THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
54th Annual Meeting
May 17-21, 2013 Orange County Convention Center Orlando, Florida
ABSTRACT SUPPLEMENT Table of Contents
Schedule-at-a-Glance ...... 2
Sunday Plenary, Video, and Quick Shot Session Abstracts ...... 6
Monday Plenary, Video, and Quick Shot Session Abstracts ...... 22
Tuesday Plenary Session Abstracts ...... 50
Sunday Poster Session Abstracts ...... 59
Monday Poster Session Abstracts ...... 112
Tuesday Poster Session Abstracts ...... 166 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
PROGRAM BOOK ABSTRACT SUPPLEMENT
FIFTY-FOURTH ANNUAL MEETING Orange County Convention Center Orlando, Florida May 17–21, 2013 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Schedule-at-a-Glance
FRI, MAY 17, 2013 SATURDAY, MAY 18, 2013
300 208ABC Other 6:30 AM 6:45 AM 7:00 AM 7:15 AM 7:30 AM 7:45 AM 8:00 AM 8:15 AM 8:30 AM 8:45 AM 9:00 AM 9:15 AM 9:30 AM 9:45 AM 10:00 AM 10:15 AM 10:30 AM 10:45 AM NAFLD DDW CCS:
11:00 AM Therapeutic 11:15 AM Approaches in 11:30 AM
11:45 AM (by invitation only) 12:00 PM 12:15 PM 12:30 PM 12:45 PM 1:00 PM 1:15 PM 1:30 PM RESIDENTS & FELLOWS RESEARCH CONFERENCE 1:45 PM 2:00 PM 2:15 PM MAINTENANCE OF CERTIFICATION COURSE:
2:30 PM Evidence Based Treatment of Colorectal Diseases 2:45 PM of IBD DDW CCS: 3:00 PM Endoscopic
3:15 PM Evaluation & Mgt 3:30 PM 3:45 PM 4:00 PM 4:15 PM 4:30 PM 4:45 PM Barrett's Esophagus 5:00 PM DDW CCS:
5:15 PM Controversies in 5:30 PM 5:45 PM
2 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Schedule-at-a-Glance
SUNDAY, MAY 19, 2013 Exhibit 300 308D 303ABC 304AB Other Hall 6:30 AM 6:45 AM 7:00 AM 7:15 AM 7:30 AM OPENING SESSION 7:45 AM 8:00 AM PRESIDENTIAL 8:15 AM PLENARY A 8:30 AM (PLENARY SESSION I)
8:45 AM DDW
9:00 AM COMBINED RESEARCH PRESIDENTIAL IBD FORUM: 9:15 AM ADDRESS 9:30 AM 9:45 AM 10:00 AM 10:15 AM PRESIDENTIAL 10:30 AM PLENARY B 10:45 AM (PLENARY SESSION II) Syndrome 11:00 AM DDW CCS: Tx of Metabolic
11:15 AM GUEST ORATOR Bariatric Surgery as 11:30 AM 11:45 AM 12:00 PM 12:15 PM 12:30 PM 12:45 PM 1:00 PM 1:15 PM MEET-THE- LUNCHEON
1:30 PM PROFESSOR 1:45 PM 2:00 PM 2:15 PM 2:30 PM
2:45 PM Mets Lesions
3:00 PM GI SURGERY of Benign Liver DEBATES A: HIPEC; HIPEC; A: DEBATES Synch Colorectal Liver CONTROVERSIES IN CONTROVERSIES 3:15 PM DDW CCS: Mgt Cancer
3:30 PM POSTER SESSION I (authors available @ posters 12:00 PM - 2:00 PM) 3:45 PM VIDEO SESSION I
4:00 PM STATE-OF-THE-ART CONFERENCE: Evolving PLENARY SESSION III 4:15 PM Management in Pancreatic 4:30 PM 4:45 PM 5:00 PM 5:15 PM 5:30 PM 5:45 PM
3 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Schedule-at-a-Glance
MONDAY, MAY 20, 2013 Exhibit 300 308D 303ABC 304AB Other Hall 6:30 AM 6:45 AM 7:00 AM
7:15 AM EXPERT WITH THE
7:30 AM BREAKFAST 7:45 AM 8:00 AM 8:15 AM
8:30 AM WARD Rectal CA CLINICAL Trans-Anal Surgery for ROUNDS I: 8:45 AM THE MOVIES BREAKFAST AT PANEL PUBLIC VIDEO SESSION II:
9:00 AM POLICY & 9:15 AM ADVOCACY 9:30 AM 9:45 AM 10:00 AM CA Parenchymal Resection for 10:15 AM SYMPOSIUM: III Metastatic Colorectal SSAT/AHPBA JOINT JOINT SSAT/AHPBA QUICK QUICK Preservation in Hepatic 10:30 AM SHOTS VIDEO PLENARY SESSION I SESSION 10:45 AM SESSION IV Post-Op IBD Patient 11:00 AM STATE-OF- DDW CCS:
11:15 AM THE-ART in Complications 11:30 AM LECTURE 11:45 AM 12:00 PM 12:15 PM 12:30 PM 12:45 PM 1:00 PM 1:15 PM 1: Writing WRITERS MEET-THE- LUNCHEON WORKSHOP 1:30 PM PROFESSOR 1:45 PM 2:00 PM 2:15 PM 2:30 PM II QUICK QUICK SHOTS
2:45 PM SESSION Cirrhosis DEBATES B: B: DEBATES 3:00 PM GI SURGERY of Sx Risk in of Sx Risk Patients with Adenocarcinoma; PLENARY DDW CCS: Mgt Resectable Pancreas SESSION V Nissen Fundoplication 3:15 PM IN CONTROVERSIES
3:30 PM POSTER SESSION II (authors available @ posters 12:00 PM - 2:00 PM) 3:45 PM Mgt of JOINT JOINT Disease
4:00 PM Diverticular SSAT/ASCRS 4:15 PM SYMPOSIUM: 4:30 PM III WARD WARD QUICK SHOTS Pancreas CLINICAL SESSION Cysts of the the of Cysts 4:45 PM ROUNDS II:
5:00 PM DDW CCS: Myotomy or POEM Botox, Balloon, Lap 5:15 PM Achalasia Treatment: 5:30 PM ANNUAL MEETING
5:45 PM BUSINESS
4 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Schedule-at-a-Glance
TUESDAY, MAY 21, 2013 Exhibit 308D 303ABC 304AB Other Hall 6:30 AM 6:45 AM 7:00 AM
7:15 AM EXPERT WITH THE
7:30 AM BREAKFAST 7:45 AM 8:00 AM 8:15 AM 8:30 AM B'FAST GI Leaks
8:45 AM Early Cancer PLENARY 9:00 AM Pancreatic DDW CCS: SESSION VI SSAT/ISDS JOINT
9:15 AM SYMPOSIUM: Mgt of 9:30 AM 9:45 AM 10:00 AM 10:15 AM 10:30 AM 10:45 AM Necrosis
11:00 AM Pancreatic DDW CCS: 11:15 AM 11:30 AM PANEL & QUALITY OUTCOMES OUTCOMES PLENARY SESSION VII
11:45 AM HEALTH CARE 12:00 PM 12:15 PM 12:30 PM 12:45 PM 1:00 PM 1:15 PM WRITERS MEET-THE- LUNCHEON 2: Reviewing WORKSHOP 1:30 PM PROFESSOR 1:45 PM 2:00 PM 2:15 PM SSAT/SAGES JOINT GI Surgery Complications GI 2:30 PM LUNCHEON SYMPOSIUM:
2:45 PM Upper from Complex Mgt & Rescue Tumors 3:00 PM Junction
3:15 PM DDW CCS: GE
3:30 PM BEST OF DDW POSTER SESSION III (authors available @ posters 12:00 PM - 2:00 PM) 3:45 PM . 4:00 PM 4:15 PM 4:30 PM 4:45 PM 5:00 PM 5:15 PM 5:30 PM 5:45 PM
5 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
SSAT PLENARY, VIDEO, AND QUICK SHOT ABSTRACTS Printed as submitted by the authors. indicates a paper that is also being presented at the Residents & Fellows Research Conference. Participation in and attendance at this conference is by invitation only. 6indicates a video presentation scheduled during a Plenary Session. Sunday, May 19, 2013 8:00 AM – 9:00 AM 303ABC PRESIDENTIAL PLENARY A (PLENARY SESSION I)
199 benign lesions. E3 and E4 were expressed at extremely low levels in all patients. Compared to IPMN alone, E6 levels Clinical Signifi cance of Serum COL6A3 Isoforms in were signifi cantly higher in PDA (p = 0.0036). There were Pancreatic Ductal Adenocarcinoma no signifi cant differences between E6 levels in IPMN and Christopher Y. Kang, Dierdre Axell-House, Pranay Soni, Normal sera (p = 0.59). Using a logistic regression model, Galina Chipitsyna, Konrad Sarosiek, Mazhar Al-Zoubi, Hwyda A. we found that for each increasing unit of log E6 COL6A3, Arafat, Charles J. Yeo patients are 9.5 times more likely to harbor a cancer rather Surgery, Thomas Jefferson University, Philadelphia, PA than a benign lesion, 95% CI (2.4, 38.1), p = 0.002. The area under the ROC curve, AUC, was 0.72. Knocking down E3 INTRODUCTION: Type VI collagen (COL6) forms a or E4 or E6 with isoform-specifi c siRNA resulted in reduced microfi brillar network associated with type I collagen fi brils PDA cell migration and invasion and concomitant reduc- and constitutes a major component of the prominent des- tion of the expression of several infl ammation and angio- moplastic reaction in pancreatic ductal adenocarcinoma genesis-related genes, such as MMP-9, OPN, MCP-1 and (PDA). We have demonstrated recently that a subunit of VEGF. Interestingly, knocking down any of the 3 isoforms COL6, COL6A3, is expressed in high levels in PDA tissue. resulted in increased expression of TNF-alpha. We also showed that COL6A3 gene undergoes tumor-spe- cifi c alternative splicing to produce 3 isoforms E3, E4 and CONCLUSIONS: Our data show for the fi rst time the E6 that are tumor tissue-specifi c. The aim of this study is to potential clinical signifi cance of circulating E6 COL6A3 lev- investigate the diagnostic value and clinical signifi cance of els in the diagnosis of pancreatic malignancy. Our in vitro circulating COL6A3 isoforms mRNA in PDA. data suggests a role for COL6A3 isoforms in PDA progres- sion and metastatic potential. METHODS: Serum samples were obtained from patients that underwent pancreatic resection at a single institution between 2006 and 2009. COL6A3 levels in the sera from 200 patients with pathologically confi rmed PDA (n = = 40), Treatment Sequencing for Resectable Pancreatic intraductal papillary mucinous neoplasms (IPMN) (n = 20), Cancer: Infl uence of Early Metastases and Surgical and chronic pancreatitis (n = 10) were analyzed by real time Complications on Multimodality Therapy Completion PCR using isoform-specifi c primers for E3, E4 and E6. In Rates and Survival addition, sera from age-matched healthy volunteers were 1 2 1 analyzed (n = 30). The prediction levels for malignancy Ching-Wei D. Tzeng , Daniel E. Abbott , Jeffrey E. Lee , 1 1 1 were determined by the area under the receiver operating Peter W. Pisters , Jason B. Fleming , Jean-Nicolas Vauthey , characteristic curve (AUC). In vitro, wound healing, cell Matthew Katz1 proliferation and soft-agar colony formation assays evalu- 1. Department of Surgical Oncology, The University of Texas MD ated the functional impact of each isoform in PDA cells Anderson Cancer Center, Houston, TX; 2. Surgical Oncology, (MIAPACA-2 and ASP-C-1) transfected with isoform-specifi c University of Cincinnati, Cincinnati, OH siRNA. A panel of infl ammation- and invasion/angiogene- INTRODUCTION: Multimodality therapy (MMT) is impor- sis-related genes was also evaluated. tant to the long-term survival of patients with resectable RESULTS: Circulating E6 mRNA levels were signifi cantly pancreatic adenocarcinoma (PDAC), but its completion can (p = 0.006) elevated in PDA patients when compared to all be hindered by early cancer progression or by treatment
6 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
complications. We sought to compare the infl uence of each CONCLUSIONS: Completion of multimodality therapy of these factors on the MMT completion rates of operable is strongly associated with improved survival of operable patients with resectable PDAC treated with either a neoad- patients with resectable PDAC. Even in the highly selected Abstracts Sunday juvant (NT) or surgery-fi rst (SF) sequencing strategy. cohort evaluated in this study, early cancer progression METHODS: We retrospectively evaluated all patients and PMC negatively impacted MMT completion rates and with PDAC at our institution from 2002–2007, who had OS, particularly among SF patients. Thus, NT sequencing 1) a radiographically resectable pancreatic head tumor, 2) remains a valuable alternative to SF sequencing for tumor a performance status (PS) and comorbidities suitable for biology evaluation and patient selection. immediate surgery, and 3) a carbohydrate antigen (CA) 19-9 <1000 U/ml. MMT was defi ned as resection before or 201 after completion of planned pre- or post-operative therapy. Early Referral for 24-Hour Esophageal pH Monitoring Postoperative major complications (PMC) were defi ned as Is More Cost-Effective Than Prolonged Use of Clavien Grade ≥3. Disease progression was considered early when it developed within 3 months in SF patients or prior Proton Pump Inhibitors in Patients with Suspected to planned resection in NT patients. Reasons for and rates Gastroesophageal Refl ux Disease of failure to complete MMT, 90-day PMC, and overall sur- David Kleiman1, Toni Beninato1, Brian P. Bosworth2, vival (OS) were compared between the two cohorts. Laurent Brunaud4, Thomas Ciecierega3, Carl V. Crawford2, 2 1 1 RESULTS: 112 NT and 58 SF patients met inclusion cri- Brian G. Turner , Thomas J. Fahey , Rasa Zarnegar teria. 92/112 (82%) NT and 33/56 (59%) SF patients with 1. Surgery, New York Presbyterian Hospital – Weill Cornell Medical complete follow-up completed MMT (p < 0.001). NT College, New York, NY; 2. Medicine, New York Presbyterian patients did not complete MMT due to early progression (n Hospital – Weill Cornell Medical College, New York, NY; 3. = 13, including 8 nontherapeutic laparotomies) and PS (n = Pediatrics, New York Presbyterian Hospital – Weill Cornell Medical 7). SF patients did not complete MMT due to early progres- College, New York, NY; 4. Surgery, University Hospital Nancy, sion (n = 10), PMC (n = 6), and PS (n = 3); 4 SF patients also Brabois, France underwent nontherapeutic laparotomy due to metastases. BACKGROUND: Gastroesophageal refl ux disease (GERD) affects nearly 25% of adults, but the diagnostic algorithm remains controversial. Most guidelines recommend an empiric 8-week trial of proton-pump inhibitors (PPIs), but many patients remain on PPIs for much longer periods. Twenty-four hour esophageal pH monitoring can help rule out GERD and avoid the unnecessary cost and risks of pro- longed PPI use. We hypothesized that performing pH moni- toring promptly after an 8-week PPI trial would be a more cost-effective strategy than prolonged courses of PPIs. METHODS: A cost model was created from a third-party payer perspective over a 10-year period. Average wholesale unit prices of generic and name-brand PPIs were obtained from the 2012 Micromedex Redbook®. Low-dose (20 mg daily) generic PPI and high-dose (40 mg twice daily) name- brand PPI costs were used as the low-end and high-end PPI costs, respectively. The cost of 24-hour pH monitoring and manometry was obtained from 2012 Medicare fees. Manometry was necessary to determine esophageal func- Among all patients, those who completed MMT lived lon- tion and for appropriate positioning of the pH probe. A ger than those who did not (36 vs. 11 mo, p < 0.001). The cohort of 100 patients who underwent pH monitoring at median OS durations of all NT and SF patients (NT 28 vs. an academic institution was retrospectively reviewed for SF 21 mo, p = 0.082), the subset in each cohort who com- type of GERD symptoms and duration of PPI use prior to pleted MMT (NT 36 vs. SF 36 mo, p = 0.565), and the subset pH monitoring. The sensitivity of 24-hour pH monitoring in each cohort who did not complete MMT (NT 11 vs SF 13 by literature review ranged from 30% to 96%. The cost of mo, p = 0.325) were not statistically different. unnecessary PPIs was subtracted from the cost of pH moni- toring for all patients to evaluate the cost-effectiveness. The rate of PMC did not differ between NT and SF groups (19% vs. 17%, p = 0.782). SF patients with no PMC had a RESULTS: The weekly cost of PPIs ranged from $29.06 71% (31/44) MMT completion rate vs. 25% (2/8) after PMC to $107.70, and the cost of 24-hour pH monitoring was (p = 0.014). When resected NT patients suffered PMC, there $690. The cost of PPI therapy reaches equivalence with pH was no signifi cant decrease in OS (36 vs. 30 mo, p = 0.934), monitoring after 6.4 to 23.7 weeks, depending on the PPI in contrast to the negative effect of PMC in SF patients (26 regimen. Patients who experienced esophageal and extra- vs. 10 mo, p < 0.001). esophageal GERD symptoms reported a median of 208 and
7 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
52 weeks of PPI use, respectively, prior to referral. The cohort was prescribed a total of 21,411 weeks of PPIs beyond the 202 initial 8-week trial, 32% of which were for patients who had Risk Factors Associated with 30-Day Readmissions in a negative 24-hour pH monitoring study and were therefore Major Gastrointestinal Resections unnecessary. If the sensitivity of pH monitoring was 100%, performing pH monitoring on all patients after an 8-week Kristin N. Kelly, James C. Iannuzzi, Aaron S. Rickles, PPI trial would have saved between $1,966 and $7,285 per Veerabhadram Garimella, John R. Monson, Fergal Fleming patient over 10 years. This strategy remains cost-effective Surgical Health Outcomes & Research Enterprise, Division of as long as the sensitivity of pH monitoring is above 35% Colorectal Surgery, Department of Surgery, University of Rochester (Figure 1). In this model, since patients with extra-esopha- Medical Center, Rochester, NY geal GERD symptoms were referred a median of 156 weeks PURPOSE: Preventable readmissions represent a major sooner than patients with esophageal symptoms, the cost burden on the health care system and by risk stratifying savings were less (Table 1). patients resources can be directed to prevent these costly complications. This study examines patient characteristics, surgical factors, and postoperative complications associated with 30-day readmissions in gastrointestinal (GI) resections. METHODS: Inpatients undergoing major GI surgery were selected from the 2011 ACS National Surgical Qual- ity Improvement Program prospectively collected database. Procedures were classifi ed into esophageal, gastric, small bowel, large bowel, liver, and pancreatic resections using Common Procedural Terminology codes. Postoperative complications were divided into pre- and post-discharge groups by comparing time to complication and discharge. Operative times were grouped by 75th percentile (≥4 hours). Univariate analysis using Chi-square, Mann Whitney-U, and Student’s T-test were used to compare patient comor- bidities, surgical characteristics, and postoperative compli- Table 1: Estimated Range of Cost Savings Over 10-Years cations with 30-day unplanned readmission rates. Factors (Per Patient) of Performing Early 24-Hour pH Monitoring with a p < 0.1 were included in multivariate logistic regres- Across the Range of Reported Sensitivity for Diagnosing GERD sion. Odds ratios (OR) and 95% confi dence intervals (CI) are reported and p-value < 0.05 was considered statistically (1) (1) 30% Sensitivity 96% Sensitivity signifi cant. (2) All patients –$100.31 to $1,495.45 $1,196.99 to $6,303.43 RESULTS: For 43,894 patients undergoing GI resection, the Esophageal –$59.57(2) to $1,646.44 $1,327.37 to $6,786.61 overall 30-day unplanned readmission rate was 12.0% rang- symptoms ing from 11.4% for colorectal resections to 15.7% for pan- Extra-esophageal –$299.23(2) to $758.25 $560.47 to $3,940.09 creatic resections. Median postoperative length of stay was symptoms longer in the readmission group (7 vs. 6 days p < 0.0001). (1)Range of savings from low-dose generic PPI to high-dose name-brand PPI, Major predictors of 30-day readmissions included pre-dis- (2)Negative values refl ect additional cost over 10-years. charge major complications (OR = 1.28; CI: 1.14,1.44, p < 0.0001), preoperative steroid use (OR = 1.62; CI: 1.39,1.89, CONCLUSIONS: Most patients are maintained on PPIs for p < 0.0001), operative time ≥4 hours (OR = 1.61; CI: periods greatly surpassing the cost-equivalence point with 1.45,1.78, p < 0.0001) and discharge to a facility other than 24-hour esophageal pH monitoring. Early referral for pH home (OR = 1.48; CI: 1.28,1.70, p < 0.0001). Other factors monitoring after a brief empiric PPI trial may result in sub- associated with increased readmission included dependent stantial cost savings for patients with both esophageal and functional status, open surgery, pulmonary comorbidity, extra-esophageal GERD symptoms. neurologic comorbidity, higher ASA score, diabetes, and preoperative anemia (table 1). Post-discharge major and minor complications were highly correlated with 30-day readmission rates (OR = 59.3; CI: 52.2,67.3, p < 0.0001 and OR = 6.3; 95% CI: 5.8,6.9, p < 0.0001) and not included in the fi nal model.
8 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Table 1: Factors Associated with Unplanned 30-Day Readmissions Following GI Resection
Risk Factor % Readmitted Adjusted OR 95% CI p-Value Abstracts Preoperative Steroid use 18.6 v 11.5 1.62 1.39, 1.89 <0.0001 Sunday Operative time ( 4 hrs vs. <4hrs) 15.4 v 12.8 1.61 1.45, 1.78 <0.0001 Discharge Destination (Facility vs. Home) 15.1 v 11.6 1.48 1.28, 1.70 <0.0001 Open Surgery 13.1 v 9.1 1.46 1.30, 1.63 <0.0001 Pre-discharge Major Complication 14.9 v 11.1 1.28 1.14, 1.44 <0.0001 Neurologic Comorbidity 15.2 v 11.9 1.26 1.06, 1.51 0.01 Pulmonary Comorbidity 14.2 v 11.5 1.22 1.03, 1.45 0.02 Dependent Functional Status 15.8 v 11.8 1.24 1.01, 1.51 0.037 ASA (3/4 vs 1/2) 13.5 v 9.9 1.17 1.06, 1.30 0.002 Pre-discharge Minor Complication 12.8 v 11.9 1.16 0.98, 1.38 0.077 Diabetes 13.8 v 11.6 1.13 1.01, 1.27 0.032 Preoperative Anemia (HCT < 36) 13.5 v 11.3 1.11 1.01, 1.22 0.024
CONCLUSIONS: Unplanned 30-day readmissions repre- could inform resource utilization and post-operative care sent a major medical and fi nancial concern, but some may to help prevent readmissions in select high-risk GI surgical be foreseeable and thus preventable. Although previous patients. studies have identifi ed major complications as a strong risk Logistic regression model also controlled for age, hepatic factor for readmissions, this might represent an overesti- insuffi ciency, cardiac comorbidity, renal insuffi ciency, mate of the risk due to confounding by including post-dis- wound class, smoking, bleeding disorder, chemotherapy/ charge complications that may in fact cause, not predict, radiation, weight loss, preoperative sepsis, and LOS. readmissions. This model provides insight into factors that
8:00 AM – 9:30 AM 203AB DDW COMBINED RESEARCH FORUM (AGA-ACCREDITED) IBD
203 205 Mesenchymal Stem Cell Transplantation Improves Intestinal Barrier Dysfunction Measured with Confocal Chronic Colitis-Associated Cholangitis Through Endomicroscopy in Macroscopically Normal Mucosa Inhibiting the Activity of LPS/TLR4 Can Predict Requirement for Treatment Escalation X. Zhang1, G. Niu1, L. Liu1, H. Li1, J. Guo1, J. Song1, Y. Liu1, K. Liu1, J. Mill1, B. Wong2, C.P. Selinger1, V.C. Kariyawasam1, S. Chen2 N. Merrett1,3, R.W. Leong1, 2 1. Department of Gastroentology, The Second Hospital of Hebei 1. Gastroenterology and Liver Services, Bankstown and Concord Medical University, Shijiazhuang City, China; 2. Division of Hospitals, Concord, NSW, Australia; 2. Faculty of Medicine, The Pediatric Infectious Diseases and Immunology, Burns and Allen University of New South Wales, Sydney, NSW, Australia; Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA 3. Gastroenterology and Liver Services, University of Western Sydney, Sydney, NSW, Australia 204 IL-10-Producing Mucosal B Cells Attenuate T Cell- Mediated Colitis Through Induction of Tr-1 Cells Y. Mishima1, B. Liu1, C. Karp2, R.B. Sartor1 1. CGIBD, University of North Carolina, Chapel Hill, NC; 2. Cincinnati Children’s Hospital Medical Center, Cincinnati, OH
9 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
206 Link Between VEGF Expression, Angiogenesis and Infl ammation in Pediatric Crohn’s Disease Jennifer L. Knod1, Kelly M. Crawford1, Mary R. Dusing1, Artur Chernoguz1, Margaret H. Collins2, Jason Frischer1 1. Division of Pediatric General and Thoracic Surgery, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH; 2. Division of Pathology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH PURPOSE: Early-onset Crohn’s disease (CD) accounts for 25% of cases but is distinct from adult-onset CD by a more severe disease activity index, increased immunosuppressant requirement, and more extensive intestinal involvement. The pathogenic link between chronic infl ammatory dis- eases and angiogenesis prompted investigations into its role in infl ammatory bowel disease. We hypothesize that VEGF driven angiogenesis plays a signifi cant role in Crohn’s dis- ease infl ammation. METHODS: Pediatric patients (n = 13), ages 12 to 16, at our institution having undergone resection involving the terminal ileum for CD were compared to controls (n = 5) with non-infl ammatory indications for resection. Addition- ally, from each Crohn’s pathology specimen, infl amed and non-infl amed ileum were obtained for comparison. Sam- ples were evaluated for infl ammation using the Crohn’s Histology Index of Severity (range 0–13) and for microves- sel density by quantitative endothelial cell immunohisto- Figure 1: Infl ammation score (range 0–13) of infl amed pediatric Crohn’s chemistry using CD31. Corresponding tissues were assessed for VEGF-A mRNA and protein expression by RT-PCR and disease ileum increased compared to both non-infl ammed Crohn’s Western blot respectively. Results expressed as mean ± SEM diseae and control. Results expressed as mean ± SEM (*P < 0.001). were analyzed for signifi cance (P ? 0.05) by ANOVA and Stu- dent’s t-test. RESULTS: Infl ammation scores were signifi cantly increased (Figure 1) between infl amed CD and controls (5.8 ± 0.7 vs 0.62 ± 0.38, P < 0.001), and between paired infl amed and non-infl amed ileum (5.8 ± 0.7 vs 1.2 ± 0.6, P < 0.001). Increased microvessel density was observed in both infl amed and non-infl amed CD groups compared to con- trols (infl amed 24,955 ± 3,202 μm2, non-infl amed 18,719 ± 2,050 μm2, control 9,032 ± 1,474 μm2), with statistical signifi cance (P = 0.008) only present between infl amed CD and control subjects (Figure 2). Expression of tissue VEGF- A mRNA was upregulated in CD (CD 8.5 ± 2.51 vs control 2.32 ± 0.58, P = 0.034), and was associated with an increased trend in VEGF-A protein levels (VEGF/GAPDH, CD 3.96 vs control 2.20, P = 0.53). CONCLUSION: Angiogenesis is associated with pediatric Crohn’s disease as observed by increased microvessel den- sity that correlates with greater infl ammation in resected ileal specimens. At the molecular level, we demonstrate elevated VEGF transcription and protein levels, which implicates a VEGF pathway for angiogenesis associated infl ammation in early-onset Crohn’s disease. Further inves- tigations regarding mechanism of angiogenesis, its relation- ship to infl ammation, and effectiveness of anti-angiogenic therapies are warranted. Figure 2: Microvessel density (MVD) in pediatric Crohn’s disease ileum (infl amed and non-infl amed) increased compared to control, detected by CD 31 quantitative immunohistochemical staining. Results expressed as mean ± SEM (*P = 0.008).
10 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
10:15 AM – 11:00 AM
303ABC Abstracts Sunday PRESIDENTIAL PLENARY B (PLENARY SESSION II)
363 CONCLUSIONS: PVE is safe and effective in inducing hypertrophy in patients with small FLR and allows 2/3 of Safety and Effi cacy of Portal Vein Embolization patients with inadequate FLR the opportunity for curative Before Planned Major Hepatectomy: An Institutional resection. Experience of 358 Patients Junichi Shindoh1, Ching-Wei D. Tzeng1, Thomas Aloia1, 364 Steven Curley1, Giuseppe Zimmitti1, Steven Y. Huang2, Armeen Mahvash2, Sanjay Gupta2, Michael J. Wallace2, Gastric Emptying, Ensuing GLP-1 Release and Insulin Jean-Nicolas Vauthey1 Sensitivity After Partial Pancreaticoduodenectomy: 1. Surgical Oncology, University of Texas MD Anderson Cancer Improved Glycemic Control in Cases Without Pylorus Center, Houston, TX; 2. Diagnostic Radiology, University of Texas Preservation (Whipple Procedure) 1 2 1 MDAnderson Cancer Center, Houston, TX Johannes Miholic , Marlene Wewalka , Stefan Harmuth , Jens J. Holst3 INTRODUCTION: Portal vein embolization (PVE) induces 1. Department of Surgery, Medical University of Vienna, Vienna, hypertrophy of the future liver remnant (FLR) in patients Austria; 2. Gastroenterology, Department of Internal Medicine with unfavorable tumor distribution and low calculated standardized FLR (sFLR). We sought to evaluate the safety III – Medical University of Vienna, Vienna, Austria; 3. The and effi cacy of PVE. Panum Institute, Department of Medical Physiology, University of Copenhagen, Copenhagen, Denmark METHODS: We evaluated 358 consecutive patients who underwent PVE before intended major hepatectomy from OBJECTIVE: Investigate the relationship between gastric 1995–2012. Diagnoses, morbidity, degree of hypertrophy emptying, postprandial GLP-1 and insulin sensitivity after (DH), and post-PVE resectability were evaluated in the pancreaticoduodenectomy (PD). whole study period and compared over time. BACKGROUND: Abnormal glucose regulation is highly RESULTS: The diseases treated included colorectal liver prevalent in patients with pancreatic neoplasm, and metastases (CLM, 217, 61%), hepatocellular carcinoma (49, resolves in some after PD, the cause of which is unclear. The 14%), extrahepatic biliary cancers (31, 9%), neuroendocrine procedure is carried out with pylorus preservation (PPPD) metastases (25, 7%), intrahepatic cholangiocarcinoma (13, or with distal gastrectomy (Whipple procedure). Acceler- 3%), and others (23, 6%). Right PVE alone was performed in ated gastric emptying, and ensuing enhanced release of 31% of cases; due to tumor distribution and to the necessity glucagon-like peptide-1 (GLP-1) conceivably play a role of resecting segment IV, right PVE with segment IV PVE was in glucose metabolism after PD. Any procedure associated required in 66% of patients. The fi rst-session PVE success with accelerated gastric emptying might improve glycemic rate was 98%. Post-PVE complications occurred in 12/358 control. It was the purpose of this study to shed light on the patients (3%), with portal vein thrombosis occurring in relationship between gastric emptying, GLP-1 and glycemic 6 (2%) patients. Median pre-PVE standardized FLR (sFLR) control after PPPD and the Whipple procedure. was 19% (inter-quartile range, IQR, 15.0–25.9). Median METHODS. A 75 g oral glucose tolerance test was carried out post-PVE sFLR was 30% (IQR, 22.5–38.2). Of 358 patients in tumor free subjects, 13 having undergone PPPD, and in 13 who underwent PVE, 282 (79%) were taken to the operat- after the Whipple procedure (Table1). Gastric emptying was ing room with 240/358 (67%) undergoing curative hepa- measured by the paracetamol absorption method. Plasma tectomy. Post-hepatectomy major complications occurred concentrations of glucose, insulin, GLP-1, and paracetamol in 62/240 (26%) patients, with postoperative hepatic insuf- were measured at baseline, 10, 20, 30 60, 90, 120, 150, and fi ciency (PHI) in 20/240 (8%) and a 90-day liver-related 180 minutes. Homeostasis model assessment-estimated insu- mortality rate of 9/240 (4%). Over the 18-year study period, lin resistance (HOMA-IR) and oral glucose insulin sensitivity the rate of PVE performed for CLM increased from 39% were calculated from glucose and insulin concentrations. before 2005 to 78% in 2010–12. The use of preoperative chemotherapy and long-duration (>12 weeks) chemother- RESULTS. Patients with Whipple procedure as compared apy increased from 26% to 86% and from 16% to 43%, to PPPD had accelerated gastric emptying (p = 0.01) which respectively, in that time frame (all p < 0.001). However, correlated with early (0–60 min.) integrated GLP-1 (AUC30; despite increased preoperative chemotherapy usage, PHI r2 = 0.61; p = 0.02) and insulin sensitivity (r2 = 0.41; p = and 90-day liver-related mortality rates improved over the 0.026), and inversely with HOMA insulin resistance (r2 = last decade (11% and 4%, respectively before 2010 vs. 3% 0.17; p = 0.033). 2 of 13 Whipple patients (15%) as com- and 3%, in 2010–12). pared to 7 of 13 after PPPD (54%) had postload glucose
11 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
concentrations (i.e. 120 minutes postmeal) ≥200 mg/dl (p < METHODS: A cost-effectiveness decision tree was cre- 0.05). None of 13 (0%) after Whipple procedure but 4 of 13 ated using TreeAge (Figure 1). Assigned probabilities were (31%) after PPPD had fasting glucose concentrations ≥126 derived from published literature. The decision point com- mg/dl (p < 0.05). pared extended duration thromboprophylaxis with low CONCLUSIONS. Gastric emptying was accelerated after molecular weight heparin for 21 days after discharge to Whipple procedure as compared to patients who have inpatient-prophylaxis alone, with base case assumptions undergone pylorus preserving PD, resulting in higher post- (Table 1) based on an abdominal oncologic resection with- prandial GLP-1 concentrations and insulin sensitivity and out complications in a 45 year-old male. The end points improved glycemic control. were pulmonary embolism or deep vein thrombosis with attendant costs and assigned effectiveness evaluated by Whipple PPPD P-Value Quality Adjusted Life Years (QALY). Willingness to pay was Age (yr) 61 (32–70) 62 (48–66) NS set at $50,000/QALY. Sensitivity analyses were performed to Interval (mo) 31 (7–199) 19 (5–107) NS assess uncertainty within the model, with particular inter- est in the threshold for cost-effectiveness based on VTE Gastr. Emptying (Integr.parac. 495 (309–860) 319 (230–601) 0.01 30 min) incidence. Early integ. GLP-1 (30 min) 2880 (920–9205) 1740 (340–3215) 0.03 Fasting glucose (mg/dl) 91 (75–123) 108 (83–170) 0.02 Insulin resistance (HOMA-IR) 0.6 (0.22–1.75) 0.8 (0.6–5.8) 0.02 Insulin sensitivity (OGIS 180) 488 (310–568) 406 (265–500) 0.009
365 Extended Duration Thromboprophylaxis Cost- Effectiveness in Abdominal Surgery J.C. Iannuzzi1, A.S. Rickles1, J.G. Dolan2, F. Fleming1, J.R. Monson1, K. Noyes1 1. Surgical Health Outcomes & Research Enterprise, Division of Colorectal Surgery, Department of Surgery, University of Rochester Medical Center, Rochester, NY. 2. Community and Preventative Figure 1: Decision Tree. Health, University of Rochester Medical Center, Rochester, NY RESULTS: Given base case assumptions with VTE probabil- BACKGROUND: Post-discharge thromboprophylaxis is ity of 4%, extended duration thromboprophylaxis had an the practice of prescribing antithrombotic therapy for 21 incremental cost effectiveness ratio of $8123/QALY, which days after discharge, commonly used in surgical patients was considered cost-effective. The results were robust to who are at high risk for venothromboembolism (VTE). sensitivity analysis with the highest uncertainty associated Multiple consensus guidelines recommend extended dura- with VTE incidence and medication cost. The threshold for tion thromboprophylaxis (EDTPPX) after major abdominal the relative cost-effectiveness was a VTE incidence exceed- oncologic resections based on randomized clinical trials ing 2.53%. demonstrating a signifi cant reduction in VTE events after surgical discharge in these patients. While the National CONCLUSIONS: Given the base case assumptions, Comprehensive Cancer Network suggests all major abdom- extended prophylaxis is more cost effective than inpatient inal oncologic resections receive EDTPPX, the American prophylaxis alone, and the threshold for its use should be College of Chest Physicians suggests an individualized cases where the estimated VTE risk exceeds 2.53%. These risk assessment, with only high risk patients undergoing fi ndings should inform future guidelines’ defi nition of oncologic resections suggested to receive EDTPPX, how- “high risk” and individualized risk scores should be devel- ever, high risk is not currently defi ned. The threshold for oped to predict patient likelihood of post-discharge VTE. high risk ought to be informed by when it is cost-effective These results can lead to specifi c individualized EDTPPX to provide EDTPPX, which has not previously been estab- application. lished. In order to further inform current guidelines this study sought to determine the VTE incidence threshold for the cost-effectiveness of low molecular weight heparin for 4 weeks after surgery as compared to inpatient prophylaxis only.
12 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Table 1: Baseline Model Assumptions and Sensitivity Analysis
Variable Baseline Probability Sensitivity Range Cost ($,2010) Cost Range ($) Utility (Range) Abstracts VTE 0.04 0.001–0.3 N/A N/A N/A Sunday DVT 0.923 0–.0923 10,804 3371–22,748 0.84 (0.7–0.931) PE 0.077 0.077–1 16,644 6443–25,554 0.76 (0.6–0.89) LMWH 0.22 0.22–0.7 885.99 357–885.99 0.98 (0.9–.099) No VTE 0.96 0.7–0.999 680 0–680 1 VTE = Venothromboembolism, DVT = Deep Vein Thrombosis, PE = Pulmonary Embolism, LMWH = Low Molecular Weight Heparin
2:00 PM – 4:00 PM 300 VIDEO SESSION I
437 439 Robotic Assisted Median Arcuate Ligament Release Endoscopic Removal of a Laparoscopic Adjustable Martin J. Dib, Mark P. Callery, Marc Schermerhorn, Gastric Band That Is Eroded A. James Moser Aurora D. Pryor, Dana A. Telem, Joshua Karas, Surgery, BIDMC, Boston, MA Georgios Spentzouris, Eleanor Fallon, Jonathan Buscaglia 40-year-old female with chronic abdominal pain and pre- Surgery, Stony Brook University Medical Center, Stony Brook, NY operative aortography consistent with median arcuate This is a case of a 52-year-old male with history of morbid ligament syndrome. Ports and a laparoscopic liver retrac- obesity status-post Laparoscopic Band at an outside hospi- tor are placed. After docking the robot, the left gastric vein tal complicated by port infection and band erosion. Fol- is divided. The left gastric artery is encircled with a vessel lowing port removal, the patient presented for removal of loop to apply inferior traction and identify the common the laparoscopic band. Due to the extent of the intra-gastric hepatic artery of the celiac trunk. The left lateral border of band erosion, total endoscopic removal was attempted suc- the celiac trunk is dissected. Hook cautery and LigaSure is cesfully. This case highlights the mechanism of endoscopic used to divide the left crus of the diaphragm. Finally, cir- removal of the band, challenges encountered, and tech- cumferential skeletonization of the aorta at the entrance of niques to navigate these obstacles. The patient tolerated the celiac trunk is achieved. removal of the band, was started on a liquid diet immedi- ately, and was discharged on post-operative day one. 438 Enucleation of Hepatic Neuroendocrine Tumor 440 Metastases Totally Laparoscopic Left Colonic Resection with Nicholas N. Nissen, Vijay G. Menon Intracorporeal Anastomosis Cedars-Sinai Medical Center, Los Angeles, CA Laura Doyon, Celia M. Divino, Scott Q. Nguyen, Edward Chin Neuroendocrine tumors (NET) represent a unique type of Surgery, The Mount Sinai School of Medicine, New York, NY hepatic metastasis. These tumors tend to be well encapsu- This video demonstrates two complementary laparoscopic lated and generally carry a favorable prognosis. Many of cases, each focusing on techniques for intracorporeal anas- these patients will require repeated hepatic interventions tomosis. The fi rst is an elective sigmoid resection for history over a period of several decades. Surgical enucleation of of uncomplicated diverticulitis. It uses an end-to-side intra- hepatic NETs is a technique that is not often employed corporeal anastomosis performed with a circular stapler. but that holds great potential for preservation of maximal The second is a left hemicolectomy, performed for descend- hepatic parenchyma, while carrying a low risk of injury ing colon cancer. It employs a side-to-side intracorporeal to underlying vascular and biliary structures. This video anastomosis performed with a linear stapler and sewn com- describes the application of enucleation to patients with mon enterotomy. Totally laparoscopic colonic resection NET metastases and addresses patient selection, surgical with intracorporeal anastomosis can facilitate resection techniques and management of complications. in obese patients, as well as improve cosmesis and wound complications by reducing incision length for extraction.
13 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
441 443 Laparoscopic Central Pancreatectomy and Laparoscopic Repair of a Large Right Sided Morgagni’s Pancreaticogastrostomy for the Management of Hernia a Proximally Migrated Pancreatic Stent David Lawrence1, Yuhsin V. Wu2, Michael J. Rosen1,2 Marc G. Mesleh1, Frank Lukens2, Michael B. Wallace2, 1. Case Western Reserve University, Cleveland, OH; 2. Surgery, Horacio J. Asbun1, John Stauffer1 University Hospitals, Cleveland, OH 1. General Surgery, Mayo Clinic Jacksonville, Jacksonville, FL; Morgagni’s hernias are rare congenital anterior diaphrag- 2. Gastroenterology, Mayo Clinic Jacksonville, Jacksonville, FL matic hernias for which the optimal method of repair is A 43 year old female had a pancreatic stent placed during unknown. This video presents a morbidly obese patient with ERCP for elevated LFTs. The stent migrated proximally into oxygen dependent chronic obstructive pulmonary disease the pancreas and was unable to be retrieved with multiple and a Morgagni’s hernia that compresses her entire right endoscopic attempts. After several episodes of pancreatitis, lung. Omentum and colon are seen herniating through she was evaluated for surgical retrieval. A laparoscopic cen- the 10x15cm defect. Through a laparoscopic approach the tral pancreatectomy was performed to remove the stent, and intra-abdominal contents were reduced, the defect primar- a pancreaticogastrostomy was created for reconstruction. ily closed, and re-enforced with mesh. After the repair, the patient had signifi cant improvements in pulmonary status. 442 Laparoscopic repair with mesh re-enforcement is a viable and easily accomplished approach for Morgagni’s hernia Diffi cult Diverticulits and Failed Anastomosis: Troubles repair. and More Troubles Barry Salky Surgery, Mount Sinai Hospital, New York, NY This is a 68 year old female with mulitiple episodes of diverticulitis documented on CT scans. Dyspareunia is a recent symptom. This video demonstates several techni- cal challenges assoiciated with chronic diverticulitis. After completion of the descending rectal anastomosis, a leak was detected and the video demostrates one technique of recovery in a diffi cult clinical situation.
2:00 PM – 4:45 PM 308D PLENARY SESSION III
445 staging system, lymph node ratio (LNR), and overall sur- vival (OS) from date of diagnosis were analyzed. Median Modern Chemotherapy Mitigates Adverse Prognostic follow-up was 19 months (range, 1–211 months). Effect of Regional Nodal Metastases in Stage IV Colorectal Cancer RESULTS: The number of positive regional nodes and LNR 1 1 2 correlated with the presence of multiple sites of distant Yun Shin Chun , Steven Cohen , John H. Donohue , metastases (p < 0.001). Survival was signifi cantly associ- 1 1 2 Barbara Burtness , Michael J. Hall , David M. Nagorney ated with number of positive nodes and LNR, with median 1. Fox Chase Cancer Center, Philadelphia, PA; 2. Mayo Clinic, OS of 36 months with negative regional nodes, compared Rochester, MN to 17 months with ≥7 positive nodes (p < 0.001). Among BACKGROUND: In colorectal cancer, the involvement 315 patients treated with modern oxaliplatin- or irinote- of regional lymph nodes with metastasis is an established can-based chemotherapy after colorectal resection, survival prognostic factor. However, the impact of the number of was not signifi cantly associated with number of positive positive regional nodes on patient outcome with stage IV regional nodes (p = 0.072) or LNR (p = 0.34). The number disease is not well-defi ned. of regional nodal metastases correlated with OS among 249 patients who underwent resection of liver metastases but METHODS: A retrospective review was performed of 869 lost prognostic signifi cance in the subset of 105 patients patients at two tertiary referral centers with synchronous who underwent hepatectomy with perioperative modern stage IV colorectal cancer who underwent resection of their chemotherapy. primary tumors. Associations between number of positive regional lymph nodes stratifi ed by the 7th edition AJCC
14 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL Abstracts Sunday
CONCLUSIONS: In stage IV colorectal cancer, increasing 447 number of positive regional lymph nodes and LNR corre- late with multiple sites of distant metastases and poorer sur- Night Time Is Not the Right Time: Increased Risk of vival. The number of metastatic regional lymph nodes loses Complications After Laparoscopic Cholecystectomy at prognostic signifi cance with modern chemotherapy, partic- Night ularly in patients undergoing resection of liver metastases. Uma R. Phatak, Curtis J. Wray, Debbie Lew, Richard Escamilla, Winston M. Chan, Tien C. Ko, Lillian S. Kao Surgery, University of Texas Health Science Center, Houston, TX 6 446 Evidence from a large national database has shown that Totally Laparoscopic Cytoreductive Surgery and performance of non-emergent general surgery procedures at night does not predispose patients to increased morbid- Hyperthermic Intraperitoneal Chemotherapy for ity or mortality. However, these results may not be gener- Mucinous Adneocarcinoma of the Appendix alizable to high risk populations of medically underserved Cherif Boutros, Nader Hanna patients. We hypothesized that performance of laparoscopic Division of Surgical Oncology, University of Maryland, cholecystectomy (LC) at night in such a population would Baltimore, MD be associated with increased post-operative complications. Open cytoreductive Surgery (CRS) and heated intraperito- We conducted an IRB approved single center retrospective neal chemotherapy (HIPEC) has emerged as the procedure review of consecutive LC patients between October 2010 of choice for mucinous adenocarcinoma of the appendix and May 2011 at a safety-net hospital in Houston, TX. (MAA), however is associated with substantial morbidity. Data were collected regarding demographics, date and site We present a case of a totally laparoscopic R0- CRS-HIPEC of diagnosis (defi ned as fi rst imaging study demonstrat- for MAA. CRS included: right hemicolectomy, omentec- ing gallstones), number of biliary-related admissions and tomy, cholecystectomy, bilateral salpingo-oopherectomy, emergency room (ER) visits between diagnosis and surgery, excision of the round and falciform ligaments and strip- length of stay (LOS) for each admission, dates and types of ping of the peritoneum of the right diaphragm; followed procedures, dates and types of imaging studies, and 30-day by HIPEC through single infl ow and outfl ow catheters. OR postoperative complications (bile leak/biloma, common time was 380 mns and EBL was 100 mL. There was no post- bile duct injury, retained stone, superfi cial surgical site operative morbidity. The patient was discharged home on infection, organ space abscess, pneumonia, readmission, postoperative day 8. and death). We defi ned “night” as 7PM to 7AM. Statistical analyses were done using STATA 12 (College Station, TX).
15 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
During the 8-month period, 580 patients received LC and hazards. We compared survival of patients with thrombosis incision times were available for 549. Of these 38% (n = of the mesenteric venous system to that of patients with 208) were elective and 62% (n = 341) were non-elective. A patent reconstructions. majority were female (n = 460, 84%) and Latino (n = 456, 83%). There were 196 LC performed at night of which 186 were non-elective and 10 were elective. Of the 353 daytime LCs, 198 were elective and 155 were non-elective. There were 35 complications in 22 patients (4 elective, 18 non- elective). Multivariate analysis revealed age (OR 1.05, 95% CI 1.01 to 1.08, p = 0.003) and LC at night (OR 3.1, 95% CI 1.3 to 7.6, p = 0.012) to be associated with increased risk of complications. The predicted probability of a complication increased three fold for older patients who received LC at night (Figure). Age and performance of LC at night were predictive of an increased risk of complications among medically under- served patients treated at a high volume safety net hospi- tal with limited resources. Restricting performance of LCs to the daytime in high risk patients, such as the elderly, may lead to improved outcomes in this challenging clinical setting.
Predicted probability of complication after LC at night by age.
448 Short-Term But Not Long-Term Patency of Venous Reconstruction During Pancreatic Resection Predicts Survival Irmina Gawlas, Irene Epelboym, Megan Winner, Joseph DiNorcia, Yanghee Woo, James A. Lee, Beth Schrope, John A. Chabot, John D. Allendorf Department of Surgery, Columbia University, New York, NY BACKGROUND: Pancreatic surgery with concomitant vascular reconstruction is being performed with increas- RESULTS: Between 1994 and 2011, 203 pancreatic opera- ing frequency, and offers the benefi ts of surgical resection tions requiring venous reconstruction were performed. Of to patients with locally advanced disease. The technique these, 106 (52.2%) included resection of the portal vein (PV), is not standardized, however, and the short and long-term 59 (29.1%) included the superior mesenteric vein (SMV) patency rates and the clinical signifi cance of thrombosis of only, and in 38 (18.7%) patients, the confl uence of the PV a reconstructed venous system are unknown. and SMV was resected. Segmental resection was performed in 131 (64.5%), and 72 (35.5%) underwent tangential resec- METHODS: We reviewed clinical and operative character- tion. Ninety-seven veins (47.8%) were repaired primar- istics as well as follow up records of patients who under- ily, 67 (33.0%) were repaired using a venous interposition went pancreatic resections requiring venous resection and graft, and 34 (16.8%) were repaired using an autologous reconstruction from 1994 to 2011. We sought to identify vein patch. Acute thrombosis occurred in 9 (4.4%) cases, predictors of acute (occurring within 30 days) thrombosis and was signifi cantly associated with increased periopera- of the venous reconstructions using logistic regression, and tive mortality (22.2% versus 4.6%, p = 0.023). After exclud- predictors of late loss of patency using Cox-proportional ing cases of perioperative mortality, acute thrombosis was
16 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
associated with decreased median survival (7.1 versus 15.9 RESULTS: Nineteen patients had carcinoid syndrome and months, p = 0.011) and increased hazard of death (HR 8.6, all had dramatic improvement after surgery, with complete
CI 3.7–19.9, p < 0.001). These events were more common resolution in 11 (58%) cases. Overall there were 8 complica- Abstracts Sunday in cases of total or subtotal resection compared to Whipple tions of Clavien grade >2 including bile leak requiring ERCP or distal resections (22.2 versus 2.7%, p < 0.001). Long- (n = 2) and repeat laparotomy (n = 4). There was no 30-day term follow-up imaging was available for 138 patients at a mortality. Median length of stay was 7 days. Overall sur- median of 11.7 months. Of these, 43 (31.2%) experienced vival for the entire cohort at 1, 3 and 5 years was 95%, 82% a loss of patency of the portal venous system at a median and 82%, while the progression free survival at 1, 3 and 5 of 9.5 months; the majority of these were associated with years was 77%, 37% and 28%. In the 22 patients under- tumor recurrence. Independent predictors of late loss of going near total surgical CR, no patient required repeat patency were age under 65 (HR 2.2, CI 1.2–4.1, p = 0.015) hepatic intervention within 12 months. In the 25 patients and segmental resection (HR 3.3, CI 1.5–7.2, p = 0.002). undergoing partial surgical CR, 18 (72%) went on to receive Later loss of patency was not associated with decreased postoperative hepatic treatments within 12 months. Pro- median survival (18.1 versus 16.8 months, p = 0.455) or gression free survival was similar in patients whether they increased hazard of death (HR 1.3, CI 0.8–2.1, p = 0.375). underwent total CR or partial CR with staged hepatic treat- CONCLUSIONS: Acute thrombosis of the reconstructed ment. Patients who failed to undergo postoperative hepatic portal venous system after pancreatic surgery is clinically therapy (n = 7) were at increased risk of progression com- signifi cant; it is associated with increased perioperative pared either to patients with near total CR (HR = 3.10, P = mortality, and even when non-fatal, is associated with 0.044) or partial CR and staged liver treatment (HR = 3.37, decreased survival. Late loss of patency occurs in one-third P = 0.029) (Figure). of patients but does not affect survival.
449 Simultaneous Surgical Resection of Primary and Metastatic Carcinoid and Neuroendocrine Tumors Is Both Safe and Effective Nicholas N. Nissen1, Vijay G. Menon1, Edward M. Wolin2, Run Yu2, James M. Mirocha3, Alagappan Annamalai1, Deepti Dhall4, Ashley Wachsman5, Marc L. Friedman5, Steven D. Colquhoun1 1. Hepatobiliary and Pancreatic Surgery, Cedars-Sinai Medical Center, Los Angeles, CA; 2. Carcinoid and Neuroendocrine Tumor Program, Cedars-Sinai Medical Center, Los Angeles, CA; 3. Biostatistics, Cedars-Sinai Medical Center, Los Angeles, CA; 4. Pathology, Cedars-Sinai Medical Center, Los Angeles, CA; 5. Radiology, Cedars-Sinai Medical Center, Los Angeles, CA INTRODUCTION: Management strategies for patients Progression Free Survival Related to Type of Cytoreductive Treatment. with carcinoid and neuroendocrine tumors (CNETs) gen- erally include removal of the primary tumor and cytore- CONCLUSION: To our knowledge this series represents the duction (CR) of metastatic tumor burden, both to improve largest single center report of simultaneous resection of pri- survival and control symptoms. Patients with synchronous mary abdominal CNETs and hepatic metastases in the lit- presentation of primary tumors and hepatic metastases erature. Our results demonstrate that this surgical approach present a unique challenge. We reviewed our experience is safe and effective in expert hands. In patients undergo- with simultaneous surgical removal of primary abdominal ing near total hepatic CR, which made up almost half of CNETs and hepatic metastases. our series, no additional hepatic treatments were required PATIENTS: Forty-seven patients underwent simultaneous over the next year, which in turn demonstrates the effective hepatic resection and removal of either small bowel carci- consolidation of treatments into a single surgical endeavor. noid (n = 32) or pancreatic NET (n = 15) by two experi- In remaining patients, resection of the primary tumor com- enced hepatobiliary surgeons as part of a multidisciplinary bined with partial hepatic CR combined with postoperative CNET treatment group. Surgical details are shown in the hepatic therapy was equally effective. A multidisciplinary Table. In 22 patients, surgery was undertaken with a goal of and multimodal approach is essential in these patients. near total surgical CR, while in 25 patients partial surgical CR was performed as part of a plan to include postoperative hepatic arterial or ablative therapy. Tumor progression was categorized using RECIST criteria.
