Database. Patients were stratified according to whether they were first readmitted at days 0-30 post-discharge (early), 31-90 (intermediate), 91-365 (late), or had no readmissions within the first year. Descriptive statistics were used to characterize each readmission group. Multinominal logistic regression was used to identify risk factors for readmission by timeframe.

Results: 3,118 patients underwent pancreatectomy. 1,507 (48%) patients were readmitted in the year following discharge (21.6% early, 8.9% intermediate, 17.9% late). The median time to first readmission was 43 days (IQR 9 – 165). The median number of readmissions within the first year for readmitted patients was 2 (IQR 1-3).

Demographically, age (p < 0.0001), insurance status (p = 0.0001), diagnosis (p < 0.0001), and malignancy (p < 0.0001) were the most significant predictors of readmission. A similar percentage of patients with malignant and benign disease were first readmitted within 30 days (21.5% and 22.0%, respectively). However, 20.7% of patients with malignancy versus only 8.8% of patients with benign disease were first readmitted at day 91-365. Comorbidities that had the most significant associations with readmissions were hypertension, CHF, metastatic disease (including lymph nodes), renal failure, and weight loss (p < 0.0001 for all). The complication that was most strongly associated with readmission was postoperative shock (OR 26.7 [95% CI 5.2 – 489] early, OR 11.8 [95% CI 1.1 – 253] intermediate, OR 11.6 [95% CI 1.7 – 228] late). Factors not significantly associated with readmission included obesity (p = 0.496), race (p = 0.055), income (p = 0.549), wound complications (p = 0.244), and GI complications (p = 0.524).

Complications related to surgical/medical care were the primary reason for 39.9% of early readmissions but only 15.9% of intermediate and 6.1% of late readmissions. However, the primary procedure performed on readmission was similar in all groups. Either an abdominal drainage procedure or an upper endoscopy was the primary procedure in 22.3% of early readmissions, 17.9% of intermediate readmissions, and 14.2% of late readmissions.

Conclusions: Reducing readmissions after pancreatectomy remains challenging. Time to first readmission after pancreatectomy can vary widely, and 30-day readmission rates can therefore underestimate the extent of the problem. Reasons for readmission vary with time, but invasive interventions are often necessary at any time point.

10.02 Complicated Diverticulitis: Management and Trends in an Aging Population

Posted on January 18, 2016

T. Galbreath1, B. Palachick1, T. Bell1, R. Grim1, V. Ahuja1 1York Hospital,Surgery,York, PA, USA

Introduction: Though classically an operative disease, complicated diverticulitis has started to trend towards nonoperative management, which is primarily achieved with percutaneous drainage. The aim of our study was to evaluate the trends and discover its effects on management and whether an aging population is spared from operations.

Methods: 158,251 cases from HCUP-NIS discharge data (2007-2011) were collected with primary admission of complicated (abscess, ICD-9 569.5 or perforated 569.83) diverticulitis (ICD-9 562.11 or 562.13). Patients were assessed for percutaneous drainage (ICD9 54.91). Intervention groups were non-operative, non-operative with percutaneous drainage or operative (all other surgeries). Age of patients was evaluated by two groups (younger 18-59 and older 60+). Exclusion criteria were: < 18 years of age, malignancy, and inflammatory bowel disease. Descriptive, x2, and test of proportion statistics were used to evaluate incidence rate of admission, and demographic changes.

Results: Complicated diverticulitis hospitalizations have increased from 19 to 21.5% from 2007 to 2011 (p<.001). The majority of patients were younger (55.5%), however the highest rate increase was in older patients with complicated diverticulitis (2.8%, p<.001). The abscess rate increased from 91.8% to 98.6%, while at the same time perforations decreased from 10.6 to 1.9 (all p<.001). Patients were managed non-operatively (42.3%), non-operatively without percutaneous drainage (13.5%) and with operative intervention (44.2%). However, there was an increase in non-operative intervention (4.1%) and non-operative percutaneous drainage (3.9%) and a decrease in operative treatment (-8.0%) over the study time period (Figure 1, all p<.001). By age, in older patients there was an increasing trend in operative treatment (4.2%, p = 0.825). While the younger group saw a significant increase in non-operative percutaneous drainage (0.4%, p<.001) over the study time period.

Conclusions: Our findings do support the rising trend in overall hospitalizations with nonoperative management and use of percutaneous drainage for abscess. However the elderly are presenting with an increasing rate of complicated diverticulitis requiring operative intervention. It is possible the nonoperative management in the younger population is increasing the need for future hospitalizations and operative management as they age. Further studies could aim at determining long-term outcomes of elective colectomies after effective nonoperative management in the younger population and if this reduces emergent operations in the elderly.

10.01 Lymph Node Retrieval is Decreased after Neoadjuvant Chemoradiation Therapy for Esophageal Cancer

Posted on January 18, 2016

D. Giugliano1, A. Berger1, M. J. Pucci1, E. Rosato1, C. Lamb1, H. Meidl1, F. Palazzo1 1Thomas Jefferson University,Philadelphia, PA, USA

Introduction: Induction chemoradiation therapy for the treatment of locally advanced esophageal cancer has been used with increasingly successful outcomes. Lymph node retrieval after induction therapy is known to be decreased for rectal cancer and breast cancer. The aim of this study was to compare the differences in total lymph nodes (TLN), positive lymph nodes (PLN), and the ratio between positive lymph nodes and total lymph nodes (LNR) in patients who underwent esophagectomy with or without induction therapy.

Methods: We queried our IRB-approved prospective esophageal surgery database to identify a total of 175 patients who underwent minimally-invasive (MIE) or open esophagectomy (OE) from 2008 to 2015 for esophageal cancer. TLN, PLN and LNR were reviewed and compared using Student’s t-test. Significance was established at p<0.05.

Results:There were 70 patients (40.0%) who underwent esophagectomy without induction therapy and 105 patients (60%) with induction chemoradiation therapy. The majority of patients (n=157, 89.7%) underwent MIE while 18 (10.3%) had OE. Most patients (n=155, 88.5%) were diagnosed with adenocarcinoma. Complete pathologic response was seen in 28.6% of patients (n=30) who underwent induction therapy. Median TLN retrieval was significantly less (p=0.027) in those patients who underwent induction therapy (range: 0-53, median=18 ) compared to no induction (range: 9-57, median=21). Average PLN was less in induction patients (range: 0-13, average=0.74) compared to those without (range: 0-12, average=0.96; p=0.262), though this was not statistically significant. There was no difference in average LNR between the two groups (0.05 vs. 0.04, p=0.31).

Conclusion: Lymph node retrieval was significantly less in patients who underwent induction chemoradiation therapy for esophageal cancer. Though not statistically significant, there was a trend towards a lower number of positive lymph nodes in patients who underwent induction chemoradiation therapy, while the lymph node ratio did not differ. Further research is needed to determine whether this finding affects overall survival.