17 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Variables Associated with Simultaneous Resection >0.8). MRE accuracy for infl ammation, thickening, steno- sis, abscess and fi stula were all above 85% in per-patient Type of Primary Surgery analysis. In 68/75 cases (90.7%) both approach and strategy Small Bowel Resection 21 were correctly predicted by MRE. Conversely, in 7/75 cases Right Hemicolectomy 12 (9.3%, 3 false positives: 2 enterocolic fi stulas and 1 anasto- Segmental or Distal Pancreatic Resection 11 motic stricture; and 4 false negatives: 3 enteric fi stulas with Pancreaticoduodenectomy 3 colon, duodenum and bladder and 1 enteromesial abscess) Type of Liver Surgery surgical strategy (type of resection or strictureplasty, n = 5) Major Resection of 3 or more segments 9 and/or surgical approach (conversion from laparoscopy to Segmental Resection (1 or 2) 11 open surgery, n = 2) changed due to discordance with MRE Multiple Wedge Resections and Enucleations 27 fi ndings. Largest Liver Tumor size (cm) 4.9 (mean); 3.5 (median); CONCLUSION: Preoperative MRE correctly predicts surgi- 0.6–17 (range) cal strategy in the majority of patients undergoing surgery No. of liver lesions resected; 5.5 (mean); 3 (median); for complicated CD. MRE is especially valuable before lapa- 1–28 (range) roscopic surgery, since unrecognized lesions may lead to Grade; High: Intermediate: Low: N/s 4.3%: 27.7%: 51.1%: 17% conversion to open surgery. Differentiation; Poor: Well: N/s 4.3%: 82.9%: 12.8% Positive Lymph nodes (%) 78.6% 451 R0 Resection (%) 57.4% Bilobar Resection (%) 55.3% Tumor Size Does Not Dictate Prognosis After Resection for Hepatocellular Carcinoma: Results from a Large Western Series 450 Michael D. Kluger1,2, Andrea Belli2, Alexis Laurent2, Daniel Azoulay2, Daniel Cherqui1,2 Value of Preoperative Magnetic Resonance Enterography 1. Division of Hepatobiliary Surgery and Liver Transplantation, to Predict Surgical Findings and to Guide Decisions in New York-Presbyterian Hospital Weill Cornell Medical College, Crohn’s Disease: A Prospective Study New York, NY; 2. Service de Chirurgie Digestive et Hépatobiliaire, 1,2 3 1,2 Antonino Spinelli , Gionata Fiorino , Piero Bazzi , Hôpital Henri Mondor, Assistance Publique-Hôpitaux de Paris – 4 1 1 Cristiana Bonifacio , Matteo Sacchi , Sarah De Bastiani , Université Paris-Est, Créteil, France Andrea Gatti1, Alberto Malesci2,3, Luca Balzarini4, INTRODUCTION: Operative management remains the Laurent Peyrin-Biroulet5, Marco Montorsi1,2, Silvio Danese3 gold standard approach for hepatocellular carcinoma 1. Department of Surgery, Istituto Clinico Humanitas, Rozzano (HCC). Resection is the preferred treatment in patients Milano, Italy; 2. Dip. di Biotecnologie Mediche e Medicina without cirrhosis, with transplantation being the best Traslazionale, Università degli Studi di Milano, Milano, Italy; option for decompensated cirrhotics. This study evaluated 3. Department of Gastroenterology, Istituto Clinico Humanitas, underlying liver disease, operative factors and histopatho- Rozzano Milano, Italy; 4. Department of Radiology, Istituto logical characteristics on overall and recurrence-free sur- Clinico Humanitas, Rozzano Milano, Italy; 5. Department of vival in 313 patients undergoing liver resection for HCC at Hepato Gastroenterology, University of Nancy, Nancy, France a single Western center. BACKGROUND: Surgery is still required for many patients METHODS: Patients who underwent liver resection for with Crohn’s disease (CD). Intraoperative detection of new HCC between 3/89 and 9/10 were studied. Patients were lesions is common and may lead to a change in the planned not excluded based on tumor size, extent of fi brosis, or approach (laparoscopic or open surgery) and strategy (type etiology of underlying liver disease. As indications for of resection or strictureplasty). Whether magnetic reso- treatment are mostly based on tumor size, patients were nance enterography (MRE) can be used to optimize surgi- stratifi ed by diameter: <50 mm, 50–100 mm and >100 mm. cal planning and to guide decision-making in CD patients Patients with Child’s A cirrhosis, no esophageal varices, and undergoing surgery is currently unclear. a platelet count ≥100 × 10^9/L were directed toward resec- tion. Kaplan-Meier and Cox regression methodology were METHODS: Seventy-fi ve consecutive patients with com- utilized. plicated CD who were candidates for surgery were pro- spectively enrolled. MRE was performed according to RESULTS: 36% had tumors <50 mm, 36% had tumors a standardized protocol within 30 days before surgery. 50–100 mm, and 28% had tumors >100 mm. Patients with Two experienced radiologists blindly and independently larger tumors were more likely to have normal underlying assessed MRE images. Radiological fi ndings were correlated liver parenchyma: 43% >100 mm, 15% 50–100 mm and 1% with intraoperatively detected lesions. Analysis included <50 mm (p < 0.001). 77% underwent an open and 23% a MRE accuracy (per-segment and per-patient) and change in laparoscopic procedure (p < 0.001). Major hepatectomies surgical strategy due to discordance with MRE fi ndings. comprised 56%, anatomic resections 87%, and R0 88% of resections. There was no signifi cant difference in Clavien RESULTS: Surgery was performed laparoscopically in 39/75 3–5 complications among the groups (p = 0.78), 16% over- pts (52%; conversion to open surgery 6/39, 15%). Concor- all. This rate decreased in the second decade of our expe- dance rate among observers was excellent (kappa value rience. For example, the mortality rate between 3/89 and
18 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
12/99 was 14%, and 5% through 9/10 (p < 0.008). Median overall survival was 60 months, with 1- and 5-year over- all survival rates of 76% and 50%. On multivariate analy- Abstracts Sunday ses, intra-operative transfusion (HR = 2.60), cirrhosis (HR = 2.42), salvage transplantation (HR = 0.23), poorly differ- entiated tumor (HR = 2.04), satellite lesions (HR = 1.68), microvascular invasion (HR = 1.48), and AFP > 200 (HR = 1.53) were signifi cant predictors of survival. Median time to recurrence was 20-months, with 1- and 5-year recurrence- free survival rates of 61% and 28%. By multivariate analyses intra-operative transfusion (HR = 2.15), poorly differenti- ated tumor (HR = 1.87), cirrhosis (HR = 1.69) and micro- vascular invasion (HR = 1.71) independently impacted recurrence-free survival. CONCLUSION: It is demonstrated that resection is a safe and readily available treatment for any size HCC in properly selected patients in the modern era of liver surgery. Tumor size did not independently impact recurrence or survival on multivariate analyses, whereas tumor histopathology and background parenchyma did. The current investiga- tion adds to a growing body of literature supporting that HCC tumor biology and the condition of the non-tumor parenchyma should be given greater consideration in con- sidering resection in this era of organ shortage.
Table 1: Clinical Characteristics, Operative Details and Pathologic Characteristics of Patients with HCC Undergoing Resection
Figure 1: Overall and recurrence-free survival among patients resected for hepatocellular carcinoma stratifi ed by tumor size.
19 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
452 6 453 Analysis of a Single-Series Learning Curve for Peroral Peroral Endoscopic Myotomy (POEM) Feasible as Esophageal Myotomy (POEM) Reoperation Following Heller Myotomy Ezra N. Teitelbaum, Byron F. Santos, Fahd O. Arafat, Amy K. Yetasook1, Jin-cheng Zhao1,2, Woody Denham1,2, Nathaniel J. Soper, Eric S. Hungness John G. Linn1, Michael B. Ujiki1,2 Northwestern University, Chicago, IL 1. Minimally Invasive Surgery, NorthShore University BACKGROUND: Peroral esophageal myotomy (POEM) is a HealthSystem, Evanston, IL; 2. Department of Surgery, University novel endoscopic operation for the treatment of achalasia. of Chicago, Chicago, IL The operator learning curve for POEM and patient factors Peroral Endoscopic Myotomy (POEM) is a promising new associated with operative diffi culty are not known. treatment for achalasia. We present three cases of recurrent METHODS: A single-institution prospective POEM out- achalasia after failed therapy with reoperation by POEM. comes database was analyzed. All POEM procedures were Additionally, we also demonstrate our technique with performed conjointly by the same two surgeons. Associa- a patient who underwent POEM after failed endoscopic tions between preoperative patient variables (series case intervention and laparoscopic Heller myotomy (LHM) with number, gender, age, BMI, ASA class, prior treatment with reoperation by POEM. This case series illustrates feasbility dilation or Botox, symptom duration, manometric pres- of completing a peroral minimally invasive approach in the sures, achalasia subtype) and operative outcomes (proce- management of recurrent achalasia. dure time, tunnel length, myotomy length, number of clips used for closure, EBL, mucosal perforation, need to decom- 454 press pneumoperitoneum) were tested using bivariate lin- ear correlation. To assess for changes in effi ciency over the Idiopathic Pulmonary Fibrosis and Gastroesophageal course of the series, the total procedure time and the time Refl ux: Implications for Treatment required for each individual procedural step (submucosal Marco E. Allaix1, P. Marco Fisichella2, Fernando A. Herbella1, access, tunnel creation, myotomy, and mucosal closure) Marco G. Patti1 were tested for best fi t to linear, logarithmic, and exponen- 1. Department of Surgery, Center for Esophageal Diseases, tial regression curves using case number as the dependent variable. A subgroup analysis of treatment-naïve patients University of Chicago Pritzker School of Medicine, Chicago, IL; was secondarily performed. 2. Department of Surgery, Swallowing Center, Loyola University Chicago, Stritch School of Medicine, Maywood, IL RESULTS: 30 patients underwent POEM, of which 26 were treatment-naïve. Preoperative symptom duration was posi- BACKGROUND: While the pathogenesis of idiopathic tively associated with increased operative time (r^2 = .55, pulmonary fi brosis (IPF) is multifactorial, it has been shown p < .01). Prior achalasia treatment and EGJ resting pressure that the prevalence of abnormal refl ux (GERD) is very high, were both positively associated with operative time at a and that antirefl ux surgery may affect the progression of trend level (p = .08 for both). Case number correlated nega- this disease. tively with the number of clips required for closure (r^2 = AIMS: The aims of this study were to compare in a group of –.51, p < .01), whereas duration of symptoms was positively patients with GERD and a group of patients with GERD and correlated with clip number (r^2 = .40, p < .05). Case num- IPF: a) the clinical presentation; b) the esophageal function ber correlated negatively, whereas duration of symptoms as defi ned by high resolution manometry; and c) the refl ux correlated positively, with occurrence of a mucosal perfo- ration, both at a trend level (p = .06 and .07). Myotomy profi le by dual sensor pH monitoring. length proximal to the EGJ increased over the course of the PATIENTS AND METHODS: We compared the clinical series (r^2 = .44, p = .02), whereas there was no change presentation, the esophageal function and the refl ux profi le in myotomy length distal to the EGJ. Total procedure time in 80 patients with GERD and in 22 patients with GERD and time to perform tunnel creation, myotomy and muco- and IPF. sal closure did not change over the course of the series. Sub- RESULTS: Data are expressed as mean ± SD. mucosal access time decreased over the course of the series (r^2 = .22, p < .01) with best fi t to a logarithmic curve. CONCLUSIONS: The results of this study show that in In treatment-naïve patients only, mucosal closure time patients with GERD and IPF: a) heartburn is present in less decreased over the series (r^2 = .17, p = .03) with best fi t to than 60% of patients; b) with the exception of a weaker an exponential curve. UES, the esophageal function is preserved; and c) proximal CONCLUSIONS: In this POEM series, the time needed to refl ux is more common, and in the supine position it is access the submucosa and the number of clips required to coupled with a slower acid clearance. Because these factors close the mucosotomy both decreased with experience. expose IPF patients to the risk of aspiration, antirefl ux sur- Myotomy length proximal to the EGJ increased with expe- gery should be considered early in the course of the disease. rience. Total procedure time did not change over the course of the series, and may not be an important marker of proce- dural skill for POEM. Longer symptom duration and prior endoscopic treatment may result in increased operative diffi culty.
20 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
GERD GERD + IPF sigmoid-shaped mega-esophagus (stage 4) were excluded. (80 Patients) (22 Patients) P Value Symptoms were scored using a detailed questionnaire for dysphagia, regurgitation, and chest pain; barium swallow, Abstracts Age (years) 55.7 ± 15 61.3 ± 8.9 0.098 Sunday Gender (male), N (%) 31 (38.8) 13 (59.1) 0.143 endoscopy, and esophageal-manometry were performed, Heartburn, N (%) 67 (83.8) 13 (59.1) 0.028 before and 6 months after the treatment. Patients were clas- sifi ed in three age brackets: group A (≤45 years), group B LES pressure (mmHg) 19.9 ± 9.7 20.5 ± 11.9 0.807 (45–70) and group C (≥70). Treatment was defi ned as a fail- Normal peristalsis, 51 (63.8) 14 (63.6) 0.810 th N (%) ure if the postoperative symptom-score was >10 percentile of the preoperative score (i.e. > 8). Hypotensive UES, 6 (7.5) 7 (31.8) 0.008 N (%) RESULTS: We consecutively performed the LHD as pri- % time <4, total Distal 9.7 ± 6.6 12.7 ± 13.6 0.149 mary treatment in 514 achalasia patients, 272 (53%) in Proximal 0.9 ± 1.1 2.5 ± 6.9 0.047 group A, 208 (40.4%) in group B and 34 (6.6%) in group C. Acid clearance, total 85.3 ± 65.2 137.5 ± 112.4 0.006 The mortality was nil; the conversion and morbidity rates (seconds) Distal 42.4 ± 67.1 169.9 ± 406.8 0.008 were both 1.2% with no-difference in the 3 groups. Proximal % time <4, supine 10.5 ± 12.8 8.6 ± 17.1 0.569 Group C patients had higher preoperative symptom scores Distal Proximal 0.5 ± 1.2 2.9 ± 7.6 0.007 (p = 0.02), while the symptom duration was similar in all Acid clearance, supine 181.1 ± 270.1 171.5 ± 259.9 0.882 groups. At a median follow-up of 40 months (IQR 15–80), (seconds) Distal 47.6 ± 72.3 899.1 ± 1668.1 <0.001 the median of symptom scores was signifi cantly lower after Proximal surgery (18 [IQR 14–20] vs 0 [IQR 0–3]; p < 0.0001). The median of resting LES pressure decreased from 27 mmHg (IQR 19–36) to 11 mmHg (IQR 8–14) (p < 0.001) and the 455 residual LES pressure from 10 mmHg (IQR 5–17) to 3 mmHg (QR: 1–5) (p < 0.001). No statistically signifi cant differences Laparoscopic Heller Myotomy Can Be Used as Primary emerged between the 3 groups in any of these aspects. Therapy for Esophageal Achalasia Regardless of Age Mucosal tears occurred in 16 patients (3%): 5 (1.8%) in Renato Salvador1, Mario Costantini1, Francesco Cavallin1, 1 1 1 group A; 8 (3.9%) in group B; and 3 (8.9%) in group C (p Elena Finotti , Cristina Longo , Michela Di Giunta , = 0.06). The postoperative hospital stay was slightly lon- 1 1 1 Nicola Passuello , Loredana Nicoletti , Giovanni Capovilla , ger for group C (p = 0.06). The treatment failure rate was Stefano Merigliano1, Ermanno Ancona1, Giovanni Zaninotto1 quite similar: 31 failures in group A (11.4%), 19 in group B 1. Department of Surgical and Gastroenterological Sciences, (9.1%) and 2 in group C (5.9%) (p = 0.55) (table). The fail- Clinica Chirurgica 3, University of Padova, Padova, Italy ures were seen more in manometric-pattern III (22.2%, p = BACKGROUND: Laparoscopic Heller-Dor (LHD) surgery is 0.002). All the patients whose surgical treatment failed were the current treatment of choice for patients with esopha- treated with pneumatic dilations. The overall success rate geal achalasia, but elderly patients are generally referred of this combined treatment was therefore 98.4% (507/515). for less invasive treatments (pneumatic-dilations or botu- Postoperative 24-hour pH-monitoring was abnormal in 16 linum-toxin injections). The aim was to assess the effect of patients (6.6%): 7 patients were in group A, 6 in group B age on the surgical outcome of patients receiving laparo- and 3 in group C (p: n.s.). scopic Heller-Dor as primary treatment. DISCUSSION: LHD is often performed in old patients as METHODS: We evaluated the patients who underwent sur- a “last resource”, after other treatments have failed. Given gery from 1992 to January 2012 . Patients who had already our high success and low complication rate, this study sup- been treated for esophageal achalasia and patients with ports the use of LHD as the fi rst treatment of achalasia in elderly patients with an acceptable surgical risk.
Table: Postoperative Findings in the Three Groups. Data Are Shown as Median and IQR (in Brackets)
Group A ( 45 Yrs) Group B (45–70 Yrs) Group C ( 70 Yrs) n = 272 n = 208 n = 34 p Value Postoperative symptom score 0 (0–3) 0 (0–3) 0 (0–3) 0.89 Postoperative chest pain score 0 (0–0) 0 (0–0) 0 (0–0) 0.11 LES resting pressure (mmHg) 10 (8–13) 12 (8–17) 10 (7–14) 0.07 LES residual pressure (mmHg) 3 (1–5) 3 (2–6) 2 (1–4) 0.21 Esophageal diameter (mm) 20 (18–27) 22 (20–25) 22 (20–25) 0.95 Mucosa tear 5 (1.8%) 8 (3.9%) 3 (8.9%) 0.06 Postoperative hospital stay (days) 3 (3–4) 3 (3–4) 3 (3–6) 0.06 Failures 31 (11.4%) 19 (9.1%) 2 (5.9%) 0.55
21 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Monday, May 20, 2013 7:30 AM – 9:15 AM 303ABC VIDEO SESSION II: BREAKFAST AT THE MOVIES
507 510 Laparoscopic Enucleation of Benign and Low Grade Transanal Minimally Invasive Surgery Assisted Single Hepatic Lesions Incision Low Anterior Resection with Total Mesorectal Nicholas N. Nissen, Vichin Puri, Vijay G. Menon Excision (TAMIS Assisted LAR TME) in a Cadaver Model Cedars-Sinai Medical Center, Los Angeles, CA Elisabeth C. McLemore1, Alisa M. Coker1, Bikash Devaraj1, 1 1 1 1 Enucleation is a technique which can be applied to benign Jeffrey Chakedis , Ali Maawy , Tazo Inui , Mark A. Talamini , and low grade lesions of the liver such as select neuroendo- Santiago Horgan1, Michael R. Peterson2, Patricia Sylla3, crine tumors (NET), cysts, hemangiomas and focal nodular Sonia Ramamoorthy1 hyperplasia. The benefi ts of enucleation include the preser- 1. Surgery, UC San Diego, La Jolla, CA; 2. Pathology, UC vation of maximal hepatic parenchyma, as well as the low San Diego, La Jolla, CA; 3. Surgery, Massachusetts General, likelihood that underlying vascular or biliary structures will Boston, MA be compromised. A laparoscopic approach to enucleation not only offers the benefi ts of minimal access surgery, but The purpose of this video is to demonstrate the feasibility also allows simultaneous access to multiple regions of the of an innovative technique for the surgical management of abdomen. This may be ideal for managing certain scenarios rectal cancer: trans anal minimally invasive surgery assisted such as the patient with distal pancreatic NET and synchro- low anterior resection with total mesorectal excision nous liver metastases. Illustrative cases are shown. (TAMIS assisted LAR TME) in a cadaver model. Trans anal LAR via natural orifi ce translumenal endoscopic surgery (NOTES) has been reported in cadaveric series using rigid 508 transanal platforms. This procedure has not been described Use of Fluorescence Angiography During 2-Field using a combination of a single incision laparoscopy and Minimally Invasive Esophagectomy TAMIS trans anal endoscopic platform. Herein, we describe C. Daniel Smith, Steven P. Bowers the fi rst cadaveric series of TAMIS assisted laparoscopic LAR with TME. Surgery, Mayo Clinic Florida, Jacksonville, FL This video depicts the use of fl uorescence angiography 511 using the SPY technology to assess the perfusion of the gas- tric conduit during 2-fi eld minimally invasive technology. Central Pancreatectomy with Pancreatogastrostomy for The SPY allowed real-time visualization of the perfusion of Traumatic Transection of the Pancreas the gastric conduit and subsequent esophagogastrostomy. Farzad Alemi, Jonathan Carter, Carlos U. Corvera The tip of the conduit was found to have poor perfusion Surgery, UCSF, San Francisco, CA based on the fl uorescence imaging and was resected. The distal end of the gastric conduit at the anastomosis showed A 22 year-old man sustained abdominal trauma resulting good perfusion. in complete transection of the pancreas at the neck of the gland. Pancreatic ascites and mesenteric hematoma was found at exploration. Given the normalcy of the distal 509 pancreatic remnant, a central pancreatectomy and pan- Surgical Treatment Options for Delayed Gastric creatogastrostomy was done. The operation entailed 1) Emptying partial mobilization of the distal pancreatic remnant, 2) Nathan Lytle, Juan Toro, Ankit Patel, Jahnavi Srinivasan, cannulation and stenting of the pancreatic duct, 3) oppos- ing anterior and posterior gastrotomies, and 4) a two-layer, S. Scott Davis, Edward Lin interrupted anastomosis. The proximally transected seg- Surgery, Emory University, Atlanta, GA ment was treated with biologic adhesive and wide drainage. Delayed gastric emptying is a common problem that is Postoperatively the patient exhibited normal digestion and seen by both gastroenterologists and general surgeons. glucose homeostasis. Poor gastric emptying from outlet obstruction or diabetic, post-surgical, and idiopathic gastroparesis can be diffi cult to treat. This video demonstrates four surgical options for treatment. Gastric stimulator, duodenojejunostomy, pylo- roplasty, and distal gastrectomy are shown demonstrating technique, and indications for each are discussed.
22 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
512 513 Laparoscopic Hilar Resection with Roux-en-Y Robotic Assisted Laparoscopic Total Proctocolectomy Hepatico-Jejunostomy with Ileal Pouch Anal Anastomosis Juan Toro, Nathan Lytle, Ankit Patel, S. Scott Davis, Mehraneh D. Jafari, Alessio Pigazzi, Michael J. Stamos Edward Lin, Juan M. Sarmiento Surgery, University of California, Irvine, Orange, CA Surgery, Emory University, Atlanta, GA This is a 38 year old female with polyposis syndrome. A Several series demonstrate the safety and feasibility of laparoscopic total colectomy is performed in a medial to laparoscopy for complex hepatobiliary procedures. These lateral fashion. The ielocolic vessels, middle colic vessels, reports show the results of laparoscopic liver resections for and inferior mesenteric vessels are divided. The four arm different types of neoplasms and benign diseases such as Di Vinci robot is docked and a total proctocolectomy is car- choledochal cysts. However, the adoption of laparoscopic ried out to the level of the dentate line. The distal rectum Abstracts resection with common bile duct excision is still uncom- is transected and the specimen removed through a Pfan- Monday mon due to technical complexity and longer operative nenstiel incision. An ileal J-pouch is created and an end to times. We perform laparoscopic extended hepatectomies end ileo-anal anastomosis is performed. Pathology revealed with biliary reconstructions using minimally invasive surgi- colonic polyposis with no evidence of malignancy. cal technique. In this video, we demonstrate our technique for right and left extended hepatectomies with Roux-en- Y hepaticojejunostomy. This approach allows superior visualization.
9:45 AM – 11:00 AM 308D PLENARY SESSION IV
587 addition, data on hospital setting (teaching-research hospi- tals vs. community hospitals) were collected and analyzed. Trends in the Surgical Treatment of Pancreatic The Cochran-Armitage test for trend was used to assess Adenocarcinoma changes in treatment over time. Siavash Raigani1, John Ammori2, Julian Kim2, Jeffrey Hardacre2 RESULTS: 47, 086 patients with stage 1–2 pancreatic ade- 1. Department of Surgery, CWRU School of Medicine, Cleveland, nocarcinoma were included in the analysis. Between 2003– OH; 2. Department of Surgery, University Hospitals Case Medical 2009, the use of surgery alone as fi rst course treatment of Center, Cleveland, OH stage 2 disease decreased signifi cantly at both teaching- INTRODUCTION: Multiple prospective, randomized trials research hospitals and community hospitals by nearly 25% have demonstrated that the addition of adjuvant therapy (p < 0.0001 for both cases). In the same period, the use of after surgical resection of pancreatic cancer improves sur- chemotherapy in addition to surgery as treatment of stage vival compared to surgery alone. However, the optimal type 1 and 2 disease increased two-fold at both types of hospi- of adjuvant therapy, chemotherapy alone or chemotherapy tals (p < 0.0001 for all cases). Treatment with surgery plus combined with chemoradiation therapy, remains con- chemoradiation decreased signifi cantly for both stages in troversial. Our aim was to determine whether the type of both hospital settings by approximately 30% (p < 0.05 for adjuvant therapy for pancreatic cancer given in the United all cases). Non-surgical treatment for stage 2 disease was States has changed by examining treatment trends using surprisingly high and signifi cantly increased over time (p the National Cancer Data Base. < 0.0001 for both), ranging from approximately 30–37% at teaching-research hospitals and 40–49% at community METHODS: The National Cancer Data Base (NCDB) is a hospitals. national oncology outcomes database for over 1,500 Com- mission on Cancer-accredited cancer programs. Patients CONCLUSION: Data from the NCDB from 2003–2009 diagnosed with stage 1–2 pancreatic adenocarcinoma illustrate changes in the adjuvant treatment of pancreatic between 2003–2009 were selected from the NCDB Hospital cancer. There is an alarmingly high rate of non-surgical Comparison Benchmark Reports. Attention was paid to the therapy for stage 1 and 2 disease. The use of chemother- initial treatment regimen, such as surgery alone, surgery apy alone as adjuvant therapy increased whereas the use of plus chemotherapy, or surgery plus chemoradiation. In multimodality therapy decreased.
23 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Surgery Surgery Plus Surgery Plus No Surgical Only Chemotherapy Chemoradiation Therapy Percent Percent Cochran- Change Cochran- Percent Cochran- Change Cochran- Armitage Test Between Armitage Test Change Armitage Test Between Armitage Test Percent Change for Trend 2003 and for Trend between 2003 for Trend 2003 and for Trend Between 2003 and p Value 2009 p Value and 2009 p Value 2009 p Value 2009 Stage 1 Teaching-Research 0.1069 –3.14% <.0001 235.48% 0.0006 –27.70% 0.162 –1.42% Hospitals Community Hospitals 0.6323 –10.86% <.0001 261.12% 0.0002 –35.31% 0.145 3.53% Stage 2 Teaching-Research <.0001 –23.05% <.0001 229.31% <.0001 –34.21% <.0001 16.89% Hospitals Community Hospitals <.0001 –23.97% <.0001 174.81% <.0001 –30.35% <.0001 15.94% No Surgical Therapy includes no fi rst course therapy, chemotherapy only and chemoradiation only
588 RESULTS: The overall incidence of adenocarcinoma of the esophagus and the gastric cardia increased from 13.4 per Does the Incidence of Adenocarcinoma of the million in 1973 to 51.4 per million in 2009, a nearly 400% Esophagus and Gastric Cardia Continue to Rise in increase. Jointpoint analysis demonstrated that the yearly the 21st Century? increase in incidence has slowed somewhat from 1.27 per Attila Dubecz1, Norbert Solymosi2, Michael Schweigert1, million before 1987 to 0.97 between 1987–1997 and 0.65 Rudolf J. Stadlhuber1, Hubert J. Stein1, Jeffrey H. Peters3 after 1997. Stage-specifi c analyses suggests, that incidence 1. Surgery, Klinikum Nurnberg, Nuremberg, Germany; of early stages has actually declined after 2001 with a yearly 2. Faculty of Veterinary Science, Szent Istvan University, Budapest, decrease of 0.22. The percentage of patients diagnosed with Hungary; 3. Division of Thoracic and Foregut Surgery, Department early cancer declined after 2000 and remained under 2.5% through the study period. Regression analysis showed a of Surgery, University of Rochester School of Medicine and substantially higher correlation of incidence of adenocarci- Dentistry, Rochester, NY noma of the esophagus and the gastric cardia with popula- BACKGROUND: The rising incidence and histologic tion (r2 = 0.95) than with time (r2 = 0.65). change to adenocarcinoma in esophageal cancer over CONCLUSION: The incidence of esophageal adenocarci- the past four decades has been among the most dramatic noma continues to rise in the 21st century in the United changes ever observed in human cancer. Recent reports States. A signifi cant linear correlation of incidence with have suggested that its increasing incidence may have pla- total population was found. teaued over the past decade. Our aim was to examine the latest trends in esophageal adenocarcinoma incidence and analyze its correlation with time and population density. PATIENTS AND METHODS: We used the Surveillance Epidemiology and End Results (SEER) database of the National Cancer Institute to identify all patients with ade- nocarcinoma of the esophagus and gastric cardia between 1973 and 2009. Both overall and stage specifi c trends in incidence were analyzed using joinpoint regression. The correlation of incidence with time and total population within the geographic areas covered by SEER was analyzed by linear regression.
24 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
589 590 Reoperative Intervention in Patients with Mesh at Infl uence of Ethnicity on the Effi cacy and Utilization Hiatus Is Associated with High Morbidity and High of Bariatric Surgery in the United States Incidence of Esophageal Resection: Single Center Ranjan Sudan1, Deborah Winegar2, Steven Thomas3, Experience John M. Morton4 Kalyana C. Nandipati, Maria Bye, Se Ryung Yamamoto, 1. Department of Surgery, Duke University Medical Center, Pradeep K. Pallati, Tommy H. Lee, sumeet K. mittal Durham, NC; 2. Department of Clinical Affairs, LipoScience, Creighton University, Omaha, NE Raleigh, NC; 3. Department of Biostatistics and Bioinformatics, BACKGROUND: Increasing use of mesh for hiatus repair Duke University, Durham, NC; 4. Department of Surgery, Stanford during anti-refl ux surgery has been reported. Re-operative University, Palo Alto, CA intervention with previously placed mesh is technically BACKGROUND: In the US more blacks than whites are Abstracts more challenging. The aim of this study is to present a severely obese (26% vs. 15%) and suffer from hypertension Monday single Center experience with reoperative intervention in (40% vs. 27%). Prior studies examining the infl uence of patients with previous mesh at hiatus and outcomes in this race on bariatric surgery have been from single-institution subset of patients. or small cohorts. This is the fi rst study to examine dispari- METHODS: After Institutional review board approval pro- ties in national patterns of utilization and the infl uence of spectively maintained database was retrospectively queried ethnicity on outcomes after Roux-en-Y gastric bypass sur- to identify patients who underwent re-operative interven- gery (RYGB) from the large multi-institutional prospective tion between 2003 to 2012 and had mesh placed at a previ- database for the American Society for Bariatric and Meta- ous hiatal hernia procedure. Patient charts were reviewed bolic Surgery. and variables collected included demographics, indications, METHODS: All research-consented white, black or His- operative details (initial and reoperative) and postoperative panic patients undergoing RYGB between 6/2007 and complications. 10/2011 and eligible for one year of follow-up were RESULTS: Twenty-six patients met inclusion criteria included. Other races were excluded. Descriptive statistics and form the cohort for the study. There were 14 females were used for demographic information. Multivariate logis- with a mean age of 58.3 + 29.2 years. Synthetic mesh was tic and normal regression models examined relationships placed in 15 (58%) patients, while the remaining 11 had between race and outcomes, controlling for age, gender, bio-prosthetic mesh. Mean duration of re-operative inter- baseline BMI and comorbid conditions. Races were com- vention since the last surgery was 22 (1–52) months. Dys- pared using a t-test for continuous variables and Pearson phagia (57%) was the most common presentation while 4 chi-square test for categorical variables. Reported p-values patients had mesh erosion. Recurrent hiatus hernia (2 to 7 were adjusted for the false discovery rate (FDR) to control cm) was noted in 16 (62%) patients. Nine patients (35%) for multiple testing. underwent redo fundoplication, 8 (31%) were converted RESULTS: The racial distribution of the 135,262 study to Roux en Y gastrojejunostomy, 3 (12%) underwent dis- patients was 79% white, 12% black, and 9% Hispanic. tal esophagectomy with esophago-jejunostomy, 5 (19%) Among the blacks undergoing RYGB only 15% were male had subtotal esophagectomy with gastric pull-up and one whereas 22% of the white and Hispanic patients were patient underwent substernal gastric pull-up for esophageal men. Compared to whites, blacks were younger (42.8 ± bypass with interval esophagectomy. The mean operative 10.6 vs. 46.3 ± 11.6 yrs.), heavier BMI (50.2 ± 9.2 vs. 47.6 time was 250 + 70.1 min, the median blood loss was 150 ± 8.0 kg/m2 and more often hypertensive (58% vs. 53%) ml (50–1650 ml). Reoperative intervention was performed at baseline. Although mortality rates within 30 days were with laparoscopic approach in 50% (13/26) of the patients, equivalent for all races (0.23–0.26%), serious adverse events laparoscopy converted to laparotomy in 12% (3/26) of the were higher for blacks (3.65%) versus whites (3.19%) and patients, laparotomy was performed in 34% (9/26) and tho- Hispanics (2.01%). At 1 year, mean BMI decreased markedly racotomy was performed in 1 patient. There was no post- to 35.0 ± 7.5 for blacks, 31.6 ± 6.73 for whites and 32.6 ± operative mortality. Major complications were noted in 6 7.0 kg/m2 for Hispanics. However, the percentage decrease patients. Mean ICU stay was 6 days and hospital stay was in BMI from baseline was lower for blacks (-30%) compared 14 days. to whites (-34%) and Hispanics (–32%). Similarly, hyper- CONCLUSION: Reoperative intervention in patients with tension decreased from 57% to 37% (blacks), 53% to 27% mesh at hiatus is associated with a high (>35%) need for (whites) and 42% to 29% (Hispanics) but, the percentage esophageal resection. More than half the patients also had decline was less for blacks (-35%) versus whites (–49%) and a recurrent hiatal hernia. Caution is advised in liberal use of Hispanics (–50%). Resolution of diabetes also demonstrated mesh for hiatoplasty. a similar pattern for blacks (59%) versus whites (65%) and Hispanics (61%). Racial differences in outcomes for weight loss and major comorbid conditions persisted after adjust- ment for baseline characteristics (p values and odds ratios are in Table1).
25 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Effect of Ehnicity on Outcomes at 1 Year
Black vs. Hispanic Black vs. White Hispanic vs. White Outcomes P-value PPM (95% CI) P-value PPM (95% CI P-value PPM (95% CI BMI 1 <.0001 1.23 (1.04, 1.42) <.0001 1.77 (1.65, 1.89) <.0001 0.54 (0.38, 0.71) Excess Body Weight (kg) <.0001 2.00 (1.59, 2.42) 0.0155 0.50 (0.14, 0.87) <.0001 –1.50 (–1.73, –1.28) Outcomes P-value OR (95% CI) P-value OR (95% CI) P-value OR (95% CI) GERD 0.7875 1.05 (0.91, 1.20) 0.0095 0.87 (0.79, 0.96) 0.0056 0.83 (0.74, 0.93) Diabetes 0.8914 0.98 (0.82, 1.17) 0.0258 1.15 (1.03, 1.28) 0.0661 1.17 (1.02, 1.35) Hypertension <.0001 1.66 (1.44, 1.90) <.0001 1.69 (1.57, 1.83) 0.7339 1.02 (0.90, 1.16) Obstructive Sleep Apnea Syndrome 0.8914 1.02 (0.86, 1.21) 0.9332 .00 (0.89, 1.11) 0.7339 0.97 (0.84, 1.13)
CONCLUSIONS: Race exerts a signifi cant infl uence on out- clinical database to supplement perioperative data. Read- comes after RYGB. Despite lower effi cacy in blacks, overall missions unrelated to the index admission were omitted. benefi ts from RYGB were signifi cant. Given the higher prev- RESULTS: We identifi ed 173 (16%) patients who required alence of obesity in blacks, bariatric surgery is underutilized readmission after PD within the study period. The readmis- by this group (particularly males). Higher baseline BMI and sion rate was higher in the 2nd half of the decade when com- more frequent hypertension in blacks indicate need for ear- pared to the 1st half (18.6% vs 12.3%, p = 0.003), despite a lier surgical intervention. stable 7 day median length of stay. Readmitted patients were Outcomes are fi t with a generalized linear model control- analyzed against those without readmissions after PD. The ling, sex, age, current tobacco use, prior medical history, demographics and tumor pathology of both groups did not and current BMI unless noted. Reported p-values were differ signifi cantly. In the multivariate logistic regression adjusted with FDR. 1 The covariate BMI was replaced with analysis, preoperative albumin ≤3.5 (19% vs 11%, OR 1.6, p baseline BMI PPM: Predicted population marginal mean = 0.046), multi-visceral resection at time of PD (3% vs 0.6%, difference. OR 11.9, p = 0.031) and a length of initial hospital stay >7 days (59% vs 43%, OR 1.6, p = 0.043) were independently 591 associated with readmissions. There were no postoperative biochemical variables that were predictive of readmissions. Understanding Hospital Readmissions After Fifty percent (n = 87) of the readmissions occurred within 7 Pancreaticoduodenectomy: Can We Prevent Them? days from initial operative discharge. The reasons for early A 10-Year Contemporary Experience with 1173 Patients and late (>7 days) readmissions differed; ileus, delayed gas- at the Massachusetts General Hospital tric emptying and pneumonia were more common in early Zhi Ven Fong, Klaus Sahora, Seefeld J. Kimberly, readmissions, whereas wound infection, failure to thrive Cristina R. Ferrone, Sarah P. Thayer, Andrew L. Warshaw, and intraabdominal hemorrhage were associated with late Keith D. Lillemoe, Matthew M. Hutter, Carlos Fernandez-del readmissions. The incidence of readmissions due to pan- creatic fi stulas and intraabdominal abscesses were equally Castillo distributed between both time frames. General Surgery, Massachusetts General Hospital, Boston, MA CONCLUSION: The frequency of readmission after PD is INTRODUCTION: The morbidity and mortality of pancre- 16%, and has been on the uptrend over the last decade. aticoduodenectomy (PD) have signifi cantly decreased over Poor preoperative nutritional status and the complexity of the past decades to the point that they are no longer the initial resection were independently associated with hospi- sole indicators of quality and safety. In recent times, hospi- tal readmissions after PD. Further efforts should be centered tal readmission is increasingly used as a quality metric for on preventing early readmissions, which constitute half of surgical performance, and has direct implications on health all readmissions. care costs. We sought to delineate the natural history and predictive factors of readmissions after PD. METHODS: The clinicopathologic and long-term follow- up data of 1173 consecutive patients who underwent PD between August 2002 and August 2012 at a single institu- tion were reviewed. The NSQIP database was linked with our
26 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
9:45 AM – 11:00 AM 300 QUICK SHOTS SESSION I
592 594 Does Intramesorectal Proctectomy Affect Overall Adenovirus-Mediated Interferon Therapy Sensitized Complication Rates Compared to Standard Total Chemotherapy and Radiation for Pancreatic Cancer in Mesorectal Excision in Patients with Ulcerative Colitis? Vitro and in Vivo Models Abstracts Caitlin W. Hicks1,2, Richard A. Hodin1, Lieba R. Savitt1, Joohee Han1, Yoshiaki Miura1, Leonard Armstrong1, Monday Liliana Bordeianou1 Ryan M. Ryan M Shanley2, Xianghua Luo2, Eric H. Jensen1, 1. Department of Surgery, Massachusetts General Hospital, Edward W. Greeno3, Selwyn M. Vickers1, Masato Yamamoto1, Boston, MA; 2. Department of Surgery, The Johns Hopkins Julia Davydova1 Hospital, Baltimore, MD 1. Surgery, University of Minnesota, Minneapolis, MN; PURPOSE: In patients with ulcerative colitis (UC), intrame- 2. Division of Biostatistics, University of Minnesota, Minneapolis, sorectal proctectomy involves close dissection along the MN; 3. Medicine – Hematology, Oncology and Transplantation, rectal wall with concomitant rectal eversion (IMP/RE). In University of Minnesota, Minneapolis, MN contrast, standard total mesorectal excision (TME) involves Interferon- (IFN) in conjunction with chemoradiother- close dissection along the pelvic wall with an intraabdom- apy has emerged as a promising treatment for pancreatic inal stapled rectal transection above the levator muscles. adenocarcinoma. However, despite encouraging survival Our goal was to compare surgical outcomes among UC results (e.g. a 5-year survival rate of 55% in a phase II trial patients following IMP versus TME (Figure). by the Virginia Mason study group evaluating adjuvant METHODS: All patients undergoing IPAA surgery for active chemotherapy, immunotherapy and external-beam radia- UC at a tertiary referral hospital over a 10.5-year period tion for resected PDAC), utilization of this regimen has (09/2000-04/2011) were included in analysis. Univari- been impeded by systemic toxicity of IFN. ate analysis (T-tests and chi square tests) and step-wise fi t To circumvent these problems, we engineered a novel infec- regression modeling were used to compare complications tivity-enhanced oncolytic adenoviral vectors for high-level rates among patients undergoing IMP vs. TME procedures. targeted IFN expression (Ad-IFN). We hypothesized that RESULTS: Of 201 patients identifi ed for inclusion in a new therapeutic modality combining an Ad-IFN with the study, 119 (59%) underwent IMP/RE. Age, race, gen- chemoradiation would overcome the major drawbacks of der, smoking status, disease comorbidity, steroid or other IFN-based regimens. The adenovirus-mediated tumor-selec- immunomodulator use, surgical urgency, severity of disease tive expression of IFN will eliminate systemic toxicity of on pathology, and surgical staging were similar between cytokine, while massive IFN expression via replication- groups (p = ns). IMP/RE patients underwent fewer laparo- competent vector will yield an extended response. In this scopic procedures (2% vs. 37%, p < 0.0001) based on sur- study, we combined this vector with chemo- and radiother- geon preference. On univariate analysis, IMP/RE patients apy and analyzed its therapeutic ability in vitro and in vivo had fewer total perioperative complications (0.9 ± 0.1 vs. models. 1.4 ± 0.1, p = 0.02), but no differences in abdominal sepsis, The in vitro assays revealed that combination of Ad-IFN post-operative length-of stay, or hospital readmissions (p = with chemotherapeutics (5-FU, gemcitabine, cisplatin) and ns). However, in a step-wise regression model accounting X-ray radiation killed human and hamster pancreatic can- for age, co-morbidities, disease severity, pre-operative medi- cer cells signifi cantly better than either of the single treat- cations, operative technique, and follow-up time (mean 5.5 ments. Furthermore, we established pancreatic tumors in ± 0.2 years), both anastamotic leak rate [OR –0.56 (95% CI immunocompetent hamsters and discovered that combi- 0.33, 0.99); p = 0.04] and overall post-operative combined nation of Ad-IFN with either 5FU or radiation (8 Gy and pouch-related and infectious complications were lower in 20 Gy were tested) resulted in remarkable tumor shrinkage the IMP/RE group (2.0 ± 0.2 vs. 2.6 ± 0.2, p = 0.03). and was signifi cantly superior to radiation and 5-FU alone CONCLUSIONS: IMP/RE appears to be associated with or both of these combined. The triple-therapy (Ad-IFN+X- fewer overall post-operative complications than TME in ray+5-FU) outperformed all treatment groups. The evalua- patients with UC. This may be a refl ection of smaller free tion of the survival rate also showed statistically signifi cant space within the pelvis and/or the ability to invert the rec- improvement in groups treated with dual (Ad-IFN+X-ray) tal stump staple line during concomitant rectal eversion. and triple (Ad-IFN+X-ray+5-FU) therapies versus conven- However, further studies on functional and long-term out- tional approaches (radiation or/and 5FU). comes are needed.
27 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Our results support the impact of Ad-mediated IFN to 596 sensitize chemotherapy and radiation for pancreatic can- cer. This strategy may expand clinical use of the robust and Laparascopic Pyloroplasty: A Promising Treatment for promising IFN-based multimodal therapy to meet the press- Refractory Gastroparesis ing continued need for PDAC treatment. Katie Farah, Elie Aoun, Elizabeth Dovec, Sheri A. Mancini The Western Pennsylvania Allegheny Health System, 595 Pittsburgh, PA Planned Delay of Contrast Swallow Study and Oral INTRODUCTION: Gastroparesis is a chronic debilitating Intake After Minimally Invasive Esophagectomy digestive disorder characterized by a delay in gastric emp- tying. Nausea, vomiting, abdominal pain, and bloating Reduces the Anastomotic Leak Rate and Hospital are characteristic of the disease and signifi cantly affect the Length of Stay patient’s quality of life. Medical therapy is limited to very John S. Bolton, William C. Conway few agents most of which are limited by their side effect Surgery, Ochsner Clinic, New Orleans, LA profi le. Various surgical techniques for the treatment of INTRODUCTION: With the increasing use of minimally gastroparesis are invasive, not as effective, and have high invasive esophagectomy (MIE) and early hospital discharge, complication rates. We describe our experience with lapa- the timing and role of postoperative contrast swallow study roscopic pyloroplasty as a therapeutic alternative for the (SS) has become increasingly problematic. We systemati- treatment of gastroparesis. cally evaluated a policy of delayed SS and oral intake after PATIENTS AND METHODS: 22 patients with refractory MIE until the second postoperative week, approximately gastroparesis were enrolled in this prospective study and one week after hospital discharge. underwent minimally invasive pyloroplasty. Pre- and post- METHODS: Between Sept 2007 and October 2012, 143 operative gastric emptying study (GES), Gastroparesis Car- consecutive patients undergoing MIE were evaluated for dinal Symptom Index (GCSI©2003 Johnson & Johnson), inclusion in the study. Patients with obvious clinical or and complications were recorded. radiographic leak by d 7 were excluded from the study. Our RESULTS: 21 patients underwent laparoscopic pyloro- study group consisted of 30 patients (Late Eaters) whose SS plasty and one patient was converted to open laparotomy. and po intake were intentionally delayed (20 patients) until The mean duration of the procedure was 106 ± 25 minutes. a week after hospital discharge or were delayed by postop The average length of stay was 2.9 ± 1.2 days. There were events (10 patients) which made early SS and institution or no major complications linked to the surgical procedure. At oral feeds impractical. The study group was compared to a one month follow up, 16/22 (72.7%) patients noted over- control group who were deemed ready to have SS done and all improvement in their symptoms. At six months, 19/22 po intake started while in the hospital on postop day 5-7. (86.3%) patients showed improvement in their GES. The Primary endpoints studied were the anastomotic leak rate GES results normalized in 18/22 (81.8%) patients at one (ALR) and the hospital length of stay (LOS). year. The T1/2 decreased from 392 to 110 minutes (p = RESULTS: Data are shown in Table 1. 0.001). Signifi cant improvements were noted on all items of the GCSI score at one year post-operative visit (Table 1). CONCLUSION: After MIE, a policy of early hospital dis- charge nil per os, delaying SS and resumption of oral intake Gastroparesis Cardinal Symptom Index (GCSI) Pre and to about two weeks postoperatively, signifi cantly reduces Post-Op Data hospital LOS and anastomotic leak rate. Early institution or oral feeds after MIE appears to increase ALR. Symptom Pre-Op Score Post-Op Score p- Value Nausea 4.45 1.41 <0.001 Table 1 Retching 3.00 0.55 <0.001 Postop Day on which Vomiting 3.14 0.64 <0.001 SS Done and Oral Intake Anastomotic Hospital Length Fullness 4.73 2.32 <0.001 Begun (Median) Leak Rate of Stay Unable to fi nish a meal 4.41 1.55 <0.001 Early Eaters D6 22% (20/91) D8 Feeling full 4.68 2.05 <0.001 (n = 91) Loss of appetite 4.41 1.36 <0.001 Late Eaters D12 3% (1/30) D6 Bloating 4.73 2.05 <0.001 (n = 30) Stomach visibly bigger 3.73 1.45 <0.001 p < 0.05 p < 0.05 p < 0.05 CONCLUSION: Laparoscopic pyloroplasty is an effective, minimally invasive, and safe surgical alternative in patients with refractory gastroparesis who either fail or are intoler- ant of medical therapy. While our results are very promis- ing, larger studies are needed to further evaluate its role.
28 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
597 RESULTS: Between 2006 and 2010, 316 pancreatectomy cases were reported to NSQIP by our institution. Two hun- Limitations of NSQIP in Reporting Complications for dred and forty-nine were reviewed in detail, among them Patients Undergoing Pancreactectomy: Underscoring 145 (58.2%) Whipples, 19 (7.6%) total pancreatectomies, the Need for a Pancreas-Specifi c Module 65 (26.1%) distal pancreatectomies, and 15 (6.0%) central Irene Epelboym, Irmina Gawlas, James A. Lee, Beth Schrope, or partial resections. Median age was 65.7, males comprised John A. Chabot, John D. Allendorf 41.5% of the group, and 74.3% of patients were Cauca- Surgery, Columbia University Medical Center, New York, NY sian. Overall rate of complications reported by NSQIP was 44.0%, compared with 55.0% in our review, however dis- BACKGROUND: Administrative databases are used with cordance was observed in 73 (29.3%) cases (p < 0.001), increased frequency for reporting hospital-specifi c and including 24 cases of reporting a complication where nationwide trends and outcomes after various surgical pro- there was not one, and 49 cases of missed complication. cedures in order to improve quality of surgical care. NSQIP
Most frequently reported event was postoperative bleed- Abstracts is a risk-adjusted case-weighted complication tracking ini- ing requiring transfusion (22.7%), however true incidence Monday tiative that reports 30-day outcomes from more than 400 of postoperative bleeding was actually 19.0%, with NSQIP academic and community institutions in the United States missing 27 (57.5%) and incorrectly reporting 36 (64.3%), p alone. However, the accuracy of reported events specifi c to < 0.001. Four procedures unrelated to the index operation pancreatic surgery has never been reported in depth. were recorded as reoperation events. While a pancreas-spe- METHODS: We retrospectively reviewed a randomly cifi c module does not yet exist, NSQIP reports a 7.6% rate selected subset of patients, the information on whose post- of organ-space surgical site infections; when compared with operative course was originally reported through NSQIP. our institutional rate of Grade B and C postoperative fi stula Preoperative characteristics, operative data, and postopera- (8.8%), we observed discordance 6% of the time, p<0.001. tive events were recorded after review of electronic medical Delayed gastric emptying, a common post-pancreatectomy records including physician and nursing notes, operative morbidity, was not captured at all. Additionally, there were room records and anesthesiologist reports. We compared signifi cant inaccuracies in reporting urinary tract infec- categorical variables using chi-square or Fischer’s exact test tions, postoperative pneumonia, wound complications, and continuous variables using Student’s t-test. and postoperative sepsis, with discordance rates of 4.4%, 3.2%, 3.6%, and 6.8%, respectively. CONCLUSIONS: NSQIP data is an important and valu- able tool for evaluating quality of surgical care, however pancreatectomy-specifi c postoperative events are often mis- classifi ed, underscoring the need for a hepatopancreatobil- iary-specifi c module to better capture key outcomes in this complex and unique patient population.
29 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
598 respectively compared to 72.5, 60.1, and 80.5 for M-Ab (p < 0.001). The Chi-Ab demonstrated improved specifi city as Chimeric Antibodies to CEA Improve Detection of evident by the lower signal intensity in normal human tis- Human Colon Cancer in Orthotopic Mouse Models sue samples compared to M-Ab (normal colon tissue: 4.3 vs Cristina A. Metildi1, Sharmeela Kaushal1, George A. Luiken2, 5.4; normal pancreas tissue: 1.5 vs 2.7) indicating decreased Mark A. Talamini1, Robert M. Hoffman1,3, Michael Bouvet1 binding of Chi-Ab. The chimeric CEA antibody was also 1. Surgery, University of California San Diego, La Jolla, CA; accurate in labeling human colon cancer in mouse xeno- 2. OncoFluor, San Diego, CA; 3. AntiCancer, Inc., San Diego, CA grafts enabling improved detection of tumor margins for more effective resection, increasing the R0 resection rate Positive surgical margins after colorectal cancer surgery from 86% to 96%. are strong predictors for higher local recurrence rates and poor overall survival. Currently, no real-time, reliable detec- The chimeric form of our fl uorophore-conjugated CEA anti- tion assays for positive surgical margins at the time of sur- body has more effective labeling of human CEA-expressing gery exist. We have previously shown improved detection cancer in tissue arrays and in our xenograft mouse models and resection of primary pancreatic cancer with a mouse- of human colon cancer. The improved sensitivity and speci- derived fl uorophore-conjugated antibody against the tumor fi city of the chimeric fl uorophore-conjugated antibody is antigen CEA in open laparotomies in mouse models. The clinically translatable. aim of this study was to demonstrate improved sensitivity The top left panel is an illustration of the steps required and specifi city of a new chimerized antibody against CEA to convert a mouse antibody to a human antibody. Before in detection of CEA-expressing colon cancer for improved fully humanizing the antibody, we tested the chimeric anti- resection in xenograft mouse models. body on normal tissue and CEA-expressing colon tissue Mouse models of human colon cancer were established with samples, comparing its labeling sensitivity and specifi city fragments of a CEA-expressing patient colon tumor. Two to the mouse antibody. The bottom left panel shows that a to four weeks after implantation, mice were randomized brighter signal is obtained by labeling the tumor with the to fl uorescence-guided surgery (FGS) or bright-fi eld surgery chimeric antibody, as compared to the mouse antibody. (BS). Mice in the FGS group received tail vein injections of Also, there is less labeling on normal tissue with the chi- the chimeric anti-CEA-Alexa-488 antibody 24 hours prior meric antibody. The two panels on the right illustrate the to resection. Pre- and postoperative images were obtained improved detection of CEA-expressing colon tumor in our to assess for completeness of resection. Mice were then fol- mouse models with the chimeric antibody. lowed for 6 months postoperatively to assess for recurrence and overall survival. At termination, all tumor lesions were 599 harvested and evaluated histologically. The chimeric anti- body was also tested on frozen tumor and normal tissue Endoscopic Submucosal Dissection for Early Neoplasia arrays comparing it to the mouse antibody. of the Foregut: A North American Perspective Jonathan Cools-Lartigue, Lorenzo E. Ferri Surgery, McGill University, Montreal, QC, Canada INTRODUCTION: Endoscopic resection as an organ spar- ing option in the management of early cancers of the fore- gut is becoming increasingly accepted. In North America, endoscopic mucosal resection (EMR) is the technique pri- marily employed. However lesions greater than 1 cm fre- quently require piecemeal resection with EMR, resulting in a high rate of local recurrence. Endoscopic Submucosal Dissection (ESD) allows for the en-bloc removal of larger tumors, however there is very limited data of this procedure in North America. We present our experience as one of the only centers in North America routinely performing ESD for neoplasia of the foregut. METHODS: A prospectively maintained database of all patients with early neoplasia of the foregut managed in a busy North American centre was reviewed for patients undergoing ESD. Patient characteristics, endoscopic/post- The chimeric antibody directed against CEA demonstrated endoscopy outcomes, pathologic features, and oncologic improved sensitivity and specifi city in labeling CEA- outcomes were captured. Data presented as median(range). expressing tumor compared to the mouse antibody. The fl uorophore conjugation effi ciency to the chimeric CEA RESULTS: From 5/2009–11/2012 twenty patients (74 antibody (Chi-Ab) was 2-fold higher than the mouse CEA (38–85)yrs: 16M/4F) underwent ESD for neoplasia in the antibody (M-Ab). On frozen tumor tissue arrays, the signal gastric antrum (10), body (2), cardia (6), or esophagus (2). intensity of the Chi-Ab was signifi cantly brighter compared General anesthesia was performed in the majority (19/20) to the M-Ab indicating improved binding to tumor tis- and endoscopy time was 75 (30–330) minutes. The fi rst sue. For colon, pancreas and lung tumor samples, the sig- 5 cases were longer than the last 15 (235(132–330) vs 75 nal intensity with the Chi-Ab was 94.1, 85.3, and 106.1, (30–240) minutes). Median lesion size was 2.25 (0.6-5) cm
30 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
and most underwent en-bloc resection (18/20). Perforation CONCLUSION: In this study, CAW concentrated mainly in occurred in 3 patients, which was repaired by endoscopy (2) surgical specimen extraction sites, and port-related compli- or by laparoscopy (1). Bleeding requiring repeat endoscopy cations were uncommon. occurred in 1 pt. Length of stay was 2 (1–7) days, there were no re-admissions. Final pathology revealed invasive cancer 601 in 13 (ADC = 9, SCC = 3, NET = 1: T1a = 9, T1b = 3, T2 = 1)) and adenoma with dysplasia in 7. Complete resection (R0) Intraoperative Testing Following EEA Stapling, and the was achieved in 18/20, the 2 incomplete resection cases Implications for Postoperative Anastomotic Leaks underwent laparoscopic resection. There is no evidence of Jake G. Prigoff, Adam C. Fields, Sapna Rustagi, recurrence at 5 (1–41) months follow-up. Celia M. Divino CONCLUSIONS: Although technically challenging, endo- Mount Sinai School of Medicine, New York, NY scopic submucosal dissection for neoplastic lesions of the INTRODUCTION: Anastomotic leaks will occur in roughly Abstracts upper GI tract is effective, feasible, and can be safely applied 3-10% of surgical procedures that include an end-to-end Monday in a North American setting. anastomosis (EEA). To prevent leaks, surgeons evaluate the doughnuts removed from the autosutures and perform 600 maneuvers to assess the viability of the anastomosis includ- ing air-leak tests, beta-dine leak tests, and rigid sigmoidos- How Frequent Are Complications of the Abdominal copies. This study will evaluate the effi cacy of these tests to Wall After Laparoscopic Colorectal Surgery? determine if they are valid. Alejandro J. Zarate, Camila Estay, Udo Kronberg, METHODS: A cohort of 44 surgical patients (ages 18–92; Claudio Wainstein, Francisco López-Köstner 25 male) who presented to The Mount Sinai Medical Center Colorectal Unit, Clinica las Condes, Santiago, Chile between 2005 and 2012 with anastomotic leaks were com- BACKGROUND: At present, advantages of laparoscopic pared to a case control group of 86 patients without postop- colorectal surgery (LCRS) over a conventional approach erative leaks. Only anastomoses distal to the splenic fl exure have been demonstrated, mostly due to a reduced morbid- and created by employing an EEA stapler were included, ity and hospital stay. This has encouraged the development as those anastomoses are testable via beta-dine, air-leak, of new minimally invasive techniques like single incision rigid sigmoidoscopy, and doughnut observation. Air-tests laparoscopic surgery, which are thought to have further and beta-dine tests for left hemicolectomies were excluded. benefi ts over LCRS especially due to reduction of the num- Patient demographics, surgical procedures, reoperations, ber of incisions on the abdominal wall. However, there is pathologic fi ndings, and outcomes of the intraoperative only few information available regarding complications on tests were reviewed. All statistical analyses were performed the abdominal wall (CAW) after LCRS. using SPSS v20 software. PURPOSE: To analyze CAW in patients undergoing LCRS. Table 1 METHODS: Patients were selected from our prospectively Post-Operative Leak No Post-Operative Leak maintained database of LCRS, operated between July 2007 Test Result (Number Of Patients) (Number Of Patients) and July 2012. Following a standardized protocol, the surgi- Positive Air-Leak Positive 1 0 1 0 2 1 2 0 cal specimen was extracted using an Alexis retractor to pro- Beta-dine Incomplete tect surgical site, and incisions of 10 mm were sutured both Doughnuts Positive Rigid aponeurosis and skin, while 5 mm incisions only had skin Sigmoidoscopy closure. Patients with anastomosis leak and/or deep surgical Negative Air-Leak 25 9 43 15 66 29 81 13 site infection were included. Information of demographic, Negative Beta-dine operative and follow-up data was analyzed using chi square Complete Doughnuts and t of Student tests. Negative Rigid Sigmoidoscopy RESULTS: In 455 patients that underwent LCRS during the above-mentioned period, 16 (3.7%) had ≥ 1 CAW. Eight patients (1.9%) had an incisional surgical site infection Table 2 (SSI), six (1.4%) had an abdominal wall hematoma, three (0.7%) presented an incisional hernia, and two (0.5%) had Positive Negative a covered evisceration during early follow-up. Frequency of Specifi city Predictive Predictive CAW was similar in patients operated secondary to diver- Test Sensitivity (%) (%) Value (%) Value (%) ticular disease and those with malignancy. Every SSI, hema- Air-Leak 3.85 97.06 33.33 72.53 toma and incisional hernia developed in surgical specimen Beta-dine 0.0 96.67 0.0 76.32 extraction site, with no signifi cant differences between EEA Doughnut 2.27 97.59 33.33 65.32 periumbilical and suprapubic incision (SSI p = 0.15; Hema- Integrity toma p = 0.990; Incisional hernia p = 0.08). Two out of three Rigid 0.0 100 - 46.43 patients with incisional hernia had a prior SSI. When ana- Sigmoidoscopy lyzing morbidity associated with ports, there was 1 (0.08%) covered evisceration in 1,180 incisions of 5 mm ports, and another one in the 890 incisions of 10 mm ports (0.11%).
31 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
RESULTS: The procedures included low anterior resection histopathological diagnosis. Available published data (50.0%), left hemicolectomy (15.9%), sigmoid resection seems to indicate that there is an incongruency between (29.5%), and ilio-anal pull through (13.6%). Procedures of sonographic diagnosis and the actual surgical pathology of the control group were equivalent ±7%. Beta-dine and air- gastric subepithelial lesions with a reported accuracy that leak tests, as well as doughnut observation, were more likely ranges between 20% and 84%. in the control group, and rigid sigmoidoscopies less likely. AIM: To compare the diagnosis suggested by sonographic Table 1 displays the true positives, true negatives, false posi- evaluation of various subepithelial lesions throughout the tives, and false negatives. Table 2 has their corresponding GI tract with the histological description of tissue speci- sensitivities and specifi cities. mens obtained by endoscopic mucosal resection (EMR). CONCLUSIONS: A combination of EEA doughnut integ- METHODS: Patients who underwent endoscopic mucosal rity and an air-leak test show the highest predictive value. resection for a suspected subepithelial lesion were retro- However, the high rates of false negatives and correspond- spectively identifi ed from 2009 to 2012 at two university ing low sensitivities of these tests show their inability to hospitals. The diagnosis made at the time of the endoscopic correctly identify the patients who will ultimately have ultrasound prior to EMR was compared with the diagnosis anastomotic leaks. made from the specimen obtained after EMR. 602 RESULTS: Twenty-three patients (12M, 11F) with a mean age of 58yrs (range 36–82 yrs) were identifi ed who under- The Pathologic Diagnosis of Subepithelial Lesions of went both EUS and EMR for subepithelial lesions of the GI the GI Tract Based Solely on Sonographic Features tract. The location of the lesions were: esophagus 1, stom- Correlates Poorly with Histopathological Diagnosis ach 13, duodenum 6, rectum 3. EUS based diagnosis was Obtained by Mucosal Resection only 61% accurate when compared with the histopatho- Andrew Jatskiv1, Gabriel H. Lee1, Laura Rosenkranz1, logical diagnosis obtained via EMR. (Table 1) Sandeep Patel1, Kenneth Sirinek2 CONCLUSION: Except for the rectum, there is poor cor- 1. Medicine, UTHSCSA, San Antonio, TX; 2. Surgery, UTHSCSA, relation between the sonographic and histopathological San Antonio, TX diagnosis of subepithelial lesions of the GI Tract. Based on these fi ndings, clinical decision making protocols that rely BACKGROUND: Subepithelial lesions are incidentally solely on endoscopic sonographic fi ndings for pathologic found during routine endoscopy of the gastrointestinal diagnosis are fl awed and may lead to unnecessary surgical (GI) tract. Endoscopic ultrasonography (EUS) has become procedures. All subepithelial lesions of the GI Tract should the preferred non-invasive technique in evaluating these undergo endoscopic mucosal resection for a defi nitive his- subepithelial lesions. It is assumed that it provides valuable topathological diagnosis to direct appropriate treatment. information such as size, echotexture, layer of origin and features of invasion which may correlate with a particular
EUS Diagnosis Histologic Diagnosis Age Sex Esophagus (1) Duplication Cyst Granular Cell Tumor 36 F
Stomach (13) Carcinoid Carcinoid 48 F Carcinoid Carcinoid 82 F Carcinoid Carcinoid 59 M Carcinoid Carcinoid 60 M Carcinoid Oxyntic Gastric Mucosa 46 F Carcinoid Gastric Adenoma 55 M Granular Cell Tumor Granular Cell Tumor 61 M Granular Cell Tumor Pancreatic Heterotopia 54 M Granular Cell Tumor Lipoma 81 F Infl ammatory nodule Infl ammatory nodule 63 F Pancreatic Heterotopia GIST 63 F Pancreatic Heterotopia Pancreatic Heterotopia 44 F Pancreatic heterotopia Pancreatic Heterotopia 71 M
Duodenum (6) Pancreatic Heterotopia Carcinoid 44 F Carcinoid Carcinoid 75 M GIST Leiomyoma 58 M Infl ammatory nodule Ectopic Salivary gland tissue 52 F Infl ammatory nodule Infl ammatory nodule 54 M Carcinoid Carcinoid 51 M
Rectum (3) Infl ammatory nodule Infl ammatory nodule 73 M Carcinoid Carcinoid 47 F GIST GIST 61 M
32 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
603 CRC and 249 (11.6%) had early-onset CRC. 1,447 patients (67.4%) were Caucasian, 111 (5.2%) African American, and Characteristics of Early-Onset Colorectal Cancer 589 (25.4%) Asian. Tumor was located in the appendix in Kidist Yimam, Richard E. Shaw, Christine Wong, 21 (1%), cecum in 275 (12.8%), ascending colon in 207 Joyce Louie, Edward W. Holt, Michael S. Verhille, (9.6%), hepatic fl exure in 102 (4.8%), transverse colon in Taehyun P. Chung, Michael Abel 134 (6.2%), splenic fl exure in 79 (3.7%), descending colon California Pacifi c Medical Center, San Francisco, CA in 90(4.2%), sigmoid colon in 431 (20.1%), rectosigmoid junction in 166 (7.1%), and rectum in 592 (27.7%). At BACKGROUND: Colorectal cancer (CRC) diagnosed at diagnosis, 170 patients (7.9%) had carcinoma in situ, 553 or before age 50 (early-onset) is increasing in the United (25.8%) stage I, 57 (24.1%) stage 2, 489 (22.8%) stage 3, States. Early-onset CRC is associated with more advanced and 291 (13.6%) stage 4. The prevalence of early-onset CRC stage disease at diagnosis compared to CRC diagnosed at or increased from 11.4% to 16% during the study period (p after 50 years of age (late-onset).
= 0.157). Patients with early-onset CRC had more rectal Abstracts PURPOSE: To compare the occurrence of early-onset and tumors than patients with late-onset CRC (48.6% vs. 33.3%, Monday late-onset CRC at our center from 2000 to 2011 and iden- p < 0.001), higher rate of recurrence (34.7% vs. 23.6%, p < tify characteristics associated with early-onset CRC. 0.001), and more advanced tumor stage at diagnosis (p < METHODS: We retrospectively studied all patients diag- 0.001). Independent predictors of early-onset CRC included nosed with CRC at our center from January 2000 to Janu- 1st (aOR 1.8 (1.1–2.9), p = 0.016) and 2nd (aOR 5.4 (2.9– ary 2011 using our cancer registry database. Patients were 10.1), p < 0.001) degree family history of CRC and receiving defi ned as early-onset or late-onset CRC based on age at chemotherapy (aOR 3.5 (2.44–5.43), P < 0.001). History of diagnosis. Additional variables were recorded including smoking, cancer in sigmoid colon, and stage 1 and 2 dis- demographic data, personal or family history of CRC or eases were less associated with early-onset CRC on the mul- other cancers, alcohol and tobacco use, tumor location by tivariate analysis. colonic subdivision and tumor stage at diagnosis. Univari- CONCLUSION: Patients with early-onset CRC had more ate analysis (Pearson’s Chi-square or Kendall’s tau-b tests) rectal tumors, more advanced stage disease at diagnosis and was used to identify factors associated with early onset CRC. a higher rate of recurrence. These patients more frequently Multivariate analysis (Cox proportional hazards regression) had a family history of CRC but less frequently had a his- determined independent predictors of early-onset CRC. tory of smoking. Early-onset CRC is an aggressive disease RESULTS: We identifi ed total 2,147 patients, of these, that portends a poor prognosis. Further work is merited to 1,057(49.2%) were male, 1,898 (88.4%) had late-onset identify additional risk factors for this disease.
10:00 AM – 11:00 AM 304AB VIDEO SESSION III
604 605 Robotic Assisted Single Incision Ileocolic Resection Dual-Scope Endoscopic Resection of Benign GE Using Standard Robotic Instrumentation and a Junction Tumors (with Video) Single Incision Laparoscopy Surgery (SILS) Port Edwin O. Onkendi1, Larissa Fujii2, Michael J. Levy2, Jennifer Hrabe, Anthony R. Cyr, John W. Cromwell, Christopher J. Gostout2, Juliane Bingener1 John Byrn 1. Surgery, Mayo Clinic, Rochester, MN, Rochester, MN; Surgery, University of Iowa, Iowa City, IA 2. Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN, We present a representative case from our initial experi- Rochester, MN ence with robotic single incision colectomy. This ileoco- Surgical resection for benign subepithelial tumors near lic resection was performed in a 20 year old female with the gastroesophageal junction is diffi cult, often leading to medically refractory fi brostenotic Crohn’s disease. The per- esophagectomy. Here we demonstrate the feasibility of a ceived advantages of robotic single incision surgery over novel technique of dual endoscopic resection using retro- laparoscopic single incision surgery are: improved visual- fl exed standard adult upper endoscope and second smaller ization, a surgeon-controlled camera platform, and wristed caliber (baby) endoscope to resect benign GE junction leio- instrumentation. myomas (2-6 cm size) in four patients. Maneuvering the small caliber endoscope allowed off-axis retraction of the mass while the adult endoscope was used to carry out the dissection from the submucosal tissue. Our experience highlights the feasibility of this minimally inva- sive approach by enabling triangulation using endoscopic tools.
33 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
606 607 Pass the Courvoisier’s: Laparoscopic Extrahepatic Bile eTAMIS: Transanal Minimally Invasive Surgical Duct Resection with Roux-en-Y Hepaticojejunostomy Submucosal Excision of a Large, Circumferential, Rebecca Kowalski1, Niket Sonpal2, Jennifer Montes1, Rectal Adenoma with Endoscopic Visualization Paresh C. Shah1 Elisabeth C. McLemore, Alisa M. Coker, Peter T. Yu, 1. Surgery, Lenox Hill Hospital, Northshore-LIJ Health System, Garth R. Jacobsen, Mark A. Talamini, Sonia Ramamoorthy, Hofstra Medical School, New York, NY; 2. Medicine, Lenox Hill Santiago Horgan Hospital, Northshore-LIJ Health System, Hofstra Medical School, Surgery, UC San Diego, La Jolla, CA New York, NY TECHNICAL POINTS ADDRESSED: Transanal endolu- We present an 82 year old man with obstructive jaundice, minal surgical techniques can be employed to safely and dark urine and pale stools. On presentation his total biliru- completely remove a large, circumferential rectal adenoma. bin was 12.4. CT of the abdomen and pelvis showed dilated This video demonstrates the benefi ts of combining endo- intrahepatic and common bile ducts with a 1.4 x 2.3 × 2.0 scopic visualization and submucosal injection with mini- cm mass at the junction of the gallbladder neck and cys- mally invasive endoluminal surgery using a soft, disposable tic duct. He underwent laparoscopic extrahepatic bile duct transanal access platform. resection with Roux-en-Y hepaticojejunostomy. Pathology CASE HISTORY: A 51 year old female was found to have demonstrated a 2.5 × 1.5 x 1.0 cm moderately-differentiated a circumferential rectal mass extending from 3cm to 11 cm adenocarcinoma of the gallbladder with 4 negative lymph on fl exible endoscopy and non-invasive transrectal ultra- nodes. Staging was determined to be T2 N0 (Stage II). sound (uT0N0). Final pathology revealed a 8.5 × 6.2 × 1.5 cm tubulovillous adenoma with high grade dysplasia and negative margins.
2:00 PM – 3:00 PM 303ABC QUICK SHOTS SESSION II
707 RESULTS: 664 patients (m:f = 136:526; median age 44.2 years [range 18–66], average BMI 45.6 [range 33.1–76.9]) Is Esophago-Gastro-Duodenoscopy Prior to Roux-en-Y underwent preoperative EGD. In 341 cases no abnormalities Gastric Bypass Mandatory? were found (A), 115 patients had fi ndings that did not have Usha K. Coblijn, Arvid Schigt, Sjoerd D. Kuiken, consequences (B1), 112 patients needed HP eradication Pieter Scholten, Sjoerd M. Lagarde, Bart A. van Wagensveld therapy (B2), 87 patients needed preoperative treatment by Sint Lucas Andreas Ziekenhuis, Haarlem, Netherlands proton pump inhibitors (B3), and 6 patients needed follow up EGD prior to surgery (C). For one patient the operation BACKGROUND and study aims: Roux-Y Gastric Bypass was cancelled because preoperative EGD showed Barrett’s is one of the most frequently used techniques in surgery esophagus with carcinoma (D). When all abnormalities for morbidly obese patients. Postoperative anatomy is are taken into account, baselines show a signifi cant differ- altered by exclusion of the remnant stomach which makes ence for age, gender, hypertension and alcohol consump- this organ inaccessible for future Esophago-gastro-duode- tion. The number of performed EGD’s to fi nd one serious noscopy (EGD). There is no consensus about preoperative abnormality (requiring a follow up EGD and/or postponing assessment of the stomach. Some institutions choose to or cancelling the operation) is 94,5. The estimated costs of investigate the future remnant stomach by EGD, others do one EGD (including personnel costs but without sedation, not. Aim of the present study is to quantify the yield of pre- admission and possible complications) is approximately operative EGD in a bariatric center of excellence. 385 US dollar. METHODS: Patients, who were planned for laparoscopic CONCLUSION: Based on our results and those in literature Roux-Y Gastric Bypass (LRYGB) from December 2007 until it can be concluded that routine assessment by EGD prior to August 2012, were all screened by EGD in advance. These laparoscopic Roux-Y Gastric Bypass should be abandoned. fi les were retrospectively reviewed for EGD outcome, co- In this selected series, risk factors for abnormalities are age, morbidities, medication and other patient characteristic. gender, hypertension and alcohol consumption. The num- All these data were analyzed using a statistical program. ber of EGD’s needed to perform to fi nd one abnormality A two sided P value of <0.05 was considered statistically that requires treatment is high, with equal high costs. signifi cant.
34 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
708 709 Upper Endoscopy Prior to Bariatric Surgery: Do Visual Perioperative Blood Transfusion Reduces Survival Findings Accurately Predict Mucosal and Anatomical in Patients with Pancreatic Adenocarcinoma: A Pathology? Multi-Institutional Study of 698 Patients Craig D. Kolasch, Kristian T. Dacey, Eric Boyle, Jeffrey M. Sutton1, David Kooby2, Gregory C. Wilson1, Amanda Walters, Keith S. Gersin, Dimitrios Stefanidis, Dennis J. Hanseman1, Shishir K. Maithel2, David J. Bentrem3,4, Timothy Kuwada Sharon M. Weber5, Clifford S. Cho5, Emily Winslow5, Carolinas Medical Center, Charlotte, NC Charles R. Scoggins6, Robert C. Martin6, Hong Jin Kim7, 8 8 1 BACKGROUND: Evaluation of upper GI mucosa and anat- Nipun Merchant , Alex Parikh , Daniel E. Abbott , omy is important prior to bariatric surgery. Esophagogastro- Michael J. Edwards1, Syed A. Ahmad1 duodenoscopy (EGD) can diagnose H.pylori infection (HP), 1. Surgery, University of Cincinnati College of Medicine, Cincinnati, Abstracts mucosal infl ammation and hiatal hernia (HH). HP can be OH; 2. Surgery, Emory University School of Medicine, Atlanta, GA; Monday treated preoperatively and the degree of GERD and HH may 3. Surgery, Northwestern University Feinberg School of Medicine, infl uence the choice of bariatric procedure. Mucosal biopsy Chicago, IL; 4. Surgery, Jesse Brown Veterans Affairs Medical adds to the cost of EGD and some endoscopists do not “rou- Center, Chicago, IL; 5. Surgery, University of Wisconsin School of tinely” biopsy for HP if the gastric mucosa appears normal. Medicine and Public Health, Madison, WI; 6. Surgery, University The goal of this study was to determine the relationship of Louisville School of Medicine, Louisville, KY; 7. Surgery, between gross visual fi ndings on EGD and histopathology. We also examined the ability of EGD to accurately diagnose University of North Carolina School of Medicine, Chapel Hill, NC; hiatal hernia. 8. Surgery, Vanderbilt University School of Medicine, Nashville, TN METHOD: A retrospective review of prospectively col- INTRODUCTION: In this multi-institutional study of lected data of a single surgeon (TSK) series of laparoscopic patients undergoing pancreaticoduodenectomy (PD) for non revisional bariatric procedures at a center of excellence pancreatic adenocarcinoma, we sought to identify factors between 2010-2012. Preoperative EGD was performed on associated with a perioperative blood transfusion require- all patients. Patients without a gastric biopsy were excluded ment. In addition, we investigated the hypothesis that from analysis. Endoscopic appearance (gross) was consid- receiving blood transfusion reduces long-term survival in ered positive if there were any signs of infl ammation or this patient population. hiatal hernia. The gross and histological appearances were METHODS: A retrospective chart review was performed compared. Biopsy results (histology) and laparoscopic eval- across six high-volume institutions to identify patients uation of the hiatus were considered the gold standard Sen- who underwent PD between 2005 and 2010. Data collec- sitivity (SS) and specifi city (SP) of the gross EGD appearance tion included patient demographics, perioperative fac- were calculated. tors, transfusion status, and survival data. For statistical RESULTS: There were 274 patients in the study group. analysis, patients were then grouped according to whether Mean age and BMI were 42.8 and 43.3 respectively. The they received 0, 1-2, or >2 units of packed red blood cells majority of the patients were female (88%). Procedures (pRBCs). included: 189 laparoscopic gastric bypass, 69 laparoscopic RESULTS: Among 698 patients identifi ed, 168 (24%) sleeve gastrectomy and 16 laparoscopic adjustable gastric required blood transfusion. 105 (15%) received 1–2 units bands. There were 57 HH confi rmed at the time of surgery and 63 (9%) received >2 units (range 0–25 units). Patient (20%). Preoperative EGD identifi ed 21 of these (SS = .37, SP demographics associated with an increased transfusion = .86). H. pylori was identifi ed in 34 patients (12.5%); 19 of requirement included age, smoking status, and heart dis- these patients had gross infl ammatory changes on EGD (SS ease (all p < 0.03). Operative variables associated with an = .56, SP = .58). Gross gastric infl ammatory changes were increased transfusion requirement included operative time, identifi ed in 63/125 patients that had histological gastritis estimated blood loss, tumor size, and R1/R2 margin status (SS = .50, SP = .64). There were 12 patients with histological (all p < 0.03). Postoperative complications were not associ- GERD, 8 of these patients had grossly infl amed esophageal ated with transfusion requirement. However, those patients mucosa (SS = .67, SP = .24). who received transfusions experienced a longer length of CONCLUSION: EGD prior to bariatric surgery can provide stay (p = 0.0009) as well as increased rate of readmission important information that may alter preoperative inter- within 90 days (p = 0.002). The median survival of patients ventions and infl uence the choice of bariatric procedure. who received >2 units of pRBCs was signifi cantly less than Our fi ndings suggest that the gross appearance during EGD those who received either 0 or 1–2 units (10.2 months vs. is unreliable for detecting histological infl ammation, infec- 18.4 or 18.9 months, p = 0.0002). A multivariate model tion (HP) and HH. Thus, we recommend routine gastric including margin status, nodal involvement, tumor size, biopsies to maximize H. pylori detection during EGD. Fur- and transfusion status identifi ed the transfusion of >2 units thermore, if the presence of a hiatal hernia could change of pRBCs as an independent predictor of reduced survival a procedural recommendation, a complimentary upper GI (HR 1.56, p = 0.03). evaluation should be considered.
35 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
CONCLUSIONS: This multi-institutional study represents tiveness of modern chemotherapeutic regimens and the the largest series to date analyzing the effects of pRBC trans- high operative mortality in this population, further studies fusion in patients undergoing PD for pancreatic adenocar- are needed to evaluate the role of and timing of resection of cinoma. The transfusion rate in this series is less than what the primary tumor. has been previously reported. Our data confi rm that blood transfusion confers a negative impact on long-term survival 711 in this patient population. These results can be utilized as a benchmark for future studies. Transplant Versus Resection for the Management of Hepatocellular Carcinoma in the Post-2006 MELD 710 Exception Era at a Single Institution in the Southeast UNOS Region Trends in Resection and Chemotherapy in Patients Malcolm H. Squires1, Steven Hanish2, Sarah B. Fisher1, with Stage IV Colorectal Cancer Cristen Garrett2, David Kooby1, Juan M. Sarmiento3, 1 1 1,2 Gabriela Vargas , Kristin Sheffi eld , Abhishek Parmar , Kenneth Cardona1, Stuart J. Knechtle2, Maria C. Russell1, 1 1 1 Yimei Han , Taylor S. Riall , KImberly M. Brown Joseph F. Magliocca2, Andrew B. Adams2, Charles A. Staley1, 1. General Surgery, University of Texas Medical Branch, Galveston, Shishir K. Maithel1 TX; 2. General Surgery, USCF East Bay, Oakland, CA 1. Department of Surgery, Division of Surgical Oncology, Winship INTRODUCTION: Patterns and trends in the use of mod- Cancer Institute, Emory University, Atlanta, GA; 2. Department ern chemotherapeutic regimens, primary tumor resection, of Surgery, Division of Liver Transplantation, Emory Transplant and the relative timing of chemotherapy and resection in Center, Emory University, Atlanta, GA; 3. Department of Surgery, older patients with stage IV colorectal cancer (CRC) have Division of General and GI Surgery, Emory University, Atlanta, GA not been evaluated. BACKGROUND: Optimal management of hepatocellular METHODS: We used Texas Cancer Registry-Medicare carcinoma (HCC) in the post-2006 MELD (Model for End linked data (2001–2007) to identify patients 66 and older Stage Liver Disease) exception era remains controversial with stage IV colorectal cancer (N = 3,343). Time trends and is regionally dependent. We compared outcomes for in resection of the primary tumor and receipt of chemo- patients undergoing liver transplant versus resection at a therapy were determined. We defi ned chemotherapy regi- single institution in a UNOS region with short wait times mens as “standard” (5-fl uorouracil (5-FU)/leucovorin (LV)) for organ availability. or “modern” (oxaliplatin or irinotecan plus 5-FU/LV or bevacizumab). METHODS: All patients who underwent resection of HCC between 1/00 and 8/12 were identifi ed. Inclusion of RESULTS: The mean age of patients was 76.9 ± 7.2 years, patients who underwent transplant was limited to those 53.3% were female, and 80.9% were white. 87.7% of can- after 1/06, when the MELD exception policy for HCC based cers were in the colon and 25.7% of tumors were poorly dif- on the Milan criteria (MC) was universally incorporated ferentiated histologically. Liver metastases, lung metastases, into UNOS organ allocation. Primary outcomes were over- and carcinomatosis were documented in 72.8%, 32.3% and all survival (OS) and recurrence-free survival (RFS). 36.3% of patients, respectively. 37.4% were treated with both chemotherapy and resection, 26.4% had resection RESULTS: 259 patients were identifi ed, of whom 133 only, 11.8% had chemotherapy only, and 24.4% had no underwent transplant and 126 underwent resection. Trans- treatment. Resection of the primary tumor was performed plant patients had a higher incidence of hepatitis C (67% vs in 63.8% of patients, of which 24.3% were emergent. After 29%, p < 0.001), a greater median raw MELD score (15 vs 8, excluding emergent procedures, resection decreased from p < 0.001), and smaller tumor size (2.4 vs 7.0 cm, p < 0.001). 59.9% to 53.0% between the early (2001–2002) and late All 133 patients who underwent transplant met MC, while (2006–2007) study periods (P = 0.007). In patients undergo- 37 (29%) who underwent resection met MC. Of these 37 ing elective resection and chemotherapy (N=1015), resec- patients, 26 had preserved liver function with a raw MELD tion was done prior to chemotherapy in 88.5% of patients. score ≤8. Median follow-up time was 30 mos. Median wait 30-day post-operative mortality for all patients undergoing time to transplant was 55 days (1-321); no patients dropped resection of the primary tumor was 13.5% and 10.7% for off the waitlist while awaiting an organ. elective resection. Chemotherapy was given to 49.2% of Transplant compared to resection was associated with patients and was stable over time. However, in patients who improved OS (median not reached (MNR) vs 28.7 mos, p received chemotherapy, the use of oxaliplatin or irinotecan < 0.001) and greater RFS (MNR vs 17.4 mos, p < 0.001). plus 5-FU/LV increased from 53.3% in 2001 to 89.7% in When compared to the 37 patients within MC who under- 2007 (P < 0.0001). After approval of bevacizumab for meta- went resection, transplant demonstrated a trend towards static CRC in 2004, its use increased from 0.5% to 30.8% improved OS (MNR vs 57.4 mos, p = 0.065) and greater from 2001–2004 and from 30.8% to 55.8% from 2004-2007 RFS (MNR vs 35.2 mos, p < 0.001; Figure). Of these 37 (P<0.0001). patients who underwent resection, 11 (30%) have under- CONCLUSIONS: In patients with stage IV colorectal can- gone salvage procedures for recurrence of HCC versus only cer, modern chemotherapeutic regimens have been rapidly 4 (3%) transplant patients. Compared to resection patients adopted over the last decade. Concomitantly, there has within MC with a raw MELD score ≤8 (n = 26), transplant been a decrease in resection rates. Given the increased effec- demonstrated similar OS (MNR vs 57.4 mos, p = 0.84) but
36 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
greater RFS (MNR vs 17.6 mos, p < 0.001). For patients RESULTS: Forty-one incisional hernia repairs met our with hepatitis C, those undergoing transplant (n = 89) had inclusion criteria. Alloderm was used in 21 (51.2%) cases improved outcomes compared to the 19 patients who met and Permacol was utilized in 20 (48.8%) cases. Seventeen MC and underwent resection (OS: MNR vs 47.9 mos, p = patients (41.5%) developed a recurrent hernia at a mean 0.04; RFS: MNR vs 16.2 mos, p < 0.001). interval of 10.4 months from surgery. Hernias repaired with Alloderm recurred in 47.6% (10 of 21) patients, while Per- macol repairs recurred in 35% (7 of 20) of cases (p = 0.412). Infectious complications necessitating surgical interven- tion developed in 9 cases (22%). Hernias repaired using the bridging technique revealed an 87.5% recurrence rate (7 of 8), while underlay fi xation of the mesh with native fascial reapproximation led to recurrence in only 31.3% of the
cases (10 of 32). Abstracts Monday CONCLUSION: Our results demonstrate relatively high rates of recurrence when performing a single stage ventral hernia repair in a contaminated fi eld with biologic mesh. Permacol and Alloderm showed similar results in this series. This rate is signifi cantly higher than typically reported in literature, likely due to longer follow-up and relatively high patient acuity, and calls into question the cost-effectiveness of the use of biologic mesh in a single stage repair for con- CONCLUSION: In a region with short wait times for organ taminated recurrent hernias vs the older approach of using availability, liver transplant is associated with improved a lightweight absorbable synthetic with a second-stage survival compared to resection for HCC. For patients defi nitive repair. within Milan criteria, transplant appears to confer an onco- Table 1: Characteristics of Patients Undergoing Hernia Repair with logic advantage as well, even in those with preserved liver Biologic Mesh in an Infected Field function. For patients within Milan criteria with hepa- titis C, transplant is associated with improved survival Overall Alloderm Permacol p-Value and decreased recurrence when compared to resection. N 412120— Transplant should be considered for all patients meeting Gender, Males 39% 30% 48% 0.248 Milan criteria, particularly those with hepatitis C, when Age, years 58.7 (11) 56.8 (8) 60.7 (14) 0.094 being managed in a region with short wait times for organ Mesh positioning 78/20/2 62/33/5 95/5/0 0.037 availability. (%underlay/inlay/ onlay) 712 Length of stay, days 6.1 (2) 6.8 (2) 5.4 (2) 0.024 Recurrenc rate 41.5% 47.6% 35% 0.412 Biologic Mesh in a Contaminated Field: Infected Mesh Interval to recurrence, 10.4 (7) 9.0 (7) 12.4 (8) 0.370 Removal and Hernia Repair in a Single-Stage months Jeffrey Mino, Rosebel Monteiro, Steven Rosenblatt Duration of follow-up, 16.1 (15) 17.0 (15) 15.3 (16) 0.790 General Surgery, Cleveland Clinic, Cleveland, OH months PURPOSE: High rates of recurrence and infectious compli- Data presented as means (SD), or percentages where indicated cations are associated with the repair of hernias with syn- thetic mesh in a contaminated surgical fi eld. Biologic mesh Table 2: Mesh Positioning and Recurrence Rates: is believed to reduce the rates of these complications. We Permacol vs. Alloderm compared the performance of two widely available bioma- terials, Permacol and Alloderm, in a single-stage procedure Overall Alloderm Permacol of infected mesh removal and hernia repair. N 412120 Underlay 32 13 19 METHODS: All patients who underwent a single-stage Recurrence (%) 10 (31.3) 4 (30.8) 6 (31.6) incisional hernia repair with replacement of an infected Inlay 8 7 1 synthetic mesh by a biologic mesh were identifi ed. Data retrieved included patient demographics, details of current Recurrence (%) 7 (87.5) 6 (85.7) 1 (100) hernia repair with biologic mesh, post-operative complica- Onlay 1 1 0 tions, and hernia recurrence. Recurrence (%) 0 (0) 0 (0) 0 (0)
37 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
713 714 Predictors of Bile Duct Injury During Laparoscopic Molecular Pathological Phenotypes and Outcome in Cholecystectomy: Experience from Two Hospitals Pancreatic Ductal Adenocarcinoma Curtis J. Wray1, John A. Harvin1, Uma R. Phatak1, Nigel B. Jamieson1, Mohamed A. Mohamed1, Karin Oien2, Lillian S. Kao1, Tien C. Ko1, Taylor S. Riall2 Fraser Duthie2, Euan J. Dickson1, Ross Carter1, Colin McKay1 1. Surgery, University of Texas Medical School at Houston, 1. West of Scotland Pancreatic Unit, Glasgow University Houston, TX; 2. Surgery, University of Texas Medical Branch, Department of Surgery, Glasgow Royal Infi rmary, Glasgow, Galveston, TX Lanarkshire, United Kingdom; 2. Department of Pathology, INTRODUCTION: Bile duct injuries (BDI) during laparo- Southern General Hospital, University of Glasgow, Glasgow, United scopic cholecystectomy (LC) are a devastating complica- Kingdom tion. Due to the infrequent occurrence of BDIs, predictive INTRODUCTION: Individuals with pancreatic ductal factors are not well-understood. We hypothesized increased adenocarcinoma (PDAC) demonstrate a generally poor out- age is a risk factor for BDI. come following resection. Molecular profi ling has previ- METHODS: IRB-approved review of LC at two safety- ously enhanced the identifi cation of phenotypic subtypes net hospitals from 2005-2011. LC cases were coded elec- of ampullary adenocarcinoma. Furthermore an intestinal tive (same day surgery) or non-elective (admitted through subtype of PDAC has been described however the prognos- emergency room). Morbidity/mortality conferences, hospi- tic impact of this variant has not been described in detail. tal records and prospective data were used to identify BDIs. We sought to better characterize the intestinal subgroup of The Strasberg classifi cation was used to defi ne BDI. We cre- PDAC and assess the impact on outcome. ated a multivariate regression model to identify predictors METHODS: We assessed the potential clinical utility of of BDI. molecular pathological phenotypes defi ned using a com- RESULTS: 2896 LCs (n = 2370 female n = 526 male) were bination of histopathology and protein expression (CDX2 performed at two hospitals. 52% of cases were elective. 120 [caudal-type homeodomain transcription factor 2]—an cases were converted to an open operation. Males had a intestinal marker and MUC1—a pancreaticobiliary marker) higher mean age (years) than females (47 ± 15 vs 39 ± 14, assessed by immunohistochemistry (Figure 1) in 95 patients p = 0.01). 40 BDIs were identifi ed. The most common BDI who underwent operative resection for PDAC by pancreati- was type A (N = 27, women = 14 vs men = 13), followed by coduodenectomy at a single institution over a 12 year time E3 (N = 4), D (N = 4), E2 (N = 3) and C (N = 2). Predictors of period. A tissue microarray was used with at least 4 cores BDI included: Age (OR 1.44, 95%–CI 1.10-1.89), male gen- evaluated for each tumor for protein expression analysis in der (OR 3.07 95%–CI 1.83–5.12) and non-elective operation addition to whole section analysis of tumor morphology. (OR 5.11 95%–CI 1.16–22.5). The predicted probability of Care was taken to exclude all other periampullary malig- BDI increased with advancing ages, but more so for men nancies from the analysis. who underwent non-elective LCs (see graph). RESULTS: In addition to prognostic impact of T stage, lymph node status, resection margin status, perineural inva- sion and vascular invasion, a small proportion of tumors had features of an intestinal histological subtype (13%) and a more favorable prognosis. CDX2 and MUC1 expression were signifi cant prognostic variables. Patients with CDX2 negative tumors had a signifi cantly shorter survival (Hazard ratio [HR] = 2.77, 95%CI: 1.5–5.2, P = 0.002 as did those with MUC1 positive tumors (HR = 2.89, 95%CI: 1.7–4.9, P < 0.0001 – no survivors at 24 months). Patients with CDX2 negative/MUC1 negative tumors had an intermediate out- come (Figure 1). In a multivariate analysis lymph node involvement, vascular invasion, positive MUC1 expression and loss of CDX2 expression were independent predictors of poor outcome. CONCLUSION: Morphological determination of intesti- CONCLUSIONS: The risk of major or minor BDIs is signifi - nal subtype of PDAC has clinical relevance. Furthermore cant in older males undergoing non-elective LCs. Increased maintenance of CDX2 expression identifi es a group of risk of BDIs, due to increase age and emergency surgery, PDAC patients with a relatively good outcome while MUC1 should be risk factors considered when discussing the need expression identifi ed patients with a very poor outcome. and timing of elective LC, especially with male patients. When combined histopathological and molecular criteria Earlier referral and interventions aimed at decreasing the defi ne clinically relevant phenotypes of PDAC with signifi - percentage of emergency LCs may decrease the incidence cant implications for prognostication, current therapeutic of BDIs. strategies and may facilitate future trial design.
38 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
develop VTE after discharge. We aim to characterize the risk of post-discharge VTE after colorectal surgery and risk fac- tors that may suggest the need for the consideration of an extended postoperative VTE prophylaxis beyond the hos- pital phase. METHODS: The ACS-NSQIP dataset from 2005–2007 was used to identity patients undergoing colectomy or proctec- tomy. Patients who developed deep venous thrombosis or pulmonary embolism were identifi ed and sorted into pre or post discharge events. Univariate and multivariate analysis was done to identify risk factors for post-discharge VTE. RESULTS: 30,900 patients undergoing laparoscopic and Abstracts open resection of the colon and rectum for benign and Monday malignant conditions were identifi ed. 567 patients devel- oped DVT with 149 (26%) diagnosed post-discharge. 232 cases of pulmonary embolism were identifi ed with 82 Figure 1. Kaplan-Meier survival curves demonstrating stratifi cation of (35%) diagnosed post-discharge. Factors associated with the the 95 patient PDAC cohort according to CDX2 and MUC1 expression post-discharge risk for VTE included open vs. laparoscopic surgery (0.77% vs. 0.47%, p < 0.05), no resident vs. presence assessed by immunohistochemistry. of resident (0.91% vs. 0.62%, p < 0.05), steroid use (1.5% vs. 0.61%, p < 0.05), reoperation (1.3% vs. 0.65%, p < 0.05), 715 BMI > 30 (p < 0.05) and higher ASA class (p < 0.05). Venous Thromboembolism (VTE) After Colorectal CONCLUSION: A substantial fraction of overall VTE (DVT Surgery: Making the Case for Continuing Prophylaxis and PE) occurs post-discharge in patients undergoing After Discharge in High-Risk Patients colorectal resection, this risk higher in patients with higher Vikram Attaluri, Jeffrey Hammel, Pokala R. Kiran ASA class, on perioperative steroids and undergoing open Department of Colorectal Surgery, Cleveland Clinic, Cleveland, OH surgery and reoperation. These fi ndings strongly support the consideration of extension of VTE prophylaxis to the PURPOSE: SCIP measures target prophylaxis for venous post-discharge (at home) period after colorectal resection in thromboembolism (VTE), during the hospital phase for patients with these identifi ed risk factors. patients undergoing surgery; some patients nevertheless
Univariate Analysis for Post-Discharge VTE
Variable Overall N=30900 No Post-Discharge VTE Post-Discharge VTE p Value Age <70 yrs Age >70 ys 20150 (65.2%) 10750 (34.8%) 20018 (99.3%) 10672 (99.3%) 132 (0.66%) 78 (0.73%) 0.47 Female Male 15961 (51.7%) 14936 (48.3%) 15855 (99.3%) 14832 (99.3%) 106 (0.66%) 104 (0.70%) 0.73 Colectomy Proctectomy 28328 (91.7%) 2572 (8.3%) 28131 (99.3%) 2559 (99.5%) 197 (0.70%) 13 (0.51%) 0.26 Laparoscopic Open 8966 (29.2%) 21764 (70.8%) 8924 (99.5%) 21596 (99.2%) 42 (0.47%) 168 (0.77%) 0.004* No Resident Resident 7044 (22.9%) 23679 (77.1%) 6980 (99.1%) 23533 (99.4%) 64 (0.91%) 146 (0.62%) 0.009* No Metastatic Cancer Metastatic Cancer 29410 (95.2%) 1490 (4.8%) 29213 (99.3%) 1477 (99.1%) 197 (0.67%) 13 (0.87%) 0.35 No Steroid Use Steroid Use 28437 (92.0%) 2463 (8.0%) 28263 (99.4%) 2427 (98.5%) 174 (0.61%) 36 (1.5%) <0.001* No Sepsis SIRS Sepsis Septic Shock 26460 (85.6%) 2729 (8.8%) 26287 (99.3%) 2702 (99.0%) 173 (0.65%) 27 (0.99%) 0.049* 826 (2.7%) 885 (2.9%) 818 (99.0%) 883 (99.8%) 8 (0.97%) 2 (0.23%) No Operation within 30 days Operation 23421 (96.6%) 819 (3.4%) 23269 (99.4%) 808 (98.7%) 152 (0.65%) 11 (1.3%) 0.019* within 30 days No emergency surgery Emergency surgery 25904 (83.8%) 4996 (16.2%) 25733 (99.3%) 4957 (99.2%) 171 (0.66%) 39 (0.78%) 0.34 ASA Class 1-No Disturb 2-Mild Disturb 3-Severe 1047 (3.4%) 14338 (46.4%) 1042 (99.5%) 14256 (99.4%) 5 (0.48%) 82 (0.57%) 0.037 Disturb 4-Life Threat 5-Moribund 12591 (40.8%) 2692 (8.7%) 12483 (99.1%) 2678 (99.5%) 108 (0.86%) 14 (0.52%) 220 (0.71%) 219 (99.5%) 1 (0.45%) * signifi cant
39 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
2:00 PM – 3:45 PM 308D PLENARY SESSION V
778 independent function (OR-1.92, p < 0.001), and intraopera- tive transfusion≥4 units (OR-2.14, p < 0.001). Independent Morbidity and Mortality After Pancreatico- factors associated with mortality included: anesthesia risk duodenectomy in Patients with Borderline score > 3 (OR-2.27, p = 0.025), age ≥ 80 (OR-2.83, p < 0.001), Resectable Type C Clinical Classifi cation lack of independent function (OR-2.89, p = 0.002), and Ching-Wei D. Tzeng1, Matthew Katz1, Jason B. Fleming1, intraoperative transfusion ≥4 units (OR-2.80, p = 0.003). Holly M. Holmes3, Jeffrey E. Lee1, Peter W. Pisters1, Jean-Nicolas Vauthey1, Gauri R. Varadhachary2, Robert A. Wolff2, James Abbruzzese2, Thomas Aloia1 1. Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX; 2. Department of GI Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX; 3. Department of General Internal Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX BACKGROUND: We previously described the clinical classifi cation of patients with resectable pancreatic tumor anatomy but marginal performance status (PS) or revers- ible comorbidities as “borderline resectable type C” (BR-C for condition/comorbidity). This study was designed to analyze the nationwide incidence and risk factors for post- pancreaticoduodenectomy (PD) morbidity/mortality in patients who could be classifi ed as BR-C. METHODS: All elective PDs were evaluated in the 2005–10 ACS-NSQIP database. BR-C was defi ned by the following: age ≥ 80, lack of independent function, pulmonary disease, ascites/varices, recent myocardial infarction/angina, stroke CONCLUSIONS: These data confi rm that a large number history, steroid use, weight loss >10%, and/or preoperative of medically high-risk patients are being treated with PD. sepsis. Clinical variables potentially associated with 30-day These BR-C patients were at higher risk for and less able morbidity/mortality were analyzed, with a focus on the to be rescued from major morbidity, with higher rates of development of postoperative major complications defi ned mortality from potentially reversible risk factors. These data as: pneumonia, re-intubation/ventilation >48 hr, renal fail- suggest the need for optimization of comorbidities and for ure, cardiovascular event, sepsis, re-operation, dehiscence, increased utilization of prehabilitation to address nutri- organ space infection, and venous thromboembolism. tional and conditioning defi cits before PD. RESULTS: Of 8,266 PDs, 3,033 (36.7%) involved patients with BR-C classifi cation. Analysis of preoperative variables 779 determined that BR-C patients were more likely to have abnormal preoperative lab values (albumin, liver function Value of Frailty and Nutritional Status Assessment in tests, leukocytosis, coagulation, hematocrit, uremia, creati- Predicting Perioperative Mortality in Gastric Cancer nine, all p≤0.002) and need for preoperative hospitaliza- Surgery tion (23.6% vs. 12.3%, p < 0.001). Despite similar operative Juul Tegels, Michiel de Maat, Karel Hulsewé, times (≥360 min in 47.2% BR-C vs. 49.2% non-BR-C, p = Anton G. Hoofwijk, Jan H. Stoot 0.081), BR-C patients were more likely to suffer major Surgery, Orbis Medical Center, Sittard-Geleen, Netherlands complications (30.8% vs. 25.9%, p < 0.001) and mortality BACKGROUND: Reported perioperative mortality in gas- (4.1% vs. 2.3%, p < 0.001). In addition, BR-C patients with tric cancer surgery is relatively high (4–16%). This may major complications suffered a 50% higher mortality rate be due increased patient age and poor condition related compared to non-BR-C patients with major complications to advanced tumor stage at time of diagnosis. Comorbid- (11.5% vs. 7.7%, p < 0.001). For BR-C patients, multivari- ity and age are currently the main factors associated with ate analysis identifi ed the following risk factors for major the risk of surgical mortality however their predictive complications: albumin < 3.5g/dL (odds ratio, OR-1.24, p value is limited. Better preoperative evaluation tools have = 0.036), dyspnea (OR-1.71, p < 0.001), preoperative sepsis become warranted for better patient selection and prevent- (OR-1.89, p = 0.001), age ≥ 80 (OR-1.56, p < 0.001), lack of ing unnecessary surgery related mortality. The aim of the
40 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
study was to investigate the additional value of frailty and METHODS: Three hundred and fi fteen patients who nutritional status assessment for predicting perioperative underwent laparoscopic anti-refl ux surgery at the Univer- performance. sity of Nebraska Medical Center between 2002 and 2012 METHODS: Patients in our hospital are screened for frailty were included in this study. Patient data including pre and at admission using a scoring questionnaire to assess their post-operative studies and symptom questionnaires were GFI (Groningen Frailty Indicator) which entails Activities of prospectively collected and the database was used to ana- Daily Living, self-perceived physical and mental fi tness and lyze postoperative outcomes. Statistical analysis includ- multipharmacy. Further, nutritional status is assessed using ing multivariate regression models were used to compare SNAQ (Short Nutritional Assessment Questionnaire). We patient factors and their effect on outcomes. evaluated in a retrospective analysis whether these scores RESULTS: A total of 302 Nissen fundoplications, 24 redo were associated with 30-day mortality considering standard fundopliations, 11 Toupet and 2 Dor procedures were per- clinicopathological parameters. formed. Mean BMI was 31%. The median follow up was Abstracts RESULTS: From January 2005 to September 2012 180 68 (6–130) months. There were 3 reoperations (0.9%) for Monday patients underwent surgery for gastric cancer with an over- recurrent symptoms. Mesh was used in 210 cases where all 30-day mortality of 8.3%. Complete GFI and SNAQ hiatal hernia was larger than 2 cm. Median preoperative scores were available in 127 (71%) and 160 (89%) cases DeMeester score was 60.5 which decreased to 6 (p < 0.05). respectively. Univariate analysis showed a cumulative mor- Heartburn (N = 264) improved in 245 (92%), regurgitation tality risk in association with increasing GFI score (P < (N = 264) improved in 256 (97%) and dysphagia (N = 253) 0.001). Patients with a GFI ≥ 3 (n = 30, 24%) had a mortal- improved in 227 (89%). Atypical presentation such as pul- ity rate of 23.3% versus 5.2% in the lower GFI group (OR monary and throat symptoms only, improved in 82% of 4.0, 95% CI 1.1 to 14.1, P = 0.03). This was independent patients. Radiographic studies were available in 60% of the from patient age, ASA classifi cation, tumor stage and type patients with mean follow up of 1.5 years with evidence of surgery. SNAQ score ≥1 (n = 98, 61%) was associated with of overall recurrence of 7% (21% in hiatal hernia >5 cm). a mortality rate of 13.3% versus 3.2% in the group that Of those with recurrence over 91% were asymptomatic at scored 0 (OR 5.1, 95% CI 1.1 to 23.8, P = 0.04). This was follow up. Male gender, advanced age, size of hiatal hernia, also independent for patients age, type of surgery, tumor and preoperative BMI are independent predictors of worse stage and ASA classifi cation. Patients who were in the group outcomes. of both GFI ≥3 and SNAQ ≥1 (n = 25, 19%) had a mortality CONCLUSIONS: Antirefl ux surgery is effective in control- rate of 28% versus 6% in the remaining patients (OR 6.1, ling symptoms of refl ux and correction of hiatal hernia in 95% CI 1.7 to 22.4, P = 0.006). This fi nding was also inde- long term follow up of greater than 5 years. The durability pendent for patients age, type of surgery, tumor stage and of this procedure can be affected by large hiatal hernia and ASA classifi cation. high BMI and male sex. Overall patients were very satisfi ed DISCUSSION: This is the fi rst study that shows a signifi - with the operation even after 10 years of follow up. cant relation between gastric cancer surgical mortality and geriatric frailty as well as nutritional status using a simple scoring questionnaire. This information may be of value in 6 781 the preoperative decision making for selecting patients who would optimally benefi t from surgery for gastric cancer. Retroesophageal Fundic Wrap Incarceration: A Late Complication After Nissen Fundoplication Margaret M. McGuire, Erik G. Lough, Donald R. Czerniach, 780 John J. Kelly, Philip Cohen Surgery, University of Massachusetts, Worcester, MA Long Term Patient Outcomes After Laparoscopic This video illustrates the laparoscopic treatment of a late Antirefl ux Procedures retroesophageal fundic wrap incarceration in a patient who Anton Simorov, Ajay Ranade, Jeremy P. Parcells, had undergone laparoscopic Nissen fundoplication. Our Dmitry Oleynikov patient presented to the emergency room 3 years after a Surgery, UNMC, Omaha, NE Nissen fundoplication with epigastric pain, back pain and BACKGROUND: Laparoscopic antirefl ux surgery with or early satiety. Imaging revealed gastric outlet obstruction. without large hiatal hernia has been shown to have good Endoscopy showed a friable mucosa with patchy necrosis. short term outcomes. However, limited data are available On laparoscopic exploration the fundic wrap was found on long term outcomes of greater than 5 years. The aim incarcerated, due to rotation beyond 360 degrees through of this study is to review functional and symptomatic out- the retroesophageal space, and lying on top of the greater comes of anitrefl ux surgery in a large tertiary referral medi- curve of the stomach. After dissection and reduction of the cal center. wrap the incarcerated portion was necrotic and required resection.
41 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
782 783 The Association of Proton Pump Inhibitor Use on the Pancreatic Morphological Changes in a Long-Term Incidence of Erosive Esophagitis Follow-Up After First Episode of Acute Alcoholic Steven P. Bowers1, Armando Rosales-Velderrain1,2, Pancreatitis Marc G. Mesleh1, Horacio J. Asbun1, John Stauffer1, Eric J. Lam1, Jussi Nikkola2, Irina C. Rinta-Kiikka1, Sari Raty1, Johanna Mauricia Buchanan1, Jeffrey Ferrell1, Li-Ling Iem1, Laukkarinen1, Riitta Lappalainen-Lehto1, Satu Järvinen1, Ross F. Goldberg1, C. Daniel Smith1 Hanna Seppänen1, Isto Nordback1, Juhani Sand1 1. Surgery, Mayo Clinic, Florida, Jacksonville, FL; 2. General 1. Department of Gastroenterology and Alimentary Tract Surgery, Surgery, Cleveland Clinic Florida, Weston, FL Tampere University Hospital, Tampere, Finland; 2. School of BACKGROUND: Erosive esophagitis (EE) is experimen- Medicine, University of Tampere, Tampere, Finland tally and epidemiologically linked to Barrett’s esophagus OBJECTIVE: Long-term morphological changes induced (BE) and esophageal adenocarcinoma (EAC). The authors by a single episode of alcohol pancreatitis are not known. speculated that the increased availability of proton pump Our aim was to study these morphological changes in secre- inhibitor (PPI) medications may alter the epidemiology of tin stimulated MRCP (S-MRCP) after the fi rst episode of erosive refl ux disease. alcohol associated acute pancreatitis, and to evaluate the METHODS: Between 1991 and 2009, 18,000 inhabitants risk factors and the possible protective factors that might be of Olmsted County, Minnesota were identifi ed by the Roch- associated with later chronic fi ndings. ester Epidemiology Project as having undergone diagnostic PATIENTS AND METHODS: In this prospective follow- upper endoscopy (EGD). A review of medical records was up study S-MRCP-imaging was performed for 44 (41 M, 3F, performed on a random 10% sample of cases. Incidence of mean age 46 (25–68) years) patients who survived their fi rst EE was calculated based on the initial EGD of each case, and episode of alcohol associated pancreatitis. Pancreatic mor- was correlated with demographic and medical history data. phology was evaluated at 3 months, and at 2, 7 and 9 years RESULTS: Of the 1792 records reviewed, the incidence of after hospitalization. Recurrent attacks of pancreatitis were EE was 22.3% (by LA Classifi cation: Grade D, 0.9%; Grade studied and pancreatic function was monitored by labo- C, 2.8%; Grade B, 6.8%; Grade A, 11.8%). EE was associ- ratory tests. Patients’ alcohol consumption was evaluated ated strongly with male gender, but not with patient age or with questionnaires, laboratory markers and self-estimated obesity. PPI use history among cases signifi cantly increased alcohol consumption via interview. Smoking and BMI were throughout the study, from 7% in the period 1991–1997 annually recorded. to 25% in 1998–2003 to 42% in the period 2004–2009. RESULTS: At 3 months 32% of the patients had nor- Rates of EE in the time periods were 26%, 22%, and 21% mal fi ndings in S-MRCP, 52% had acute and 16% chronic respectively. Cases with history of PPI use had signifi cantly changes. At seven years, S-MRCP was performed to 36 lower incidence of EE (PPI use, 18%; no PPI use, 26%; p patients, with normal fi ndings in 53%, the rest (47%) hav- = 0.004, Fisher’s exact test), despite higher likelihood of ing chronic fi ndings. Pancreatic cyst was present in 36%, refl ux-related complaints. Including analysis of 1354 sub- parenchymal changes in 28% and atrophy in 28% of the sequent EGD in 676 cases, 7.5% of all cases had fi nding of cases. 7/36 (19%) patients demonstrated new chronic fi nd- endoscopically suspected esophageal metaplasia (ESEM) on ings after two years. There were no changes in the pan- any EGD; BE/EAC was found in 3.5% of all cases. ESEM and creas in the attending patients between 7 and 9 years (18 BE/EAC were each present at initial EGD or followed EE in patients). If chronic changes were present at 3 months after 90%. diagnosis, they would show in later S-MRCPs also. 45% CONCLUSIONS: Early treatment of refl ux symptoms of the patients who had only acute fi ndings at 3 months with PPI medication may be protective of development resolved to normal in 7 years, but the rest (55%) showed of erosive esophagitis and may decrease the likelihood of chronic changes later on. 22% of the patients who attended future development of Barrett’s esophagus and esophageal the seventh year S-MRCP had gone through a recurrent epi- adenocarcinoma. sode of acute pancreatitis (mean 22 (2–60) months), and 8% had a clinical diagnosis of chronic pancreatitis. At 7 years, 88% of patients with recurrences had chronic fi ndings in S-MRCP versus 36% with non-recurrent pancreatitis (p = 0.02). 6 (17%) patients maintained abstinence through the follow-up (mean 8.7 (7–9.1) years), but even one of these developed pancreatic atrophy. Out of the non-abstinent patients who didn’t have recurrences, 4/22 (18%) developed new fi ndings in the follow-up S-MRCP (NS). Heavy smok- ing didn’t show correlation to increased chronic changes compared to non-smoking in univariate analysis. CONCLUSIONS: Morphological pancreatic changes increase with recurrent episodes of acute pancreatitis. How- ever, even a single episode of acute alcoholic pancreatitis may induce chronic morphological changes in a long-term follow-up.
42 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
were directly attributable to pancreatic fi stula. Increasing 784 FRS scores (0–10) correlated well with CR-POPF develop- A Multi-Institutional External Validation of the Fistula ment (p < 0.001), with an area under the curve of 0.716. Risk Score for Pancreaticoduodenectomy When segregated by FRS risk groups, CR-POPFs occurred 1 2 in Low, Moderate and High Risk patients 6.6%, 12.9% and Benjamin C. Miller , John D. Christein , 28.6% of the time respectively (fi gure). Clinical outcomes 3 1 4 Stephen W. Behrman , Jeffrey A. Drebin , Wande B. Pratt , Mark including complications, length of stay, and readmission P. Callery4, Charles M. Vollmer1 rates, also increased across risk groups (Table). 1. Hospital of the University of Pennsylvania, Philadelphia, PA; 2. University of Alabama, Birmingham Medical Center, Birmingham, AL; 3. University of Tennessee Health Science Center, Memphis, TN; 4. Beth Israel Deaconess Medical Center, Boston, MA Abstracts
BACKGROUND: Accurate prediction of postoperative Monday pancreatic fi stula (POPF) after pancreaticoduodenectomy (PD) would help tailor optimal intra- and postoperative management of this morbid complication. Distinct risk factors for ISGPF clinically relevant fi stulas (CR-POPF), previously identifi ed as small duct size, soft gland texture, high-risk pathology, and increased blood loss, are best dis- cerned intraoperatively. The Fistula Risk Score (FRS), a 10 point scale derived at a single institution, relies on weighted infl uence of these four variables and has been shown to effectively predict (area under the curve of 0.942) CR-POPF development and its consequences. External validation of this tool would confi rm its universal applicability. CONCLUSION: This multi-institutional experience con- METHODS: From 2001–2012, 594 PDs with pancreatojeju- fi rms the Fistula Risk Score as a valid tool for predicting the nostomy reconstruction were performed by four pancreatic development of CR-POPF in patients undergoing pancre- surgical specialists at three institutions. POPFs, when they aticoduodenectomy. Patients devoid of any risk factors did occurred, were graded by ISGPF standards as biochemical not develop a CR-POPF, and the rate of CR-POPF approxi- (Grade A) or clinically relevant (Grades B and C). The FRS mately doubles with each subsequent risk zone. The lower was calculated for each patient and clinical outcomes were value of the area under the curve in this analysis is attribut- evaluated across four discrete categories (Negligible Risk, able to the decreased rate of CR-POPF observed in the high 0 points; Low Risk, 1–2 points; Moderate Risk, 3–6 points; risk group (29% vs. 89% originally). This difference might High Risk, 7–10 points). Receiver operator curve analysis be ascribed to variations in operative technique, postopera- was performed to judge model validity. tive management styles, patient characteristics, and a larger sample size in the current study. Despite this, the FRS is RESULTS: 142 patients developed any sort of POPF, of validated as an accurate prediction tool, with widespread which 68 were CR-POPF (11.4% overall: 8.9% Grade B, applicability, which can be readily translated into common 2.5% Grade C). There were 21 overall deaths, six of which practice.
Risk Profi le (Model Score) Variable Negligible Risk (0 points) Low Risk (1–2 Points) Moderate Risk (3–6 Points) High Risk (7–10 Points) p-Value Patients, n (% Total) 63 (10.6) 166 (27.9) 302 (50.9) 63 (10.6) – POPF, n (%) 1 (1.6) 19 (11.4) 90 (29.8) 32 (50.8) <.001 ISGPF Classifi cation, n (%) No fi stula 62 (98.4) 147 (88.6) 212 (70.2) 31 (49.2) <.001 Grade A 1 (1.6) 8 (4.8) 51 (16.9) 14 (22.2) <.001 Grade B – 9 (5.4) 29 (9.6) 15 (23.8) <.001 Grade C – 2 (1.2) 10 (3.3) 3 (4.8) <.001 CR-POPF, n (%) – 11 (6.6) 39 (12.9) 18 (28.6) <.001 Any complication, n (%) 32 (50.8) 97 (58.4) 216 (71.5) 54 (85.6) <.001 Length of stay, median 8 9 10 11 <.001 Readmission, n (%) 10 (15.9) 24 (14.5) 51 (16.9) 21 (33.3) <.001
43 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
4:00 PM – 5:00 PM 308D QUICK SHOTS SESSION III
785 Preexisting comorbidities such as chronic renal insuffi - ciency and infl ammatory bowel need to be considered in Impact of Key Factors of Enhanced Recovery Pathway predicting adverse outcomes after minimally invasive and Preexisting Comorbidities on Complications and colorectal surgery, while ASA was not suffi cient as a risk- Length of Stay Following Colorectal Surgery adjustment factor. Marianne Huebner1,2, David W. Larson1, Robert R. Cima1, Elizabeth Habermann1 786 1. Surgery, Mayo Clinic, Rochester, MN; 2. Statistics, Michigan Impact of a Standardized Surgical Safety Checklist on State University, East Lansing, MI Operative Effi ciency, Direct Operative Cost and Patient BACKGROUND: Patient and case complexity infl uence Outcomes Following Laparoscopic Incisional Hernia colorectal surgery outcomes. Success using enhanced recov- Repair ery pathways (ERP) after surgery requires assessing both Claire L. Isbell, Rahila Essani, Harry T. Papaconstantinou patient-factors risk adjustment as well as compliance with pathway elements. Surgery, Scott & White Memorial Hospital, Temple, TX METHODS: During 2011, 535 minimally invasive colorec- INTRODUCTION: The Surgical Safety Checklist (SSC) has tal surgery patients enrolled in an ERP protocol at a single been introduced as a proven tool to signifi cantly improve institution were reviewed. Patient comorbidities at admis- patient safety and outcomes through effective communica- sion and compliance with key ERP elements were captured tion of the surgical team. It has been suggested that the SSC using billing data and prospectively-collected data, respec- can reduce healthcare associated cost through reduction tively. The association of American. in postoperative complications. However, the impact of a SSC on operating room (OR) cost is not known. This study Society of Anaesthesiologists Physical Status classifi cation compares outcomes and direct OR costs for laparoscopic (ASA), comorbidities, and ERP element compliance were incisional hernia repair (LIHR) before and after implemen- considered in logistic regression models to predict length tation of a standardized SSC. of stay (LOS). A prolonged LOS was defi ned as 9 days or longer. Competing risk models were used to examine the METHODS: In September 2010, our institution imple- impact of factors on in-hospital outcomes. Surgery was mented a standardized SSC. We retrospectively reviewed the initial state, discharge the endpoint, and occurrence of all patients that underwent LIHR at our hospital for 1-year complications a time-dependent intermediate state. prior (PRE) and 1-year after (POST) implementation. Demo- graphic data included age, sex, BMI, ASA score and previous RESULTS: Compliance with the ERP protocol diet and fl uid laparotomy. Measures of OR effi ciency and cost included management was 76%. Surgical complications occurred in operative times (OT), implant cost and total direct OR cost 16% of the patients, with Ileus being the most common (TDORC). There was no change in vendor contract pric- (12%), and 9% of the patients had a prolonged LOS. The ing during the timeframe of this study. Outcomes included majority of patients had at least one comorbidity, includ- length of hospital stay (LOS), and 30-day morbidity and ing infl ammatory bowel diseases (IBD, 36%), chronic renal mortality rates. Statistical analysis by student’s t-test and insuffi ciency (5%) heart disease (9%), diabetes (9%), or Fisher’s exact test where appropriate. COPD (11%). An ASA score 3 or 4 was present in in 19%. Chronic renal insuffi ciency, IBD, conversion to open, and RESULTS: A total of 154 patients were identifi ed; 79 PRE non-compliance with ERP diet/fl uid protocol were risk fac- and 75 POST group. There were no signifi cant differences tors for occurrence of complications (c-index = 0.74) and between groups for age (p = 0.9), gender (p = 0.7), BMI (p = prolonged length of stay (c-index = 0.78). Using ASA in 0.7), ASA score (p = 0.4) and previous laparotomy (p = 0.9). place of other comorbidities or excluding diet/fl uid compli- Although mean OT was 12 min shorter in the POST group, ance reduced the predictive value of the models (c-index the difference was not signifi cant (176.1 vs. 164.5; min p = 0.67 for complications and 0.70 for prolonged LOS). In a 0.2). However, was a signifi cant reduction in implant cost competing risk model chronic renal insuffi ciency, IBD, ($2081 vs. $879; p = 0.02) and TDORC ($3630 vs. $2463; p non-compliance with diet/fl uid ERP protocol were predic- = 0.03) in the POST group. There was no difference in LOS tors of a longer LOS. (1.5 vs.1.6 days p = 0.8), surgical site infection rate (2.5% vs. 4%; p=0.9, total complication rates (18.1% vs. 12.8%; p = CONCLUSION: In the era of ERP diet/fl uid management 0.4), readmission (3.8% vs. 9.3%; p = 0.28) and reoperation compliance leads to predictably earlier recognition and rates (3.8% vs. 5.3%; p = 0.9) between groups. There was treatment of complications and thus shorter LOS. one death reported in the study.
44 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
CONCLUSIONS: Our data indicate that implementation of METHOD: The capacity of CSC sphere formation was a standardized SSC is associated with a signifi cant reduction tested in a human HCC cell line, Hep3B (derived from in implant cost and TDORC for patients undergoing LIHR. an 8-year-old black patient). The CSC sphere condition We speculate that these cost savings are a direct result of an medium was the DMEM/F12 medium (1:1) supplied with 20 improvement in effective communication between surgical ng/ml of EGF and 20 ng/ml of bFGF. Flow cytometery were team members regarding special equipment and implant performed using CSCs surface antigens (including, CD133, needs for these complex cases. Further studies are required CD90, CD44, and EpCAM). To test the ability of CSCs for to determine the impact of our SSC on direct OR cost at an tumor formation, an orthotopic model was developed using institutional level. nude BALB-B/C. Tumor inoculation were performed using Hep3B cells and Hep3B CSCs at 2 × 106 per injection. The 787 mice were assessed for tumor formation at 4 weeks. Flow cytometery were performed in the cells isolated from tumor Molecular Predictors of Recurrent Hepatocellular
tissue to test CSCs surface antigens mentioned above. Abstracts Cancer: Role of Cancer Stem Cells Monday RESULTS: Sphere forming Hep3B cells were successfully Prejesh Philips, Xuanyi Li, Yan Li, Suping Li, induced and confi rmed to contain CSCs morphology. Com- Erik M. Dunki-Jacobs, Robert C. Martin pared to the regular cultured Hep3B cells, Hep3B sphere Surgical Oncology, University of Louisville, Louisville, KY cells demonstrated signifi cantly higher surface antigens. BACKGROUND: Recurrence rates after either resection of CD44 was expressed by 74.25% of CSC vs. 59.79% regular hepatocellular carcinoma (HCC) or liver transplantation Hep3B cells (difference of 24.46%, p value = 0.029). With occur in 25 to 75% of patients. HCC recurrence has been regards to EPCAM expression the CSC cells expressed 1.6% thought to be driven by cancer stem cells (CSCs). Under- versus 0.51% expressed by regular Hep 3b cells (p value = standing the role CSCs play in HCC recurrence will provide 0.02). This enhanced expression dropped down to near the important information to improve prognosis and better baseline when the Hep3B sphere re-cultured in standard defi ne adjuvant therapy. nutrient rich medium. CD44 and EPCAM expression was noted at 60.1% and 0.6%, which was not signifi cantly dif- AIM: To demonstrate that HCC cells can dedifferenti- ferent compared to regular Hep3b cells (p = 0.87 and 0.8) ate into CSCs, which contribute to HCC is an important but was signifi cantly lower compared to CSC cells (p = resource. 0.03 and 0.0.05). In the orthotpoic injection liver, Hep3B sphere demonstrated a signifi cantly higher tumor prolifera- tion rate compared to non-sphere Hep3B cells. The tumor weights are as follows: 389 mg ± 65 (Hep3B sphere) vs 94 mg ± 32 (Hep3B cells). CONCLUSIONS: Hep3B cells show the capacity of CSC induction in nutritionally stressed phase. Hep3B derived CSC can not only differentiate into HCC cells when sup- plied with nutrient rich medium, but also form tumor when inoculate into mouse liver. The study is on the way to investigate the signaling such as Wnt pathway to evaluate the clinical relevant biomarkers of CSCs in HCC patients.
Hep3B expression of CD 44 (sphere forming) left (FITC +ve 74.25%) versus control Hep3B cells right (59.79%).
45 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
788 789 Long-Term Quality of Life After Oesophagectomy for Surgery for Gallbladder Cancer in the US: Greater Need Cancer: Comparison of Cervical Versus Mediastinal for Radical Cholecystectomy and Maximal Lymph Anastomoses Node Clearance John M. Bennett1, Justin C. Wormald2, Marc Van Leuven1, Thuy B. Tran, Vijay G. Menon, Nicholas N. Nissen Michael P. Lewis1 Cedars-Sinai Medical Center, Los Angeles, CA 1. General Surgery, Norfolk and Norwich University Hospital, BACKGROUND: Gallbladder cancer (GBC) is an uncom- Norwich, Norfolk, United Kingdom; 2. Norwich Medical School, mon cancer with poor overall survival and frequent local Norwich, Norfolk, United Kingdom and metastatic relapse. GBC is often identifi ed incidentally BACKGROUND: With recent improvements in neoadju- after cholecystectomy, in which case reoperation with por- vant therapy and earlier diagnosis, long-term survival after tal lymph node dissection (LND) is frequently performed. oesophagectomy for adenocarcinoma is becoming more The value of LND both for its staging and therapeutic value frequent. With longer survival the quality of life (QOL) continues to be debated. In particular, the optimal extent of of patients post resection has thus become a greater prior- lymph node clearance is unclear. ity. There has been extensive debate focusing on the long METHODS: The Surveillance, Epidemiology, and End term effects of different sites for surgical anastomosis. We Results (SEER) database was queried for patients diagnosed aimed to examine if long-term post-oesophagectomy QOL with GBC. Overall survival was analyzed using the Kaplan- is affected by the site of the surgical anastomosis. Meier method and compared using Log rank testing. Cox METHOD: Following local ethics Committee approval proportional hazard modeling was used in multivariate QORTC C-30 and OG-25 QOL questionnaires were sent to analysis to identify predictors of survival using age, type patients who had survived post-oesophagectomy patients of surgery (simple vs radical cholecystectomy), adjuvant for greater than 3 years. The data was analysed in subsets treatment, stage, and number of lymph nodes examined. dependant on the site of oesophago-gastric anastomosis In addition, the contribution to survival of minimal LND – either thoracic or cervical. No patients in either group (min-LND; 1–3 LN removed) vs maximal LND (max-LND; underwent formal pyloroplasty. Data was analysed using >3 LN removed) was evaluated separately within each of the student’s t-test on SPSS statistical software. QORTC tumor Stages I-IIIB. Predictors that patients would undergo C-30 data was compared against the reference tables for max-LND were also evaluated. oesophageal cancer pre-treatment. RESULTS: A total of 12,962 patients with gallbladder can- RESULTS: A total of 60 patients responded (82%) with a cer were identifi ed. We included 11,113 patients without median time post-surgery of 6.1 years (range 3–12 years). distant metastases in our analysis. Multivariate analysis Cervical and thoracic anastomosis subgroups were equiva- of Stage I-IIIB GBC demonstrated that strong predictors lent in terms of age at time of surgery, time post op and can- of improved survival are early tumor status, negative LNs cer stage. No signifi cant QOL difference was noted between and adjuvant treatments (p < 0.001 for all). Predictors of cervical or chest anastomosis groups for any functional or worse overall survival are simple cholecystectomy (HR 1.74; symptom score, especially focusing on dysphagia (OG25, p < 0.001) and minimal LND (HR 2.56; p < 0.001). When p = 0.24), odynophagia (OG 25, p = 0.68) and swallowing evaluated by tumor stage, the extent of LN removal did not problems (OG25, p = 0.73). The patients’ overall general signifi cantly affect mean overall survival of Stage I GBC. health (QL2) score was 72.0 ± 19.43 compared with 71.2 ± In contrast, the extent of LND for GBC was signifi cantly 22.4 for the general population (QORTC reference tables). associated with improved overall survival of patients with Functional indices and symptom scores are improved for Stage II (T2N0), Stage IIIA (T3N0) and Stage IIIB (T1-3,N1) our cohort compared to the QORTC oesophageal cancer disease (p < 0.01 for all; see Table). reference baseline except symptom scores for diarrhoea and Comparison of Survival by Stage and Extent of Lymph Node dyspnoea which worsen post-operatively. Dissection DISCUSSION: There is no signifi cant difference in QOL scores between oesophagectomy patients with cervical or Stage Minimal LND – OS Maximal LND – OS p Value thoracic anastomosis long term post-surgery. QOL in long- I 50.5 mo 58.2 mo 0.196 term survivors after oesophagectomy compares favourably II 40.8 mo 57.4 mo <0.001 with QORTC reference data for both pre-treatment oesoph- IIIA 27.2 mo 43.2 mo 0.01 ageal cancer and baseline general population data in our IIIB 21.2 mo 31.9 mo <0.001 cohort, possibly due to the absence of pylorplasty. Further OS – Overall Survival, LND – Lymph Node dissection, mo – months. prospective QOL data collection is required to elucidate any Overall Survival of Stage II Disease Related to Extent of Lymph Node Dissection long-term differences between the two anastomosis sites.
46 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
790 Factors Associated with Healing of the Perineal Wound After Proctectomy in Crohn’s Disease Patients Faisel Elagili, Scott A. Strong, Pokala R. Kiran Colorectal surgery, Cleveland Clinic, Cleveland, OH BACKGROUND: For patients with Crohn’s disease (CD) undergoing proctectomy with or without colectomy and end ileostomy, while the procedure may relieve debilitat- ing symptoms and improve quality of life, postoperative perineal wound complications can be a persistent problem. The aim of our study is to assess perineal wound healing Abstracts
in patients with CD who undergo proctectomy or procto- Monday colectomy with end ileostomy and to evaluate the infl uence of various factors including types of perineal dissection on eventual wound healing. PATIENTS AND METHODS: Data for patients with CD who underwent total proctocolectomy or proctectomy with end ileostomy from 1995–2012 were reviewed. Peri- neal wound healing was classifi ed as follows: Early healing The likelihood that max-LND was performed was predicted (within 12 weeks), delayed healing (between 12 weeks and by younger age (p < 0.001), as well as T4 tumor status com- 6 months) and persistent sinus (unhealed >6 months). pared to T1-3 (p < .001). No LNs were recovered in 74%, RESULT: For 139 patients (63% female), mean age 41 ± 60% and 50% of patients with Stage I, II and III disease 13 years, perineal wound healed by 12 weeks in 74 (53%) respectively. patients, delayed healing occurred in 36 (25.9%) patients, CONCLUSIONS: This is the largest population-based study and in 29 (20.9%) patients, there was a persistent sinus. Per- of patients with GBC in the literature. Not surprisingly, ineal dissection was either intersphincteric or extrasphinc- early tumor stage and adjuvant therapy correlate with sur- teric depending upon the extent of perineal Crohn’s/sepsis, vival. In addition we found radical cholecystectomy and extensive disease managed by leaving the wounds open. maximal LND correlate with survival even in node negative Factors associated with a signifi cantly greater risk for peri- patients (Stage II and IIIA). Our results support an approach neal sinus were age (p < 0.001), surgical management of of radical cholecystectomy and extensive LN dissection for perineal wound by open drainage (p = 0.04), high fi stula Stage II and III disease, and also suggest that many patients (p=0.01), preoperative perineal sepsis/disease (p = 0.001) in the US with GBC are currently surgically undertreated. and smoking at time of surgery (p = 0.03). On multivari- These patients may not be receiving the best chance for ate analysis, the only factor associated with delayed healing control of an otherwise diffi cult disease. Finally, the finding and persistent sinus was preoperative perineal sepsis and that LND benefi ts patients with N0 disease raises questions disease (P = 0.001). about current methods of LN analysis. CONCLUSION: The perineal wound after proctectomy or proctocolectomy for CD is associated with poor healing and poses a particular challenge for patients with exten- sive preoperative perineal disease or sepsis. These fi ndings support a preoperative discussion that examines potential outcomes and the consideration of measures such as the initial creation of a defunctioning ostomy, control/drain- age of local sepsis and appropriate medical treatment prior to proctectomy in CD patients considered to be at high risk for perineal wound problems.
47 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
791 CONCLUSIONS: Patients with increased esophageal acid exposure at both the proximal and distal probes tended to Dual Probe pH Monitoring Is Not Useful in Patients have more severe refl ux disease with hernias larger than with Refl ux Disease and Respiratory Symptoms 3 cm in size, erosive esophagitis and Barrett’s esophagus. Stephanie G. Worrell, Steven R. DeMeester, Evan T. Alicuben, Isolated abnormal proximal acid exposure was uncommon, Christina L. Greene, Daniel S. Oh, Jeffrey A. Hagen and was not associated with different symptoms. These Keck School of Medicine of Univeristy of Southern California, fi ndings suggest that dual probe pH monitoring does not Los Angeles, CA signifi cantly improve the ability to detect patients with refl ux disease that might be related to respiratory symptoms INTRODUCTION: The etiology of respiratory symptoms compared to standard pH monitoring in the distal esopha- such as cough, hoarseness, and asthma is often multi-fac- gus. Consequently, there is little added benefi t to monitor torial. Gastroesophageal refl ux can cause or contribute to patients with a dual probe pH system. these symptoms. Typically pH monitoring is done to iden- tify patients in whom refl ux may be related to respiratory symptoms. A dual probe study has been recommended for 792 these patients based on prior studies showing that approxi- Hospital Center Effect for Laparoscopic Colectomy mately 17% will have abnormal acid exposure only at the Among Elderly Stage I-III Colon Cancer Patients proximal probe. The aim of this study was to determine Zhiyuan Zheng2, Nader Hanna1, Eberechukwu Onukwugha2, the frequency of isolated abnormal proximal acid exposure, Kaloyan A. Bikov2, C. Daniel Mullins2 and to evaluate symptoms and endoscopic fi ndings associ- ated with distal versus proximal acid refl ux. 1. Surgery, University of Maryland School of Medicine, Baltimore, MD; 2. Pharmaceutical Health Services Research Department, METHODS: We reviewed the records of all patients that University of Maryland School of medicine, Baltimore, MD had dual probe pH testing from January 1999 to Novem- ber 2012. Only patients with complete foregut evaluation OBJECTIVE: To investigate hospital level variation in including endoscopy, video esophagram, and motility were short-term laparoscopic colectomy outcomes among stage included. Increased esophageal acid exposure was defi ned I-III elderly colon cancer patients. as a DeMeester composite score of >14.76 in the distal probe BACKGROUND: Surgical outcomes are associated with and >16.4 in the proximal probe. Dual probe pH catheters patient and surgeon characteristics. If outcomes are also with sensors spaced 10, 15 or 18 cm apart were selected impacted by the specifi c hospital where the surgery occurs, such that the proximal probe would be as close as possible there is a hospital center effect (HCE). Previous studies of to the upper esophageal sphincter in each patient. laparoscopic colectomies focus on patient, provider and RESULTS: From 425 total patients 256 (60%) had increased hospital characteristics, ignoring potential HCE. esophageal acid exposure on dual probe pH testing. Pre- Subgroup Analyses for Hospital Center Effect on Short-Term senting symptoms in these patients were heartburn (73%), Outcomes of Laparoscopic Colectomy regurgitation (60%), cough (54%), hoarseness (50%), or asthma (24%). The location of abnormal refl ux was at the Not Affi liated distal probe only in 133 patients (31%), at the proximal High Volume Colorectal Affi liated with with Medical probe only in 11 patients (3%) and at both probes in 112 Hospitals (≥30) Surgeons Medical School School patients (26%). There was no signifi cant difference in the N of Hospitals 43 119 196* 281 prevalence of cough, hoarseness or asthma based on loca- N of Patients 1661 1020 2397 2220 tion of the abnormal acid exposure. Abnormal acid expo- MIRR for LOS 1.24 (<0.001) 1.62 (<0.001) 1.21 (<0.001) 1.46 (<0.001) sure at both the proximal and distal probes was most likely (P-value) to occur in patients with a hiatal hernia larger than 3 cm MOR for 1.96 (0.004) NA 2.09 (0.022) 1.63 (0.132) in size [Table]. In-hospital Mortality (P-value) * There are 12 hospitals who changed their medical school affi liation during the study period. Therefore, these hospitals appeared in both affi liated and not affi liated with medical schools. METHODS: The Surveillance, Epidemiology and End Results (SEER)-Medicare dataset was used to identify stage I-III colon cancer patients in 2003 to 2007 with laparo- scopic colectomies. Multilevel model regressions were uti- lized to study potential HCE for length of stay (LOS), 30-day re-hospitalization, and in-hospital mortality, adjusting for patient, surgeon and hospital level characteristics. To quan- tify the impact of HCE, we calculated median instantaneous rate ratio (MIRR) for LOS and median odds ratio (MOR) for in-hospital mortality and 30-day re-hospitalization. Sensitivity analyses were also conducted for high volume/ medical school affi liated hospitals and colorectal surgeons.
48 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
793 Usefulness of Bowel Sound Auscultation: A Prospective Evaluation Seth Felder, Zuri A. Murrell, Phillip Fleshner Surgery, Cedars-Sinai Medical Center, Los Angeles, CA BACKGROUND: Although the auscultation of bowel sounds is considered an essential component of an ade- quate physical examination, its clinical value remains largely unstudied and subjective. The aim of this study was to determine whether an accurate diagnosis of mechani- cal small bowel obstruction, postoperative ileus, or normal Abstracts
controls is possible based on bowel sound characteristics. Monday METHODS: Using an electronic stethoscope (3M Littmann Model 3200) with sound amplifi cation capabilities and a computer Bluetooth interface, bowel sounds of healthy vol- unteers (n = 10), patients with a CT diagnosed mechanical small bowel obstruction demonstrated by a transition point and confi rmed at surgery (n = 10), and patients with post- operative ileus (n = 10) were recorded for 30 seconds while lying in the supine position. The bowel sounds were trans- ferred into a computer and then replayed randomly through high defi nition speakers to study physicians blinded to the clinical scenario. A total of 45 recordings were played con- secutively, with 15 of the recordings duplicated. A survey was taken just prior to the recording playback assessing each physician’s perceived level of expertise interpreting bowel sounds. Study physicians were instructed to catego- rize the patient recording as normal, obstructed, ileus, or not sure. RESULTS: Study physicians (n = 28) included 4 medical students on a surgical service, 8 surgical interns, 4 senior surgical residents, and 12 surgical attendings. Most par- ticipants (64%) stated they rarely listened to bowel sounds in their training or clinical practice. Almost all (96%) Joint impact of HCE and selected patient level characteristics on LOS (A) responded they knew what normal bowel sounds should and in-hospital mortality (B). sound like, but were less confi dent in what obstructive bowel sounds should sound like (71%). Study participants correctly diagnosed the clinical scenario in a median of RESULTS: The multilevel analyses based on 4,617 patients 11 (range, 5–16) of the 45 recordings (23%). A median of from 465 hospitals documented signifi cant HCEs for LOS 2 (range, 0–22) responses were ‘not sure.’ Normal bowel (MIRR = 1.36; p < .001) and in-hospital mortality (MOR = sounds were correctly identifi ed 29% of the time with a 1.72; p = 0.037), but no HCE for 30-day re-hospitalization. positive predictive value (PPV) of 22%. Obstructive bowel For patients with CCI = 3+, MIRR rose to 2.27 for LOS and sounds were correctly identifi ed 24% of the time with a PPV MOR rose to 6.87 for in-hospital mortality. The sensitivity of 26%. Postoperative ileus was correctly identifi ed 20% of analyses confi rmed our fi ndings. HCE was signifi cant for the time with a PPV of 39%. No difference was found in LOS in all subgroup analyses, and was signifi cant for in- diagnostic accuracy between levels of training. For par- hospital mortality for high volume/medical school affi li- ticipants responding he/she was ‘always able’ to identify ated hospitals. normal or obstructive bowel sounds if present, accuracy of CONCLUSION: HCE is an important source of variation diagnosis was 32% and 29%, respectively. Fixed-marginal for laparoscopic colectomy short-term outcomes, and it is inter-observer kappa value was only 0.17. still signifi cant when patient, provider and hospital level CONCLUSION: Auscultation of bowel sounds is not a use- characteristics are adjusted. HEC exists for both LOS and in- ful clinical practice when trying to differentiate normal hospital mortality. The fi ndings are robust to high volume/ versus ileus versus obstruction. Based on our results, the lis- medical school affi liated hospitals and colorectal surgeons. tener usually arrives at an incorrect diagnosis, and the low HCE is a potential area to improve the quality of care for inter-observer agreement further suggests the inaccuracy of stage I-III laparoscopic colon cancer patients. utilizing bowel sounds for clinical purposes. Routinely lis- tening to bowel sounds should be abandoned.
49 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Tuesday, May 21, 2013 8:00 AM – 9:30 AM 308D PLENARY SESSION VI
858 0.007) and MODS manifestations (P = 0.005). These fi nd- ings held in the subgroup with biliary bacteria; and HDL Plasma Lipids and Biliary Infections: Decreased Levels independently correlated with illness severity even among of HDL Are Associated with More Severe Biliary patients with bacterial-laden CBD stones. Infections: A Multivariate Analysis Lygia Stewart1,2, Gary Jarvis3, Lawrence Way1 1. Surgery, UCSF, San Francisco, CA; 2. Surgery, SF VA Medical Center, San Francisco, CA; 3. Infectious Disease and Laparotory Medicine, SF VA Medical Center, San Francisco, CA We, and others, have previously identifi ed factors associated with severe biliary infections: biliary bacteria, age, sex, cho- ledocholithiasis, pigment stones. Recently we reported that a low/normal BMI was associated with more severe biliary infections. To understand this association, we studied cor- relations between plasma lipid levels and biliary infections. METHODS: 475 patients with gallstones were studied; 410 men, 65 women; average age 62 (range 17–104). Gallstones, bile, and blood (as applicable) were cultured, Stone type recorded. Illness severity was classifi ed as: none (no infl am- matory manifestations), SIRS (fever, leukocytosis), severe (abscess, cholangitis, empyema), or MODS (bacteremia, hypotension, organ failure). Using bivariate and multivari- ate analysis, we examined associations between lipid levels (HDL, LDL, Triglycerides – obtained prior to the acute ill- ness) and: BMI, biliary bacteria, bacteremia, gallstone type, illness severity. In the multivariate analysis we used factors associated with biliary infections (age, sex, biliary bacteria, CONCLUSION: This study demonstrates the importance pigment stones, choledocholithiasis). of plasma lipids to biliary infections. Increased LDL levels RESULTS: On bivariate analysis (Table), BMI correlated favored cholesterol (rather than pigment) gallstone forma- with HDL and Trig (P < 0.003). LDL inversely correlated tion, and were more often associated with a sterile biliary with biliary bacteria (P = 0.002) and pigment stones (P = tract. HDL levels, however, seemed to dictate the course 0.0001); while HDL and LDL inversely correlated with bac- of the biliary infection; patients with low HDL levels had teremia (P < 0.03). HDL and LDL inversely correlated with signifi cantly more severe biliary infections. Even among increasing infection severity (P < 0.02) and MODS mani- patients with bacterial-laden CBD stones, those with high festations (P < 0.006). On Multivariate analysis of factors HDL levels had a mild illness, while those with low HDL associated with pigment stones and biliary bacteria, LDL levels more often had MODS manifestations. This data sug- inversely correlated with pigment stones (P = 0.001) and gest that HDL may be protective in biliary infections; and biliary bacteria (P = 0.017). Multivariate analysis of factors that patients with biliary infections and low HDL levels associated with biliary infection severity revealed HDL as may need a more aggressive clinical approach. This is the the most important lipid (Figure); only HDL inversely inde- fi rst study to demonstrate the importance of plasma lipid pendently correlated with biliary infection severity (P = levels to the severity of biliary infections.
50 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Bivariate Analysis: Lipids and Biliary Infections
HDL LDL Triglycerides N mg/dl P Value mg/dl P value mg/dl P Value Stone Type Pigment 244 42 NS 96 0.0001 156 NS Cholesterol or Mixed Chol 231 43 120 164 Biliary Bacteria Bacteria 238 42 NS 101 0.002 150 0.046 Sterile 237 43 116 174 Bacteremia Bacteremia 39 36 0.017 85 0.006 131 NS Sterile blood 436 43 109 163 MODS fi ndings (pts with biliary bacteria) Present 72 35 0.0001 89 0.006 149 NS None 166 45 105 150 MODS fi ndings (pts with bacterial-laden CBD stones) Present 42 33 0.005 86 0.074 153 NS None 65 46 103 151
6 859 in mice have been diffi cult to establish due to the techni- cal diffi culty and high mortality associated with esophageal Laparoscopic-Assisted ERCP After Gastric Bypass surgery in such small animals. The limited mouse models Surgery for Choledocholithiasis that have been described lack adequate molecular charac- Jonathan Carter, Jennifer Kaplan, Steve Elliott, terization to confi rm the development of Barrett’s metapla- Stanley J. Rogers, John P. Cello sia. We now describe the development and characterization Abstracts
Department of Surgery, UCSF, San Francisco, CA of Barrett’s-like columnar metaplasia in mice that have Tuesday refl ux esophagitis induced by esophagojejunostomy (EJ). Gastric bypass is common in the United States and is often performed without synchronous cholecystectomy. The METHODS: To induce refl ux, we performed EJ in twenty resultant rapid weight loss can lead to gallstones in up C57Bl/6 mice weighing 15 to 33 grams. At various time to 40% of patients, some of whom will develop choledo- points thereafter, the distal esophagus was removed, par- cholithiasis. Treatment of choledocholithiasis after gastric affi n-embedded, sectioned, and mounted on slides, which bypass is problematic because the long Roux limb is impos- were stained with H&E and with Alcian blue. Immuno- sible to navigate with a standard side-viewing endoscope histochemistry was performed to determine expression of passed through the mouth. Sox-9 (a columnar cell transcription factor expressed in human Barrett’s metaplasia) and the columnar cell cyto- We demonstrate a simple technique to perform ERCP after keratin (CK) 18. CK14 (an esophageal squamous cell cyto- gastric bypass by laparoscopically providing access to the keratin) was used as a control. We evaluated the specimens gastric remnant. The purpose is to familiarize surgeons and for squamous basal cell and papillary hyperplasia typical of endoscopists with the technique, as many will encounter refl ux esophagitis, as well as for columnar metaplasia. patients such as this in everyday practice. RESULTS: Procedural mortality was 40% for the fi rst 10 860 animals, but dropped to 20% for the next 10 animals. At 13 weeks after EJ, erosive esophagitis with prominent Development and Characterization of a Surgical, Mouse squamous basal cell and papillary hyperplasia was pres- Model of Refl ux Esophagitis and Barrett’s Esophagus ent in all animals. Columnar metaplasia, with goblet Thai H. Pham1, David H. Wang2, Robert M. Genta3, cells that stained with Alcian blue, developed by week 34. Rhonda F. Souza2, Stuart J. Spechler2 The columnar metaplasia expressed CK18, but not CK14. 1. Surgery, North Texas VAMC; UT Southwestern Medical Center, Intense expression of Sox-9 was detected in areas of colum- Dallas, TX; 2. Medicine, North Texas VAMC; UT Southwestern nar metaplasia. In the squamous epithelium close to the EJ anastomosis, furthermore, Sox-9 expression was seen in Medical Center, Dallas, TX; 3. Pathology, North Texas VAMC; UT scattered basal cells, whereas squamous epithelium further Southwestern Medical Center, Dallas, TX from the anastomosis did not exhibit Sox-9 expression. INTRODUCTION: To study the molecular mechanisms CONCLUSIONS: EJ can be performed successfully in underlying how refl ux esophagitis causes Barrett’s metapla- C57Bl/6 mice, causing refl ux esophagitis and, later, goblet sia, an appropriate animal model is desirable. A number of cell-containing columnar metaplasia that expresses CK18 surgical, rat models of GERD and Barrett’s esophagus are and Sox-9. These data suggest that this surgical, mouse available, but genetic engineering of rats is not accom- model recapitulates the phenotypic and molecular changes plished readily. In contrast, constitutive and conditional seen in human Barrett’s esophagus. Thus, we have estab- transgenic mice as well as knockout allele mice can be lished a relevant and genetically-modifi able model for engineered readily and, therefore, mouse models would be studying the molecular pathogenesis of Barrett’s esophagus. highly advantageous for studying the molecular pathogen- esis of GERD and Barrett’s esophagus. Surgical refl ux models
51 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
(normal, smooth shift, unilateral narrowing, bilateral nar- 861 rowing without collaterals, bilateral narrowing with collat- A Standardized Radiographic Assessment of the Tumor- erals). Findings were correlated to the need for venous Vein Interface Predicts the Need for Venous Resection resection at pancreatectomy and to the presence of histo- logic venous invasion. and the Presence of Histologic Venous Invasion in Borderline Resectable Pancreatic Cancer RESULTS: 266 patients underwent pancreaticoduodenec- Hop S. Tran Cao1, Aparna Balachandran2, Huamin Wang3, tomy and met inclusion criteria, of whom 99 required con- comitant resection of the SMV-PV. Greatest sensitivity for Jason B. Fleming1, Jeffrey E. Lee1, Peter W. Pisters1, Matthew Katz1 predicting SMV-PV resection was achieved by an interface 1. Surgical Oncology, U.T. MD Anderson Cancer Center, Houston, threshold of abutment (sensitivity 91.9%, negative predic- TX; 2. Diagnostic Radiology, U.T. MD Anderson Cancer Center, tive value 87.9%), whereas excellent specifi city was reached Houston, TX; 3. Pathology, U.T. MD Anderson Cancer Center, with a threshold of encasement (97.6%, positive predic- Houston, TX tive value 89.7%). Among patients who underwent venous BACKGROUND: Venous resection may be required to resection, vessel encasement was associated with a 78.3% achieve complete resection of pancreatic cancers (PC). We rate of histologic SMV-PV invasion; this rate increased to have previously shown that histologic invasion of the supe- 90% when the vein was occluded. The Ishikawa system, rior mesenteric vein-portal vein (SMV-PV) is associated with while more detailed, offered no advantage in predicting the poor prognosis following resection. Using high-defi nition need for SMV-PV resection and was less accurate in predict- multidetector computed tomography (CT), we sought to ing histologic venous invasion. Subset analyses performed evaluate the ability of two commonly-used sets of radio- for patients who received neoadjuvant chemoradiation and graphic criteria to predict the need for SMV-PV resection for those who did not yielded similar fi ndings. at pancreatectomy and the histologic presence of SMV-PV CONCLUSIONS: A simple radiographic classifi cation invasion. system that categorizes the extent of the tumor-SMV-PV METHODS: All patients who underwent pancreaticoduo- interface accurately predicts the need for SMV-PV resec- denectomy for PC between 2004 and 2011 at the authors’ tion at pancreatectomy, and correlates with the pathologic institution were identifi ed. Preoperative pancreatic protocol involvement of the resected vein. To assist in treatment CT images were re-reviewed to characterize the interface planning, a standardized description of this anatomic rela- between the tumor and SMV-PV (no interface, abutment tionship should be routinely performed for patients with [≤180 degrees], encasement [>180 degrees], occlusion) and borderline resectable tumors. the appearance of the SMV-PV using Ishikawa criteria
Correlation of Radiographic Assessment of Tumor-Vessel Relationship to Surgical and Pathologic Outcomes
Tumor-Vessel Interface* Vessel Appearance** Radiographic-Surgical Correlation CT cut-off 0 1 2 3 1 2 3 4 5 Number of patients 266 200 39 11 266 166 96 30 29 Sensitivity N/A 91.9 35.4 10.1 N/ A 85.9 66.7 27.3 27.3 Specifi city N/A 34.7 97.6 99.4 N/A 51.5 82.0 98.2 98.8 PPV N/A 45.5 89.7 90.9 N/A 51.2 68.8 90.0 93.1 NPV N/A 87.9 71.8 65.1 N/A 86.0 80.6 69.5 69.6 Accuracy N/A 56.0 74.4 66.2 N/A 71.8 76.3 71.8 72.2 Radiographic-Pathologic Correlation CT value 0 1 2 3 1 2 3 4 5 Number of venous resection 8 56 25 10 14 19 39 0 27 Histologic venous invasion (%) 33.3 64.6 78.3 90.0 54.5 52.9 69.7 -- 84.6 CT – computed tomography; PPV – positive predictive value; NPV – negative predictive value. *Tumor-SMV-PV interface scale – 0: no interface, 1: abutment ( 180°), 2: encasement (>180°), 3: occlusion. **SMV-PV appearance based on the Ishikawa system –1: normal, 2: smooth shift, 3: unilateral narrowing, 4: bilateral narrowing without collaterals, 5: bilateral narrowing or occlusion with collaterals.
52 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
862 863 Depth of Submucosal Tumor Infi ltration and Its High Resolution Manometry Classifi cations for Relevance in Lymphatic Metastasis Formation for T1b Idiopathic Achalasia in Patients with Chagas Disease Squamous-Cell and Adenocarcinomas of the Esophagus Esophagopathy Michael F. Nentwich1, Katharina von Loga1,2, Matthias Reeh1, Fernando P. Vicentine1, Fernando A. Herbella1, Guido Sauter2, Thomas Rösch3, Jakob R. Izbicki1, Luciana C. Silva1, Marco E. Allaix2, Marco G. Patti2 Dean Bogoevski1 1. HSP – Unifesp, São Paulo, Brazil; 2. University of Chicago 1. General, Visceral and Thoracic Surgery, University Clinic Pritzker School of Medicine, CHicago, IL Hamburg-Eppendorf, Hamburg, Germany; 2. Department of BACKGROUND: Idiopatic achalasia (IA) and Chagas dis- Pathology, University Clinic Eppendorf, Hamburg, Germany; ease esophagopathy (CDE) share several similarities; how- 3. Clinic for Interdisciplinary Endoskopy, University Clinic ever, some differences between the 2 diseases have been Eppendorf, Hamburg, Germany noticed. The comparison between IA and CDE is important BACKGROUND: Surgical resection for early esophageal to evaluate if treatment options and their results can be carcinoma has been challenged by less invasive endoscopic accepted universally. High-resolution manometry (HRM) approaches. As lymph node involvement, one of the major has proved a better diagnostic tool compared to conven- factors infl uencing patients’ overall survival cannot be tional manometry. The study of IA patients with the aid of assessed by endoscopic resection, selecting patients in need HRM allowed the creation of new classifi cations of the dis- for surgical intervention according to their risk of lym- ease with apparent correlation with treatment outcomes, as phatic spread is mandatory. proposed by the Chicago and Rochester groups. The clinical application of HRM parameters in patients with CDE is still OBJECTIVE: The aim of this study was to evaluate submu-
elusive. This study aims to evaluate HRM classifi cations for Abstracts cosal layer thickness, depth of submucosal tumor infi ltra- idiopathic achalasia in patients with CDE. Tuesday tion and tumor length as well as lymphatic invasion in T1b esophageal carcinomas for its predictiveness on lymphatic METHODS: We studied 86 patients with achalasia: 45 metastasis formation. patients with CDE (54% females, mean age 55.8 ± 14.7 years) and 41 patients with IA (58% females, mean age 49.0 METHODS: Histopathological specimens following sur- ± 19 5 years). All patients underwent a HRM when Chicago gical resection for T1b esophageal carcinomas were re- and Rochester classifi cations for achalasia were applied and evaluated for overall submucosal layer thickness, depth a barium esophagram to measure esophageal dilatation. of submucosal tumor infi ltration, tumor length as well as lymphatic and vascular infi ltration. A ratio of overall sub- RESULTS: The Chicago classifi cation was present in IA: mucosal layer thickness and depth of submucosal tumor Chicago I: 32%, Chicago II: 66% and Chicago III: 2%; In infi ltration was calculated and this proportion of submuco- CDE: Chicago I: 49%, Chicago II: 51% and Chicago III: 0% sal invasion was used to form sub-categories either in thirds (p = 0.178). The Rochester classifi cation was present in IA: or in halfs of total submucosal gauge. Infl uence of submu- Rochester I: 2%, Rochester II: 66% and Rochester III: 32%; cosal invasion as well as tumor length on lymphatic metas- In CDE: Rochester I: 0%, Rochester II: 51% and Rochester tasis formation and overall survival was assessed. III: 49% (p = 0.178). CDE patients had more pronounced degrees of esophageal dilatation (p < 0.0001). The degree of RESULTS: A total of 67 Patients with pT1b tumors were esophageal dilatation did not correlate with neither classi- analyzed, including 36 adenocarcinomas (53.7%) and 31 fi cation (p = 0.2); however, an indirect correlation between squamous-cell carcinomas (46.3%). Lymph node involve- esophageal body pressure amplitude and the degree of ment was seen in 20.9% (14/67) patients. Overall mean esophageal dilatation was noticed (p = 0.001). In 9 (10%) thickness of submucosal layer was 5.07 mm (SD 1.53 mm). patients the HRM pattern changed during the test from Overall proportion of submucosal infi ltration was calculated Chicago I to II. as 64.79% (SD 29.2%). Comparison of overall proportion of submucosal infi ltration between patients with (62.81%, CONCLUSION: Our results show that: (a) HRM classifi ca- range 17–97%) and without (65.31%, range 2–99%) lymph tions for IA can be applied in patients with CDE and (b) node involvement did not show signifi cant differences (p = HRM classifi cations did not correlate with the degree of 0.698 Mann-Whitney-U). On log-regression models, only esophageal dilatation. The secondary fi ndings of our study the presence of lymphangioinvasion and tumor length suggest that HRM classifi cations may refl ect esophageal was signifi cantly associated with positive lymph node repletion and pressurization instead of muscular contrac- involvement. tion. The correlation between manometric fi ndings and treatment outcomes for CDE needs to be answered in a near CONCLUSION: As depth of submucosal tumor infi ltration future. did not correlate with the formation of lymph node metas- tases and in regard of the risk of lymphatic spread in these cases, surgical resection is warranted whenever the tumor invades the submucosal layer.
53 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
9:30 AM – 12:00 PM 308D PLENARY SESSION VII
910 6 911 Does Preoperative Imaging Accurately Predict Main Spleen Preserving Laparoscopic Distal Pancreatectomy Duct Involvement in Intraductal Papillary Mucinous for a Solid Pseudopapillary Tumor in a Male Patient Neoplasm (This Video Demonstrates Technical Details of a Rare Morgan R. Barron1, Joshua A. Waters1, Janak A. Parikh1, Tumor of the Pancreas in a Male Patient. Only Few John DeWitt2, Mohammad A. Al-Haddad2, Eugene P. Ceppa1, Cases Have Been Reported in Males) Michael G. House1, Nicholas J. Zyromski1, Attila Nakeeb1, Bestoun H. Ahmed, Reginald L. Griffi n, Ziad Awad, Henry A. Pitt1, C. Max Schmidt1 Carmine Volpe, Michael S. Nussbaum 1. Surgery, Indiana University School of Medicine, Indianapolis, Surgery, University of Florida College of Medicine/Jacksonville, IN; 2. Gastroenterology, Indiana University School of Medicine, Jacksonville, FL Indianapolis, IN A 31-y-old patient had a blunt abdominal trauma. CT scan OBJECTIVE: Main pancreatic duct (MPD) involvement showed an incidental tumor in the body of the pancreas. is a well-demonstrated risk factor for malignancy in intra- EUS-guided cytology was Solid pseudopapillary tumor. ductal papillary mucinous neoplasm (IPMN). Preoperative Patient in right semi-lateral position. Division of gastrocolic radiographic determination of IPMN type (main, mixed, or omentum. Release of splenic fl exure of the colon. Transec- branch) is relied upon heavily in preoperative oncologic tion of the body of the pancreas after separating splenic risk stratifi cation. We hypothesize that preoperative radio- vessels. Separation of the pancreas from the vessels. Extrac- graphic assessment of MPD involvement in IPMN is an tion of the specimen in a pouch. Operative time: 170 min- accurate predictor of pathologic MPD involvement. utes. Blood loss: 50 ml. Tolerated food on day 2. Discharged on day 4. Pathology: 6x5 cm tumor like FNA result with METHODS: Data regarding all patients undergoing resec- clear margins .In conclusion: Laparoscopic spleen preserv- tion for IPMN at a single, academic institution between ing approach is feasible in the management of this tumor. 1992 and 2012 were gathered prospectively. Retrospec- tive analysis of imaging, clinical, and pathologic data was undertaken. Preoperative classifi cation of IPMN type was 912 based on cross-sectional imaging (CT or MRI). High Fat Diet Enhances Villus Growth During RESULTS: Three-hundred and sixty four patients under- Adaptation After Massive Small Bowel Resection went resection for IPMN. Of these, 335 had adequate data Pamela M. Choi, Raphael C. Sun, Jun Guo, on both radiographic and pathologic parameters for com- Christopher R. Erwin, Brad Warner parison. Of 184 suspected branch duct (BD) IPMN, 35 Department of Pediatric Surgery, Washington University, (19%) demonstrated MPD involvement on fi nal pathology. St. Louis, MO Of 84 mixed-type (MT) IPMN 16 (19%) demonstrated no MPD involvement. Of 68 suspected main duct (MD) IPMN BACKGROUND: Adaptation is a compensatory process 13 (19%) demonstrated no MPD involvement. Of 35 of 184 following small bowel resection (SBR) that results in vil- (19%) that had a suspected BD IPMN but were found to lus growth and enhanced mucosal surface area. In prior have MPD involvement on pathology, 12 (34%) had inva- studies, High Fat Diet (HFD) had been shown to enhance sive carcinoma. Alternatively, in patients with suspected adaptation responses if fed immediately following SBR. The MD or MT IPMN who ultimately were found to have no purpose of this study was to determine if HFD could further main duct involvement on pathology 2 (7%) demonstrated enhance villus growth after resection-induced adaptation invasive carcinoma. had already taken place. CONCLUSION: In resected IPMN, MPD involvement has METHODS: C57/Bl6 mice, aged 6–8 weeks, underwent been demonstrated as an independent risk factor for inva- a 50% proximal SBR or sham operation (bowel transec- sive cancer. Preoperative radiographic IPMN type correlates tion with reanastomosis alone) and then provided a stan- with fi nal pathology in 81% of patients. In addition, risk of dard rodent liquid diet (LD) ad lib. After a typical period invasive carcinoma correlates with pathologic presence (or of adaptation (7 days), SBR and sham-operated mice were absence) of main duct involvement. Consequently, preop- randomized to receive either LD or HFD (42% kcal/fat) erative imaging for oncologic risk stratifi cation may over or for an additional 7 days. Mice were individual caged, and under weigh risk in up to one in fi ve patients. food intake and feces output were measured daily. Mice were then harvested, and small intestine was collected for analysis.
54 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
METHODS: Prospectively collected clinicopathologic and perioperative data on adult patients undergoing liver resec- tion between 1/1/2003–7/31/2011 were retrospectively reviewed to assess incidence of and risk factors for postop- erative VTE within 30 days. Risk factors for PP were ana- lyzed using multivariable logistic regression. RESULTS: Of 2198 patients undergoing hepatectomy, median age was 60 years, and 49% were female. Median BMI was 27, preoperative chemotherapy was given to 997 patients (45%), and a history of prior VTE was present in 67 patients (3%). Major hepatectomy (MH, defi ned as ≥4 seg- ments) was performed in 716 patients (33%) and another concomitant organ resection in 556 (25%). EBL was ≥600 cc in 580 patients (27%), and liv±er steatosis was noted in 142 (18%). Median peak INR within 7 days after surgery was 1.4 (peak INR ≥1.5 in 32%), and median platelet count nadir was 154k (platelet nadir <100k in 12%). PP was started on Figure 1: Villus Height Measurements after SBR or Sham Operations. day 0/1 (immediate) in 815 patients (37%), day 2–5 (early) RESULTS: There were no differences in caloric intake or in 481 (22%), and later or never (late/none) in 902 (41%). stool output between any of the groups. However, Sham Use of any (immediate or early) PP was less common with mice had increased weight gain compared to SBR mice MH (50% vs 63%, P < 0.001), EBL ≥ 600 cc (54% vs 61%, P independent of diet. As shown in Figure 1, adaptation = 0.002), and peak INR ≥ 1.5 (54% vs 62%, P = 0.001). VTE
occurred in both SBR groups, however the SBR/HFD had occurred in 57 patients (overall: 2.6%; immediate: 2.2%; Abstracts signifi cantly increased villus height compared to SBR/LD. early: 1.9%; late/none: 3.3%; P = 0.2). VTE was associated Tuesday When compared to their sham counterparts, there was a with age ≥60 (3.9% vs 1.3%, P < 0.001), MH (4.2% vs 1.8%, 102.3% increase in villus height in the HFD group com- P = 0.001), EBL ≥ 600 cc (4.8% vs 1.7%, P < 0.001), and peak pared to only 42.6% in the LD group. Real-Time PCR was INR ≥ 1.5 (5.2% vs 1.5%, P < 0.001), but not gender, BMI, performed from mRNA of isolated intestinal villus cells, preoperative chemotherapy, history of VTE, other organ and CD36 expression was markedly elevated after high fat resection, liver steatosis, or nadir platelet count <100k (all diet (greater than 50-fold) in the SBR/HFD group compared P > 0.05). There was no signifi cant time trend in VTE inci- with SBR/LD mice. dence. On multivariable analysis, age, EBL, and peak INR remained signifi cant predictors of VTE (Table). CONCLUSION: While a week-long exposure to increased enteral fat alone did not affect villus morphology in sham- CONCLUSIONS: Counterintuitively, higher INR, but not operated mice, HFD signifi cantly increased villus growth in use of postoperative PP, was associated with VTE within 30 the setting of resection-induced adaptation. These data sup- days after hepatectomy. INR alone may not be an accurate port the clinical utility of enteral fat in augmenting adap- indicator of coagulation status after hepatectomy. The role tation responses in patients who have been subjected to of PP after hepatectomy requires prospective validation. massive SBR. Increased expression of CD36 suggests a pos- Multivariable Logistic Regression Analysis of Risk Factors for VTE sible mechanistic role in dietary fat metabolism and villus growth in the setting of short gut syndrome. Variable Odds Ratio 95% CI P-Value Age 60 years 3.06 1.65–5.70 0.001 913 Pharmacologic prophylaxis Pharmacologic Prophylaxis, Postoperative INR, and Late/None Ref. 0.4 Early (Day 2–5) 0.63 0.29–1.35 Risk of Venous Thromboembolism After Hepatectomy Immediate (Day 0/1) 0.78 0.42–1.45 Hari Nathan, Matthew J. Weiss, Ronald P. DeMatteo, Major hepatectomy 1.19 0.63–2.22 0.6 Peter J. Allen, T.P. Kingham, Yuman Fong, William R. Jarnagin, EBL 600 cc 2.09 1.18–3.69 0.01 Michael D’Angelica Peak INR 1.5 3.03 1.58–5.79 0.001 Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, MD Ref.: Referent INTRODUCTION: Pharmacologic prophylaxis (PP) against venous thromboembolism (VTE) is often withheld after hepatectomy due to bleeding risk or perceived coagulopa- thy related to INR and platelet count, but its role has been inadequately studied. We sought to characterize VTE risk and defi ne the role of PP after hepatectomy.
55 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Table 1: Outcomes of Patients with Curative Esophagectomy Alone 914 Compared to Those with Esophagectomyand Lung Resection. Data Clinical Signifi cance of Incidental Pulmonary Nodules Presented as Median (Range) and N (%). * = p < 0.05 in Esophageal Cancer Patients Esophagectomy Esophagectomy and 1,2 1,2 1,2 Amin Madani , Lorenzo E. Ferri , Jonathan Spicer , Only (E) Lung Resection (EL) 2 David S. Mulder (N = 275) (N = 33) 1. General Surgery, McGill University, Montreal, ON, Canada; Age 62 (43–78) 65 (24–91) 2. Thoracic Surgery, McGill University, Montreal, QC, Canada Gender (% M) 73% 76% Incidental pulmonary nodules are frequently identi- Smoking History 130 (47%) * 25 (77%) * fi ed during staging investigations for esophageal cancer Pulmonary Complications 70 (25%) 6 (19%) patients. However, the clinical signifi cance of such nodules Anastamosis Leak 26 (9%) 2 (6%) is unclear and may bias treatment decisions towards pallia- Overall Complications 151 (55%) 15 (45%) tive options. This study is aimed to determine the value of Estimated Blood Loss 310 mL (100–3500) 300 mL (150–2000) an aggressive surgical approach in patients with esophageal Length of Stay 11 days (5–185) 10 days (7–86) cancer and incidental pulmonary nodules. Operative Time 285 min (100–600) 310 min (220–510) From 2005–2012, a prospectively entered institutional Post-Operative Mortality 9 (3%) 1 (3%) clinical database of esophageal cancer patients was accessed to identify patients with incidental pulmonary nodules. Those patients who underwent combined esophageal and 915 lung resection (EL) were compared to those who had esoph- Internal Hernia After Laparoscopic Roux-en-Y Gastric agectomy alone (E) in terms of demographics, tumor char- Bypass acteristics and peri-operative outcomes. Fishers exact and Ayman Obeid1, David M. Breland1, Richard Stahl1, MWU-test determined signifi cance (*p < 0.05). Ronald H. Clements2, Jayleen M. Grams1 During the study period, 424 patients were treated for 1. Surgery, University of Alabama at Birmingham, Birmingham, esophageal cancer, of which 93 (22%) had lung nodules. AL; 2. Surgery, Vanderbilt University, Nashville, TN Of these, 29 (31%) were treated non-surgically either due to their poor performance status or extra-pulmonary distant INTRODUCTION: Although laparoscopic Roux-en-Y gas- metastasis on CT and/or PET. The remaining 64 patients tric bypass (LRYGB) has decreased morbidity compared to had no evidence of extra-pulmonary metastasis and under- the open approach, it was initially associated with a higher went neo-adjuvant therapy (35 (55%)) followed by curative rate of internal hernia (IH). This study investigated the en-bloc esophagectomy (with lung resection, 33 (50%), or impact of mesenteric defect closure on the rate and charac- without lung resection, 31 (47%) as per a tumor board con- teristics of IH after LRYGB. sensus). Of 33 lung resections, there were 27 benign lesions METHODS: Retrospective review was conducted on all (mostly granulomas or fi brotic scars), 4 primary stage I lung patients undergoing LRYGB from 2001–2011. Only patients cancers and 2 metastases (1 esophageal cancer and 1 renal who had all defects closed (DC) or all defects not closed cell carcinoma). Of the 31 patients with lung nodules who (DnC) were included. Patients with an incidentally iden- underwent curative esophagectomy without lung resec- tifi ed IH during another operation were excluded. Data tion, only 1 (3.2%) showed interval size increase on fol- collected included demographics, clinical presentation, low-up imaging (median 9 months (3–40)). A total of 308 operative details, and postoperative course. Data were ana- patients underwent a curative esophagectomy, of which lyzed using SPSS (version 16) statistical software. 33 had a combined esophagectomy and wedge lung resec- tion (EL) and 275 had an esophagectomy alone (E). There RESULTS: Of 1160 patients who underwent LRYGB from were no differences in age or gender, but the EL group had 2001–2011, 914 met inclusion criteria [DC = 663 (72.5%) more smokers (EL:25 (77%) vs E:130 (47%) *). There was no patients and DnC = 251 (27.5%)]. Median follow-up was difference in pulmonary complications, anastomotic leak, 24.3 (range 0.5–93.3) vs 31.7 months (range 0.5–131) in DC overall complications, operative time, blood loss, length-of- vs DnC, respectively (p < 0.0001). A total of 46 patients stay, or post-operative mortality (Table 1). (5%) developed a symptomatic IH [25 (3.8%) in DC vs 21 (8.4%) in DnC group, p = 0.005]. This remained statistically The presence of incidental pulmonary nodules in the signifi cant on multivariate analysis (p = 0.0098, OR 0.44; absence of extra-pulmonary metastases in esophageal can- 95% CI 0.24–0.82). Nineteen patients (42.2%) presented for cer patients are rarely distant metastases, and should not emergent or urgent repair and 26 (57.8%) for elective repair. bias caregivers towards palliative therapy. In addition, The most common symptom was chronic post-prandial should a nodule be of uncertain etiology, resection of lung abdominal pain (53.4%), followed by abdominal pain with nodules during the esophagectomy is safe. nausea ± vomiting (35.6%), acute abdominal pain ± nau- sea and vomiting (8.8%), and an acute abdomen (2.2%).
56 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Median time to presentation from LRYGB was 16.6 (range (10%), or high (13%). Positive margins for EMR, LR, and PD 3.1–71.9) vs 33.5 months (range 10–103) in the DC vs DnC were identifi ed in 66%, 29% and 3%, respectively. Median group, respectively (p < 0.001). At the time of IH repair there follow-up was 27 months. The 3 year recurrence-free sur- was no signifi cant difference in BMI or % EWL between vival (RFS) rate was 83% and there were no differences in the two groups. All patients underwent CT scan which RFS between the three different treatment groups. Tumors was consistent with IH in 26 patients (57.5%), suggestive were smaller in the EMR group (p = 0.005) and more likely in 7 (15.6%), showed small bowel obstruction in 4 (8.9%), to have a positive margin compared to the LR and PD group and was negative in 8 (17.8%). The majority of IH repairs (p < 0.001). In all patients, RFS was better in low grade were performed laparoscopically (86.7%) vs open (13.3%). tumors as compared to those that were high or intermedi- Intra-operatively, 71 herniation sites were identifi ed. In the ate (p = 0.04). Negative margin status and negative lymph DC group, there were 23 (67.6%) pseudo-Peterson’s and 11 nodes were not associated with better RFS. Morbidity after (32.4%) meso-mesoenteric defects. In the DnC group, there EMR, LR, and PD was 0%, 24%, and 41% respectively. were 5 (13.5%) mesocolic, 15 (40.5%) Peterson’s, 2 (5.4%) pseudo-Peterson’s, and 15 (40.5%) meso-mesenteric defects. Table I Median OR time was 104 minutes (range 75–180). Median EMR LR PD length of stay was 1 day (range 0.5–32). One patient who Variable (n = 12) (n = 35) (n = 30) p Value presented in extremis died after being hospitalized else- Tumor Size (cm) 0.6 + 0.5 1.8 + 1.4 1.9 + 1.1 0.005 where for 3 days with the incorrect diagnosis. One patient 18/28 Low Grade Tumor* 9/10 (90%) 28/33 (84%) NS had IH recurrence 11.5 and 14.2 months after initial repair. (64%) CONCLUSIONS: Complications of IH can be devastating Positive Resection Margin 8/12 (66%) 9/31 (29%) 1/30 (3%) <0.001 and closure of mesenteric defects during LRYGB signifi - 18/29 Positive Lymph Nodes NA 5/17 (29%) 0.03 cantly lowers IH rate. A high index of suspicion must be (62%)
maintained since symptoms may be nonspecifi c and imag- Recurrence 1/12 (8%) 6/35 (17%) 5/30 (17%) NS Abstracts Tuesday ing may be negative in nearly 20% of patients. *pathology specimens were unavailable for review of grade in 6 patients 916 CONCLUSIONS: EMR, LR, and PD are all effective treat- Tumor Grade, Not Extent of Resection, Is Associated ment approaches for duodenal neuroendocrine tumors. with Recurrence-Free Survival in Patients with Tumor grade is associated with recurrence-free survival but Duodenal Neuroendocrine Tumors not lymph node or margin status. When feasible, a less Brian Untch, Laura H. Tang, Keisha Bonner, Kevin K. Roggin, aggressive surgical approach to treat duodenal neuroendo- crine tumors should be considered. Michael D’Angelica, Ronald P. DeMatteo, William R. Jarnagin, T.P. Kingham 6 917 Surgery, Memorial Sloan-Kettering Cancer Center, New York, IL BACKGROUND: Duodenal neuroendocrine tumors are Laparoscopic Pancreas Sparing Segmental Resection of rare and few studies exist to guide surgical management. the Distal Duodenum for GIST Endoscopic mucosal resection (EMR), local duodenal resec- Robert Sung, Diana J. McPhee, Paresh C. Shah tion (LR), and pancreaticoduodenectomy (PD) are typically Lenox Hill Hospital, New York, NY performed as primary treatment. This study identifi es fac- This is a laparoscopic pancreas sparing, segmental resection tors associated with recurrence after resection. of the distal duodenum for a GIST. A 65 year old female METHODS: A retrospective, single institution review was presented with GI bleeding, the diagnosis and initial bleed- performed between 1987 and 2011 on patients with a patho- ing control were done endoscopically. Imaging confi rmed logic diagnosis of duodenal neuroendocrine tumor. Biopsy location and vascular supply. We begin with an extended and surgical specimens were independently reviewed by a Kocher maneuver to the ligament of Treitz. The tumor is pathologist. Tumor grade was assigned based on WHO 2010 identifi ed and the duodenum freed from the superior mes- criteria (KI-67 and/or mitoses per high power fi eld). enteric vessels. An extraserosal dissecton off the pancreas is RESULTS: Seventy-seven patients with a median age of done using ultrasonic shears. The jejunum is divided at the 60 had resectable duodenal neuroendocrine tumors. Based ligament, the duodenum divided just distal to the ampulla. on pathologic review, there were 9 somatostatinomas, 18 A two-layer handsewn anastomosis is created. Pathol- gastrinomas, and 49 not otherwise specifi ed. In the entire ogy demonstrated a 3.9 cm low-grade GIST with negative group, 12 underwent EMR, 35 had LR, and 30 underwent margins. PD (Table). Tumors were graded as low (77%), intermediate
57 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
918 919 Overexpressing TNF-alpha in Pancreatic Ductal Adenocarcinoma Cells and Fibroblasts Modifi es Does Resident Experience Affect Outcomes in Complex Cell Survival and Reduces Fatty Acid Synthesis via Abdominal Surgery? 1 1 1 Downregulation of Sterol Regulatory Element Binding Daniel Relles , Richard Burkhart , Michael J. Pucci , 2 1 1 Protein-1 and Activation of Acetyl COA Carboxylase Jocelyn Sendecki , Renee Tholey , Ross E. Drueding , 1 1 1 Mazhar Al-Zoubi, Galina Chipitsyna, Konrad Sarosiek, Patricia K. Sauter , Eugene P. Kennedy , Jordan M. Winter , 1 1 Christopher Y. Kang, Charles J. Yeo, Hwyda A. Arafat Harish Lavu , Charles J. Yeo Surgery, Thomas Jefferson University, Philadelphia, PA 1. Surgery, Thomas Jefferson University, Philadelphia, PA; 2. Biostatistics, Thomas Jefferson University, Philadelphia, PA INTRODUCTION: The effect of TNF- on pancreatic tumorigenesis is controversial due to the differential sig- BACKGROUND: For complex abdominal operations, the naling pathways initiated after binding its receptors TNFR1 infl uence of provider and hospital volume on surgical out- and TNFR2. TNFR1 activation by TNF- leads to cell apop- comes has been described. The impact of resident experi- tosis, whereas TNFR2 signaling is believed to be involved ence is less well understood. in cell survival through the activation of NF kappa B. TNF- METHODS: We reviewed perioperative outcomes after gene delivery has been suggested as a potentially useful pancreaticoduodenectomy (PD) at a single high-volume therapeutic approach to improve gemcitabine treatment center between 2006 and 2012. Resident participation and of pancreatic ductal adenocarcinoma (PDA), but its exact outcomes were collected in a prospectively maintained mechanism of action is not clearly understood. Although database. Resident experience was defi ned as post-graduate TNF- has been shown to increase the expression of the year (PGY) and number of PDs performed. lipogenesis promoting enzyme, fatty acid synthase (FAS) in RESULTS: Twenty-nine residents and four attending sur- liver steatosis, its impact on de novo lipogenesis in tumor geons completed 681 PDs. The overall complication rate cells has not been determined. In this study, we investigated was 44%; PD-specifi c complications (defi ned as pancreatic effect of TNF- on fatty acid synthase (FAS) in PDA cells and fi stula, delayed gastric emptying, bile leak, abscess, and in fi broblasts as part of the tumor micro-environment. wound infection) occurred in 28% and were signifi cantly METHODS: PDA cells (MIAPACA-2 and AsPC-1) and the more common when the fi rst assistant was a PGY 4 rather fi broblast cell line, hTERT-BJ were transfected with TNF than a PGY 5 or 6 (44% vs. 27%, p = 0.016). Logistic regres- gene by lentivirus-vector transduction. Control cells were sion demonstrated that as residents perform more cases, transfected with the empty vector. FAS mRNA and pro- PD-specifi c complications decrease (OR = 0.97, p < 0.01). tein were analyzed by real time PCR and Western blot- For a resident’s fi rst case, the predicted probability of a PD- ting, respectively. Total- and phospho-AMPK, total- and specifi c complication is 27%; this rate decreases to 19% by phospho-Acetyl CoA carboxylase (ACC), FAS, and LKB/ case 15 (Figure 1). STK11 were analyzed by Western immunoblotting. The effects of TNF- on sterol regulatory element binding pro- tein-1, SREBP-1, the transcription factor responsible for FAS transcription, LKB1/STK11 (a tumor suppressor and the established upstream regulator of AMPK) and ACC (the downstream target of AMPK and the rate-limiting enzyme of fatty acid synthesis) were evaluated by real time PCR. MTT and Wound healing assays were used to determine cell survival and migration, respectively. RESULTS: TNF- signifi cantly (P < 0.05) reduced PDA and fi broblast cell survival and migration. This was associ- ated with signifi cant reduction of FAS mRNA and protein expression levels in PDA cells (P = 0.02) but not the fi bro- blasts. Cells overexpressing TNF- also showed signifi cantly (p < 0.05) reduced SREBP-1 and ACC. Reduction of FAS by TNF-was inhibited when either SREBP-1 or ACC was knocked down by siRNA. No signifi cant differences were seen in AMPK phosphorylation in cells that overexpress CONCLUSIONS: We highlight the impact of resident TNF-. involvement in complex abdominal operations, demon- CONCLUSION: Our data demonstrate a previously unknown strating that as residents build experience with PD, patient involvement of TNF- in PDA and microenvironment lipo- outcomes improve. This is consistent with volume-outcome genesis and suggest that targeted introduction of intratu- relationships for attending physicians and high-volume mor TNF- can have the potential as a novel therapeutic hospitals. Complex cases provide unparalleled learning anti-lipogenic agent in human PDA. opportunities and remain an important component of sur- gical training. Maximizing resident repetitive exposure to complex surgical procedures benefi ts both the patient and the trainee.
58 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
POSTER SESSION DETAIL Printed as submitted by the authors. indicates a poster featured in a Poster Tour (ticketed session with complimentary but limited registration):
Sunday, May 19, 2013, 11:00 – 11:45 AM: Esophageal and Stomach
Monday, May 20, 2013, 11:00 – 11:45 AM: HPB
Tuesday, May 21, 2013, 11:00 – 11:45 AM: Small Bowel and Colon-Rectal indicates a Poster of Distinction. Sunday, May 19, 2013 Authors available at their posters to answer questions 12:00 PM – 2:00 PM; posters on display 8:00 AM – 5:00 PM.
12:00 PM – 2:00 PM West Hall A POSTER SESSION I (NON-CME)
Basic: Colon-Rectal completed the tasks using the same grasper but with scissors capable of shaft articulation up to 85 degrees. Outcomes were the standard FLS metrics of time and error (deviation as a percentage of total circle area). Instrument switches Su1794 Poster Abstracts between hands and TEM position adjustments were also Effect of Instrument Type on Transanal Endoscopic recorded. Overall group outcomes were compared using Microsurgery (TEM) Learning Curves t-tests. Mixed models were used to compare changes over Sunday Ezra N. Teitelbaum, Fahd O. Arafat, Brittany Lapin, the 10 runs. Anne M. Boller Northwestern University, Chicago, IL BACKGROUND: The transanal endoscopic microsurgery (TEM) proctoscope is used to resect benign and early-stage malignant rectal tumors, and has received recent attention as a potential platform for transanal natural orifi ce surgery. No study has evaluated the effectiveness of different instru- mentation types for TEM surgery. We tested whether learn- ing curves for surgical novices using a TEM proctoscope would be improved with the use of scissors with shaft artic- ulation. Additionally, we compared TEM and laparoscopic learning curves for the same task. METHODS: Medical students were randomized into three study groups: laparoscopic (LAP), TEM rigid (TEM-R), and TEM articulating (TEM-A). All groups completed the Fun- damentals of Laparoscopic Surgery (FLS) circle-cut task 10 times. The LAP group completed the task using an FLS box-trainer and a standard laparoscopic grasper and rigid laparoscopic scissors. The TEM-R group completed the task using the same instruments but through a TEM procto- Procedure time versus run number. scope within a custom TEM box-trainer. The TEM-A group
59 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
RESULTS: 33 subjects were randomized, 11 to each group. Expression intensity ranged from weak to strong, whereas Overall mean task time was shorter for the LAP group than VEGFR3 and EGFR showed only weak expression in esopha- both the TEM-R and TEM-A groups (200 vs. 362 and 417 geal samples. PDGFR expression was observed in esopha- seconds; p < .001) (See Figure). Subjects in all three groups geal and gastric samples. Specimen showed intermediate to showed improvement in time over the 10 runs (coeffi cient strong expression. PDGFR expression was seen in esopha- estimates –33, –62, and –61; p < .001 for each group). The geal, gastric and colonic samples. Intensities varied from LAP group made fewer errors than the TEM-R group, but weak to strong. KGFR was expressed in all intestinal samples not the TEM-A group (7% vs. 10% vs. 8% circle area; p = and revealed expression intensities from weak to strong. .02 LAP vs. TEM-R). The LAP and TEM-A group made fewer CONCLUSION: Our results reveal a high expression rate errors over time (coeffi cient estimates –0.6 and –0.7; p < .01 of growth factor receptors in the rat intestine and facilitate for both), whereas the TEM-R group had a trend towards methodic experimental studies on gastrointestinal anasto- increased errors (coeffi cient estimate 0.4; p = 0.1). The LAP motic healing in rat models using the positive impact of group switched instruments between hands during fewer specifi c growth factors. runs than both the TEM-R and TEM-A groups (9% vs. 30% vs. 24% of runs; p < .01). The TEM-A group adjusted the proctoscope position during fewer runs than the TEM-R Basic: Pancreas group (36% vs. 54% of runs; p = .01). CONCLUSIONS: A LAP approach results in faster circle-cut Su1796 task times than a TEM approach. While times were simi- lar, TEM procedures using articulating scissors may result in Histone Deacetylase Inhibition (HDAC) by Vorinostat fewer errors and less need to adjust proctoscope position as Sensitizes Pancreatic Cancer Cells to TRAIL Induced compared with rigid scissors. These results can be used to Cell Death develop specifi c curricula and training strategies for TEM surgery. Rohit Chugh, Vikas Dudeja, Osama Alsaied, Sulagna Banerjee, Veena Sangwan, Ashok Saluja, Selwyn M. Vickers Surgery, Basic and Translational Research Lab, Minneapolis, MN Basic: Esophageal INTRODUCTION: Pancreatic cancer is one of the most lethal human malignancies with fi ve-year survival of less Su1795 than 5% because of its resistance to most conventional che- motherapies like gemcitabine and other novel anti-cancer Growth Factor Receptors in the Gastrointestinal Tract therapies like TRAIL. Histone deacetylase (HDAC) inhibitors of the Rat: New Targets for Improved Anastomotic are a new and promising drug family with strong antican- Healing? cer activity. The aim of the current study was to evaluate Daniel G. Drescher, Laura Kulzer, Carl Christoph Schimanski, whether inhibition of histone deacetylase sensitizes pancre- atic cancer to TRAIL induced cell death. Hauke Lang, Ines Gockel University of Mainz, Mainz, Germany METHODS: Highly aggressive metastatic pancreatic cancer cell lines (S2VP10, Capan-1) were treated with the HDAC BACKGROUND: Anastomotic leakage after gastrointestinal inhibitor, Vorinostat (0–5μM), TRAIL (0–40 ng/ml) or a com- surgery is a signifi cant cause of morbidity and mortality. bination of Vorinostat and TRAIL for 12–72 h. The effect on In particular, esophagogastric and colorectal anastomoses cell viability was evaluated using a WST-8 cell viability assay are vulnerable to leakage, resulting in an increased need (Dojindo Labs), apoptosis (caspase 3, 8 and 9 activation) for reoperation and a high risk of subsequent anastomotic was evaluated using Caspase Glo assay kit (Promega). stenosis formation and fi stula. Studies in well-established experimental rodent models showed a positive impact RESULTS: HDAC inhibition markedly increased TRAIL of growth factors on anastomotic wound healing. So far, induced cell death in both pancreatic cancer cell lines eval- methodic investigations on the expression profi le of growth uated. Viability, data expressed as % of Control (untreated factor receptors in the gastrointestinal tract do not exist. cells), mean ± SEM. S2VP10 (48 h): Vorinostat (5μM) – 64.5 ± 0.1%, TRAIL (20 ng/ml) – 95.13 ± 0.825%, Vorinostat MATERIAL AND METHODS: We investigated the co- (5μM) + TRAIL (20 ng/ml) – 41 ± 0.8%. HDAC inhibition expression pattern of vascular growth factor receptor markedly augmented Caspase 3 activation in response to (VEGFR1-3), epidermal growth factor receptor (EGFR), plate- TRAIL. Caspase 3, data expressed as % of Control, mean let-derived growth factor receptor (PDGFR / ) and kerati- ± SEM. S2VP10 24 h: Vorinostat (5μM) – 206.1 ± 12.07%, nocyte growth factor receptor (KGFR) in the rat intestine. TRAIL (20 ng/ml) – 159.6 ± 1.2%, Vorinostat (5μM) + TRAIL Additional, IHC staining was applied for confi rmation of (20 ng/ml) – 2187.4 ± 77.62%. expression and analysis of growth factor receptor localisation. CONCLUSION: Inhibition of Histone deacetylases sensi- RESULTS: VEGFR1-3, EGFR, PDGFR / and KGFR expres- tizes pancreatic cancer cells to TRAIL induced apoptosis and sion in rat intestinal samples revealed varying transcription cell death. Combination of HDAC inhibition and TRAIL has intensities. VEGFR1 expression was observed in all samples immense potential to emerge as novel therapeutic strategy and varied from intermediate to strong. VEGFR2 expres- against pancreatic cancer. sion was found in esophageal, gastric and colonic samples.
60 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1797 Clinical: Biliary Concepts for the Periampullary Carcinoma Enigma from Clinico-Pathologic Analysis of 198 Patients Su1590 Peter Bronsert1, Ilona Kohler1, Martin Werner1, Frank Makowiec2, Laura H. Tang3, Ulrich T. Hopt2, Tobias Keck2, Laparoscopic Cholecystectomy in Patients on Ulrich F. Wellner2 Clopidogrel: Is It Safe? 1Pathological Institute, University of Freiburg, Freiburg, Germany; Richard Frazee, Stephen Abernathy 2Department of Surgery, University of Freiburg, Freiburg, Germany; Surgery, Scott & White Healthcare, Temple, TX 3Department of Pathology, Memorial Sloan-Kettering Cancer Center, BACKGROUND: Clopidogrel is a common antiplatelet New York, NY medication for patients with coronary, peripheral, and cere- brovascular disease. Many surgeons recommend cessation AIMS: Periampullary adenocarcinomas comprise pancre- of Clopidogrel prior to surgery to avoid bleeding complica- atic ductal (PDAC), distal bile duct (DBDAC), ampullary tions. Clopidogrel cessation however, is associated with an (AMPAC) and duodenal (DUOAC) adenocarcinoma. The increased risk of thrombotic events up to 3 months after epithelia of these anatomical structures share a common cessation. We review our experience with laparoscopic cho- embryologic origin from the foregut. While there seem to lecystectomy in patients who remained on Clopidogrel in be signifi cant differences regarding tumor biology, the clas- the perioperative period. sifi cation, grading, staging and treatment of these entities remains a matter of substantial debate. Due to the anatomi- METHODS: An IRB approved retrospective review of cal complexity of the periampullary region, there is still patients having laparoscopic cholecystectomy from 2008– considerable debate on how carcinomas and their precur- 2012 while on Clopidogrel was performed. Patient demo- sor lesions arising in this region should be classifi ed. Our graphics, indication for surgery, ASA score, operative time, study aimed at a detailed analysis of clinical, pathological conversion to open cholecystectomy, estimated blood loss, and immunohistochemical parameters for assessment of length of stay, morbidity and mortality were reviewed. tumor biology and identifi cation of prognostic factors after RESULTS: Thirty-one patients (13 women and 18 men) resection of periampullary carcinomas. underwent laparoscopic cholecystectomy while on Clopi- MATERIAL AND METHODS: 198 patients who had resec- dogrel. Sixteen were performed in an elective setting and tion of periampullary adenocarcinoma from 2001 to 2011 fi fteen were done as emergency/urgent operations. ASA were identifi ed. All tissue samples were processed by a stan- score was 2 in four patients (13%), 3 in twenty-four patients dardized protocol for pathological workup of pancreatoduo- (77%), and 4 in three patients (10%). Two patients (6%) denectomy specimen. Archived Hematoxylin-Eosin stained were converted from laparoscopic to open cholecystectomy Poster Abstracts slides were reevaluated by three experienced pathologists due to indistinct anatomy in acute cholecystitis. Average
for accuracy of diagnosis. For the growthpattern, three operating time was 71 minutes (27–129 minutes). Average Sunday typical subtypes were defi ned: intestinal, pancreatobiliary, estimated blood loss was 48 ml (1–300 ml). Morbidity was mixed intestinal-pancreatobiliary and Poorly-differentiated experienced in 22.5% of patients, and two 30 day mortali- carcinomas. Additionally for immunohistochemical sub- ties occurred (6%) secondary to cardiovascular complica- typing of the growth pattern CK7, CK20 and CDX2 staining tions. Length of stay averaged 3 days (outpatient – 15 days). were performed for each slide. Furthermore we established CONCLUSIONS: Laparoscopic cholecystectomy performed a modifi ed tumorgrading system. on patients taking Clopidogrel did not produce clinically RESULTS: 127 patients had PDAC, 39 had AMPAC, 23 signifi cant operative blood loss. Conversion to open cho- had DBDAC and nine had a DUOAC. The distribution of lecystectomy, morbidity and mortality were higher in this subtypes was signifi cantly different among the carcinoma patient population but appear to be more related to patient groups. Tumor location, histological subtype and grading comorbidities than the effects of the Clopidogrel. Recom- were highly signifi cant predictors of survival (p < 0.001). In mendations for Clopidogrel cessation prior to laparoscopic accordance, a high CK7 expression and a low CDX2 expres- cholecystectomy should be reconsidered. sion, which characterize PB differentiation, were signifi cant predictors of poor survival. Only histological subtype, grad- ing and lymph node ratio were found to represent indepen- dent predictors of survival in multivariate analysis. CONCLUSIONS: Our results demonstrate that there should be a change in the pathological management of making diagnoses in periampullary carcinomas. By multivariate analysis, traditional parameters as tumorlocation, TNM classifi cation lost their prominence as a source of prognos- ticating survival of periampullary carcinoma. Therefore, we recommend comprising the histological subtype and our adjusted histological grading for a better valuation of survival.
61 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1591 Su1592 Diagnostic Accuracy of Preoperative Multidetector- Elevated Perioperative Serum CA 19-9 Level Is an Row Computed Tomography Imaging in Predicting Independent Predictor of Poor Outcome in Patients Microscopic Curative Resection of Hepatobiliary with Resectable Cholangiocarcinoma and Pancreatic Malignancy: A Prospective Naru Kondo, Yoshiaki Murakami, Kenichiro Uemura, Multi-Institutional Study Takeshi Sudo, Yasushi Hashimoto, Hayato Sasaki, Kenjiro Okada, Kazuaki Shimada1, Yoshito Takeuchi2, Masaru Konishi3, Taijiro Sueda Tatsushi Kobayashi4, Akio Saiura5, Kiyoshi Matsueda6, Surgery, Hiroshima University, Hiroshima, Japan Tsuyoshi Sano7, Hideyuki Kanemoto8, Katsuhiko Uesaka8 BACKGROUND: Prognosis of cholangiocarcinoma is still 1Hepatobiliary and Pancreatic Surgery, National Cancer Center unsatisfactory, and identifi cation of predictive marker of Hospital, Tokyo, Japan; 2Diagnostic Radiology, National Cancer survival after surgical resection is important to establish the Center Hospital, Tokyo, Japan; 3Division of Digestive Surgery, perioperative therapeutic strategy for cholangiocarcinoma. National Cancer Center Hospital East, Kasiwa, Japan; 4Diagnostic Impact of perioerative serum carbohydrate antigen 19-9 Radiology, National Cancer Center Hospital East, Kasiwa, Japan; (CA19-9) levels on survival of patients with resectable chol- angiocarcinoma is still unclear. 5Gastrointestinal Surgery, Cancer Institute Hospital, Japanese Foundation for Cancer Research, Tokyo, Japan; 6Diagnostic Radiology, PURPOSE: The purpose of this study was to investi- Cancer Institute Hospital, Japanese Foundation for Cancer Research, gate whether perioerative serum CA19-9 levels can pre- Tokyo, Japan; 7Gastroenterological Surgery, Aichi Cancer Center dict survival of patients underwent surgical resection for cholangiocarcinoma. Hospital, Tokyo, Japan; 8Hepato-Biliary-Pancreatic Surgery, Shizuoka Cancer Center Hospital, Tokyo, Japan METHODS: One hundred and six patients with cholangio- carcinoma including 33 with intrahepatic, 48 with perihi- PURPOSE: To assess the accuracy of preoperative diagno- lar and 25 with distal cholangiocarcinoma who underwent sis of hepatobiliary and pancreatic malignancy with mul- surgical resection between 2002 and 2012 were eligible for tidetector-row computed tomography (MDCT) to predict this study. Preoperative biliary drainage was performed for microscopic curative resection. the patients with obstructive jaundice. Preoperative serum DESIGN AND SETTINGS: Prospective observational study CA19-9 levels were measured after biliary drainage, and of hepatobiliary and pancreatic malignancy resected postoperative serum CA19-9 levels were measured about 4 between November 2007 and December 2008, in 5 Cancer weeks after operation. The relationships between clinico- Center Hospitals in Japan. pathological factors including perioperative serum CA19-9 PARTICIPANTS: 271 consecutive patients with highly levels and overall survival (OS) were analyzed with univari- suspected and potential resectable hepatobiliary and pan- ate and multivariate analyses. creatic malignancy undergoing MDCT judged fi t for lapa- RESULTS: Preoperative CA19-9 levels were signifi cantly rotomy were studied. higher in patients with moderately and poorly differenti- MAIN OUTCOMES MEASURES: Sensitivity and specifi c- ated adenocarcinoma than in those with well differentiated ity of MDCT predicting a microscopic curative resection adenocarcinoma (P = 0.009), and in patients with UICC based on the histopathological examination of presence or stage I/II than those with III/IV (P = 0.008). In contrast, absence of tumors at the margin of the specimen. there was no signifi cant difference between postoperative CA19-9 and any other clinicopathological factors. Univari- RESULTS: 164 patients of 217 macroscopic resectable ate analysis revealed postoperative adjuvant chemotherapy patients (75.6%) with hepatobiliary and pancreatic malig- (P = 0.03), residual tumor factor status (P = 0.01), patho- nancy underwent microscopic curative resection. MDCT logical differentiation (P = 0.02), UICC pT stage (P = 0.009), predicted clear margin resections in 146 patients (89.0%). lymph node metastasis (P < 0.001) and UICC fi nal stage Sensitivity for prediction of microscopic curative resection (P = 0.001) were signifi cantly associated with OS. In addi- by MDCT in perihilar cholangiocarcinoma, gallbladder car- tion, differences in OS were signifi cant between groups cinoma, middle/lower bile duct carcinoma, and pancreatic divided on the basis of two preoperative CA19-9 cutoff val- carcinoma was 64.7% [CI,52.3–78.9%], 90.9% [CI,90.9– ues (37 and 200 U/ml), and three postoperative CA19-9 cut- 97.29%], 95.5% [CI,97.7–99.1%], and 89.7% [CI,86.3– off values (37, 100 and 200 U/ml). In multivariate analysis, 93.1%], respectively. On the other hand, specifi city was no postoperative adjuvant chemotherapy (odds ratio [OR], 30.8% [CI, 14.5–49.3%], 0%, 33.3% [CI, 14.4–42.4%], and 3.02: 95% confi dence interval [CI], 1.54–5.89; P = 0.001), 36.4% [CI, 21.5–51.6%], respectively. lymph node metastasis (OR, 3.96; 95% CI, 1.91–8.48; P < CONCLUSIONS: Expert radiologists in hepatobiliary and 0.001), preoperative CA19-9 (≥200 IU/ml) (OR, 2.27; 95% pancreatic disease could not predict microscopic curative CI, 1.10–4.61; P = 0.03) and postoperative CA19-9 (≥37 IU/ resection in patients with perihilar cholangiocarcinoma. ml) (OR, 6.88; 95% CI, 3.36–14.41; P < 0.001) were identi- Even if MDCT predict a possibility of surgical margin posi- fi ed as independent predictors for OS. tive resections, surgery seems to be not always contraindi- CONCLUSION: Perioperative serum CA19-9 levels predict cated in hepatobiliary and pancreatic malignancy, because the survival of patients with resectable cholangiocarci- the accurate preoperative diagnosis with MDCT has still noma, and they may contribute to establishment of new remained diffi cult. therapeutic strategy, as perioperative treatment can be opti- mized based on its value.
62 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1593 Su1594 Stapling the Cystic Duct During Laparoscopic Diagnostic Utility of Magnetic Resonance Cholecystectomy Results in Increased Rates of Cholangiopancreatography (MRCP) in Patients Unintended Post-Operative ERCP with Intermediate Probability of Cholidocholithiasis Irene Epelboym2, Florita Martin1, Megan Winner2, Hari P. Sayana, Dany Jacob, Mir Fahad Faisal, An-Lin Cheng, Zachary L. Gleit2, Michael D. Kluger1,2 Sreenivasa S. Jonnalagadda 1Division of Hepatobiliary Surgery and Liver Transplantation, New University of Missouri Kansas City School of Medicine, Kansas City, MO York-Presbyterian Hospital Weill Cornell Medical College, New York, BACKGROUND: Patients with symptomatic cholelithia- NY; 2Surgery, Columbia University Medical Center, New York, NY sis and suspected choledocholithiasis can be risk stratifi ed BACKGROUND: Since the advent of laparoscopic cholescys- into a low (<10%), intermediate (10–50%) or high probabil- tectomy in 1987, there have not been noteworthy changes ity (>50%) of having CBD stone disease based on clinical in technique for ligation and transection of cystic artery and predictors. Guidelines recommend laparoscopic cholecys- duct: metal clips and sharp transection. Laparoscopic sta- tectomy for patients with low probability of common bile plers (LS) have been suggested as a safe alternative in severe duct (CBD) stone, pre-operative Endoscopic retrograde infl ammation or when the cystic duct appears too wide for cholangiopancreatography (ERCP) for high probability of complete clip occlusion. We hypothesized an increased rate CBD stone and pre-operative Endoscopic ultrasound (EUS) of adverse postoperative events following use of LS. or Magnetic resonance cholangiopancreatography (MRCP) or Intra-operative cholangiography (IOC) for intermediate METHODS: All patients who underwent laparoscopic probability of cholidocholithiasis. In patients with interme- cholecystectomy for biliary colic, cholecystitis, pancreati- diate probability, ERCP is often deferred due to its potential tis or choledocholithiasis at our institution were identifi ed complications and MRCP is commonly performed as EUS using billing records. Operative notes were reviewed for use is not widely available. However, the diagnostic utility of of LS. A 2:1 control group was selected using propensity MRCP in this sub set of patients is not well defi ned in clini- score matching on age, gender and operative diagnosis. cal practice. Presenting features, operative characteristics and postop- erative outcomes were analyzed. Continuous variables were METHODS: Charts of all patients admitted with symptom- compared using Student’s t-test. Categorical variables were atic cholelithiasis that had cholecystectomy and underwent compared using chi-square or Fisher’s exact test. Prediction prior MRCP for cholidocholithiasis between the periods of models were constructed using logistic regression. Jan 2007 and Oct 2012 at an academic tertiary referral cen- ter were reviewed. Of these, patients who met the criteria RESULTS: Between 1997 and 2009, LS was used in 58
for intermediate likelihood of CBD stone and underwent Poster Abstracts (0.9%) of 6272 patients. These were matched to 116 patients preoperative MRCP, IOC or pre/post-operative ERCP were in whom cystic duct was divided between metal clips (MC). included in the study. Patients with any intrinsic liver dis- Sunday Differences in age, gender, race, ASA status, admission ease, or hepato-biliary malignancy or <18 years of age were diagnosis, as well as in presence of leukocytosis, hyperbili- excluded. Pertinent demographic, clinical, biochemical and rubinemia, or elevation in pancreatic enzymes were not sta- ultrasound parameters were collected by three investigators. tistically signifi cant (p > 0.05) between LS and MC groups, though LS was used more often in acute compared with RESULTS: Of a total of 330 patients, 125 met the inclusion elective cases (40% vs. 24%, p = 0.05). Compared with MC, criteria for intermediate probability and were included in average intraoperative blood loss (50 vs 25 ml, p < 0.001) fi nal analysis. Mean age of all patients was 52 ± 21 years and postoperative length of stay (2 vs 1 day, p = 0.016) with 37% males (n = 46). Eighty four patients had IOC and were both signifi cantly greater for LS. When intraoperative sixty patients had ERCP. MRCP was positive for CBD stone cholangiography (IOC) was attempted, successful cannula- in only 26.4% of patients (n = 33/125). CBD stone was pres- tion was achieved in only 2 of 8 (25%) LS cases, versus 28 ent in 33% (n = 41/125) patients as confi rmed by either of 31 (90%) controls (p < 0.001). Patients in the LS group IOC (n = 11/84) or ERCP (n = 32/60). False positive rate of required post-operative ERCP for clinically evident post- MRCP was 36% (12/33) and false negative rate was 21% (n operative choledocholithiasis at twice the rate of those in = 20/92). Sensitivity and specifi city of MRCP in detection of the MC group (p = 0.009). Controlling for preoperative and impacted stone was 51% and 85% respectively. Positive pre- demographic factors, LS remained the only statistically sig- dictive and negative predictive values were 63% and 78% nifi cant predictor of requiring postoperative ERCP (OR = respectively. 4.0, p = 0.03). There were no bile duct injuries. CONCLUSION: MRCP has a poor sensitivity in patients CONCLUSIONS: Stapling of the cystic duct during lapa- with intermediate likelihood of cholidocholithiasis. Intra- roscopic cholescystectomy is associated with an increased operative cholangiography is recommended for defi nitive need for unintended postoperative ERCP. We suspect this evaluation for a residual bile duct stone in this sub group. is secondary to passage of stone fragments into the com- mon bile duct after crushing by the stapler, or leaving a remnant infundibulum/neck after incomplete dissection and stapling. Prior to using a stapling device, we advocate for more meticulous dissection or conversion to open cho- lescystectomy in order to complete the operation safely and with minimal postoperative complications.
63 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1595 and underwent either pre/post-operative endoscopic retro- grade cholangiography (ERC) or intraoperative cholangiog- Trends in Liver Biochemistries: Are They a Better raphy (IOC) were included in the study. Patients with any Predictors Than MRCP in Evaluation of Patients with intrinsic liver disease, or hepato-biliary malignancy were Intermediate Probability of Choledocholithiasis? excluded. Hari P. Sayana, Dany Jacob, Mir Fahad Faisal, An-Lin Cheng, RESULTS: Of a total of 330 patients, 125 met the crite- Sreenivasa S. Jonnalagadda ria for intermediate risk group and were included in fi nal University of Missouri Kansas City School of Medicine, Kansas analysis. Mean age of all patients was 52 ± 21 years with City, MO 37% males (n = 46). MRCP was positive for CBD stone in 26.4% of patients (n = 33/125). CBD stone was present in BACKGROUND: Patients with symptomatic cholelithiasis 33% (n = 41/125) patients as confi rmed by either IOC (n and suspected choledocholithiasis can be risk stratifi ed into = 11/84) or ERCP (n = 32/60). Sensitivities, specifi cities, a low (<10%), intermediate (10–50%) or high probability positive and negative predictive values and accuracy of the (>50%) based on clinical predictors. Liver biochemistries tests were calculated for preoperative MRCP, trends in total are the fi rst line tests and any abnormal liver test raises sus- bilirubin alone, alkaline phosphatase (ALP) alone, AST/ALT picion for impacted common bile duct stone. A bilirubin alone, total bilirubin in combination with ALP, total bili- level >4 mg/dl is considered a very strong predictor with rubin in combination with AST/ALT, ALP in combination high probability of CBD stone (>50%). Likewise, bilirubin with AST/ALT, and total bilirubin in combination with ALP level between 1.8–4 mg/dl along with dilated ducts on and AST/ALT for patients with intermediate risk group and ultrasound is considered a very strong predictor. All other for all patients (Table 1). abnormal liver biochemistries other than bilirubin are con- sidered moderate predictor with low probability (<10%). CONCLUSION: The sensitivity of an upward trend in However, predictive value of liver biochemistry trends in hepatic transaminases, alkaline phosphatase and total bili- detecting choledocholithiasis in intermediate probability rubin alone and in different combinations is low although group is not known. some of them are comparable to that of MRCP in detecting CBD stone. MRCP has a high specifi city in detecting CBD METHODS: Charts of all patients admitted with symp- stone and this is matched by an upward trend in AST/ALT tomatic cholelithiasis that had cholecystectomy and under- with a similar specifi city. Thus in the intermediate group, went work up including MRCP for CBD stone evaluation proceeding with ERC may be a better option when an between the periods of Jan 2007 and Oct 2012 at a tertiary upward trend in the AST/ALT, ALP and/or total bilirubin are referral center were reviewed. All patients who received pre- noted, instead of performing a more expensive alternative operative work up for suspected CBD stone including liver diagnostic MRCP testing. biochemistries on 2 occasions with at least 12 hours apart,
Table 1
MRCP T. Bili ALP AST/ALT Bili + ALP Bili + AST/ALT ALP + AST/ALT Bili + ALP + AST/ALT Sensitivity 51 27 49 22 52 43 55 60 Specifi city 85 76 67 83 57 70 60 51 PPV 63 35 43 39 38 40 40 38 NPV 78 67 72 68 71 70 60 51 Accuracy 74.4 60 61 62 56 60 58 54
Su1596 two groups based on days from surgical admission to cho- lecystectomy within 48 hours (Group 1) and after 48 hours Procedure Timing in Cholecystectomy Infl uence (Group 2). Patient demographic, comorbidities and out- Outcome in Patients Admitted for Acute Gallstone come were compared using t-test and chi-square as appro- Disease priate. Regression models were used to adjust for patient Muhammad Asad Khan, Roman Grinberg, John Afthinos, and operative risk factors. Karen E. Gibbs RESULT: Patient demographics and preoperative charac- Staten Island University Hospital, Staten Island, NY teristics of patients are detailed in Table 1. Patients who had delayed procedure (>48 hours) have generally higher post- BACKGROUND: Objective of this study was to exam- operative complications including pneumonia, unplanned ine the effect of delay in cholecystectomy on outcome in intubation, acute renal failure, MI, DVT, sepsis and bleed- patient admitted for acute gallstone disease. ing requiring transfusion (Table 2). Delayed procedure was METHODS: Patient with acute gallstone disease admit- related to higher incidence of re-operation (2.3 vs. 1.7), ted inpatient between 2007–2009 were identifi ed from require prolong postoperative stay (mean 3.6 vs. 2.4 days) American College of Surgeons National Surgical Quality and higher cumulative morbidity (). Mortality was signifi - Improvement Program (NSQIP) database using Interna- antly higher in delayed surgery group even after adjusting tional Classifi cation of disease (ICD-9) diagnosis codes for co-morbidities (A Something missing here? (574-574.91) and (575–575.2). Patients were divided into
64 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Table 1: Patient Characteristics and Co-Morbidites Su1597 Duration of Surgery Within 48 Hours After 48 Hours P Value Gallbladder Perforation and Grade Do Not Affect Number 19484 (66.9%) 9621 (33.1%) Survival in Patients with Incidental Gallbladder Cancer Age 51.8 ± 18 55.1 ± 19 <0.001 Vadim P. Koshenkov1, Tulay Koru-Sengul2, Danny Yakoub1, Diabetes on oral 1614 (8.3%) 958 (10%) <0.001 Alan S. Livingstone1 hypoglycemic 1Surgery, University of Miami, Miami, FL; 2Epidemiology, University of Diabetes on Insulin 870 (4.5%) 732 (7.6%) <0.001 Miami, Miami, FL BMI 30 ± 36 29 ± 9 0.51 Smoking 3840 (19.7%) 1868 (19.4%) 0.561 INTRODUCTION: Gallbladder cancer is the fi fth most Severe COPD 670 (3.4%) 535 (5.6%) <0.001 common malignancy of the gastrointestinal tract and car- CHF 90 (0.5%) 240 (2.5%) <0.001 ries a poor long-term survival, unless the disease is iden- History of MI 69 (0.4%) 86 (0.9%) <0.001 tifi ed early. Most frequently, it is diagnosed incidentally Prior PCI 829 (4.3%) 623 (6.5%) <0.001 after a laparoscopic cholecystectomy for benign gallbladder CABG 977 (5.0%) 820 (8.5%) <0.001 disease. ESRD 163 (0.8%) 211 (2.2%) <0.001 METHODS: A retrospective review was performed for Hypertension 7599 (39.0%) 4658 (48.4%) <0.001 patients with incidental gallbladder cancer (IGC) at two ter- Prior peripheral 189 (1%) 184 (1.9%) <0.001 tiary care referral centers who underwent cholecystectomy revascularization for symptomatic cholelithiasis or cholecystitis from 1/1996 Prior operation within 138 (0.7%) 239 (2.5%) <0.001 to 8/2011. Of the 26572 gallbladders that were removed 30 days during the study period, 67 (0.25%) harbored cancer. Clini- Partially dependent 666 (3.4%) 932 (9.7%) <0.001 copathologic variables such as age, sex, grade, gallbladder Totally dependent 137 (0.7%) 260 (2.7%) <0.001 perforation and stage were assessed for impact on overall Bleeding disorder 787 (4%) 926 (9.6%) <0.001 survival. Steroid Use 389 (2%) 331 (3.4%) <0.001 RESULTS: A total of 67 patients with IGC were identifi ed. Open Cholecystectomy 3931 (20.2%) 2012 (20.9%) 0.146 Laparoscopic cholecystectomy was performed in 58 of these patients. Most patients were female (70.1%), had gallstones Table 2: 30-day Moratliy and Postoperative Complications (91.0%), and 31 were 70 years old or older (46.3%). Tumors were most commonly poorly differentiated (35.8%). Met- Within 48 Hours After 48 Hours astatic disease was detected in 13 (19.4%) patients, while N = 19484 N = 9621 P Value gallbladder perforation occurred in 16 (23.9%) patients. Superfi cial SSI 353 (1.8%) 142 (1.5%) 0.039 Univariate analysis determined that only metastatic disease Poster Abstracts Deep SSI 53 (0.3%) 24 (0.2%) 0.804 had effect on overall survival (HR = 2.76, p = 0.006). Both Organ space SSI 179 (0.9%) 94 (1%) 0.653 univariate and multivariate analyses failed to show the Sunday Wound Dehiscence 50 (0.3%) 34 (0.4%) 0.162 impact of age, sex, grade, and gallbladder perforation on Pneumonia 192 (1%) 143 (1.5%) <0.001 overall survival. Only early T stage independently predicted Unplanned Intubation 174 (0.9%) 144 (1.5%) <0.001 overall survival (HR = 0.06, p = 0.003). Pulmonary Embolism 31 (0.15%) 31 (0.3%) 0.006 DISCUSSION: In patients with IGC, advanced age, male Failure to wean >24 140 (0.7%) 150 (1.6%) <0.001 sex, poorly differentiated tumors and presence of gallblad- hours der perforation did not adversely affect survival. Only Acute renal failure 40 (0.2%) 45 (0.5%) <0.001 advanced stage, be it distant or locoregional, predicted a UTI 173 (0.9%) 141 (1.5%) <0.001 worse overall survival. MI 38 (0.2%) 25 (0.3%) 0.284 Bleeding required 24 (0.1%) 36 (0.4%) <0.001 transfusion DVT 31 (0.2%) 51 (0.5%) <0.001 Sepsis 207 (1.1%) 168 (1.7%) <0.001 Return to OR 337 (1.7%) 220 (2.3%) 0.001 Operative time 86.3 ± 48 88.6 ± 48 <0.001 Days from operation to 2.4 ± 3.7 3.6 ± 5.4 <0.001 discharge Cumulative morbidity 832 (4.3%) 696 (7.2%) <0.001 Mortality 83 (0.4%) 151 (0.6%) <0.001
CONCLUSION: In this retrospective study, delay in sur- gery was related to higher postoperative complications and mortality in patients with acute gallstone disease. These fi ndings demonstrate that medical optimization of patients should be sought expeditiously to decrease potential post- operative complications.
65 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1598 Clinical: Colon-Rectal Insurance Impacts Biliary Disease: A National Study Elizaveta Ragulin-Coyne1, Zeling Chau1, Elan R. Witkowski1, Su1600 Jillian K. Smith1, Sing Chau NG3,1, Mark P. Callery3, Heena P. Santry1, Shimul A. Shah2, Jennifer F. Tseng3,1 Extensive Abdominal Wall Necrotizing Fasciitis 1Department of Surgery, Surgical Outcomes Analysis & Research, and Fournier’s Gangrene Complicating Perforated University of Massachusetts Medical School, Worcester, MA; Appendicitis 2Department of Surgery, University of Cincinnati, Cincinnati, OH; Basem Azab, John Afthinos, Karen E. Gibbs 3Department of Surgery, Beth Israel Deaconess Medical Center, Surgery, Staten Island University Hospital, Staten Island, NY Boston, MA INTRODUCTION: While many randomized trials dem- BACKGROUND: Health care reform emphasizes insurance onstrated the possibility of non-operative management of coverage to improve outcomes. Biliary disease affects all uncomplicated appendicitis, appendectomy remains the population segments, and its treatment utilizes radiology, standard of care. In distinct contrast, perforated acute appen- GI, and surgical procedures. We hypothesized that insur- dicitis is widely treated non-operatively; supported by many ance affects biliary procedure rates and outcomes. prior studies. Although few reports demonstrated that Fourni- er’s gangren and necrotizing fasciitis are potential complica- METHODS: Nationwide Inpatient Sample 2004–2009 tions of perforated appendicitis, we are reporting the fi rst case was used to identify urgent biliary disease, including acute of abdominal wall necrotizing fasciitis and Fournier’s gan- cholecystitis, duct obstruction, cholangitis. Insurance was grene during the non-operative management of perforated defi ned as private, Medicaid, uninsured. To eliminate Medi- appendicitis in a young healthy gentleman. This case dem- care confounding, patients >64 were excluded. Hospital onstrates the need for close observation and the potential for type was defi ned as for-profi t vs. not-for-profi t. We com- signifi cant disease progression in complicated appendicitis. pared procedures, including cholecystectomy (OR), inter- ventional radiology (IR), and endoscopic (GI). We further CASE REPORT: Our patient is a 23 year-old Afro-Caribbean analyzed hospital characteristics and length of stay (LOS). gentleman with no signifi cant medical history. He pre- Multivariable analyses were performed. sented with a gradual onset of diffuse abdominal pain of 7 days duration, progressively localized to the right lower RESULTS: 1,269,668 weighted patients were identifi ed; abdominal region. On physical exam, the patient was nor- 197,644 uninsured, 244,538 Medicaid, 827,486 private. motensive, pulse 108/minute, temperature 100.1 F, mild Uninsured patients were signifi cantly more likely than pri- distended abdomen and right lower abdominal tenderness vately insured patients to have no intervention; Medicaid with an elevated WBC of 20k/cc. On admission, CT of abdo- patient resembled uninsured [Table]. Within OR, Unin- men and pelvis demonstrated an appendicolith, thickening sured/Medicaid patients were more likely to undergo open of the cecum, a 5.9 × 2.6 × 14.8 cm gas and fl uid containing vs. laparoscopic cholecystectomy. Disparities were more locules in the right lower abdominal quadrant compatible pronounced in for-profi t compared to not-for-profi t hospi- with perforated appendicitis. These locules were not drain- tals. On multivariable analysis, independent predictors for able, with the appearance of an appendicluar mass rather receiving no procedure included older age, nonwhite, rural, than a contained abscess. The patient was admitted to the male, for-profi t hospital, lower-volume center. After mul- hospital for non-operative management which included tivariable adjustment, uninsured patients were 1.6x more intravenous broad spectrum antibiotics and serial abdomi- likely (95% CI 1.5–1.8) to undergo no procedure than pri- nal exams. On hospital-day 4, the patient developed vom- vate patients. iting, more abdominal distension, a scrotal abscess, pulse 120, fever 102 F, WBC decreased to 16k/cc. A repeat CT Biliary Procedures and Outcomes by Insurance demonstrated stable locules of air and fl uid (mostly retro- Uninsured Medicaid Private p-Value peritoneal) and diffuse. No Intervention 11.7% 12.4% 8.1% <0.0001 Abdominal wall edema. The patient underwent a diagnos- Cholecystectomy (OR) 85.7% 84.1% 88.3% <0.0001 tic laparoscopy that was converted to open due to diffi culty Interventional Radiology developing an appropriate working domain. A perforated 1.0% 1.2% 1.3% <0.0001 (IR) appendix adherent to the right pelvic side wall and an extra- Endoscopic (GI) 13.9% 14.3% 13.6% <0.0001 peritoneal purulent collection was noted. After appropriate LOS (median) days 2.2 2.4 1.9 <0.0001 abdominal washout and appendectomy, incision and drain- age of the right hemiscrotal abscess was performed. Postop- CONCLUSIONS: Treatment of biliary tract disease, includ- eratively, the patient had a protracted hospitalization course ing cholecystectomy, varies with insurance. As health care (60 days) consistent with septic shock and multi-system reform ensues, reimbursement becomes bundled and cen- organ failure. The patient’s condition necessitated multiple ter/provider outcomes are increasingly scrutinized, stan- returns to the operative room for debridement of necrotizing dardization of care to refl ect best practices for all patients fasciitis involving the scrotum and most of the lower half of will be essential. his abdominal wall. The patient was also managed by the burn critical care unit for extensive skin loss, received appro- priate wound care (including negative pressure wound dress- ing), and later was covered successfully with skin grafts and was discharged in stable condition.
66 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1601 METHODS: A Pubmed literature search was performed using key search words “robotics”, “colorectal”, “cancer”, Robotics in Colorectal Surgery: A Paradigm Shift? and “laparoscopic”. After exclusions, 13 studies were iden- Fatima G. Wilder, Advaith Bongu, Michael Demyen, tifi ed from 2000–2012. 6 studies were a direct comparison Ravi Chokshi between robotics and laparoscopic surgery for CRC and the Surgery, UMDNJ – University Hospital, Newark, NJ remaining 7 looked only at robotic surgery for CRC. The series of resections were analyzed for demographics, type of INTRODUCTION: Laparoscopic colectomy is the stan- procedure, procedure length (PL), length of stay (LOS), esti- dard of care for primary colorectal cancer (CRC) resections. mated blood loss (EBL), complications, and oncologic out- The benefi ts of robotic techniques have been described comes (Table 1). Non-parametric statistical analyses were throughout the gynecologic and urologic literature, but performed with GraphPad software (La Jolla, CA). the data relating to colorectal cancer resections is still in its infancy. A review of the literature and analysis of outcomes will help us to determine the safety and oncologic value of this technology in CRC.
Table 1. Demographics, Surgical and Pathological Data
Robotic (R) or Male: EBL LOS Nodes PL Path Author/Study Laparoscopic (L) Age** Female (cc) (days) (Number) (Minutes) Conversions Resection (Staging) deSouza R (Hybrid*) 63 28:16 150 5 14 347 2 LAR – 30 APR – 8 Rectal CA, stages IS – 6 unspecifi ed Baik (2008) R 56 8:1 — 7.4 20.1 221 0 TME I – 3 II – 6 Rectal CA Hellan R (Hybrid) 58 21:18 200 4 13 285 1 LAR – 22 CA- 11 0 — 8 I – 13 II – 4 APR – 6 III – 13 IV – 1 All rectal cancer Pigazzi (2006) R (Hybrid) 60 4:2 104 4.5 14 264 0 TME Rectal CA, stages unspecifi ed Kwak R 60 39:20 — — 20 270 0 LAR – 54 IS –5 APR – 0 0 – 3 I – 16 II – 23 III – 13 IV – 4 Rectal CA Koh R 61 13:8 — 6.4 17.8 316 0 APR – 1 Anterior I – 3 II – 6 III – 5 IV – 3 resection – 7 LAR – No cancer found – 2 7 Ultralow anterior resection – 5 Sigmoid Poster Abstracts resection & rectopexy
–1 Sunday Patel R (Hybrid) 58.8 3:2 150 5.4 7 204 0 TME I – 14 II – 4 III – 7 Rectal CA Baek R 63.6 25:16 200 6.5 13.1 296 3 LAR – 33 CA – 2 Rectal CA 0 – 7 I – 12 APR – 6 II – 4 III – 15 IV – 3 Pigazzi (2010) R (Hybrid) 62 87:56 283 8:3 14.1 297 7 Unspecifi ed number of Rectal CA 0 – 18 I – 36 IS vs APR II – 36 III – 53 Spinoglio R 66.7 32:18 — 7.74 22.03 384 2 R Hemi – 18 L Hemi 0 – 3 I – 36 II – 24 III – – 10 Rectal anterior 28 IV – 9 resection w/ total proctectomy – 19 Choi R 58.5 38:12 — 9.2 20.6 304 0 TME Rectal CA 0 – 0 I – 10 II – 19 III – 19 IV – 2 Baik (2009) R (Hybrid) 60.3 37:19 — 5.7 18.4 190 0 TME I – 22 II – 16 III – 18 Pigazzi (2006) L 70 2:4 150 3.6 17 258 0 — Rectal CA, stages unspecifi ed Kwak L 59 42:17 — — 21 228 2 LAR – 52 APR – 6 IS- 1 0 – 3 I – 16 II – 23 III – 12 IV – 5 Bianchi L 62 17:8 — 6 17 237 1 — I – 14 II –7 III – 4 Baek L 63.7 26:16 300 6.6 16.2 315 9 LAR – 33 CA – 2 0 – 3 I – 15 II – 3 III – APR – 6 19 IV – 1 Spinoglio L 68.8 74:86 — 8.31 22.85 266 4 — 0 – 4.8 I – 19.3 II – 33.7 III – 28.9 IV – 13.3 Baik L 63.2 34:23 — 7.6 18.7 191 6 — I–14 II–19 III–24 Bianchi R 69 18:7 — 6.5 18 240 0 TME I–14 II–4 III–7 Rectal CA *Hybrid studies used laparoscope for early dissection (establishing pneumoperitoneum to mobilization of splenic fl exure), w/ robot then used for rectal mobilization and TME **Values reported as median; LAR – Low anterior resection, APR – abdominoperineal resection, CA – coloanal, IS – intersphinteric
67 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
RESULTS: Thirteen studies were identifi ed that directly Su1602 compared the outcomes of laparoscopic and robotic surgery for CRC. When hybrid resections were detailed, laparo- Cancer-Associated Infl ammation in a Cohort of scopic methods were used only for establishing pneumo- Colorectal Cancer Patients with Infl ammatory peritoneum or early dissection. There were no statistically Bowel Disease signifi cant differences between the groups in age, gender Graeme J. Guthrie, Donald C. Mcmillan, Alan K. Foulis, distribution, procedure length, EBL when reported, or LOS Paul G. Horgan, Campbell S. Roxburgh (Table 2). Academic Department of Surgery, University of Glasgow, Glasgow, Table 2: Robotic vs Laparoscopic Outcomes United Kingdom
Robotic Laparoscopic p AIMS: Patients with infl ammatory bowel disease (IBD) who Median age (years) 60.3 (56-69) 63.5 (59-70) 0.09 develop colorectal cancer (CRC) have poorer outcomes, Number of Males 25 (3-87) 30 (2-74) 0.70 reasons for which remain unclear. Cancer-associated- infl ammation is a key determinant of disease progression EBL (ml) 175 (104-283) 225 (150-300) 0.61 and survival in colorectal cancer. Infl ammation measured Nodes (number) 17.8 (7-22.03) 17.9 (16.2-22.85) 0.27 locally and in the systemic circulation, has not previously LOS (days) 6.5 (4-9.2) 6.6 (3.6-8.31) 0.78 been examined in a cohort of Infl ammatory bowel disease- Conversions 0 (0-7) 3 (0-9) 0.06 associated colorectal cancers. The aim of the present study PL (Minutes) 285 (190-384) 247.5 (191-315) 0.27 was to compare clinico-pathological characteristics and EBL – Estimated Blood Loss; LOS – Length of Stay; PL – Procedure Length survival in those with an IBD history and those without. Complications were reported in 10 out of the 13 papers and In particular the role of local and systemic infl ammatory were graded according to the Clavien-Dindo Scale. 100% responses in determining outcome was assessed. of the groups had some type of complication. Of the top METHODS: Patients were identifi ed from a database of 3 complications reported in the robotics group, 20% were colorectal cancer patients undergoing surgery between Grade I, 90% were Grade II and 80% were Grade III. In the 1997–2009. Systemic infl ammation was measured using 4 of 6 laparoscopic groups reporting, 75% of 3 most com- neutrophil:lymphocyte ratio (NLR) and Glasgow Prognos- mon complications were Grade I, 50% Grade II, and 100% tic Score (GPS: C-RP and albumin). Local tumour infl amma- Grade III. The most common complications in both laparo- tion was measured with the Klintrup criteria. scopic and robotic groups were ileus, anastomotic leak and RESULTS: 755 pts were included, 57 of which had Infl am- wound infection. Number of conversions at 3 approached matory bowel disease. IBD patients developed Colorectal signifi cance in the laparoscopic group (p = 0.06). 2 of the 13 cancer at a younger age (64 vs 70 yrs, P < 0.005). Despite papers looked at oncologic outcomes based on recurrence similar stage and tumour location to non-IBD cancers, IBD at follow-up. At 17 months follow-up, the recurrence rate associated tumours displayed higher risk pathology includ- was 5.4% in the robotics cases and 5.5% in the laparoscopic ing poor differentiation (P < 0.001), signet ring cell pathol- group. 1 report looked specifi cally at long-term survival ogy (P < 0.05), serosal involvement (P < 0.005), tumour outcomes with a reported disease-free survival of 77.9% at perforation (<0.001), and high-risk Gloucester prognostic 3 years and overall survival of 97% at 3 years in the robot- index (P < 0.001). Higher-grade local infl ammation (evi- ics groups. denced by Klintrup criteria: P < 0.05) and higher-grade CONCLUSIONS: Robotic colectomy for CRC is still in its systemic infl ammation (evidenced by NLR (P < 0.001) and infancy. However, early data indicates that it is a safe and GPS, P < 0.001) were observed in IBD patients. feasible option in comparison to laparoscopic techniques. Median follow up was 53 months (303 deaths). IBD patients Outcomes may be comparable, but there is need for longer had poorer overall survival (5-year survival 14% vs 41%, P term follow-up and prospective data. < 0.005). When considered with age (HR1.59, P < 0.001), TNM stage (HR1.94, P < 0.001) and GPS (HR1.58, P < 0.001), history of IBD was an independent prognostic fac- tor (HR1.99, P = 0.001). Even within the IBD cohort, local infl ammation (P = 0.003) and systemic infl ammation (GPS, P = 0.001) remained strong predictors of overall survival. CONCLUSIONS: Poorer survival in Infl ammatory bowel disease-associated colorectal cancer may relate to higher frequency of high risk pathological characteristics as well as higher levels of cancer associated infl ammation. Despite this close association, systemic infl ammation remains an independent prognostic factor on multivariate analysis.
68 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1604 versus 309 minutes for RAR + SR (p < 0.001, CI 60.8–135.4). Intra-operative blood loss for RAR was 97.8 mL compared Outcomes for Robotic-Assisted Rectopexy With or to 87.5 mL for RAR + SR (p = 0.924, CI –77.4–89.7). Conver- Without Sigmoid Colon Resection for Rectal Prolapse sion to open procedure occurred twice with RAR and once Hyuma Leland, Sonia Ramamoorthy, Elisabeth C. Mclemore with RAR + SR (p = 0.407). Length of hospitalization aver- Surgery, University of California, San Diego, San Diego, CA aged 3 days for RAR and 4 days for RAR + SR (p = 0.196, CI –0.2–2.2). The mean follow up was 127 days (range 1–72 PURPOSE: Our experience in robotic-assisted rectopexy months). Eight of 10 patients with preoperative constipa- with sigmoid colon resection (RAR + SR) or without sigmoid tion subjectively noted improvement with RAR versus 1 of colon resection (RAR) was reviewed to determine if there 2 patients with RAR + SR (p = 0.455). Six of 10 patients with was a signifi cant difference in intra-operative or postopera- preoperative fecal incontinence had subjectively improved tive outcomes. symptoms with RAR versus 1 of 2 patients with RAR + SR METHODS: We retrospectively reviewed 26 patients that (p = 1.00). Six patients in the RAR group experienced a underwent robotic-assisted rectopexy for rectal prolapse postoperative complication (over-sedation, corneal abra- from 2006 to 2012. Rectopexy is performed with primary sion, atrial fl utter, pulmonary edema, pelvic hematoma, suture pexy and without mesh. Continuous variables were ileus) versus 1 complication (urinary retention) after RAR + analyzed by t-test, while Fisher’s exact test was applied for SR (p = 0.924). There were no mortalities and no complica- categorical data. tions requiring operative intervention. RESULTS: Twenty-two patients underwent RAR and 4 CONCLUSIONS: The majority of patients underwent RAR patients underwent RAR + SR. The average age for study and outcomes for RAR versus RAR + SR were not signifi - participants was 56 years (range 21–78 years) and consisted cantly different in this series with the exception of increased of 4 men and 22 women. Recurrent rectal prolapse occurred operative time for RAR + SR. The recurrence rate was 8% in 2 patients at 3 and 30 months after surgery in the RAR and the complication rate was 27% with the majority of group and no recurrence occurred in the RAR + SR group (p complications being minor complications. = 1.00). The duration of surgery for RAR was 211 minutes
Subjective Subjective Length of Improvement in Improvement in Duration of Blood Loss Conversion Hospitalization Postoperative Postoperative n Recurrence Surgery (Min) (mL) to Open (Days) Constipation Fecal Incontinence Morbidity Mortality RAR 22 2 210.6 97.8 2 3 8 of 10 patients 6 of 10 patients 6 0 RAR + SR 4 0 308.8 87.5 1 4 1 of 2 patients 1 of 2 patients 1 0 Poster Abstracts p value p = 1.00 p < 0.001 p = 0.924 p = 0.407 p = 0.196 p = 0.455 p = 1.00 p = 0.924 p = 1.00 Sunday Su1605 defunctioning stoma created during the original procedure. Patient demographics and comorbidities were listed. Multi- Anastomotic Leak Following Anterior Resection for variate regression analysis was used to compare outcomes Rectal Cancer: Does the Presence of a Defunctioning between cases that had a defunctioning stoma and leaked Stoma Reduce the Burden of a Leak? and cases that leaked but did not have a stoma. Wissam J. Halabi1, Mehraneh D. Jafari1, Vinh Q. Nguyen2, RESULTS: We identifi ed 3,099 anterior resections that Joseph C. Carmichael1, Steven Mills1, Michael J. Stamos1, leaked. A stoma was present in 28.6% of cases, especially Alessio Pigazzi1 in male patients (31.7% vs. 22.7% p < 0.01). When a leak 1Surgery, University of California-Irvine, Orange, CA; 2Statistics, occurred, the presence of a defunctioning stoma did not University of California-Irvine, Irvine, CA reduce mortality (OR = 1.07; 95% CI 0.51–2.27; p = 0.85). However, the presence of a stoma was associated with an OBJECTIVE: To examine if the presence of a defunction- increased risk of infectious complications (OR = 1.50; 95% ing stoma performed during anterior resection for rectal CI 1.24–1.82; p < 0.01), sepsis (OR = 1.58; 95% CI 1.08– cancer affects outcomes when an anastomotic leak occurs. 2.32; p = 0.05), an increased length of stay by 2.46 days (p METHODS: Using the Nationwide inpatient sample 2004– < 0.01). Furthermore patients with a defunctioning stoma 2010, we performed a retrospective review of rectal cancer had lower likelihood of routine discharge (OR = 0.18; 95% cases that underwent anterior resection. We indentifi ed CI 0.14–0.22; p < 0.01). cases that leaked and divided them into two group based on CONCLUSION: The presence of a defunctioning stoma the presence or absence of a defunctioning stoma. All cases does not appear to reduce the burden of anastomotic leak that received a stoma in response to a leak were excluded following anterior resection for rectal cancer. from our analysis. We only included patients who had a
69 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1606 to have surgical management (78.1% vs. 94.8%, p < 0.0005) and less likely to have appendectomy on hospital day 0 or Delay in Management: A Nationwide Study of 1 (56.8% s. 77.6%). The presence of abscess or peritonitis Appendicitis in Patients with Cystic Fibrosis did not differ signifi cantly between the two groups. For Allan Mabardy1, Justin Lee1, Jose L. Piscoya1, Haisar E. Dao2, patients who underwent surgery, CF patients were more Kevin O’Donnell1 likely to undergo right colectomy or ileocecectomy (10.2% 1Saint Elizabeth’s Medical Center, Boston, MA; 2Rhode Island vs. 2.5%, p < 0.0005), more likely to require conversion Hospital, Providence, RI to an open procedure (8.5% vs. 4.4%, p < 0.0005), and more likely to have an iatrogenic injury (3.7% vs. 0.5%, PURPOSE: Acute appendicitis is an uncommon diagno- p < 0.0005). Signifi cantly fewer CF patients were attempted sis in patients with cystic fi brosis (CF). Small studies and laparoscopically (52.8% vs. 58.7%, p = 0.015). The median case reports have shown that affected patients often have a length of hospital stay and total hospital charges were sig- delay in diagnosis and subsequent complications. Our goal nifi cantly higher for CF patients (4 days vs. 2 days; $23,228 was to investigate the delay in diagnosis for patients with vs. $19,251). Multivariate regression analysis demonstrated CF who present with acute appendicitis, the factors that CF patients were more likely to be white, have public insur- might contribute to this delay, and the sequellae of delayed ance, and have admission to a teaching hospital. Patients operative management. with CF who had surgery during hospital day 0 or 1 were less METHODS: Using the Nationwide Inpatient Sample data- likely to require right colectomy or ileocecectomy (7.7% vs. base, all patients in the United States with a diagnosis of 17.7%, p = 0.005). Variables associated with prompt surgi- acute appendicitis were examined during the study years cal management included the non-teaching hospital setting 2005 through 2009. Patients with a diagnosis of CF were (OR 0.446–0.964), non-urban location (0.207–0.989), and compared to patients without CF for demographic and the absence of respiratory manifestations (OR 0.260–0.607). procedural variables, as well as variables related to com- CONCLUSIONS: The management of appendicitis in CF plication and cost. Patients with CF who did not undergo patients differs signifi cantly with that of the general popu- surgical management for acute appendicitis during hospital lation. CF patients are more likely to have a delay in opera- days 0 or 1 were compared against those who had prompt tive management and the resulting sequellae associated surgical management. with a more advanced disease process. Comorbid respira- RESULTS: During the study years 2005 through 2009, tory manifestations of CF are associated with a delay in sur- 1,350,995 patients nationwide were admitted to the hospi- gical management, possibly due to the use of antibiotics in tal with a diagnosis of appendicitis, and of these patients, these patients. 526 had a diagnosis of CF. Patients with CF were less likely
Appendicitis in Patient with Cystic Fibrosis
Cystic Fibrosis General Population Appendectomy on Hospital Day 0 or 1* 55.9% 76.8% OR 0.322–0.454 Iatrogenic Injury** 3.7% 0.5% OR 4.862–13.652 Ileocecectomy or Right Colectomy** 10.2% 2.5% OR 3.246–6.149 Attempted Laparoscopic Appendectomy** 52.8% 58.7% OR 0.649–0.956 Conversion to Open Procedure*** 13.9% 7.1% OR 1.480–3.023 Median Hospital Length of Stay* 4 days 2 days alpha = 0.05 (*) All patients included; (**) All surgical patients included; (***) All laparoscopic patients included
Su1607 AIMS: Literature about this problem is scarce, even unex- isting. Therefore, we want to make a survey of proctologi- Proctological Problems in Relation to Chemotherapy cal problems in patients under chemotherapy, and evaluate Tom Lagaert1, Bruno Vanduyfhuys1, Beatrijs Strubbe1, factors that promote the development of anal disease. 1 1 1 Ingrid Bruggeman , Koen Gorleer , Pieter Hindryckx , METHODS: From March, 15th until November 30th, 2012 Daan De Maeseneer2, Ine Moors3, StéPhanie Laurent1, all patients spontaneously reporting anal complaints at the Karen P. Geboes1, Tessa Kerre3, Simon Van Belle2, different departments of oncology and currently under che- Martine De Vos1, Danny De Looze1 motherapy, are selected for this study. Informed consent is 1Gastroenterology, University Hospital Gent, Gent, Belgium; 2Medical obtained from all patients. The following data are systemat- Oncology, University Hospital Gent, Gent, Belgium; 3Hematology, ically collected: performance status (grade 0 is normal-grade University Hospital Gent, Gent, Belgium 4 is severe), medical history, current oncological disease and chemotherapy, chemotherapy-related toxicity (grade BACKGROUND: In daily practice anal problems in patients 0–4), proctological complaints, diagnosis and outcome. under chemotherapy are often seen, cause signifi cant mor- bidity and are diffi cult to treat.
70 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
RESULTS: Twenty-three people, 14 women and 9 men, RESULTS: Two patients were lost during the follow up. with a mean age of 50 years (range 20–80) are collected. The RNA transcripts for MMP-7 were detected in 31/57 samples main presenting symptom is anal pain (n = 21) and in 2 (54%). Recurrence was diagnosed in 6 out of 55 patients patients anal blood loss. Proctological diagnoses were anal (11%); 4 patients eventually died because of metastases fi ssure (n = 12), external hemorrhoidal thrombosis (n = 3), or peritoneal dissemination. All the 6 patients who had anal abscess (n = 2), anal ulceration (n = 2), internal hemor- relapsed were positive for MMP-7. Sensitivity and specifi c- rhoidal bleeding (n = 2), no diagnosis (n = 2). ity of the test were 100% and 49% respectively. Mean WHO performance status was 1,72 (range 1–4), mean CONCLUSIONS: Positivity of MMP-7 in peritoneal cavity toxicity scores for respectively oral mucositis, nausea-vom- samples could be a novel biomarker for predicting disease iting, diarrhea and constipation are 0,76–0,88–0,94 and recurrence in patients with CRC. 1,50 (ranges 0–4). Patients were under chemotherapy for breast cancer (n = 8), AML (n = 3), renal cell carcinoma (n Su1609 = 2), rectal carcinoma (n = 2), ALL (n = 2), MDS (n = 2), High-Dose Circumferential Chemodenervation (HDCC) sarcoma (n = 1), testis carcinoma (n = 1), aplastic anemia (n = 1) and non-Hodgkin lymphoma (n = 1). of the Internal Anal Sphincter: A New Treatment Modality for Uncomplicated Chronic Anal Fissure CONCLUSION: Anal fi ssure is the most frequent encoun- Porter H. Glover1, James Z. Whatley1, Shou Jiang Tang1, tered proctological problem in patients under chemother- 1 1 1 2 apy. Constipation was the most commonly seen toxicity of Eric D. Davis , Kellen T. Jex , Ruonan Wu , Christopher J. Lahr 1 chemotherapy, while oral mucositis was rarely seen in this Internal Medicine, University of Mississippi Medical Center, Jackson, patient cohorte. Treatment of anal problems is most often MS; 2Surgery, University of Mississippi Medical Center, Jackson, MS conservative, but preventive measures should be directed BACKGROUND: Botulinum toxin injection into the inter- towards prevention of constipation. nal anal sphincter (IAS) is gaining popularity as a second line therapy for chronic anal fi ssures after patients fail med- Su1608 ical therapy. Although lateral internal sphincterotomy (LIS) can achieve a healing rate of 91–95%, it is associated with Peritoneal Expression of Matrilysin Helps Identify Early fecal incontinence of 11%. The dosage of Botulinum toxin Post-Operative Recurrence of Colorectal Cancer reported in the literature ranged from 20–50 IU with no Cristina Fiorani, Giuseppe S. Sica, Carmine Stolfi , more than 3 injection sites and results in a healing rate of Rosa Scaramuzzo, Giorgia Tema, Edoardo Iaculli, Achille Gaspari, 65%–73% and recurrence rate of 35%. We propose a new Giovanni Monteleone injection method of high-dose circumferential chemode- nervation (HDCC) of 100 IU in treating chronic anal fi ssure.
Tor Vergata, Rome, Italy Poster Abstracts BACKGROUND: Recurrence of colorectal cancer (CRC) AIM: We evaluate the fi ssure healing, complication, and following a potentially curative resection is a major chal- recurrence rates with this new method during 6 months Sunday lenging clinical problem. Although detection of tumoral study period after each injection. cells within the peritoneal cavity at the time of surgery METHODS: Between 2008–2012, 75 consecutive patients has been proposed as useful tool to identify patients with (32 Blacks, 42 Whites, 1 Native American) (50 Women, recurrent CRC, the overall low sensitivity of the test has 25 Men) with complete follow-up data were included in hampered its use in the management of such patients. this study. These patients presented with uncomplicated Matrilysin, also termed matrix metalloproteinase (MMP) -7, chronic anal fi ssure and underwent HDCC-IAS by a single is over-expressed by CRC cells and supposed to play a major colorectal surgeon. HDCC is an anoscopy assisted percuta- role in CRC cell diffusion and metastasis. This study was neous injection involving greater than 8 injection sites in a aimed at determining whether MMP-7 is detectable in the circumferential technique under anesthesia. Follow up data peritoneal cavity of CRC patients undergoing potentially were obtained by chart review and offi ce follow up. curative resection and assessing whether MMP-7 positivity RESULTS: Of the 75 patients, the 1st injection success marks patients who experience CRC recurrence. rate was 90.6% at 3 months follow-up (Table 1). A few MATERIAL AND METHODS: Fifty-seven colorectal can- patients developed transient fl atus or fecal incontinence, cer patients undergoing elective colorectal resection were but shortly resolved. There was no major complication fol- prospectively enrolled from June 2009 to November 2011. lowing HDCC-IAS including hematoma, infection, fl atus, During the surgery, peritoneal cavity was fl ushed with fecal, and urinary incontinence after 1 month. cold saline solution and fl uid was then harvested and used CONCLUSION: HDCC-IAS is a safe and effective method for RNA extraction. MMP-7 RNA expression was assessed for uncomplicated chronic anal fi ssure. It’s effi cacy rivals by RT-PCR using specifi c primers. After surgery, patients reported LIS healing rate without the associated fecal incon- underwent a regular follow up (range 12–26 months) for tinence. In addition, HDCC-IAS demonstrated far superior assessing recurrence. healing rates compared to standard injection method with- out increased complication rate. Key Words: Botulinum toxin, anal fi ssure, injection, high- dose circumferential chemodenvervation (HDCC), internal anal sphincter, lateral internal sphincterotomy (LIS)
71 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Table 1: Treatment of Uncomplicated Chronic Anal Fissures with High-Dose Circumferential Chemodenervation (Hdcc) of the Internal Anal Sphincter
Healing Rate Healing Rate Recurrence Rate Average Healing Time Complications HDCC Sessions (3 Months) (6 Months) (6 Months) (Weeks) (After 1 Month) First injection (n = 75) 90.6% 90.6% 8.0% 6.2 0 Second injection (n = 16) 81.3% 93.8% 0% 7.4 0 3rd Injection (n = 3) 100% 100% 33% 12 0 4th Injection (n = 1) 100% 100% 100% 12 0 5th injection (n = 1) 0% 100% 0% 20 0
Su1610 Su1611 Medication Use and the Risk of Diverticular Postoperative Bleeding After Colon and Rectal Surgery Complications: A Systematic Review by Preoperative Diagnosis: A Nationwide Analysis Charlotte Kvasnovsky1,2, Savvas Papagrigoriadis1, Nitin Kumar1, Ashok Kumar2, Christopher C. Thompson1 Ingvar T. Bjarnason1 1Division of Gastroenterology, Brigham & Women’s Hospital, Boston, 1Department of Colorectal Surgery, King’s College Hospital, London, MA; 2Surgery, Clay County Hospital, Flora, IL 2 United Kingdom; Department of Surgery, University of Maryland BACKGROUND: Colorectal surgery is performed for a wide Medical Center, Baltimore, MD array of gastrointestinal disease, in patients with varied pre- BACKGROUND AND PURPOSE: Serious complications operative fi tness and comorbidity rates. Postoperative out- of colonic diverticula, namely bleeding and perforation, are comes, including bleeding rates, are shaped by preoperative a source of morbidity and mortality. A variety of drugs have diagnosis in addition to operative factors. been implicated in these complications. We present a sys- AIMS: To determine incidence of postoperative bleeding temic review of the literature to assess the importance of after colorectal surgery and analyze outcomes in patients this relationship. with postoperative bleeding by preoperative diagnosis. DATA SOURCES: A systematic review of articles using METHODS: This is a retrospective cohort study using the PubMed and Cochrane Reviews was undertaken in August 2010 Nationwide Inpatient Sample (NIS), a nationally rep- 2012. Search terms included ‘diverticulitis, diverticulum, resentative inpatient database. Adult patients were included diverticulosis, diverticular perforation, diverticular bleed, if they had ICD-9 code for gastrointestinal malignancy, OR ‘lower GI bleed’ AND ‘acetaminophen, anti-thrombotic, ulcerative colitis (UC), Crohn’s disease (CD), ischemic coli- cyclooxygenase-2 inhibitors, cox-2, aspirin, amino salicylic tis (IC), diverticulitis, diverticulosis, or diverticular bleed- acid, ASA, nonsteroidal anti-infl ammatory drugs, NSAID, ing. Inclusion was limited to patients with ICD-9 procedure steroid, corticosteroid, OR calcium channel’. code for colectomy, proctectomy, or colostomy. Inclusion STUDY SELECTION AND DATA EXTRACTION: An ini- also required ICD-9 code for intraoperative bleeding, intra- tial search yielded 853 results that were assessed for study operative hematoma, or lower gastrointestinal bleeding design and topicality. A total off 23 articles were included during the postoperative period, as well as postoperative in the review, including 74 subgroup analyses, where drug packed red blood cell transfusion. Charlson Comorbidity usage and specifi c complications were clearly documented. Index was calculated for each patient. Outcomes included Data were extracted on these topics, as well as effect mea- rate of reoperation, incidence of shock, mortality, length of sures found. stay (LOS), and inpatient charge. Statistical signifi cance was established if p < 0.05. DATA SYNTHESIS: We performed a qualitative data syn- thesis with a forest plot when fi ve or more studies compared RESULTS: 214,933 patients met inclusion criteria. Of these, a single medication and similar patient complications. 1528 (0.7%) had postoperative bleeding requiring blood transfusion. Charlson score for patients with postopera- RESULTS: There were increased odds of bleeding with the tive bleeding was 2.4 ± 0.2 vs 2.3 ± 0.1 in patients without use of NSAIDs (range 2.01–12.6), acetaminophen (0–3.75), bleeding. Rate of postoperative bleeding, which was high- aspirin (1.14–3.70), and steroids (0.57–5.40). There were est in patients with IC, is shown by diagnosis in Table 1. increased odds of perforation and abscess formation with Clinical outcomes are shown in Table 2. Reoperation was NSAIDs (1.46–10.3), aspirin (0.66–2.40), steroids (2.17– signifi cantly more frequent after bleeding in each diagnosis 31.9), and opioids (1.80–4.51). except diverticulosis. Shock was more frequent after bleed- LIMITATIONS: Most studies did not describe duration or ing in patients who had surgery for malignancy and IC, but dosage of medications used, nor did they systematically less frequent in patients with diverticular bleeding. Mor- describe the severity of diverticular complications. tality was signifi cantly more frequent after postoperative CONCLUSIONS: A variety of common medications are bleeding in patients with malignancy, CD, diverticulitis, implicated in colonic diverticular complications, rare, but and diverticular bleeding. Hospital outcomes are shown in occasionally devastating outcome in the many patients Table 2. Length of stay was signifi cantly longer after post- with diverticular disease. operative bleeding in patients with malignancy and diver- ticulitis; charge was signifi cantly higher in patients with malignancy, CD, IC, and diverticulitis.
72 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Table 1: Rate of Postoperative Bleeding by Diagnosis CONCLUSION: Preoperative diagnosis is associated with signifi cant differences in outcome in patients with postop- Postoperative erative bleeding after colorectal surgery. Rates of shock and n Bleeding (%) mortality are signifi cantly increased in patients with gas- Malignancy 113,202 0.59 trointestinal malignancy and IC. Mortality is substantially UC 5572 0.25 increased in patients with CD, UC, diverticulitis, and diver- CD 10,032 0.76 ticular bleeding. Further study is needed to better under- IC 11,735 2.95 stand the reasons for this disparity and to develop better Diverticulitis
Table 2: Clinical Outcomes
Reoperation (%) Shock (%) Mortality (%) LOS, Days (95% CI) Charge x1000, USD (95% CI) Bleed No Bleed Bleed No Bleed Bleed No Bleed Bleed No Bleed Bleed No Bleed Malignancy 16.7 * 2.1 3.77 * 0.48 7.68 * 1.93 11 (8.3–14) * 7.5 (7.3–7.6) 110 (84.0–138) * 69.6 (66.4–72.8) UC 35.7 * 3.0 0 0.79 0 1.97 7.9 (5.1–11) 8.6 (8.0–9.3) 96.2 (–11.4–204) 94.4 (83.2–105.6) CD 26.3 * 1.3 0 0.40 13.2 * 0.54 14 (7.0–20) 6.9 (6.6–7.2) 131 (77.0–184) * 67.7 (62.1–73.2) IC 31.2 * 9.3 2.89 * 3.13 36.1 * 17.2 18 (12–25) 13 (12–14) 255 (173–337) * 150 (139–161) Diverticulitis 14.3 * 1.6 0 0.30 8.8 * 1.20 14 (10–19) * 6.9 (6.7–7.0) 140 (97.2–183) * 65.3 (62.3–68.4) Diverticulosis 0 1.2 0 0.34 0 1.46 6.0 (4.3–7.8) 5.9 (5.7–6.1) 51.6 (38.6–64.7) 52.1 (48.9–55.2) Diverticular bleeding 60.0 * 6.2 0 * 8.46 25 * 6.43 25 (10.8–39) 11 (9.9–12) 271 (78.1–465) 129 (118–140) * denotes statistical signifi cance
Su1612 defi ned as major and those that did not require bowel resec- Poster Abstracts tion were defi ned as minor revisions. CI failure was defi ned
Long Term Outcomes of Continent Ileostomy Created as excision of the pouch and formation of an end ileostomy. Sunday in the Pediatric Age Group RESULTS: 49 patients (26 male), median age 18 (12–21) 1 1 1 Erman Aytac , Victor W. Fazio , Hasan Hakan Erem , years and median body mass index 22 (16–38.6) underwent Jennifer Liang1, David W. Dietz1, Marsha H. Kay2, Pokala R. Kiran1 CI. 10 (20%) patients had a CI at the time of total procto- 1Department of Colorectal Surgery, Digestive Disease Institute colectomy. 12 (25%) patients underwent conversion of an Cleveland Clinic, Clevaland, OH; 2Pediatric Gastroenterology, ileoanal pouch (IPAA) to a CI. The majority of the patients Digestive Disease Institute Cleveland Clinic, Clevaland, OH (n = 39, 80%) had ulcerative colitis or indeterminate colitis at the time of CI creation; however Crohn’s disease were BACKGROUND/AIM: Continent ileostomy (CI) is a sur- diagnosed in 4 patients postoperatively. There were no gically created intra-abdominal pouch in patients with intra-operative or early post-operative deaths. One patients a permanent end ileostomy. CI is one of the few surgical who underwent CI excision seven years after CI creation options that may be offered to patients who were fated to due to complicated Crohn’s disease, died ten years after CI live with a permanent ileostomy, but want to avoid a stoma excision. Median follow-up time was 21 (range 1–38) years. appliance at any cost. Data about durability, clinical and Valve slippage (33%), small bowel obstruction (25%), pou- functional outcomes of CI created in pediatric patients are chitis (25%) and fi stula (23%) were the common complica- limited. In this study, we aimed to evaluate our 36-year tions (table). 37 patients (76%) underwent at least 1 revision operative experience on CI in pediatric patients with a 21 procedure after CI creation. 36 (74%) patients underwent year median follow-up. major revision and 6 (12%) patients underwent minor revi- METHODS: Pediatric (≤21 years) * patients undergoing a sions. Median pouch intubation was 6 (range 4–10) times CI procedure at a single institution from 1973–2009 were per day. Pouch failure occurred in 9 (18%) patients with identifi ed. CI revisions that required pouchotomy or re- 7 out of 9 cases being due to complications from Crohn’s construction following total or partial excision of CI were disease.
73 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Primary Diagnosis, Complications and Follow-Up Details METHODS: A systematic review was performed on 31 pub- lished series between 1994 and 2011 describing patients Primary Diagnosis n (%) with CDI requiring surgical intervention. Of the 31 stud- Ulcerative colitis 35 (71%) ies identifi ed, a meta-analysis was performed on 17 stud- Familial adenomatous polyposis 4 (8%) ies that presented comparative data between survivors and Indeterminate colitis 4 (8%) non-survivors of surgical CDI. Independent variable anal- Crohn’s disease 4 (8%) yses were performed for age, gender, preoperative comor- Motility disorder 2 (4%) bidities, preoperative laboratory values, as well as surgical Complications management (total colectomy end ileostomy or TCEI, seg- Valve slippage 16 (33%) mental colectomy, diverting ostomy, or non-therapeutic Small bowel obstruction 12 (25%) laparotomy). Pouchitis 12 (25%) RESULTS: Based on the 31 studies included in the review, Fistula 11 (23%) the overall rate of surgical intervention for patients diag- Diffi cult intubation 9 (18%) nosed with CDI was 1.9%. The mean age of surgical CDI Incontinence 8 (16%) patients was 69 years, and 54% were male. 93.2% of the Leakage 8 (16%) patients with surgical CDI had received antibiotics prior to Valve prolapse 7 (14%) diagnosis, and 59.3% were initially diagnosed with CDI in- Bleeding 4 (8%) hospital. The mean WBC was 29.6 x103/μL. 49.9% required Afferent limb stricture 3 (6%) preoperative vasopressors, and 44.5% had preoperative Ventral hernia 3 (6%) respiratory failure requiring intubation. Overall, 89.2% of Parastomal hernia 2 (4%) patients received a TCEI for CDI. Overall mortality of surgi- Exit conduit stricture 2 (4%) cal CDI patients was 42.5%. Follow up functional details # Among the 17 studies included in the meta-analysis, there Time to fi rst major revision after CI creation (years) 2 (0.5–30) were a total of 621 patients comprised of 367 (59%) survi- Time to fi rst minor revision after creation of CI (years) 2 (0.1–5) vors and 254 (40.9%) non-survivors. The mean age of non- Time to CI excision after CI creation (years) 5 (2–28) survivors was 71.6 years and of survivors was 65.2 years (p = # Results presented as median (range) 0.001). There was no signifi cant difference between dura- tion of symptom onset to surgery between survivors and CONCLUSIONS: CI is safe and durable in pediatric non-survivors. Preoperative vasopressor requirement, respi- patients. Development of Crohn’s disease after CI creation ratory failure (RF), acute renal failure (ARF), multi-organ seems to be a risk factor for failure. Since likelihood of fur- failure (MOF), and recent antibiotic use were independent ther revisions is high, patients with CI should be followed- predictors of postoperative mortality (Table). Non-survivors up regularly. had a signifi cantly lower preoperative albumin compared to survivors (1.66 vs 2.28 g/dL, p = 0.04). Heart rate, WBC, * Council on Child and Adolescent Health. Age Limits of lactate, creatinine were not signifi cantly different between Pediatrics. Pediatrics 1988;81:736. survivors and non-survivors. Finally, the initial type of sur- gical intervention was not predictive of survival. Su1613 CONCLUSION: The initial type of surgical intervention Risk Factors for Mortality for Surgical Clostridium was not predictive of survival. Factors that were predictive Diffi cile Colitis: A Review and Meta-Analysis of mortality from CDI included hypoalbuminemia, septic shock, ARF, RF, and MOF. This study suggests that expedi- Stephanie G. Wood1, Laura Skrip1,2, Hulda M. Einarsdottir1, 1 1 tious surgical intervention prior to end organ failure may Vikram Reddy , Walter Longo lead to improved survival in fulminant CDI. 1Surgery, Yale School of Medicine, New Haven, CT; 2Public Health, Yale School of Medicine, New Haven, CT OBJECTIVE: Clostridium Diffi cile infections (CDI), the most common infectious colitis, have been increasing in incidence and severity over the last decade. Severe CDI that requires surgical intervention is rare but mortality rate is high and poorly prognosticated. In light of a paucity of level I evidence available to direct clinical decision-making, this study aims to identify factors that may predict mortal- ity from severe CDI.
74 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Table 1: Meta-Analysis of Surgical Clostridium Diffi cile Infection Risk Factors for Mortality
Overall Effect I2 Test for Covariate No. of Studies (OR and 95% Confi dence Interval) P Heterogeneity (%) Total Colectomy (TCEI) 14 1.5735 (0.8711, 2.8423) 0.13 14.7 Segmental Colectomy 14 0.6526 (0.3541, 1.2026 0.17 24.5 Other Procedure (Not TCEI) 14 0.6355 (0.3518, 1.1480) 0.13 14.7 Gender 10 1.0017 (0.6720, 1.4932) 0.99 0.0 Vasopressors 10 3.8599 (2.6063, 5.7163) <0.001 12.5 Immunosuppression 9 0.7736 (0.5142, 1.1639) 0.22 0.0 Recent Surgery 7 0.4641 (0.2381, 0.9046) 0.02 17.9 Recent Antibiotic Use 6 4.1599 (1.1733, 14.7486) 0.03 0.0 CRF 6 0.8784 (0.3886, 1.9852) 0.76 0.0 Respiratory Failure 6 6.4230 (3.4633, 11.9121) <0.001 6.0 ARF 5 3.5793 (1.5789, 8.1142) 0.002 16.7 COPD 4 1.4037 (0.5680, 3.4690) 0.46 0.0 Known Cancer 4 2.0196 (0.7804, 5.2264) 0.15 0.0 MOF 4 7.6396 (3.0764, 18.9715) <0.001 20.3 Diagnosis Known Pre-Op 3 0.3884 (0.1034, 1.459) 0.16 0.0 Organ Transplant 3 0.3834 (0.1265, 1.1622) 0.09 0.0 Recurrent c diff 3 1.1434 (0.6079, 2.1505) 0.68 0.0
Su1614 of postoperative packed red blood cell transfusion. Univari- ate logistic regression models were performed to determine Predictors of Bleeding, Reoperation, and Mortality signifi cant predictors, which were entered into multivari- After Colon and Rectal Surgery by Preoperative ate logistic regression models controlling for patient demo- Diagnosis: A Nationwide Analysis graphics and hospital characteristics to obtain adjusted Nitin Kumar1, Ashok Kumar2, Christopher C. Thompson1 odds ratios (AOR). Statistical signifi cance was established if 1Division of Gastroenterology, Brigham & Women’s Hospital, p < 0.05. Boston, MA; 2Surgery, Clay County Hospital, Flora, IL RESULTS: 214,933 patients met inclusion criteria in 2010. Mean age was 60.3 ± 0.2 years and mean Charlson score was Poster Abstracts BACKGROUND: Colorectal surgery is performed for a 1.17 ± 0.04. 1528 patients (0.7%) had postoperative bleed-
wide array of gastrointestinal disease, in patients with var- Sunday ing requiring blood transfusion. 5439 patients (2.53%) ied preoperative fi tness. Preoperative diagnosis, in addition required reoperation during the inpatient admission. 5715 to patient and operative factors, may infl uence the rate of patients (2.66%) experienced mortality during the inpa- postoperative complications. tient admission. AIMS: To determine predictors for bleeding, reoperation, Adjusted odds ratios for predictors of postoperative out- and mortality after colorectal surgery. comes are shown in Table 1. Urgent or emergent admission, METHODS: This is a retrospective cohort study using the age >65, gastrointestinal malignancy, UC, and diverticulo- 2010 Nationwide Inpatient Sample (NIS), a nationally rep- sis were signifi cant predictors of postoperative bleeding. resentative database of inpatient admissions. Adult patients Urgent or emergent admission and IC were signifi cant pre- were included if they had ICD-9 CM code for gastrointes- dictors of reoperation. Urgent or emergent admission, age tinal malignancy, ulcerative colitis (UC), Crohn’s disease >65, and IC were signifi cant predictors of mortality. (CD), ischemic colitis (IC), diverticulitis, diverticulosis, CONCLUSION: Preoperative diagnosis, patient factors, and or diverticular bleeding. Inclusion was limited to patients admission factors have signifi cant association with postop- with ICD-9 CM procedure code for colectomy, proctec- erative outcomes after colorectal surgery. Urgent or emer- tomy, or colostomy. Charlson Comorbidity Index score gent admission, age >65, and ischemic colitis are signifi cant was calculated for each patient. Outcomes included postop- and independent predictors of mortality. Identifi cation of erative bleeding, reoperation, and mortality. Postoperative modifi able factors that increase operative and postoperative bleeding was defi ned as ICD-9 CM code for intraoperative risk in these patient groups would be of benefi t in improv- bleeding, intraoperative hematoma, or lower gastrointesti- ing patient outcomes. nal bleeding during the postoperative period in the setting
75 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Table 1: Multivariable Logistic Regression
Postoperative Bleeding Reoperation Mortality AOR (95% CI) p value AOR (95% CI) p value AOR (95% CI) p value Urgent/emergent 2.8 (2.53–3.12) * <0.01 2.0 (1.75–2.24) * <0.01 3.6 (3.15–4.02) * <0.01 Age >65 1.1 (1.06–1.19) * <0.01 1.1 (0.98–1.20) 0.13 2.5 (2.24–2.79) * <0.01 Malignancy 2.4 (2.17–2.57) * <0.01 0.54 (0.46–0.63) <0.01 0.21 (0.18–0.24) <0.01 UC 2.0 (1.60–2.46) * <0.01 0.90 (0.58–1.38) 0.61 0.49 (0.31–0.78) <0.01 CD 1.1 (0.87–1.29) 0.55 0.45 (0.29–0.69) <0.01 0.19 (0.10–0.34) <0.01 IC 1.1 (0.93–1.21) 0.41 2.3 (1.88–2.72) * <0.01 2.1 (1.83–2.39) * <0.01 Diverticulitis 1.1 (0.96–1.19) 0.21 0.44 (0.37–0.52) <0.01 0.24 (0.20–0.29) <0.01 Diverticulosis 1.7 (1.42–1.96) * <0.01 0.47 (0.29–0.75) <0.01 0.40 (0.26–0.61) <0.01 * denotes signifi cant increase
Su1615 in 9%, and urinary retention occurred in 21%. UTI rates were the same in stent and no stent groups (9% vs. 9%, Risk Factors for Urinary Tract Infection in Rectal p = 1). In multivariate analysis, females older than 65 Surgery Patients (OR 5.2, 95% CI 1.63–16.64), higher estimated blood loss Janet T. Lee, Mark Y. Sun, Genevieve B. Melton, (OR 1.9, 95% CI 1.07–3.37), and a diagnosis of depression Robert D. Madoff, Mary R. Kwaan (OR 4.7, 95% CI 1.58–14.0) were independently associated Department of Surgery, Division of Colon and Rectal Surgery, with UTI (model c = 0.78). UTI + patients were more likely University of Minnesota, Minneapolis, MN to have a prolonged LOS (p = 0.04) and be discharged to a rehab facility (p = 0.01), but not more likely to have associ- PURPOSE: Urinary tract infection (UTI) has been identi- ated major complications (p = 1). fi ed as the most common hospital-acquired infection in the United States. Studies have shown that UTIs are sig- RIsk Factors for UTI in Rectal Surgery Patients nifi cantly more common after colorectal surgery, especially Variable, n (%) UTI + (n = 20) UTI – (n = 204) p-value rectal surgery, compared with other procedures. We hypoth- Age, mean y (SD) 59.7 (16.2) 52.6 (15) 0.05 esized that the use of ureteral stents in rectal surgery would be associated with a higher risk of UTI. We also sought to Female 14 (70) 96 (48) 0.10 examine other risk factors for UTI after rectal surgery. Female age >65 6 (30) 17 (8) 0.01 BMI >30 5 (25) 69 (30) 0.80 METHODS: All patients undergoing rectal resection at a ASA class >3 9 (45) 62 (35) 0.359 tertiary care medical center from 2005 to 2010 were iden- Rectal Cancer 7 (35) 110 (54) 0.16 tifi ed using ICD-9 procedure codes. Patient and procedure Infl ammatory bowel disease 6 (30) 53 (26) 0.79 variables, UTI within 30 days of surgery, urinary retention Depression 7 (35) 20 (10) 0.005 (defi ned by reinsertion of a Foley catheter), major compli- cations, and length of stay (LOS) were identifi ed on retro- Diabetes mellitus 4 (20) 19 (9) 0.13 spective chart review. UTI was defi ned as a positive urine Laparoscopic surgery 3 (15) 25 (12) 0.72 culture with >105 colonies/ml urine with no more than 2 Ureteral Stent 3 (15) 33 (16) 1.0 species of organisms and one of the following: fever >38 EBL* cc (95% CI) 509 [340–760] 330 [290–376] 0.05 degrees C, urgency, frequency, dysuria, or suprapubic ten- Operative duration >300 min 12 (60) 69 (38) 0.06 derness. Prolonged LOS was defi ned as >75th percentile and Urinary retention 6 (30) 41 (20) 0.39 was >10 days in this cohort. Comparisons between groups *EBL was analyzed after logarithmic transformation (ln) to create a normal were made with Student t tests and Fisher exact tests (alpha distribution = 0.05). Multivariable analysis of signifi cant factors (p < 0.2) was performed with stepwise logistic regression. All sta- CONCLUSION: Ureteral stents are not associated with a tistical analyses were performed using SAS 9.2 (Cary, NC). higher rate of UTI in patients undergoing rectal resection. Females older than 65, a higher estimated blood loss, and a RESULTS: We identifi ed 223 patients during the study diagnosis of depression were found to be independent sig- period with age range of 18 to 96 (median 63). There were nifi cant predictors of UTI. We have no clear explanation for 110 (49%) males. Ureteral stents were used in 36 cases why depression would be associated with UTI and it should (16%). The majority of patients had rectal cancer as the be studied further in postoperative patients. primary diagnosis (52%), followed by infl ammatory bowel disease (26%), and rectosigmoid cancer (5%). UTI occurred
76 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1616 surgery, or proctectomy. A Charlson Comorbidity Index score was calculated for each patient. Outcomes included Nationwide Analysis of Postoperative Bleeding After rate of bleeding, rate of reoperation, mortality, and inpa- Colon and Rectal Surgery over the Past Decade: tient charge in 2010 US dollars. Statistical signifi cance was Incidence, Intervention, and Mortality established if p < 0.05. Nitin Kumar1, Ashok Kumar2, Christopher C. Thompson1 RESULTS: 360,091 patients had surgeries meeting inclu- 1Division of Gastroenterology, Brigham & Women’s Hospital, sion criteria in 2000 and 381,741 patients had surgeries Boston, MA; 2Clay County Hospital, Flora, IL meeting inclusion criteria in 2010. Rate of postoperative bleeding over time is shown in Table 1. Bleeding decreased BACKGROUND: Colorectal surgery has evolved over the signifi cantly among all surgery types between 2000 and past decade as the laparoscopic era has dawned. The prem- 2010. Reoperation rate and mortality in patients with post- ise of laparoscopic surgery has been that a less invasive operative bleeding in shown in Table 2. Even as bleeding surgical modality might lead to decreased morbidity and rate declined, reoperation rate and mortality rate among health care utilization. patients with postoperative bleeding increased, especially AIMS: To determine the change in rate of bleeding, rate of in open colectomy. However, overall mortality rate for reoperation, mortality, and health care utilization over the patients undergoing colorectal surgery remained stable at past decade in patients with postoperative bleeding after 0.165% in 2000 versus 0.175% in 2010 (p = 0.29). colon and rectal surgery. CONCLUSION: Incidence of postoperative bleeding after METHODS: This is a retrospective cohort study using the colorectal surgery has decreased over the past decade. Lapa- 2000 and the 2010 Nationwide Inpatient Sample (NIS), roscopic colorectal surgery, which has a low postoperative a nationally representative database of inpatient admis- bleeding rate, has accounted for much of the decrease; how- sions. Inclusion was limited to patients with postoperative ever, a signifi cant decrease in bleeding rates was seen after bleeding was defi ned as ICD-9 CM code for intraoperative open colectomy, proctectomy, and colostomy. Although bleeding, intraoperative hematoma, or lower gastrointesti- mortality rates have increased in patients with postopera- nal bleeding during the postoperative period and postop- tive bleeding, overall mortality after colorectal surgery has erative packed red blood cell transfusion. Adult patients been stable. Further development of minimally invasive were included if they had ICD-9 CM procedure code for surgical techniques holds promise for further improve- laparoscopic colectomy, open colectomy, colostomy-related ments in postoperative complication rates.
Table 1: Rate of Postoperative Bleeding by Surgery Type
2000 2010 p Value, Bleeding Rate Poster Abstracts n Postoperative Bleeding (%) n Postoperative Bleeding (%)
Laparoscopic colectomy – – 79,021 0.20 – Sunday Open colectomy 252,225 1.27 191,379 1.16 * <0.01 Proctectomy 47,438 0.57 49,524 0.44 * <0.01 Colostomy 60,428 3.19 61,817 2.13 * <0.01 Overall 360,091 1.50 381,741 1.03 * <0.01 * denotes statistically signifi cant change
Table 2: Reoperation and Mortality in Patients with Postoperative Bleeding
Reoperation (%) Mortality (%) 2000 2010 P value 2000 2010 P Value Laparoscopic colectomy — 43.8 — — 20.0 — Open colectomy 6.8 19.5 * <0.01 10.7 19.9 * <0.01 Proctectomy 4.5 22.5 * <0.01 7.4 8.7 0.60 Colostomy 4.4 15.2 * <0.01 11.9 13.2 0.27 Overall 5.8 27.3 * <0.01 11.0 17.0 * <0.01 * denotes statistically signifi cant change
77 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Clinical: Esophageal Su1618 Restaging PET-CT After Neoadjuvant Su1617 Chemoradiotherapy Can Prevent Non-Curative Most Patients with Persistent Symptoms on Acid- Surgical Interventions in Esophageal Cancer Patients Martinus C. Anderegg1, Roelof J. Bennink2, Suppressive Therapy Do Not Have Refl ux as a Cause Hanneke Van Laarhoven3, Jean H. Klinkenbijl1, Maarten C. Hulshof4, of Their Symptoms: a Single Center Study Using Jacques J. Bergman5, Mark I. Van Berge Henegouwen1 Ambulatory Impedance-pH Study 1Surgery, Academic Medical Center, Amsterdam, Netherlands; Pradeep K. Pallati, Se Ryung Yamamoto, Kalyana C. 2Nuclear Medicine, Academic Medical Center, Amsterdam, Nandipati, Tommy H. Lee, Sumeet K. Mittal Netherlands; 3Medical Oncology, Academic Medical Center, Creighton University, Omaha, NE Amsterdam, Netherlands; 4Radiation Oncology, Academic Medical BACKGROUND AND AIMS: Multichannel intra-luminal Center, Amsterdam, Netherlands; 5Gastroenterlogy and Hepatology, impedance (MII) and pH monitoring has been shown to be Academic Medical Center, Amsterdam, Netherlands effi cacious in the evaluation of patients with incomplete BACKGROUND: Esophageal cancer is notorious for its symptom control on proton pump inhibitor (PPI) therapy. rapid dissemination, both locally and to distant sites. Accu- The aim of our study is to evaluate the relationship of typi- rate staging at the time of diagnosis is of crucial importance cal and atypical gastroesophageal refl ux (GER) symptoms to identify patients eligible for curative treatment. For the to frequency of acid and non-acid refl ux (NAR) episodes in vast majority of these patients the preferred strategy con- patients on PPI therapy. sists of neoadjuvant chemoradiotherapy (nCRT) followed METHODS: Patients with persistent GER symptoms who by esophagectomy. Given the aggressive nature of esopha- underwent 24 hour combined MII-pH monitoring while on geal tumours, it is conceivable that in a signifi cant portion PPI therapy form the cohort of the study. Refl ux episodes of patients treated with nCRT, dissemination becomes man- were detected by impedance channels located 3, 5, 7, 9, 15, ifest during this preoperative course (interval metastasis). and 17 cm above the lower esophageal sphincter (LES) and Since metastatic disease is an absolute contraindication for classifi ed into acid or non-acid based on pH data from 5 cm esophagectomy, we added a post-neoadjuvant therapy PET- above the LES. Symptom Index (SI) was considered positive CT (restaging PET-CT) to the standard work-up of patients if >50% of specifi c symptom events were preceded by acid with potentially resectable esophageal carcinoma at initial or non-acid refl ux episodes within five minutes. presentation. RESULTS: Of 63 patients who underwent combined MII- AIM: Determine the value and diagnostic accuracy of PET- pH monitoring on PPI therapy, there were 46 (75%) women CT after neoadjuvant chemoradiotherapy in identifying and the mean age was 51.6 years (range 18–83). Fifty fi ve patients with interval metastases preoperatively. patients reported one or more symptoms during the study, METHODS: From January 2011 until September 2012 all of these 25 (39.7%) patients had a positive SI for at least consecutive esophageal cancer patients deemed eligible one symptom (12 with acid refl ux and 16 with NAR). For for a curative approach with nCRT and surgical resection typical GERD symptoms, 14 (19%) had a positive SI for acid underwent a PET-CT after completion of nCRT (median refl ux, 18 (24%) for NAR, and 43 (58%) had a negative SI. interval 18 days). Staging at initial presentation consisted of For atypical symptoms, 6 (14%) had a positive SI for acid endoscopy with biopsy, endoscopic ultrasonography, exter- refl ux, six (14%) had a positive SI for NAR, and 32 (72%) nal ultrasonography of the neck and a thoracoabdominal had a negative SI. CT scan. A PET scan was not part of the initial staging. Neo- CONCLUSION: Combined MII-pH shows that about 2/3rd adjuvant therapy consisted of 5 cycles of carboplatin AUC of patients complaining of symptoms on PPI therapy do 2, paclitaxel 50 mg/m2 and concurrent radiotherapy (41.4 not have positive symptom index to either acidic and/or Gy). If abnormalities on restaging PET-CT were suspect of non-acidic refl ux while remaining patients can have their metastases, histologic proof was acquired. This study was symptoms attributable to refl ux. approved by the local ethics committee. RESULTS: During the study period a total number of 280 new esophageal cancer patients were analysed at the out- patient clinic. Of these patients 148 underwent a restag- ing PET-CT. The remaining 132 patients were considered ineligible for curative esophagectomy at initial presenta- tion due to comorbidity, unresectable tumours or distant
78 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
metastases (94 cases), refused to undergo surgery (12), were CONCLUSION: Conversion of fundoplication to RYGB is operated without nCRT (13) or did not complete nCRT in performed on patients with a lower than average BMI than our centre (13). In 29 patients (19.6%) restaging PET-CT our typical RYGB cohort and experience signifi cant weight showed abnormalities suspicious for dissemination requir- loss with improvement in comorbid disease. Complications ing additional imaging and/or biopsy, resulting in 16 cases are similar to larger cohorts of patients undergoing RYGB. of proven interval metastasis (10.8%) and a false-positive Although ongoing therapy for acid reduction is common, rate of 8.8% for restaging PET-CT. Of the patients without improvements in GERD symptoms were noted. proven metastatic disease 116 patients have been operated at this time. In 4 of these 116 cases distant metastases were Su1620 detected intraoperatively, leading to a false-negative rate of 3.4%. Esophageal Cancer in the Young: A Population-Based CONCLUSION: 10.8 percent of esophageal cancer patients Analysis of 1095 Patients develop detectable distant metastases during neoadjuvant Attila Dubecz1, Norbert Solymosi2, Michael Schweigert1, chemoradiotherapy. To avoid non-curative resections we Jeffrey H. Peters3, Hubert J. Stein1 advocate restaging PET-CT as part of the standard work-up 1Surgery, Klinikum Nurnberg, Nurnberg, Germany; 2Faculty of of candidates for surgery. Veterinary Science, Szent Istvan University, Budapest, Hungary; 3Division of Thoracic and Foregut Surgery, Department of Surgery, Su1619 University of Rochester School of Medicine and Dentistry, Rochester, NY Conversion of Fundoplication to Roux-en-Y Gastric BACKGROUND: Controversy exists about the clinical Bypass: Long-Term Results presentation and prognosis of young patients with gastro- intestinal malignancies. The aim of this study was to evalu- Daniel B. Leslie, Nikolaus F. Rasmus, Bridget M. Slusarek, ate population-based demographics and survival of young Barbara K. Sampson, Henry Buchwald, Sayeed Ikramuddin patients with esophageal cancer in the United States. Department of Surgery, University of Minnesota, Minneapolis, MN METHODS: We identifi ed 1,095 patients under 40 years INTRODUCTION: Conversion of Fundoplication to Roux- of age with cancer of the esophagus and the gastric cardia en-Y gastric bypass (RYGB) results in signifi cant weight loss diagnosed between 1973 and 2008 from the Surveillance, and resolution of co-morbid illness, especially gastrointes- Epidemiology, and End Results (SEER) database. Demo- tinal refl ux disease (GERD). This procedure offers alterna- graphic variables and cancer-related survival were assessed tive therapy to patients with recalcitrant GERD following and compared to all patients >40 years old (n = 65,930). a failed fundoplication. To date, very little long-term data Infl uence of available variables on survival was analyzed exists for this revisional procedure. with logistic regression. Poster Abstracts MATERIALS AND METHODS: Patients who underwent RESULTS: Percentage of young patients with esophageal a conversion of fundoplication (Nissen or other) to RYGB cancer is less than 2% and is declining since the 1990s. Sunday between 2000 and 2011 at our academic medical center More than fi fty percent are diagnosed in metastatic stage. were identifi ed. The department’s bariatric surgery data- Only 74% of patients with potentially resectable esophageal base was reviewed for weight loss, the presence of GERD cancer underwent surgery. Median cancer-related survival symptoms and the use of GERD medications. A refl ux and (13 months vs. 11 months) and fi ve-year survival (22% vs. heartburn questionnaire was administered to assess impact 18%) was signifi cantly higher than in older patients. Mul- on quality of life (GERD-HRQL) scores and to determine tivariate-analysis identifi ed surgical treatment (OR: 5.046) postoperative GERD symptoms. as the only as independent predictor of 5-year survival. RESULTS: Twenty eight patients (female/male = 24/4) Percentage of non-white patients, distal cancer and adeno- underwent laparoscopic (n = 17) or open (n = 11) surgery carcinoma were signifi cantly higher when compared to all by 3 different surgeons and mean weight follow-up (100%) patients. was 3 ½ years. Average preoperative BMI and weight were CONCLUSION: Most young patients with esophageal can- 43.1 kg/m2 and 119 kg; 3 patients had BMI below 35 kg/ cer are diagnosed in metastatic stage in the United States. m2. Average length of stay was 4 days. Post-revisional BMI, Survival in patients under 40 years of age is better than in weight, and% excess weight loss were 32.0 kg/m2, 87 kg, older patients. Patients undergoing surgical treatment for and 61%. Resolution of type II diabetes mellitus, hyperten- locoregional cancer have better survival. sion, and hyperlipidemia were noted in 67%, 33%, and 60% of patients, respectively. No major short-term com- plications occurred and there were no mortalities. At least 13 patients (46%) continued to use daily acid reduction medication treatment, and 7 patients reported ongoing GERD symptoms (25%). Indications for GERD therapy also include nonspecifi c abdominal pain, pre-RYGB history of Barrett’s esophagitis, and documented gastrojejunal ulcer. On a ranked scale of no symptoms (0) to incapacitating symptoms (50), mean GERD-HRQL score was 9.5/ 50 fol- lowing surgery.
79 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1621 Su1622 Prevalence of Gastroesophageal Refl ux in Chronic Celiac Nodal Status as Determined by Laparoscopic Obstructive Pulmonary Disease Patients Gastric Ischemic Preconditioning Is Prognostic Henrique Abrahao1, Fernando A. Herbella1, in Locally-Advanced Esophageal Cancer and May Amilcar M. Bigatao2, Jose R. Jardim2, Luciana C. Silva1, Determine Necessity for Completion Esophagectomy Fernando P. Vicentine1, Marco G. Patti3 Sabha Ganai1,2, Michael B. Ujiki1,2, Mark Talamonti1,2, 1Department of Surgery, Federal University of São Paulo, São Paulo, John G. Linn1,2, Amy K. Yetasook1, Joann Carbray1, Brazil; 2Department of Medicine, Division of Pneumology, Federal Marco Zahedi1, Ki Wan Kim1,2, John Howington1,2 University of São Paulo, São Paulo, Brazil; 3Department of Surgery, 1Surgery, NorthShore University HealthSystem, Evanston, IL; University of Chicago, Chicago, IL 2Surgery, The University of Chicago Medicine, Chicago, IL BACKGROUND: The association of gastroesophageal refl ux INTRODUCTION: Laparoscopic gastric preconditioning disease (GERD) and chronic pulmonary disorders has been has theoretical benefi ts of reducing conduit-related morbid- a topic of great interest recently. However, little is known ity by allowing time for the stomach to adapt and/or demar- about GERD in the setting of chronic obstructive pul- cate to ischemic insults prior to staged esophagectomy and monary disease (COPD). This study aims to evaluate in reconstruction. We hypothesized that focused pathological patients with COPD: (a) the prevalence and the sensitivity assessment of celiac lymph nodes during the conditioning of symptoms to diagnose GERD; (b) the pattern of esopha- interval could predict long-term outcomes after esophagec- geal motility; and (c) the prevalence of distal and proximal tomy in patients with locally-advanced esophageal cancer. GERD. METHODS: A single-institutional retrospective review was METHODS: A total of 50 patients with DPOC (as defi ned conducted between 10/2008 and 11/2012, identifying 34 by age >40 years with a FEV1/FVC below 88% of the pre- patients with locally-advanced (clinical Stage IIB/III) esopha- dicted value after bronchodilator use, and no prior history geal cancer who completed staged esophagectomy after lapa- of asthma) underwent symptomatic assessment, high-reso- roscopic preconditioning. Median follow-up was 9 months. lution manometry and dual probe esophageal pH monitor- ing. GERD was defi ned by a DeMeester score >14.7. Proximal RESULTS: Patients were 60 ± 10 years old, 82% male, and refl ux was defi ned by ≥1 episode of proximal refl ux. had a BMI of 24 ± 5 kg/m2. The median interval from pre- conditioning to esophagectomy was 7 days (interquartile RESULTS: GERD was present in 21 (42%) of the total patients. range, IQR, 7–8). Preoperative staging was performed with GERD symptoms were referred by 20 (40%) patients more EUS in 94% and PET in 100%, with 4 (12%) and 30 (88%) than once a month. Symptoms were not predictive of the patients having clinical Stage IIB and III disease, respec- presence of GERD (sensitivity 71%; specifi city 83%). GERD tively. Ninety-one percent of lesions were located in the symptoms presence, esophageal manometry and pHmoni- distal esophagus, GE junction, and/or cardia. Histology toring according to the presence of distal GERD by pHmon- comprised 30 (88%) adenocarcinomas, 3 (9%) squamous itoring results are depicted in Table 1. cell carcinomas, and 1 (3%) adenosquamous carcinoma. Ninety-seven percent of patients completed neoadjuvant GERD + (n = 21) GERD – (n = 29) p therapy, with a pathologic complete response rate of 27%. GERD symptoms 15 (71%) 5 (17%) <0.001 All patients had microscopically-negative margins. The hypotensive LES 8 (39%) 5 (17%) 0.1 median number of lymph nodes resected was 20 (IQR Abnormal peristalsis 2 (9%) 8 (28%) 0.1 16–24), with a median of 5 (IQR 3–6) identifi ed in the Abnormal amplitude 10 (48%) 4 (14%) 0.01 celiac nodal packet harvested during the preconditioning Proximal GERD 20 (95%) 7 (24%) <0.001 stage. Patients with positive celiac lymph nodes (32%) were GERD: gastroesophageal refl ux disease LES: lower esophageal sphincter more likely to have pN2/N3 disease (3 or greater positive nodes) than those with negative celiac nodes (64% versus CONCLUSIONS: These data show that in patients with 9%, p < 0.01). Median disease-free survival by celiac lymph DPOC: 1) GERD is present in almost half of the patients; 2) node status was 18.6 versus 3.7 months (HR 0.22, 95% CI symptoms were insensitive and nonspecifi c for diagnosing 0.06–0.75, p < 0.01). Median overall survival by pN status GERD; 3) a defective LES is not more common in patients was 32.7 months for pN0, 12.2 months for pN1, and 5.6 with GERD leading to the hypothesis that the physiopa- months for pN2/N3 (p < 0.001). On multivariate analysis thology for GERD may be linked to the transthoracic pres- controlling for histology, grade, pathological response to sure gradient in this population; 4) in 95% of the patients therapy, celiac nodal status, and pT status, only pN status with GERD, acid refl uxed into the proximal esophagus. We remained an independent predictor of both disease-free conclude that patients with DPOC should be screened with (p < 0.05) and overall survival (p < 0.01). pH monitoring for GERD.
80 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1623 POEM-Based Endoscopic Treatment of Zenker’s Diverticulum: Minimal Incision Cricopharyngeal Myotomy (CPM) Luke Mccrone1, Kondal R. Kyanam Kabir Baig2, Victoria Gomez2, John D. Casler3, Timothy A. Woodward2 1Internal Medicine, Mayo Clinic Florida, Jacksonville, FL; 2Gastroenterology, Mayo Clinic Florida, Jacksonville, FL; 3Otorhinolaryngology, Mayo Clinic Florida, Jacksonville, FL BACKGROUND: Zenker’s diverticulum (ZD), a posterior outpouching of mucosa through transverse cricophargyn- geal muscle fi bers, has a reported prevalence of 0.01% to 0.11%. ZD is associated with marked morbidity, with symp- toms ranging from halitosis and food regurgitation to aspi- ration pneumonia and cachexia. Therapeutic management of ZD has evolved from open diverticulectomy to rigid endoscopy, and, most recently, fl exible endoscopy utilizing cricophayrngeal myotomy (CPM) with diverticulotomy. Traditional open surgical and rigid endoscopic methods have been associated with high rates of symptomatic reso- lution, with acceptable rates of recurrence. There are, how- ever, associated complications of bleeding and perforation, with these risks being amplifi ed in this comorbid, aged patient population. Building upon the per oral endoscopic myotomy (POEM) A POEM-based minimal incision crico- pharyngeal myotomy (CPM) technique has been devel- oped, using a needle knife to dissect the cricopharyngeal bar by way of a 8 to 10 mm incision within the confi nes of the mucosa without an extended diverticulotomy (See image). Poster Abstracts Sunday
CONCLUSIONS: While the AJCC 7th edition for staging of esophageal cancer has changed the emphasis from the loca- tion of regional lymph node metastasis to the number of positive nodes, our data suggest that esophagectomy may be avoided in patients with positive celiac nodes after neoadju- vant therapy for locally-advanced esophageal cancer. Laparo- scopic preconditioning provides an opportunity to determine celiac nodal status prior to committing to esophagectomy.
Myotomy within minimal incision.
81 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
AIMS: To evaluate the effi cacy of endoscopic minimal inci- METHODS: After IRB approval, we retrospectively queried sion CPM in the treatment of Zenker’s Diverticulum. our prospective database for patients who underwent an Ivor METHODS: Prospectively collected cohort of patients Lewis esophagogastrectomy for esophageal cancer. We iden- undergoing minimal incision needle-knife cricophargyneal tifi ed 220 patients from May 2001 to December 2012. Path- myotomy under monitored anesthesia care at a single cen- way one consisted of 110 patients, operated on between May ter tertiary referral center. 2001 and January 2007, who had contrast examination on postoperative day (POD) # 5, and if no anastomotic leak was Subjects include patients with symptomatic ZD as demon- seen, progression of oral intake from clear liquids on POD#5, strated by radiographic and/or endoscopic evaluation. The full liquids on POD#6 and soft solids on POD#7. They were main outcome measurements consisted of pre- and post discharged on a soft solid postgastrectomy diet. Pathway dysphagia scores (0–4 with 0 as no dysphagia and 4 rep- two consisted of 110 patients operated on between January resenting severe dysphagia) and any post-procedural com- 2007 and December 2012 who had no postoperative contrast plications, as determined by clinical review and telephone swallow, jejunal tube feedings starting POD #1 at 20 cc/hr follow up. advancing 10 cc/12 hours until goal and discharge NPO on RESULTS: From January 2009 to November 2012, 8 jejunal tube feedings only for 1 month then gradual increas- patients underwent minimal incision CPM (see table). The ing oral intake and eliminating tube feedings by 6 weeks mean age was 76 years (range 61–91 years; 63% male). postoperatively. Factors analyzed included demographics, Improvement in dysphagia was seen in 100% of patients, length of stay, complications and weight changes. though complete resolution was seen only in 4 patients RESULTS: Overall there were 188 (85.5%) men; median (50%). As this was a pilot program, all patients were hospi- age was 64 years (range 32–89). Table 1 show the patient talized overnight for observation. Excluding mild subcuta- characteristics overall and of the two pathways. Median neous emphysema in two patients, no procedurally related length of stay was 10days (range 7–98) in pathway one and problems. Time to follow up ranged from 2 weeks to 17 only 7 days (range 5–54) in pathway two. Complication months. rates were similar in the two groups: 37.2% in pathway one Results of Cricopharyngeal Myotomy and 42.7% in pathway two. The anastomotic leak rate was higher in pathway one compared to pathway two: 4.5% vs. Dysphagia Dysphagia 1.8% respectively. There was no difference in the median Case Age/Sex Pre-Procedure Post-Procedure Successful weight loss from discharge to the 6 week follow-up visit 1 84/Male 3 0 Yes between the two groups: 6.8 kg in pathway one patients vs. 2 68/Male 3 1 Yes 6.4 kg in pathway two patients. 3 70/Male 3 1 Yes Patient Characteristics 4 61/Female 2 0 Yes 5 75/Female 2 0 Yes Overall Pathway 1 Pathway 2 6 76/Male 3 1 Yes Characteristic (n = 220) (n = 110) (n = 110) 7 83/Male 2 0 Yes Men (%) 85.5 87.2 83.6 8 91/Female 4 2 Yes Median age (range) 64 (32–89) 64 (32–89) 63 (33–84) Dysphagia Score: 0 = no dysphagia; 1 = able to swallow some solid foods; Preop chemo/XRT (%) 69.1 61.8 76.4 2 = able to swallow only semi-solid foods; 3 = able to swallow liquids only; Clinical Stage IIIA 53.6 51.8 55.5 4 = total dysphagia. Adenocarcinoma (%) 86.8 88.2 85.5 Preop chemo/XRT – Preoperative chemotherapy and radiation therapy CONCLUSION: Minimal incision CPM is effective in man- agement of Zenker’s diverticulum, and demonstrates effec- CONCLUSION: Changing the postoperative nutritional tive resolution of dysphagia. In our case series, no serious management after an Ivor Lewis esophagogastrectomy procedural or post-procedural complications occurred. to no contrast swallow and delaying oral intake for one month results in a shorter length of stay and reduced anas- Su1624 tomotic leaks, but no change in the overall complication rate or early postoperative weight loss. Postoperative Management of Nutrition After Ivor Lewis Esophagogastrectomy for Cancer Laura Trujillo, James Taswell, Mark Allen Mayo Clinic, Rochester, MN OBJECTIVES: Esophagogastrectomy is a complex operation and the postoperative management is variable. We hypothe- sized that waiting 5 days postoperatively to obtain a contrast swallow to start oral intake and then waiting until patients were able to take suffi cient oral intake before discharge pro- longs hospitalization after an Ivor Lewis esophagogastrec- tomy. To examine this hypothesis we analyzed two methods of management of postoperative nutrition after surgery.
82 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1625 Su1626 High-Resolution Impedance Manometry Findings in Surgical Management of Esophageal Perforation: A Patients with Epiphrenic Diverticulum 10-Year Experience Se Ryung Yamamoto, Kalyana C. Nandipati, Paul Goldsmith1, Bilal Alkhaffaf1, Bart Decadt2 Pradeep K. Pallati, Tommy H. Lee, Sumeet K. Mittal 1Manchester Royal Infi rmary, Manchester, United Kingdom; Creighton University, Omaha, NE 2Stepping Hill Hospital, Stockport, United Kingdom AIM: The objective of this study was to evaluate high-resolution INTRODUCTION: The management of esophageal per- impedance manometry (HRIM) fi ndings in patients with foration and mediastinal sepsis is challenging. Treatment esophageal epiphrenic diverticulum. strategies differ between surgical units and as a conse- METHODS: Patients with esophageal epiphrenic diver- quence outcomes can vary widely in this patient group. We ticulum who underwent HRIM between October 2008 and present our 10-year experience of esophageal perforation March 2012 are included in the study. Manometric fi ndings and evolving treatment strategy for this condition. were compared to endoscopic (EGD) and barium swallow METHODS: This was a retrospective review of all esopha- (BS) fi ndings. Patients with previous foregut surgery were geal perforations including both patients with a spontane- excluded. ous or iatrogenic perforation and cases of mediastinal sepsis RESULTS: Six patients (mean age 59.0 years, 3 females) are due to anastomotic leak following cardio-esophagectomy. included in the study. On EGD the diverticulum was 1 to Patients were grouped according to their treatment strategy. 4 cm above the gastro-esophageal junction (GEJ) and the The primary outcome measures were in-hospital death and mouth of diverticulum was 2 to 9 cm in size. Mean lower length of stay (total hospital stay and Intensive Care Unit esophageal sphincter pressure (LESP) and mean Integrated (ICU) stay). relaxation pressure (IRP) were 50.7 mmHg (range 39.2 to RESULTS: In total, thirty-seven patients were included. 61.9 mmHg) and 22.4 mmHg (range 13.8 to 30.8 mmHg) Twenty-fi ve were male with a median age of 59 (range respectively. Achalasia was the most common abnormal- 21–80). Seven patients suffered iatrogenic perforations ity noted in 3 patients (type I = 2 and type II = 1). Two (3 esophageal dilatations, 2 food bolus removal, 2 gastros- additional patients had isolated EGJ obstruction with pre- copy). Eleven patients presented following spontaneous served esophageal motility. One patient had normal IRP esophageal perforations and a further Eighteen suffered but weak peristalsis. A break in peristalsis corresponding to anastomotic leaks following cardio-esophagectomy and the mouth of the diverticulum could be seen in 4 patients. one leak following cardio-myotomy. There was decreased bolus transit in three patients. Twenty-six patients underwent surgery for their perfora- CONCLUSIONS: There is a high prevalence of esophageal tions compared to eleven who were conservatively man- Poster Abstracts outfl ow obstruction and primary peristaltic abnormality in aged. Surgical management involved either thoracotomy patients with epiphrenic diverticulum. This reconfi rms the with primary repair or creation of a controlled fi stula using Sunday need to extend the myotomy down on to the GEJ even in a T-tube (20), cardio-esophagectomy (3) or thoracoscopic patients in whom the diverticulum appears to be well above washout (3). All patients had enteral feeding routes inserted the GEJ. in conjunction with surgery. Conservative management constituted either simple insertion of chest drains (8) or stent placement (3). Death following non-operative management occurred in 4 patients compared to two (36% vs 7% p < 0.05) in those who underwent surgery. CONCLUSION: Urgent operative management is a safe treatment option for patients who have oesophageal per- foration and are fi t to undergo a surgical exploration. Thoracotomy with repair of the perforation over a T-tube with defunctioning gastrostomy, feeding jejunostomy and drainage of the thorax and mediastinum, appears a safe policy and is our preferred approach. Patients with exist- ing esophageal pathology may be considered for emergency cardio-esophagectomy.
83 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1627 use were collected at years 1, 2, 3, 4 and 10 post-procedure. The data were analyzed using a repeated measures analysis Sustained Improvement in GERD-HRQL, Patient of variance to determine whether there was a signifi cant Satisfaction, and Anti-Secretory Drug Use 10-Years temporal trend in the various outcomes. Any signifi cant After Stretta for Medically Refractory GERD effect was sub-analyzed using a Bonferroni-adjusted mul- Mark D. Noar1, Patrick Squires1, Emmanuelle Noar1,2 tiple comparison procedure. 1The Heartburn and Refl ux Center, Endoscopic Microsurgery RESULTS: The total treated pool included 227 patients (136 Associates, Towson, MD; 2School of Medicine, Tulane University, females, 60%). No strictures, perforation, deaths or other New Orleans, LA signifi cant adverse events occurred. Complete follow-up was available for subjects at intervals as follows: 6 months BACKGROUND & AIMS: The Stretta procedure is an (n = 177), 1 year (n = 149), 2 years (n = 98), 3 years (n = effective endoscopic modality for control of GERD symp- 98), 4 years (n = 94), and 10 years (n = 99). Heartburn, toms. However, long term effi cacy and safety at >10 years satisfaction, HRQL and medication use were signifi cantly has not been assessed. We prospectively evaluated the long- improved across the follow-up period (6, 12, 24, 36, 48, and term effi cacy of Stretta at 10-years and assessed anti-secre- 120 months; p < 10–6 for all of the outcomes) and results tory drug use, GERD symptoms, and patient satisfaction. were superior to those achieved for baseline drug therapy METHODS: From 8/2000 to 9/2004, 227 patients with (see Table). inadequate GERD symptom control despite BID PPI under- CONCLUSION: This 10-year, open, single center, prospec- went Stretta and were prospectively enrolled in this study. tive assessment of Stretta for refractory GERD demonstrates All patients had normal esophagogastric anatomy, except a signifi cant and sustained improvement of GERD-HQoL 16 patients with failed Nissen fundoplication and 7 with scores, patient satisfaction, and improved PPI use and vali- large (>3 cm) hiatal hernia. All procedures were performed dates the long-term usefulness of this endoscopic proce- by a single endoscopist in an outpatient setting with con- dure. Patients with variant anatomy such as prior Nissen scious sedation. Baseline and follow-up GERD-HRQL scores fundoplication or large hiatal hernia had a similar response (0–50), heartburn (0–5), satisfaction (0–5) and medication compared to patients with normal anatomy.
Before Treatment, Before Procedure, Parameter off Meds on Meds 0.5 Years 1 Year 2 Years 3 Years 4 Years 10 Years Med Scores 8.3 ± 3.8 4.9 ± 3.9 3.8 ± 3.5 3.7 ± 4.2 4.6 ± 3.6 4.3 ± 3.2 4.7 ± 3.3 GERD Scores 27.8 ± 10.7 21.4 ± 11.5 11.1 ± 10.0 6.9 ± 7.5 5.0 ± 6.9 6.9 ± 8.2 7.3 ± 8.5 8.1 ± 9.9 Satisfaction Scores 1.3 ± 0.9 1.9 ± 1.1 3.4 ± 1.3 3.9 ± 1.3 4.3 ± 1.0 3.8 ± 1.3 3.8 ± 1.3 3.8 ± 1.2
Su1628 METHODS: A query was submitted to the Web of Sci- ence database to fi nd all the publications in the topic of Bibliometric Analysis of the Scientifi c Publications gastroesophageal refl ux in the time period between January About Gastroesophageal Refl ux Disease (GERD) 1954 to December 2011. A total of 18136 references were Between 1954 and 2011 retrieved, these records were then analyzed using biblio- Shahin Ayazi, Thomas J. Watson, Carolyn E. Jones, metric parameters. Virginia R. Litle, Christian G. Peyre, Jeffrey H. Peters RESULTS: Number of publications has increased from Surgery, University of Rochester, Rochester, NY only two papers per year in 1954 to nearly three papers per day in 2011, with more than 700 publications per year in INTRODUCTION: Gastroesophageal refl ux disease (GERD) 21st century. Majority of these publications are in English is an increasingly prevalent health problem. This disease (94%). United States, England and Italy are the most pro- has been the focus of the work of many researchers in the ductive countries with 40%, 7% and 6% of the literature last 50 years, these investigations has led to a transforma- respectively. Eighty percent of the literature in the fi eld is tion in the understanding and treatment of this disease. the result of the work of authors from 10 countries [North However little is known about the characteristics and trends America (2), Europe (6), Japan and Australia], this pattern of the scientifi c publications in this fi eld. The aim of this did not change when the publications from 2000–2011 study is to present a bibliometric analysis of the scientifi c were analyzed separately. publications on gastroesophageal refl ux disease.
84 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
METHODS: A retrospective database was used to capture the clinico-pathological data of all consecutive curative resections of OGJ adenocarcinomas over the last 10 years in two UK Upper GI Units. Any report with less than 12 lymph nodes was considered inadequate and denoted as (Nx). All cases were re-reported and re-staged according to the 7th TNM staging rules. We compared the impact of the 7th TNM staging rules on neo-staging. Overall survival was analysed using the 6th and 7th TNM staging respectively. Overall sur- vival was sub-stratifi ed into 2 years, 5 years and 10 years post curative resection. Mayo clinic is the institution with the highest number of publications (268), followed by the University of Southern RESULTS: Fifty seven (57) pathology reports confi rm- California (243) and Northwestern University (211). Tom ing oesophago-gastric junctional adenocarcinomas were DeMeester, Joel Richter and Donald Castell are the most reviewed. Adequate lymphadenectomy (minimum of 12 prolifi c authors in this fi eld with 194, 178 and 169 publica- nodes) was noted in 33 patients. Overall stage migration tions respectively. Gastroenterology and American Journal was noted in 36 (63%) reports with the 7th TNM staging. Of of Gastroenterology are the two journals with the highest those who had adequate lymphadenectomy (33), 20 reports number of publications about the GERD. (60.6%) had stage migration. CONCLUSION: Parallel to the rise in the prevalence of the In terms of survival, one patient was lost to follow up and is gastroesophageal refl ux disease, there is an increase in the not included in the analysis for survival. Patients with Stage research performed about refl ux evidenced by the constant 4 disease were not operated on in this cohort. increase in the number of scholarly work published in this 2 year survival (n = 56) using the 7th TNM staging, showed fi eld (Figure). Two thirds of the literature about the GERD an apparent increase in survival by 12.4% in stage 3 disease st has been published in the 21 century. Only a small frac- with a corresponding decrease in survival by 17.8% in stage tion of GERD publications (6.3%) is supported by a funding 2 disease. 5 year survival (n = 34) using the 7th TNM staging, agency. Majority of the literature (80%) is the result of the demonstrated 14.8% increase in survival for stage 3 disease work of the authors from only 10 countries. and a corresponding decrease by 17.6% for stage 2 disease. 10 year survival (n = 10) with the 7th TNM, again demon- Su1629 strated a 30% increase in apparent survival for stage 3 dis- The Better Defi nition of Nodal Staging in the th7 ease and a corresponding decrease by 40% in stage 2 disease. Edition of TNM Manual Does Not Predict Survival For stage 1 disease, there is no signifi cant change in 2 year, Poster Abstracts or Translates Into Better Prognosticating Ability in 5 year and 10 year survival. Sunday Oesophago-Gastric Junctional Adenocarcinoma CONCLUSION: The 7th edition of TNM staging provides Ramesh Y. Kannan, Matthew L. Davies, Carys Jenkins, a detailed documentation of the lymphatic staging. The Majid Rashid, Ashraf M. Rasheed apparent increase in survival in stage 3 disease appears to Minimal Access Surgery, Royal Gwent Hospital, Newport, be compensated for the apparent decrease in survival for United Kingdom stage 2 disease. This better defi ned lymphatic staging does not seem to predict survival or have a superior prognosti- INTRODUCTION: The 7th TNM staging defi nes a mini- cating ability. mum number of nodes, recommends an optimal number for each T stage, emphasizes the prognostic importance of number of regional nodes involved and upstages based on the number of metastatic lymph nodes. AIMS: To study the impact of application of 7th TNM rules on nodal staging (N) of resected and pathologically reported oesophago-gastric junctional (OGJ) adenocarcinomas dur- ing the last 10 years stratifying them according to the 7th edition TNM staging and to compare against the original staging and assess possible impact of nodal neo-staging on survival.
85 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1631 (p < 0.001) and higher stage (p < 0.001) at diagnosis and were less likely to undergo surgical resection (4.9% vs. 18%; Does Surgery Have a Role in the Treatment of Small p < 0.01). In both all-comers as well as those undergo- Cell Cancer of the Esophagus? ing surgical resection, univariate analyses showed a worse Moshim Kukar1, Adrienne Groman1, Graham W. Warren3, survival in patients with small cell esophageal cancer. Usha Malhotra2, Chukwumere Nwogu1, Todd L. Demmy1, However, multivariate analyses adjusting for age, gender, Sai Yendamuri1 grade, stage, race and number of lymph nodes examined 1Surgical Oncology, Roswell Park Cancer Institute, Buffalo, NY; did not show a statistically signifi cant association between small cell histology and overall survival in both sets of 2Medical Oncolgy, Roswell Park Cancer Institute, Buffalo, NY; 3 patients. Univariate analysis of patients with small cell can- Radiation Medicine, Roswell Park Cancer Institute, Buffalo, NY cer alone demonstrated a signifi cant association of surgery INTRODUCTION: Small cell cancer of the esophagus is with median survival (17 months vs. 7 months; p = 0.002) an uncommon malignancy with perceived poor progno- (Figure 1). sis. Due to its rarity, no large case series has been exam- CONCLUSIONS: This large study of small cell esophageal ined to guide therapeutic decisions. We examined the SEER cancer confi rms the clinical observation that small cell can- database to analyze factors determining outcome of this cer of the esophagus is an aggressive disease. In addition, unusual malignancy. we show that outcomes with this disease are associated METHODS: All patients with esophageal cancer in the with stage of disease and not histology. Therefore, surgical SEER database from 1973–2009 were included. Character- resection for esophageal cancer patients with this histol- istics of patients with and without small cell cancer were ogy should be offered based on stage rather than histology compared. Univariate and multivariate analyses examining alone. the relationship of small cell histology with overall survival (censored at 72 months) were performed in all patients as Su1632 well as those undergoing surgical resection. In addition, only patients with small cell cancer were analyzed to exam- Dissatisfaction After Laparoscopic Heller Myotomy ine the potential benefi t of surgery. Sharona B. Ross, Carrie E. Ryan, Benjamin L. Jacobi, Harold Paul, Kenneth Luberice, Paul Toomey, Alexander Rosemurgy General Surgery, Florida Hospital Tampa, Tampa, FL INTRODUCTION: Laparoscopic Heller myotomy allevi- ates symptoms of achalasia; however, we have observed a small subset of patients who are dissatisfi ed with their experience. This study was undertaken to identify causes of dissatisfaction after laparoscopic Heller myotomy and to identify predictors of dissatisfaction following myotomy. METHODS: With IRB approval, all patients undergoing laparoscopic Heller myotomy at our institution from 1992 to 2012 were prospectively followed. Using a Likert scale patients rated their frequency and severity of symptoms before and after myotomy. Patients graded their experience from “Very Satisfying” to “Very Unsatisfying”. Objective outcomes were determined by esophagography/esophagos- copy. Complaints were derived from postoperative surveys, clinic visits, and phone surveys. Median data are reported. RESULTS: Of the 597 patients undergoing laparoscopic Heller myotomy, 44 (7%) patients reported some level of dissatisfaction following myotomy with follow-up at 32 months. Dissatisfi ed patients were more likely to have Figure 1: Patients with small cell esophageal cancer undergoing undergone prior abdominal operations (45% vs. 28%, p surgical resection (green line) have a better survival than those with = 0.03) and previous myotomies (16% vs. 5%, p = 0.02). no surgical resection (blue line) on univariate analyses. Dissatisfi ed patients also had longer postoperative lengths of stay (2 days vs. 1 day, p = 0.01), generally because of RESULTS: 387 of 64,799 (0.6%) patients with esophageal postoperative complications or exacerbations of preop- cancer had small cell histology. Patients with small cell erative comorbidities. For dissatisfi ed patients, symptom histology were similar in age and race, but had a higher frequency and severity persisted after myotomy and were proportion of females compared to other histologies (40.6% more notable than for satisfi ed patients (p < 0.03 for all; vs. 25.4%; p < 0.001). These patients also had higher grade Figure).
86 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
METHODS: With IRB approval, patients were prospec- tively followed after Heller myotomy. Patients scored the frequency and severity of their symptoms before and after myotomy using a Likert scale (0 = never/not bothersome to 10 = always/very bothersome). The symptom frequency and severity of the fi rst 100 patients undergoing laparo- scopic Heller myotomy with anterior fundoplication were compared to the last 100 LESS patients undergoing LESS Heller myotomy with anterior fundoplication. Median data are reported. RESULTS: 601 Heller myotomies with fundoplications were undertaken between 1992–2012. Of these, 470 (78%) were completed via conventional laparoscopy, 130 (21%) via the LESS approach, and 1 (.1%) as an “open” operation. All of the last 100 patients underwent the LESS approach with anterior fundoplication. The frequency and severity of all preoperative symptoms signifi cantly improved with Figure depicts the disparity of symptom resolution between satisfi ed and either the conventional laparoscopic or LESS approach; dissatisfi ed patients. *denotes symptom improvement after myotomy several postoperative symptoms had superior improve- (p < 0.05). ment with the LESS approach (e.g., vomiting, choking, p = 0.01 for each; Figures 1a, 1b). Those who underwent LESS CONCLUSIONS: Dissatisfaction is fortunately uncommon Heller myotomy with anterior fundoplication also had a after laparoscopic Heller myotomy. Dissatisfaction is directly decreased length of hospital stay (2 vs. 1 day, p < 0.05) and related to persistent severe and frequent symptoms; longer no apparent scars. lengths of stay, previous abdominal operations, and “re-do” CONCLUSIONS: Laparoscopic Heller myotomy provides myotomies predict dissatisfaction. Patients with notable an effi cacious and durable treatment for achalasia. The comorbidities and/or previous abdominal operations, par- LESS technique offers a safe approach with equivalent or ticularly Heller myotomy, are more likely to be dissatisfi ed superior symptom relief and improved cosmesis. Overall after laparoscopic Heller myotomy should be counseled patient satisfaction and durable symptom relief promotes preoperatively. Comorbidities should be addressed preop- laparoscopic Heller myotomy with anterior fundoplication, eratively and for patients with previous notable abdominal particularly using the LESS approach. operations, particularly myotomy, alternatives to laparo- Poster Abstracts scopic Heller myotomy should be considered. Sunday Su1633 A Single Institution’s Journey with Heller Myotomy: Is the Laparo-Endoscopic Single Site (LESS) Approach Best? Alexander Rosemurgy, Thara Salam, Carrie E. Ryan, Graphs illustrate symptom frequency preoperative and postoperatively. Mercedez C. Cruz, Kenneth Luberice, Harold Paul, Figure 1a is specifi c for the fi rst 100 patients undergoing laparoscopic Sharona B. Ross Heller myotomy with anterior fundoplication while Figure 1b is specifi c General Surgery, Florida Hospital Tampa, Tampa, FL for the last 100 patients undergoing LESS Heller myotomy and anterior INTRODUCTION: The surgical treatment of achalasia has fundoplication. evolved from a conventional laparoscopic Heller myotomy to a Laparo-Endoscopic Single Site (LESS) Heller myotomy with anterior fundoplication. This study illustrates our jour- ney with the evolution in technology and instrumentation and details patient outcomes along our journey.
87 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1634 Su1635 Surgical Myotomy Should Be Considered as Primary Surgical Treatment for Achalasia: A NSQIP Analysis Treatment in Elderly Patients with Achalasia Stephanie G. Wood, Edward J. Hannoush, Andrew Duffy, Donald E. Low, Artur M. Bodnar, Sheraz R. Markar Robert Bell, Kurt E. Roberts Department of Thoracic Surgery, Virginia Mason Medical Center, Surgery, Yale School of Medicine, New Haven, CT Seattle, WA INTRODUCTION: There are multiple surgical treatment INTRODUCTION: Achalasia is a rare disease which pre- options for Achalasia, with the laparoscopic approach dominantly affects patients between 25 and 55 years. Older increasingly considered the treatment of choice. We review patients are often not considered for endoscopic manage- the largest number of reported cases of laparoscopic Heller ment due to the perception that they are at higher risk for myotomies to date, from the NSQIP database, and compare surgical therapy. alternative surgical treatments. METHODS: All patients undergoing surgical treatment for METHODS: Using the American College of Surgeons achalasia between 2001 and 2012 were entered into an IRB- National Surgical Quality Improvement Program (NSQIP) approved database. Presenting characteristics and periop- participant use fi les from 2005–2010, patients diagnosed erative outcomes were compared in patients younger and with Achalasia (ICD-9 530.0) who underwent a surgical older than 70 years. Heller myotomy, including open abdominal (CPT 43330), laparoscopic (CPT 43279), open thoracic (CPT 43331), and RESULTS: 99 patients underwent surgical treatment for thoracoscopic (CPT 32665) approaches. As the CPT code for achalasia during the study period. 6 patients (1 > 70, 5 < laparoscopic Heller myotomy was created in 2009, there are 70) underwent esophageal resection and were excluded. 92 none reported before this time. patients underwent Heller myotomy and either Toupet or Dor fundoplication. 17 patients >70 years of age, mean age RESULTS: A total of 978 patients diagnosed with Achalasia was 76.5, range 70–92 years, were compared to 75 patients were identifi ed, of which 663 patients received a laparocopic <70 years. Pre-operative weight loss was similar (mean 15.6 Heller, 239 open Heller, 16 open thoracic, and 60 thoraco- lbs >70 versus 18.7 lbs <70). Patients in the >70 group had scopic myotomies. Overall, 56.8% were male and mean age higher ASA scores (2.58 versus 2.14, p = 0.01) and were more is 52 years (SD ± 16.3). There was no signifi cant difference likely to undergo previous endoscopic therapy (65% versus in age between groups. There was a signifi cant difference in 60%). Patients in the >70 group were less likely to undergo total length of stay between open and laparoscopy Heller laparoscopic surgery (35% versus 79%). Length of operation myotomy groups only, 3.45 vs 2.48 days, p = 0.015. Opera- and blood loss was similar between study groups. Intraop- tive time was signifi cantly different between the open and erative complications were more common in patients <70, laparoscopy Heller myotomy groups only, 146.0 vs 136.27 conversions (3 versus 0), intraoperative mucosal injury min (SD ± 54.9), p 0.018. There only signifi cant difference (3 versus 0). Perioperative complications occurred in 13% in postoperative complications was in superfi cial site infec- of patients <70 and 23.5% of patients >70. Complications tions, with 3 in open heller and 1 in thoracoscopic groups, in the >70 group included transient delirium 2, C. diff coli- p = 0.025. There were no deaths reported. tis 1 and fall 1. Median length of stay was 2 days in patients CONCLUSION: While there is no signifi cant difference in <70 (lap 2 versus open 4.5 days) and 4 days in patients >70 serious post-operative complications, laparoscopic Heller (lap 2 versus open 4 days). There was no mortality in either myotomy has improved length of stay and operative times group. Readmissions were required in one patient in each compared to open, and thoracic approaches. cohort. At mean objective follow-up of 3.8 months, 94% of patients >70 reported excellent (normal swallowing) or good (occasional dysphagia but no regurgitation) results. Clinical: Hepatic Followup was available in 12 patients >70 years at mean 29.1 months. 83% reported they ate a completely normal diet. Residual symptoms included occasional dysphagia Su1636 34%, occasional regurgitation 8%, and periodic heartburn Difference in Outcomes Between Right and Left 42%. No patient had required additional surgical or endo- Hepatectomy in Patients Undergoing Hepatic Resection scopic therapy and all patients indicated they would have 1 1 2 the operation again. Mashaal Dhir , Lynette M. Smith , George Dittrick , Quan P. Ly1, Aaron R. Sasson1, Chandrakanth Are1 CONCLUSIONS: Elderly patients with achalasia can 1University of Nebraska Medical Center, Omaha, NE; 2Surgery, undergo surgical myotomy and partial fundoplication Nebraska Methodist Hospital, Omaha, NE safely and with excellent results comparable to younger patients. These older patients should be given the opportu- BACKGROUND: Several studies have documented the nity to discuss surgical treatment with an experienced sur- safety of liver resections. Although right hepatectomy is geon at the time of their initial presentation. felt to be associated with higher mortality and morbidity, data on the difference in outcomes between right and left hepatectomy is scarce. The aim of the current study is to analyze the difference in outcomes between right and left hepatectomy in patients undergoing hepatic resection.
88 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
METHODOLOGY: All patients undergoing right (primary Su1639 CPT code 47130) and left hepatectomy (primary CPT code 47125) were extracted from the National Surgical Qual- Modifi ed FOLFOX6 and Bevacizumab as Neoadjuvant ity Improvement Program (NSQIP) database (2005–2010). Chemotherapy for Patients with Potentially Curable Benign and malignant diagnoses (ICD-9 155.0, 155.1, 155.2 Bilobar Liver Metastases from Colorectal Cancer and 197.7) were determined from the database. The data Atsuyuki Maeda, Masatoshi Isogai, Yuji Kaneoka was analyzed to determine differences in outcomes between Digestive Surgery, Ogaki Municipal Hospital, Ogaki, Gifu, Japan right and left hepatectomy by using the chi square test or Fisher’s exact test as appropriate. OBJECTIVE: Even though patients with colorectal cancer (CRC) and liver metastases have a poor prognosis, they can RESULTS: A total of 2311 patients who underwent right benefi t from perioperative chemotherapy and complete or left hepatectomy were identifi ed of whom 1680 patients extirpation of the disease. Oxaliplatin based chemotherapy were noted to have a malignant diagnosis. (Table 1) Patients with bevacizumab has been widely reported to improve that underwent right hepatectomy were noted to be have a outcomes with metastatic CRC. However, its impact on sur- higher mortality rate when compared to left hepatectomy gical complications and survival benefi t after liver resection (all patients: right vs left—4.24% vs 1%, P < 0.001 and for remains to be determined. patients with malignancy: right vs left—4.52% vs 1.09%, P < 0.001). Right hepatectomy was also associated with sig- PATIENTS AND METHODS: Nineteen patients with nifi cantly higher incidence of several other complications potentially curable bilobar metastases from CRC were eli- such as organ space infections, pneumonia, unplanned re- gible for this single-center, nonrandomized trial during a intubation, pulmonary embolism, failure to wean off the period between September 2008 and August 2012 (NAC ventilator, renal insuffi ciency, urinary tract infection, blood group). The study group consisted of 13 men and 6 women, transfusion, deep venous thrombosis and sepsis (P value < with median age of 63 (range 52 to 79) years. Eligible cri- 0.05). teria included synchronous liver metastases and metastatic liver disease developed within one year after resection of the primary lesions. Patients received biweekly oxaliplatin, 5-fl uorouracil, and folic acid (FOLFOX6) plus bevacizumab therapy. The sixth cycle of neoadjuvant chemotherapy (NAC) did not include bevacizumab, resulting in 4 weeks window-time between the last administration of bevaci- zumab and hepatectomy. Over all survival (OS) and pro- gression free survival (PFS) were compared with 27 patients
who underwent hepateictomies for bilober metastasis dur- Poster Abstracts ing 2002 and 2008 (non NAC group).
RESULTS: Synchronous liver diseases were observed in 14 Sunday (73%). Although objective response to NAC was achieved in 6 patients (32%), 16 patients (84%) underwent liver resection. The liver surgery included 4 hemihepatecitomies, 5 sectorectomies, and 7 partial resections of the liver with median operative time of 186 minutes and median blood loss of 340 mL without blood transfusion. Any postopera- tive morbidity or morbidity was observed. One- and three- year OS of the NAC group were 100% and 56% (MST 43 months), and those of the non-NAC group were 93% and 49% (MST 31 months), respectively (P = 0.47). DFS of the two groups were not different (P = 0.50). Among the hepa- tectomized of NAC group, 10 patients (60%) developed recurrence with median relapse free time of 16.6 months. Initial recurrent deposits were observed in remaining liver in 4 patients, lung in 3, lymph nodes in 3, and peritoneum in 1 (redundant included). CONCLUSION: The results of our study demonstrate that CONCLUSION: Our data suggest that FOLFOX6 and beva- mortality and morbidity after right hepatectomy remains cizumab can be safely administered until 4 weeks before signifi cantly higher than left hepatectomy. liver resection in patients with liver metastases from CRC without increasing perioperative complications. Although no contributions to OS and DFS were observed, control of liver recurrence may be achieved. Adjuvant therapies and further study is needed to defi ne the survival benefi t of NAC with FOLFOX6 plus bevacizumab in patients with potentially curable bilobar metastases from CRC.
89 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1640 Su1641 Hepatectomy for Liver Metastases from Gastric and Portal Venous Thrombosis After Distal Pancreatectomy: Esophageal Cancer: Tumor Biology and Surgical Results Risk Factors and Outcomes Defi ne Outcome Ashwin S. Kamath1, Michael L. Kendrick1, Michael G. Sarr1, Andreas Andreou1, Luca Viganò2, Giuseppe Zimmitti2, David M. Nagorney1, Robert Mcbane2, Michael B. Farnell1, Martin Dreyer1, Jean-Nicolas Vauthey3, Peter Neuhaus1, Kaye M. Reid Lombardo1, Florencia G. Que1 Daniel Seehofer1, Lorenzo Capussotti2, Sven-Christian Schmidt1 1General Surgery, Mayo Clinic, Rochester MN, Rochester, MN; 1Department of General, Visceral and Transplant Surgery, Charité 2Cardiovascular Diseases, Mayo Clinic, Rochester MN, Rochester, MN – Universitätsmedizin Berlin, Campus Virchow Klinikum, Berlin, AIM: Outcomes of patients developing portal vein (PV) Germany; 2Department of HPB and Digestive Surgery, Ospedale thrombosis (PVT) after distal pancreatectomy (DP) are Mauriziano “Umberto I”, Turin, Italy; 3Surgical Oncology, The unknown. The goal of this study was to identify risk fac- University of Texas MDAnderson Cancer Center, Houston, TX tors for PVT and describe the long term outcomes in these patients. BACKGROUND: The role of hepatectomy for patients with liver metastases from gastric and esophageal cancer METHODS: Patients undergoing DP without repair or (GELM) is not well defi ned. The present study examined reconstruction of the PV between 2001 and 2011 were the morbidity, mortality and long-term survival after liver included. Patients that showed evidence of PVT on pre- resection for GELM. operative imaging were excluded from the study. Location and extent of thrombosis was determined by postoperative PATIENTS AND METHODS: Clinicopathological data of CT or ultrasound imaging in all patients. Evidence of sys- patients who underwent hepatectomy for GELM between temic thrombosis (if present) in addition to PVT was also 1987 and 2012 at two European high-volume hepatobiliary documented. centers were assessed and predictors of overall survival were identifi ed. RESULTS: In the study period, 991 patients underwent DP and 21 (2.1%) patients were diagnosed with PVT. Pancreatic RESULTS: Fourty-six patients underwent liver resection neoplasm was the most frequent indication for operation for GELM. The primary tumor was located in the stom- (n = 11). Thrombus occurred in the main PV in 15 and the ach and in distal esophagus in 40 and 6 cases, respectively. right branch of the PV in 8 patients. Complete PV occlusion GELM were synchronous to primary tumor in 33 patients occurred in 9 patients with a median time to diagnosis of and multiple in 18 patients. In 13 cases, major hepatec- 16 days (range 5–85 days). Seventeen patients were antico- tomy (resection ≥3 liver segments) was performed. Thirty- agulated for a median duration of 6 months (range 3.3–36 day postoperative morbidity and mortality rates were 33% months) after the diagnosis of PVT. Over a median follow and 2%, respectively. After a median follow-up time of up of 22 months, resolution of PVT occurred in 7 patients. 76 months (range 1–135), 1, 3 and 5 year overall survival Predictors of non-resolution of PVT included anesthesia rates were 70%, 40% and 27%, respectively. Outcomes time >180 minutes (p = 0.025), DM type II (p = 0.03), BMI were comparable between the two centers. At univariate >30 Kg/m2 (p = 0.03), occlusive PVT (p < 0.001), or throm- analysis, primary tumor invasion of other organs (T4) (P = bus in a sectoral branch (p = 0.02). Anticoagulation therapy 0.004), poorly differentiated carcinoma (P = 0.006), posi- did not infl uence the frequency of thrombus resolution tive lymph node metastases, (P = 0.006), the need for blood and was complicated by gastrointestinal hemorrhage in 4 transfusions at hepatectomy (P = 0.02), major hepatectomy patients. There was no mortality as a direct result of PVT or (P = 0.017) and major posthepatectomy complications (P anticoagulation. = 0.001) were associated with worse overall survival after liver resection. Independent risk factors for shorter overall CONCLUSION: PVT after distal pancreatectomy is a rare survival identifi ed in multivariate analysis included poorly complication. Serious complications as a direct result of differentiated carcinoma (hazard ratio [HR] = 3.1, 95% con- PVT in this setting are uncommon and are not dependent fi dence interval [CI] = 1.17–8.15, P = 0.022), major hepa- on thrombus resolution. Although anticoagulation does tectomy (HR = 3.0, 95% CI = 1.22–7.39, P = 0.017) and not appear to infl uence the rate of PVT resolution in this major posthepatectomy complications (HR = 4.1, 95% CI = small retrospective series, we support the use of anticoagu- 1.31–12.57, P = 0.015). lation until larger, controlled-studies defi ne clear advan- tages or disadvantages. CONCLUSIONS: liver resection should be considered in selected patients with GELM. Patients with poor differen- tiated tumor and those who require major hepatectomy because of more advanced disease derive the least benefi t from this approach.
90 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Clinical: Pancreas CONCLUSION: PET is a more sensitive modality for iden- tifying metastatic disease than CT or MRI, however, it has a lower specifi city and lower positive predictive value. While Su1642 PET identifi ed an additional 5.6% of patients with occult metastatic disease, it is likely that unresectability would Positron Emission Tomography (PET) Has Limited have been established at diagnostic laparoscopy, thus not Utility in Preoperative Staging of Pancreatic saving an unnecessary resection. We therefore conclude Adenocarcinoma that PET has limited utility in workup of patients who Peter Einersen1, Irene Epelboym1, Megan Winner1, already undergo CT or MRI as part of initial staging of pan- David Leung2, John A. Chabot1, John D. Allendorf1 creatic adenocarcinoma. 1Surgery, Columbia University Medical Center, New York, NY; 2Radiology, Columbia University Medical Center, New York, NY Su1644 BACKGROUND: Utility of positron emission tomography Human Equilibrate Nucleoside Transporter 1 (PET) as an adjunctive imaging modality to CT or MRI in Expression Predicts Survival of Pancreatic Cancer evaluating resectability of pancreatic cancer is a subject of Patients Trated with Gemcitabine-Based Adjuvant controversy. In this study, we seek to assess the utility of PET in identifying occult metastatic disease, as well as evaluate Chemotherapy After Resection 1 2 1 predictive value of maximum standard uptake value (SUV) Toshiyuki Moriya , Shigemi Fuyama , Yukinori Kamio , with respect to tumor resectability and patient survival. Koichiro Ozawa1, Shigeo Hasegawa1, Masaomi Mizutani1, Takayuki Higashi1, Moriyoshi Yokoyama1, Osamu Usuba1 METHODS: Cross sectional imaging, clinical course, oper- 1 ative outcomes, and overall survival of all patients who pre- Surgery, Okitama Public General Hospital, Kawanishi, Japan; 2 sented with pancreatic adenocarcinoma and had PET scan Pathology, Okitama Public General Hospital, Kawanishi, Japan in workup were reviewed retrospectively. Resectability was BACKGROUND: Gemcitabine is promising adjuvant che- assessed based on established criteria. Continuous variables motherapy for patients with resected pancreatic cancer. were compared using Student’s t-test or ANOVA. Categori- Human equilibrative nucleotide transporter-1 (hENT1) is cal variables were compared using chi-square or Fisher’s the major transporter responsible for gemcitabine uptake exact test. Prediction models were constructed using linear into cells. The aim of the current study is to investigate or logistic regression where appropriate. whether hENT1 expression can predict the survival of pan- RESULTS: Complete imaging and follow-up data was avail- creatic cancer patients treated with adjuvant gemcitabine- able for 123 patients evaluated from 2005 to 2011. Of this based chemotherapy after pancreatic resection. cohort, 36 patients (29%) were thought to be free of extra- METHODS: Immunohistochemical hENT1 expression was Poster Abstracts pancreatic disease and offered resection, 21 (17%) had met- analyzed in 19 resected pancreatic cancer patients received astatic disease, and 66 (53%) were deemed locally advanced gemcitabine-based adjuvant chemotherapy. Relation- Sunday and referred for neoadjuvant therapy. PET and CT/MRI were ships between various clinicopathological factors includ- concordant in 108 (88%) cases, however metastatic lesions ing hENT1 expression and patient survival were examined were identifi ed in 7 (5.6%) patients deemed resectable by using univariate and multivariate analysis. CT or MRI. Among those offered immediate resection, 5 RESULTS: Sixteen (84%) patients highly expressed hENT1. (14%) patients had occult metastatic disease identifi ed at Patients with low expression of hENT1 showed signifi cantly diagnostic laparoscopy, including 3 previously identifi ed by worth outcome than high expression group [2-year survival: nonconcordant PETs and 2 missed by false negative PETs. 0% for low expression group vs. 60% for high expression False positive PETs led to unnecessary procedures delaying group; HR 8.72, 95% confi dence interval (CI) 1.70–44.76, p surgery for 3 (8.3%) patients who went on to resection. In a = 0.009] (Figure), although low hENT1 expression was not cohort of patients thought to be free of metastatic disease, independent prognostic factor by multivariate analysis (HR in terms of detecting metastases, overall sensitivity and 4.41, 95% CI 0.71–27.39, p = 0.111). Other prognostic fac- specifi city of PET were 89.3% and 85.1%, respectively, com- tor was only AJCC stage [2-year survival 0% (III, IV) vs. 64% pared with 62.5% and 93.5% for CT and 61.5% and 100.0% (I, II), HR 6.24, 95% CI 1.35–28.85, p = 0.0192]. Tumor size, for MRI. Positive predictive value and negative predictive lymph node metastasis, and residual tumor (R1 vs R0) did value of PET were 64.1% and 96.4% respectively, compared not reach signifi cant prognostic factor, although the trend with 75.0% and 88.9% for CT and 100.0% and 91.9% for was observed. MRI. Average difference in maximum SUV of resectable and unresectable lesions was not statistically signifi cant (5.65 CONCLUSIONS: Low expression of hENT1 strongly indi- vs. 6.5, p = 0.224), nor was maximum SUV a statistically cated worth outcome of patients treated with adjuvant signifi cant predictor of survival (p = 0.18). gemcitabine-based chemotherapy after pancreatic resec- tion. Gemcitabine-based adjuvant chemotherapy may be useful for only high expression of hENT1. The new adju- vant chemotherapy except gemcitabine might be necessary for patients with low expression of hENT1.
91 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1645 central or total pancreatectomy. The median sCr value was 0.86 (0.30–14.1). 18 patients (1.7%) had severe CKD and 31 The Effect of Preoperative Renal Insuffi ciency on patients (2.9%) had sCr ≥ 1.8. Complications occurred in Postoperative Outcomes Following Pancreatic 622 patients (58.6%), major complications in 198 (18.7%), Resection: A Single Institution Experience of 1061 and respiratory failure in 48 (4.5%). Both severe CKD and Consecutive Patients sCr ≥ 1.8 were associated with any complication, major complications, and respiratory failure on UV analysis. On Malcolm H. Squires1, Vishes V. Mehta1, Sarah B. Fisher1, 1 1 2 MV analysis, severe CKD was associated with increased Neha L. Lad , David Kooby , Juan M. Sarmiento , complications (HR 5.5; 95% CI: 1.3–25.5; p = 0.02) and 1 1 1 Kenneth Cardona , Maria C. Russell , Charles A. Staley , respiratory failure (HR 6.1; 95% CI: 1.8–20.5; p = 0.03), Shishir K. Maithel1 but not major complications. Using sCr ≥ 1.8 as a surro- 1Department of Surgery, Division of Surgical Oncology, Winship gate marker for renal insuffi ciency, patients with sCr ≥ 1.8 Cancer Institute, Emory University, Atlanta, GA; 2Department of had increased risk of any complication (HR 3.5; 95% CI: Surgery, Division of General and GI Surgery, Emory University, 1.3–9.3; p = 0.01), major complications (HR 2.2; 95% CI: Atlanta, GA 1.04–4.8; p = 0.04), and respiratory failure (HR 4.7; 95% CI: 1.8–12.6; p = 0.002, Table). Among patients undergoing BACKGROUND: Chronic kidney disease (CKD) is known Whipple, sCr ≥ 1.8 remained associated with any complica- to adversely affect outcomes after cardiac and vascular sur- tion (HR 3.6; 95% CI: 1.03–12.9, p = 0.05) and respiratory gery. We examined the effect of preoperative renal insuf- failure (HR 3.9; 95% CI: 1.2–12.8; p = 0.03), and demon- fi ciency on postoperative outcomes following pancreatic strated a trend towards increased major complications (HR resection. 2.3; 95% CI: 0.9–6.0; p = 0.10). Among patients undergo- METHODS: All patients who underwent pancreatic resec- ing distal pancreatectomy, sCr ≥ 1.8 demonstrated a trend tion at a single institution between 1/2005 and 7/2012 were towards increased complications (HR 6.8; 95% CI: 0.8–54.6; identifi ed. Creatinine clearance (CrCl) was estimated by the p = 0.07), was not prognostic for major complications, but MDRD (Modifi cation of Diet in Renal Disease) formula. remained a signifi cant risk factor for respiratory failure (HR Severe CKD was defi ned as CrCl < 30 mL/min (CKD stages 15.4; 95% CI: 2.2–106.3; p = 0.006). 4–5). Renal function also was analyzed by using serum cre- CONCLUSION: Few patients with signifi cant renal insuf- atinine (sCr) dichotomized at 1.8 mg/dL. Primary outcomes fi ciency are operative candidates for pancreatic resection. were any complication within 30 days, Clavien Class III-V Severe CKD (stages 4–5) is associated with increased risk of major complication, and respiratory failure. Multivariate complication and respiratory failure, but may be of limited (MV) models for each endpoint were constructed by includ- clinical utility. Serum creatinine ≥1.8 mg/dL may serve as a ing all variables with a p-value ≤0.1 on univariate (UV) useful surrogate marker of renal insuffi ciency and identifi es analysis. patients at signifi cantly increased risk of any complication, RESULTS: 1061 patients were identifi ed; 709 underwent major complication, and respiratory failure after pancreatic pancreaticoduodenectomy (Whipple), 307 distal, and 45 resection.
Multivariate Analysis of all Pancreatic Resections (n = 1061)
Any Complications Major (Clavien III-V) Complications Respiratory Failure Variable HR (95% CI) p–value Variable HR (95% CI) p–value Variable HR (95% CI) p–value sCr 1.8 3.5 (1.3–9.3) 0.01 sCr 1.8 2.2 (1.04–4.8) 0.04 sCr 1.8 4.7 (1.8–12.6) 0.002 Age 1.01 (1.001–1.02) 0.04 Age 1.01 0.10 Age 1.02 0.17 (0.99–1.02) (0.99–1.05) HTN 1.1 (0.9–1.5) 0.41 HTN 1.3 (0.9–1.8) 0.17 HTN 1.4 (0.7–2.7) 0.35 Intra-op transfusion 1.8 (0.9–3.4) 0.06 Intra–op transfusion 1.8 (0.9–3.3 0.07 Intra–op transfusion 2.2 (0.8–6.0) 0.12 Male gender 1.4 (1.1–1.9) 0.01 COPD 2.3 (1.03–5.3) 0.04 Pre-op biliary stent 1.4 (1.1–1.9) 0.01 Albumin <3 2.1 (1.2–3.9) 0.02 Intra-op drain placement 1.5 (1.1–1.9) 0.003 HR, hazard ratio; CI, confi dence interval; sCr, serum creatinine (in mg/dL); HTN, hypertension.
92 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1647 CONCLUSION: Clinical factors affecting survival were pre- operative abdominal pain and hypertension. Intraoperative Clinicopathologic Features Infl uencing Survival in factors affecting survival were EBL and the need for blood Patients with Resected Pancreatic Adenocarcinoma transfusions. The only pathological factor affecting survival by Pancreaticoduodenectomy was LNR. Finally, an uncomplicated postoperative course Cynthia Weber1, Eileen Bock1, Michael G. Hurtuk1, was positively correlated with survival. Gerard Abood1, Margo Shoup2, Jack Pickleman1, Gerard V. Aranha1 Su1648 1 Surgical Oncology, Loyola University Medical Center, Maywood, IL; A Comparison of the 2-Year Longitudinal Impact of 2 Surgery, Cadence Health, Winfi eld, IL Surgical Versus Endoscopic Pancreatic Pseudocyst OBJECTIVE: To determine clinicopathological features Drainage on Healthcare Utilization and Morbidity that infl uence survival in patients with resected pancreatic Jennifer M. Whittington, Scott D. Stevens, Daniel L. adenocarcinoma. Davenport, Austin Ward, Andrew C. Bernard, Shaun P. Mckenzie METHODS: A retrospective review of a prospective data- University of Kentucky, Lexington, KY base was conducted for patients undergoing pancreatico- INTRODUCTION: Previous reports have concluded that duodenectomy for pancreatic adenocarcinoma at a single endoscopic drainage (endo) of pancreatic pseudocysts has institution from December 1993 to December 2010. Clini- an advantage over surgical cystgastostomy (open) in terms copathologic features and cancer related outcomes were of both costs and morbidity. No study to date has looked collected. The cohort was then analyzed for clinicopatho- longitudinally at the overall benefi t of these two strategies. logical features infl uencing survival at 6 months, 1 year, 3 The purpose of our study was to compare 2-year resource years, and 5 years. utilization and morbidity between endo and open treat- RESULTS: A total of 246 patients underwent pancreatico- ment of pancreatic pseudocyst. duodenectomy for pancreatic adenocarcinoma. The cohort METHODS: This study is a single center retrospective case was comprised of 128 males (52%) and 118 females (48%), review of patients treated between September 2004 and with a median age of 68 years. Median operative time was December 2011 for pancreatic pseudocyst. We extracted 6.25 hours with a median blood loss of 800 cc. Median hos- clinical data from the initial procedure related admission pital length of stay was 8 days. A total of 7 patients (2.8%) along with post-procedure emergency department (ED) required re-operation and 9% of patients were readmitted visits and hospital readmissions for up to two years. We within 30 days for postoperative issues. The 30-day mortal- calculated a composite morbidity scale ranging from 1) no ity rate was 2.4% (n = 6).
intervention to 2) minor intervention (antibiotics), 3) read- Poster Abstracts There was a total of 101 associated complications in the mission, 4) repeat procedure or ICU care, to 5) death. Fish-
postoperative period, with grade 3 or less accounting for er’s exact tests, t tests and Mann-Whitney U tests were used Sunday 79% of the observed complications, based on the Dindo/ to compare characteristics between the two groups where Clavien complication scoring system. In regards to pan- appropriate. creaticoduodenectomy specifi c complications, 29 (11.8%) RESULTS: We identifi ed 45 patients who had undergone experienced delayed gastric emptying, 17 (6.9%) developed drainage procedures, 17 endo and 28 open. Three endo an anastomotic leak, with ISGPF grade A/B accounting for patients who required conversion to open were classifi ed as the majority of leaks observed (6%). Overall survival of the endo by intention to treat. Median follow up for the study cohort was 85%, 63%, 25%, and 15% at 6 months, 1 year, was 24 months. The two groups had similar etiologies, age, 3 years, and 5 years respectively, with a median survival of gender and clinical risks (table). The open group had more 17 months. multicysts and cysts with debris on imaging, but not sig- Using multivariate logistic regression, clinical factors that nifi cantly so. There was a trend toward more gastric varices infl uenced survival were abdominal pain and preoperative in the endo group (29.4% vs. 7.1%, P = .09) but venous HTN, where the presence of pain preoperatively negatively thromboses were similar in both groups (58.8% vs. 57.1%). correlated with survival at 3 years (p = 0.021), and the pres- While initial morbidity was higher in the open group, read- ence of preoperative hypertension was negatively correlated mission occurred more than twice as often in endo patients with survival at 6 months, 3 years, and 5 years (p = 0.012, (70.6% vs. 32.1%, P = .02) and total 2-year hospital days p = 0.013, p = 0.019). Intraoperative estimated blood loss did not differ signifi cantly in the two groups (p = 0.23). (EBL) showed a negative correlation with survival at 3 years There was a trend towards increased procedural readmis- (p = 0.02), and the need for intraoperative blood transfu- sions in the endo group (p = 0.07). In the open group, two sion was negatively correlated with survival at 3 years and patients required subsequent repair of ventral hernias and 5 years (p = 0.012 and p = 0.019). The only pathologic fac- one patient required two surgeries for postoperative vari- tor to have a negative impact on survival was lymph node ceal bleeding. In the endo group three patients required ratio (LNR), which correlated with decreased survival at repeat percutaneous drainage and one required repeat endo 6 months, 1 year, and 3 years (p = 0.033, p = 0.035, p = drainage in addition to the three conversions to open men- 0.01). Those who had no postoperative complications had tioned above. higher odds of being alive at 6 months (p = 0.002)
93 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Patient Characteristics, Imaging and Outcomes insuffi ciency, representing fat absorptive disturbance was defi ned as percent of cumulative 7-hour 13CO2 exhalation Variable Endo Open P-value (% dose 13C cum 7 h) <5%. Relationship with histological No. Patients 17 28 degree of islet cells at cut margin and postoperative HbA1c Male/Female 11/6 20/8 0.74 alteration were analyzed. In this study, diabetic patients Mean age, y (S.D.) 47.2 (11.3) 50.3 (13.7) 0.42 were identifi ed as treatment with insulin or oral hypoglyce- Current Smoker, Diabetes, COPD 13 (76.5%) 25 (89.3%) 0.40 mic medications or HbA1c level ≥ 6.9% (NGSP). and/or Cardiac History RESULTS: Preoperatively, 14 patients (35%) were diabetes Etiology 1.00 and 26 patients (65%) were non-diabetes. In non-diabetes Anatomic 8 (47.1%) 14 (50.0%) 26 patients, 12 patients (46%) developed impaired glucose EtOH 8 (47.1%) 11 (39.3%) tolerance within one year after DP. 8 patients were adminis- Other (HLD/Trauma) 1 (5.9%) 3 (10.7%) trated oral hypoglycemic medications, one patient needed Imaging insulin treatment and other 3 patients were not adminis- Venous Thrombosis 10 (58.8%) 16 (57.1%) 1.00 trated any medication. Differences in % dose 13C cum 7 Gastric Varices 5 (29.4%) 2 (7.1%) 0.09 h were not signifi cantly between patients with DP (9.7 ± Multiple Cysts 2 (11.8%) 9 (32.1%) 0.29 3.2%) and healthy controls (13.3 ± 5.9%). No healthy con- Debris in Cysts 8 (47.1%) 18 (64.3%) 0.35 trols had pancreatic exocrine insuffi ciency if the diagno- Outcomes sis was based on a % dose 13 C cum 7 h less than 5%. In Readmitted 12 (70.6%) 9 (32.1%) 0.02 patients after DP operation, only one patient was 5% % Procedural Readmission (s) 7 (41.2%) 4 (14.3%) 0.07 dose 13 C cum 7 h less than 5%. In the 26 pre-OP non-DM ED visit (s) 5 (29.4%) 3 (10.7%) 0.23 patients the average percentage of islet cells at cut margin Total Hospital Days Overall, 13.6 (18.0) 19.7 (39.7) 0.49 was signifi cantly lower in the post-OP DM group than in mean (S.D.) the post-OP non-DM group (1.5 ± 0.7% vs 3.5 ± 1.5%, P = Median Morbidity Score 4 (1.5–4) 2 (1–4) 0.24 0.01). The average area ratio of islet cells at cut margin was (Interquartile Range) correlated with postoperative HbA1c level (P = 0.025). CONCLUSION: Differences of fat absorptive function were CONCLUSIONS: While endoscopic drainage of pancre- not signifi cant between patients with DP and healthy con- atic pseudocysts may result in less initial procedure related trols. Perioperative histological degree of islet cells at cut morbidity and length of stay, it is associated with increased margin is predictive of glucose metabolism insuffi ciency readmissions, increased procedure related admissions due after distal pancreatectomy (DP). to treatment failure and does not provide signifi cant benefi t in overall hospital days when compared to surgical drain- Su1650 age. Further studies are necessary to select which patients are optimal candidates for each approach. Preoperative Prediction of the “High-Risk Pancreas” by Artifi cial Neuronal Network Analysis of over 450 Su1649 Pancreatoduodenectomies Hryhoriy Lapshyn1, Frank Makowiec1, Dirk Bausch1,2, Histological Degree of Islet Cells at Cut Margin Ulrich T. Hopt1, Tobias Keck1,2, Ulrich Wellner1,2 Indicates Postoperative Glucose Metabolism 1Clinic for General and Visceral Surgery, University of Freiburg Medical Insuffi ciency After Distal Pancreatectomy Center, Freiburg, Germany; 2Surgery, University Hospital of Schleswig- 1,2 2 2 Masahiko Morifuji , Yoshiaki Murakami , Kenichiro Uemura , Holstein Campus Lübeck, Lübeck, Germany Takeshi Sudo2, Yasushi Hashimoto2, Taijiro Sueda2, Akio Sakamoto1 INTRODUCTION: Pancreatoduodenectomy (PD) has 1Internal Medicine, Sanmu Medical Center, Chiba, Japan; 2Surgery, become a standard operation with low mortality in high- Hiroshima University, Hiroshima, Japan volume centers, however perioperative morbidity remains BACKGROUND: Pancreatogenic diabetes after pancre- substantial, mainly due to postoperative pancreatic fi stula atectomy is of growing importance due to the increasing (POPF). Development of preoperative protective measures life expectancy of pancreatectomized patients. This study is hampered by a lack of strictly preoperative risk stratifi ca- aimed to investigate whether perioperative histological tion. Predictive power of single parameters can be enhanced degree of islet cells at cut margin are predictive of endocrine by optimally weighed combination of risk factors in an arti- pancreatic function after distal pancreatectomy (DP). fi cial neuronal network (ANN). MATERIALS AND METHODS: This study included con- METHODS: A panel of clinical and radiological parameters secutive 40 patients who underwent distal pancreatectomy were assessed retrospectively from patients with pancreato- (DP). The percent of islet cells of each pancreas was deter- duodenectomy in our institution and risk factors analysis mined via histological examination of resected specimen at for the endpoint POPF (clinically relevant Grade B/C of pancreatic cut margin. Pre and postoperative HbA1c levels ISGPS defi nition) were identifi ed. Preoperatively available were measured in blood samples to assess postoperative glu- parameters were used for prediction of a high risk pancreas cose metabolism insuffi ciency. For assessing post operative in an ANN. Internal validation of the thereby identifi ed risk fat absorptive function after DP, non-invasive 13C-mixed group was performed by testing for POPF and other rele- triglyceride breath test (13C-MTG-T), labeled long chain vant complications. triglyceride mixture was performed. Pancreatic exocrine
94 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
RESULTS: A total of 471 patients with PD operated from colonization in the surgical drain. In the patients without 2001 to 2012 were included. Out of twelve clinical and bacteria in surgical drain, only 1% of the patients devel- radiological risk factors for POPF B/C, the most powerful oped CR-POPFs, while 29% of the patients with bacteria in was a soft pancreas. When an ANN was trained to predict a surgical drain developed CR-POPFs (p < 0.01). Moreover, in soft high-risk pancreas, correct prediction was achieved in the patients without hyperamylasemia and no bacterial col- 83% in the test group. Patients predicted to have a high- onization in the surgical drain, no patients developed CR- risk pancreas had a signifi cantly higher rate of POPF and POPF, while 60% of the patients with CR-POPF had both severe complications compared to the low-risk group (POPF hyperamylasemia and bacterial colonization in the surgical B/C (38% vs 8%, p = 0.000), intraabdominal abscess (23% drain (p < 0.01). vs 10%, p = 0.000), severe complications (26% vs 13%, p CONCLUSION: Bacterial infection in addition to activa- = 0.003), severe postpancreatectomy hemorrhage (18% vs tion of pancreatic enzyme around the pancreatico-enteric 6%, p = 0.012)), as well as a fi ve-fold elevated mortality (5% anastomosis might play an important role in the patho- vs 1%, p = 0.034). genic mechanism of CR-POPF after PD. Prevention of post- CONCLUSION: Clinical and radiological parameters com- operative pancreatitis of remnant pancreas with infection bined in an ANN model can correctly predict a high-risk pan- control might be an area of focus for reducing the incidence creas and severe complications already before the operation. of CR-POPF after PD.
Su1651 Su1652 Activation of Pancreatic Enzyme Plus Bacterial Evolution of the Treatment of Gastroduodenal Artery Infection Plays an Important Role in the Pathogenic Pseudoaneurysms and Mesenteric Arterial Hemorrhage Mechanism of Clinically Relevant POPF After Following Pancreaticoduodenectomy Pancreaticoduodenectomy Joseph Chen1, Laura Findeiss2, Aram N. Demirjian1, Kenichiro Uemura, Yoshiaki Murakami, Takeshi Sudo, David K. Imagawa1 Yasushi Hashimoto, Naru Kondo, Naoya Nakagawa, 1Surgery, University of California-Irvine, Orange, CA; 2Radiology, Hayato Sasaki, Kenjiro Okada, Hiroki Ohge, Taijiro Sueda University of California-Irvine, Orange, CA Surgery, Hiroshima University, Hiroshima, Japan INTRODUCTION: Postoperative mortality in high vol- BACKGROUND: Postoperative pancreatic fi stula (POPF) ume centers for pancreaticoduodenectomy (Whipple) has after pancreaticoduodenectomy (PD) is relatively common, decreased to less than 4%. Late postoperative bleeding and remains a major cause of morbidity and surgical mor- occurs in 0.5–5% of cases, with reported mortality rates of tality. However, the underlying pathogenic mechanism of up to 60%. Poster Abstracts POPF, with the exception of technical error, still remains PATIENTS/METHODS: This is a retrospective analysis of unclear. We previously reported that postoperative pancre- 313 patients who underwent pancreaticoduodenectomy Sunday atitis after PD plays an important role in the pathogenic from 2003–2012 at our institution, a high-volume, multi- mechanism of POPF after PD. We hypothesized that the disciplinary hepato-pancreato-biliary center. The main out- bacterial infection in addition to the activation of pancre- come measure was mortality. atic enzyme around the pancreatico-enteric anastomosis could be associated with occurrence of clinically relevant RESULTS: From 2003–2012, 10 out of 313 patients (3%) POPF (CR-POPF) after PD. presented with delayed major hemorrhage following pan- creaticoduodenectomy, occurring between postoperative OBJECTIVES: We retrospectively analyzed the possible days 6–18. Visceral arteries known to be affected were the association of postoperative pancreatitis, bacterial coloni- gastroduodenal artery (GDA) (4), hepatic artery (3), and zation in the surgical drain, and CR-POPF after PD using the pancreaticoduodenal artery (1). 5 patients presented prospectively collected data base. with gastrointestinal hemorrhage and 5 patients presented METHODS: 250 consecutive patients undergoing PD were with hemoperitoneum. 1 patient underwent immediate included. All patients were administered prophylactic anti- operative intervention, 2 patients underwent immediate biotics, which were selected based on perioperative bile cul- operation followed by percutaneous intervention by inter- tures. POPFs were diagnosed by International Study Group ventional radiology (IR). Immediate IR intervention was Pancreatic Fistula (ISGPF) criteria. Hyperamylasemia was performed in 7 patients. Mortality from GDA/visceral arte- defi ned as serum amylase more than 3 times the upper limit rial hemorrhage occurred in 1 patient (10%). of the reference value. Closed suction drains were inserted CONCLUSION: Delayed mesenteric arterial hemorrhage along the pancreatico-enteric anastomosis, and surgical following pancreaticoduodenectomy requires early recogni- drains were examined bacteriologically when they were tion and management. The mortality rate in our early expe- removed. rience with immediate operative intervention was 33%. A RESULTS: Of 250 patients, 23% developed POPF; Grade modifi ed operative technique led to preservation a long A in 16%, Grade B in 6%, and Grade C in 1%. A total of GDA stump and use of a large metallic clip as a radiographic 32% of the patients had hyperamylasemia on postopera- marker/guide. This was found to aid in easier and quicker IR tive day (POD) 1, and the presence of hyperamylasemia localization and coiling or stenting of the bleeding vessel. on POD1 was closely associated with the development of The mortality rate has decreased to 0% in patients undergo- POPF (p < 0.01). A total of 43% of the patients had bacterial ing immediate IR intervention.
95 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1653 Su1655 Trends in Pancreatic Surgery: Indications, Operative Is Frozen Section Histopathology of Any Value in Techniques and Postoperative Outcome of 1120 Patients Undergoing Resection of Intraductal Papillary Pancreatic Resections Mucinous Neoplasms? Frank Makowiec, Tobias Keck, Ulrich ADAM, Hartwig Riediger, Daniel Joyce, Gavin A. Falk, Kevin M. El-Hayek, Sricharan Uwe A. Wittel, Ulrich F. Wellner, Ulrich T. Hopt Chalikonda, Gareth Morris-Stiff, Matthew Walsh Department of Surgery, University of Freiburg, Freiburg, Germany Department of General Surgery, Section of Surgical Oncology/HPB, Low mortality rates after pancreatic resection (PaRes) have Cleveland Clinic Foundation, Cleveland, OH been reported by many centers. Hospital volume, sur- INTRODUCTION: Intraductal papillary mucinous neo- geon volume and adequate management of complications plasms (IPMN) are cystic lesions of the pancreas that follow are factors contributing to a better outcome. The aim of a step-wise dysplastic sequence from adenoma to invasive our study was to evaluate trends in indications, operative adenocarcinoma. Obtaining a frozen section (FS) at the time techniques and postoperative outcome in more than 1100 of pancreatic resection could be important to determine PaRes performed in our institution since 1994. whether additional resection of the remnant is required. METHODS: 1120 PaRes were performed since 1994. The The aim of this study is to report the correlation between FS vast majority of the operations was performed by three of the pancreatic neck and fi nal histopathology for patients surgeons. The perioperative data were documented in a with IPMN including those with IPMN carcinomas. pancreatic database. For our analyses the study period was METHODS: The departmental pancreatic cyst database subclassifi ed into three periods (A 1994–2001/n = 363; B was interrogated to identify all patients with a histopatho- 2001–2006/n = 305; C since 2007/n = 452). logical diagnosis of IPMN with or without pancreatic ade- RESULTS: 81% of the PaRes were personally performed by nocarcinoma arising from within the IPMN. The degree one of the 3 principal surgeons. The average annual number of dysplasia on the fi nal pathology report was classifi ed as of PaRes increased from 52 (period A) to 80 (C; n = 107 in high (HGD), moderate (MGD), or low (LGD. Frozen section 2011). The median age increased from 51 (A) to 65 years (C; results were reviewed with particular reference to identifi - p < 0.001). In the entire group (n = 1120) indications for sur- cation of invasive carcinoma or high-grade dysplasia and gery were pancreatic/periampullary cancer (49%), chronic these fi ndings were compared to fi nal histopathological pancreatitis (CP; 33%) and various other lesions (18%). The fi ndings, and related to patient outcome. percentage of PaRes for CP decreased from over 50% in period RESULTS: During the period January 2000 to December A to 17% (C; p < 0.01). In contrast the frequency of IPMNs 2011, 121 patients underwent resection, consisting of 41 increased from below 1% (A) to 8% (C; p < 0.05). About two patients with an invasive carcinoma and 80 with IPMN thirds of the operations were pancreaticoduodenectomies alone: HGD [n = 18]; MGD [n = 14]; and LGD [n = 48] (most PPPD). Due to the lower numbers of operations for CP (on fi nal pathology). There were 70 females and 51 males the rates of duodenum-preserving resections decreased from with a median age of 68 years (IQR: 58–73). Of the patients 18% (A) to 4% (C; p < 0.05). A more aggressive approach with IPMN carcinomas, 36 (88%) had a FS. Carcinoma or in some patients with cancer and more resected IPMNs led HGD was seen at the transaction margin on FS in 4 patients to an increase in total pancreatectomies during the study undergoing pancreatoduodenectomy leading to 4 extended period from 1% (A) to 6% (C). The frequency of mesenterico- resections, 2 of which were total pancreatectomies. There portal vein resections increased from 8% (A) to 20% (C; p < was 1 false-positive for invasive cancer that was found to 0.01). Distal resections were performed in 17%. Laparoscopic be non invasive on fi nal pathology and 1 false-negative for pancreatic head and distal resections were introduced by one HGD/invasive carcinoma on frozen section that was found surgeon in period C and were performed in 4.7% of all cases to be an invasive cancer on fi nal pathology. For those with (12% of the cases in period C). Overall mortality was 2.4% IPMN alone, 64 had frozen section analysis performed. and comparable in the 3 periods (2.8%, 2.0%, 2.4%; p = 0.8). None had carcinoma/HGD at the transection margin on FS The 3 principal surgeons in our series also had comparable or on subsequent histopathology. 3 patients in this group mortality rates (1.9–3.4%; p = 0.41). Overall complication died of IPMN-related carcinomas in their remnant pan- rates increased from 42% (A) to 56% (C; p < 0.01). The rate creas. 2 had HGD on their initial resection and 1 had only of pancreatic leak grade B/C also increased from 5% (A) to LGD, and all developed the subsequent cancers away from 12% (C; p < 0.01) but the frequencies of relaparotomies were the transaction margin. comparable (10–14%; n.s.) CONCLUSIONS: Frozen section analysis allows identifi ca- CONCLUSIONS: Operative mortality in our high-volume tion of foci of carcinoma or HGD at the transection margin institutional series of more than 1100 pancreatic resections during pancreatic resection for IPMN that should result in was low throughout the study period. Mortality remained further resection. However, the development of progressive low despite a more aggressive surgical approach to (malig- disease in the pancreatic remnants of patients without ini- nant) pancreatic disease (more extended resections, more tial evidence of carcinoma means that radiological surveil- vein resections, older patients). An increased overall mor- lance is required for this cohort. bidity may be explained by more clinically relevant pan- creatic fi stulas (more patients with soft pancreas) and better documentation (many patients in randomized studies after period A).
96 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1656 Su1657 Endoscopic and Surgical Alternatives to Central Pancreatic Resection Pancreaticoduodenectomy and Distal Pancreatectomy Vichin Puri, Vijay G. Menon, Alagappan Annamalai, Jennifer K. Plichta, Eileen Bock, Michael G. Hurtuk, Nicholas N. Nissen Gerard Abood, Gerard V. Aranha Hepatobiliary and Pancreatic Surgery, Cedars-Sinai Medical Center, Surgery Department, Loyola University Medical Center, Maywood, IL Los Angeles, CA PURPOSE: While standard resections such as pancreati- INTRODUCTION: Central pancreatectomy (CP) is an coduodenectomy and distal pancreatectomy are necessary uncommon technique used to treat select pancreatic for malignant disease, low grade tumors and benign lesions pathology. We evaluated the utility and safety of CP over a of the pancreas and duodenum present a unique surgical 10 year time span at a single institution. dilemma. Select patients may benefi t from non-standard METHODS: Review of prospective database (single sur- resections (NSR) which preserve parenchyma and function, geon) from 2003–2012. CP comprised 9% of all pancreatic and thus may avoid the potential complications inherently resections during this period (total of 310). related to more traditional resections. Here, we describe our experience with NSR of various pancreatic and duodenal RESULTS: Thirty patients underwent CP for diagnoses lesions. including neuroendocrine tumor (n = 12), cystic neoplasm (n = 9) and benign stricture (n = 9). Distal pancreatic con- METHODS: A retrospective review of a prospectively col- tinuity was established by pancreaticogastrostomy (n = 7), lected database of 777 patients who underwent resections pancreaticojejunostomy (n = 13), dual pancreaticoenteros- of pancreatic and duodenal lesions between 1999 and tomy (n = 9) or primary pancreatico-pancreatostomy (n = 2012 was conducted. Of these, 45 patients underwent NSR, 1). Major complications were limited to 4 patients (13%) defi ned as pancreatic or duodenal resections excluding stan- who required re-laparotomy or percutaneous drainage. Five dard pancreaticoduodenectomy or distal pancreatectomy. patients (17%) developed postoperative pancreatic fi stu- Clinicopathologic features and outcomes were assessed. lae, of which 3 (10%) were ISGPF grade B/C. There was no RESULTS: In sum, 26 males and 19 females were evaluated; peri-operative mortality. At mean follow-up of 29 months, median age 64 years (range 30–87) and median follow-up no patients have developed recurrent tumor. Two patients 4.4 years (range 0.3–13.3 years). Preoperatively, 32 patients (7%) developed diabetes and no patient has exocrine insuf- underwent EGD, 33 EUS, and 39 CT scans. The median fi ciency. The frequency of CP has remained constant over lesion size was 2.3 cm (range 0.7–9 cm). The various types the study time period, but patients operated in the more of NSR included: 16 pancreas-sparing duodenectomies, 9 recent 5-year period were more likely to have more proxi- central pancreatectomies, 9 enucleations, 6 ampullectomies, mal pathology (pancreatic head) and to undergo dual pan- Poster Abstracts 4 transduodenal polypectomies, and 1 endoscopic polypec- creatic anastomosis (Figure).
tomy. The fi nal pathologic diagnoses included: 12 villous Sunday adenomas, 7 neuroendocrine tumors, 5 mucinous cystad- enomas, 5 stromal tumors, 4 duodenal carcinomas, 3 serous cystadenomas, 3 tubular adenomas, 2 lymphoepithelial cysts, 2 IPMNs, and 2 other pathologies. EUS was 100% accurate in predicting depth of mucosal invasion, while EGD and CT were 100% accurate in identifying the lesion location. Furthermore, the overall accuracy of pre-operative imaging in selecting appropriate patients amenable to NSR was 100%. Overall, fi ve patients developed post-procedure complications (10.9%). Of the central pancreatectomies (n = 9), three developed pancreatic fi stulas (33%), although no patients developed diabetes or steatorrhea. One patient had a subsequent episode of pancreatitis following ampullec- tomy, and one developed a pancreatic pseudocyst requiring endoscopic drainage following enucleation. There were no peri-operative mortalities. Of the four patients with duo- Central Pancreatectomy Managed with Dual Pancreatic-Enteric denal carcinomas, all underwent pancreas-sparing duode- Anastomosis. nectomy, and the overall survival was 50% at the time of CONCLUSION: Central pancreatectomy is a safe and valu- analysis (deaths occurred at 1.7 and 4 years; follow-up for able option for management of select proximal pancreatic two survivors was 4.1 and 11.7 years). pathology and is associated with a low rate of long term CONCLUSION: Based on our fi ndings, EGD, CT, and EUS endocrine or exocrine insuffi ciency. Pancreatic fi stula and were 100% accurate in selecting appropriate patients for surgical complication rates are signifi cant but not prohibi- NSR. Therefore, proper selection of patients using certain tive. Novel reconstruction techniques such as those applied imaging modalities may allow some patients to achieve on our series may allow greater application of this tech- adequate resection, while avoiding more complicated and nique to more proximal pancreatic lesions. morbid procedures, such as pancreaticoduodenectomy or distal pancreatectomy.
97 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1658 Impact of Pancreatoduodenectomy Complications on Adjuvant Therapy and Long-Term Outcomes for Pancreatic Adenocarcinoma Janak A. Parikh, Tarek Ajam, Attila Nakeeb, Nicholas J. Zyromski, C. Max Schmidt, Eugene P. Ceppa, Henry A. Pitt, Michael G. House Surgery, Indiana University Hospital, Indianapolis, IN BACKGROUND: Postoperative morbidity is associated with poor long-term outcomes for gastrointestinal cancers. The purpose of this study is to determine the extent to which postoperative complications after pancreatoduode- nectomy for pancreatic adenocarcinoma impact adjuvant therapy and overall survival. METHODS: Over a fi ve year period ending December 2009, 310 consecutive pancreatoduodenectomies for adenocarci- noma were performed at a single institution. Retrospective review of a prospective patient database including data on patient, operative, and tumor factors along with postopera- tive Clavien-Dindo (CD) classifi ed complications, receipt of adjuvant therapy, and survival was performed. Associations with overall survival (OS), estimated by the Kaplan-Meier method, were analyzed with log-rank testing. RESULTS: Patients were predominantly male (57%) and Caucasian (96%) with a median age at operation of 66 years (36–96 years). Median operative time was 325 min- utes and median blood loss was 675 mL. Sixty-three percent of patients had lymph node metastasis on fi nal pathology. Thirty-day mortality rate was 3.2%. Major postoperative morbidity, defi ned as CD III-IV complications, were recorded for 40 (13%) patients, while minor complications (CD I-II) occurred in 155 (50%) of patients. Overall, 64% of patients received adjuvant chemotherapy or chemoradiotherapy, of which 5% received neoadjuvant chemoradiotherapy. Adju- vant therapy was administered to 76% of patients with no CONCLUSIONS: Major complications after pancreatoduo- postoperative morbidity versus 63% and 42% of patients denectomy alter plans for adjuvant therapy for pancreatic with minor or major complications, respectively (p = adenocarcinoma. Unlike other gastrointestinal cancers, 0.02). Median OS for the cohort was 18.3 months (0–116 postoperative morbidity is not associated with poor long- months). Improved median overall and 5-year survival (20 term outcomes for pancreatic adenocarcinoma. months and 21%, respectively) for patients with no post- operative morbidity was not signifi cantly different to the survival observed for patients with complications (18.8 months and 18% respectively), p = 0.36 (Figure 1). Median OS and 5-year survival for patients who received adjuvant therapy was 21 months and 18%, respectively, versus 15.5 months and 17% for patient who did not receive adjuvant therapy, p = 0.27 (Figure 2).
98 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1659 Su1660 Impact of Laparoscopic Approach on Postoperative Pain Pre-Operative Lovenox Does Not Increase Blood Loss and Opioid Consumption After Pancreatoduodenectomy During Pancreaticoduodenectomy Compared to Naru Kondo, Michael B. Farnell, Florencia G. Que, Heparin David M. Nagorney, Kaye M. Reid Lombardo, John H. Donohue, Shoichiro A. Tanaka, William C. Conway, Satvik Jhamb, Michael L. Kendrick John S. Bolton Mayo Clinic, Rochester, MN Surgical Oncology, Ochsner Medical Center, New Orleans, LA BACKGROUND: Although laparoscopic approaches are BACKGROUND: Pancreaticoduodenectomy (PD) is a generally considered to result in reduced postoperative lengthy surgical procedure often done for malignancy, pain compared to open approaches, objective evaluation of both risk factors for DVT/PE, which also carries a signifi cant amount of opioid consumption has not been evaluated for bleeding risk. We sought to determine if bleeding compli- pancreaticoduodenectomy. cations were increased with pre-operatively administered AIM: The aim of this study was to investigate if total laparo- Lovenox, a drug given once daily and shown to have excel- scopic pancreaticoduodenectomy (TLPD) results in reduced lent DVT/PE risk reduction, compared with pre-operative opioid consumption over open pancreaticoduodenectomy Heparin. (OPD). METHODS: IRB approved retrospective chart review was METHODS: A single-institutional retrospective cohort undertaken to collect data on all patients undergoing PD study of all patients having undergone pancreaticoduode- from 1/1/08–12/31/11. Data points included demographic nectomy between 2007 and 2010 was performed. Postoper- information, surgical details, and peri-operative outcomes, ative pain was evaluated by calculating opioid consumption focusing on estimated blood loss (EBL) and blood usage. from postoperative day (POD) 1 to POD 5. Five forms of DVT/PE prophylaxis consisted of either 5,000units subcu- narcotic analgesics were used including: morphine, hydro- taneous Heparin or 40 mg subcutaneous Lovenox given morphine, oxicodone, hydrocodone and fentanyl. To allow pre-operatively. comparison, narcotic consumption was converted to mor- RESULTS: The 158 patients undergoing PD had a mean phine equivalents using a standard conversion. Daily and age of 65 (range 40–85), 52% were men, and 11% received total opioid consumption after PD was compared between neoadjuvant treatment. 21.5% of the cases included major the LPD and (OPD) groups. Risk factors for increased total vascular resection, and average BMI was 27.5 (15.2–47). 52% opioid consumption were evaluated using univariate and (82) of the patients had pancreatic adenocarcinoma, 5% (8) multivariate analyses. duodenal adenocarcinoma, and 7% (11) IPMN. There was RESULTS: Five hundred and twelve consecutive patients no difference in mean EBL between the 92 patients receiv- Poster Abstracts (LPD n = 123, OPD n = 390) were included in this study. ing pre-op Heparin (731 + 525 ml) and the 35 patients Six patients (5%) with conversion to OPD were included in receiving pre-op Lovenox (794 + 634 ml, p = 0.58). Bleeding Sunday LPD group based on intent-to-treat. Daily opioid consump- complications and overall blood usage were also not signifi - tion of the LPD group was signifi cantly less than that of cantly different between the 2 groups (p > 0.05). OPD group from POD 2 through POD 5, and total opioid CONCLUSIONS: Pre-operatively administered Lovenox consumption of LPD group was also signifi cantly less (LPD: was not associated with an increase in EBL during PD 5.3 ± 6.4 mg/kg, OPD: 7.3 ± 9.4 mg/kg, P = 0.007). Multivar- compared with Heparin. Overall blood usage was also not iate analysis revealed that younger age (<65 years old) (HR increased. This, along with once daily dosing and improved 1.89, 95% CI 1.29–2.79, P = 0.001), no preoperative diabe- effi cacy make Lovenox and attractive option for DVT/PE tes mellitus (HR 1.74, 95% CI 1.10–2.80, P = 0.01), PD for prophylaxis during PD. chronic pancreatitis (HR 2.87, 95% CI 1.18–7.51, P = 0.02), OPD (HR 2.01, 95% CI 1.26–3.27, P = 0.003) and postop- erative major complication (Grade III-V) (HR 2.30, 95% CI 1.36–3.91, P = 0.001) were independently associated with increased opioid consumption after PD (total opioid con- sumption >6 mg/kg). CONCLUSION: Patients undergoing TLPD have lless opioid consumption compared to those with the open approach. Younger age, absence of diabetes, chronic pancreatitis indi- cation and major postoperative complications are indepen- dent predictors of increased opioid consumption. These fi ndings warrant further evaluation as to the potential clini- cal impact of reduced pain and less opioid consumption on patient-specifi c advantages including early recovery and better quality of life after pancreaticoduodenectomy.
99 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Clinical: Small Bowel CONCLUSION: In CD patients undergoing ileocecal resec- tion, male gender, open surgery, a long course of disease, and low postoperative hemoglobin levels are risk factors Su1661 to develop postoperative complications. Other potential risk factors, such as age, body mass index, low preoperative Risk Factors for Postoperative Complications After protein or albumin levels, the use of steroids or biologicals, Ileocecal Resection in Patients with Crohn’s Disease and NOD2 mutations had no effect on the development of Michael S. Kasparek1, Sophie Zehl1, Mario Mueller1, postoperative complications in the patient group analyzed. Stephan Brand3, Martin E. Kreis2 1Department of Surgery, Ludwig-Maximilians-University Munich, Su1662 Munich, Germany; 2Department of Surgery, Charité, Berlin, Germany; The First Validated Nomogram to Predict 30-Day 3Department of Internal Medicine II, Ludwig-Maximilinas-University Munich, Munich, Germany Mortality Following Surgery for Small Bowel Obstruction BACKGROUND: After colorectal resections, patients with Wissam J. Halabi1, Mehraneh D. Jafari1, Vinh Q. Nguyen2, infl ammatory bowel diseases may develop postoperative 1 1 1 complications. Our aim was to identify specifi c risk factors Joseph C. Carmichael , Steven Mills , Alessio Pigazzi , 1 for postoperative complications in patients who underwent Michael J. Stamos ileocecal resection due to Crohn’s disease (CD). 1Surgery, University of California-Irvine, Orange, CA; 2Statistics, University of California-Irvine, Irvine, CA METHODS: We identifi ed CD patients who underwent ileocecal resection in the time period from 2001 to 2010 OBJECTIVE: Surgery for small bowel obstruction (SBO) is in our hospital’s IBD patient register. Patients’ charts associated with signifi cant mortality and surgeons are being were reviewed for details regarding the type of surgical increasingly faced with complicated cases that have several procedure performed, complications related to the sur- risk factors for fatal outcomes. To date, there have been no gical procedure and the overall peri- and postoperative studies examining the interaction of several risk factors course of the disease. In addition, all patients were geno- and their additive effect on mortality. Our aim was to con- typed for the three main CD-associated NOD2 variants struct a comprehensive and validated model that takes into p.Arg702Trp (rs2066844), p.Gly908Arg (rs2066847), and account all the factors that predict mortality in patients p.Leu1007fsX1008 (rs2066847). undergoing surgery for SBO. RESULTS: 155 CD patients (54% female, age 35 [15–69] DESIGN: Using the ACS-NSQIP database from 2005 to years (median [range]) were identifi ed who underwent 2010, we conducted a retrospective review of SBO cases ileocecal resection. In 63 patients, laparoscopic ileocecal caused by adhesions or incarcerated hernias that under- resection was performed, while 92 patients underwent an went operative management. With 30-day mortality as the open ileocecal resection. 32 minor complications (wound primary endpoint, a predictive model was built using 52 infection: n = 17 (11%); prolonged postoperative ileus: n = presurgical, 8 surgical and 16 postsurgical variables. We 10 (6%); urinary tract infection: n = 5 (3%)) and 34 major split the data into two sets: training set (75%) and a valida- complications (intraabdominal abscess formation: n = 15 tion set (25%). The LASSO algorithm for logistic regression (10%); anastomotic leak: n = 12 (8%); hemorrhage: n = 5 was applied to the training set with 10-fold cross-validation (3%), enterocutaneous fi stula: n = 2 (1%)) occurred in 41 and the 1-SE rule used to select predictive variables. The patients (26%), while 114 patients (74%) had an unevent- ROC curve and the AUC statistic were used to test our mod- ful postoperative course. Open surgery (open surgery in el’s predictive ability. complicated group: 80% vs. open surgery in uncomplicated RESULTS: A total of 17,379 cases of surgical SBO cases were group: 52%; p = 0.002) and male gender (male gender in identifi ed. The cause of SBO was attributed to Adhesions complicated group: 63% vs. male gender in uncomplicated in 74% of cases whereas incarcerated hernia accounted for group: 40%; p = 0.018) were associated with postopera- the remaining 26%. The 30-day mortality was 5.7%. LASSO tive complications. Patients with major complications had identifi ed several predictors of mortality listed in the table. a longer course of CD (8 [0-–31] vs. no complications: 4 The following factors were not found to predict mortality: [0–30] years; p = 0.024), a lower postoperative hemoglo- gender, obesity, smoking, diabetes, emergency surgery, sur- bin level (10.3 ± 0.4 vs. 11.5 ± 0.2 g/dl; mean ± SEM; p gery day, disease type, and the use of laparoscopy. Thirty- = 0.038), and a trend towards a lower postoperative total day mortality can be predicted via the following equation: protein level (3.8 ± 0.3 vs. 5.2 ± 0.4; p = 0.051). Other fac- ex/(1 + ex) where x in the sum of coeffi cients. The predic- tors such as preoperative laboratory results, age, body mass tive model performed well with a high predictive power index, nicotine consumption, use of steroids or biologicals, and an AUC = 0.92. as well as presence of CD-associated NOD2 mutations had no effect on the incidence of postoperative complications (all p = n.s.).
100 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Predictors of Mortality Following Surgery for SBO: Coeffi cients and Su1663 Odds Ratios Serum C-Reactive Protein As Predictor of Recurrence in Coeffi cient Odds Ratios Patients Undergoing Ileo-Colonic Resection for Crohn’s Intercept –6.30 Disease: Results of a Longitudinal Prospective Study Preoperative Factors: ASA V 1.66 5.27 Edoardo Iaculli, Cristina Fiorani, Sara Onali, Giorgia Tema, ASA IV 0.82 2.27 Roberto Pezzuto, Livia Biancone, Rosa Scaramuzzo, Disseminated cancer 0.80 2.24 Khrystyna Porokhnavets, Achille Gaspari, Giuseppe S. Sica Ventilator dependence 0.78 2.18 Tor Vergata, Rome, Italy Septic shock 0.75 2.12 Dialysis dependence 0.40 1.50 BACKGROUND: Previous studies have evaluated the abil- Sepsis 0.38 1.46 ity of biological markers to detect disease relapse in Crohn’s Peripheral vascular disease 0.28 1.32 disease (CD). Yet no studies have targeted a method to BUN > 40 0.24 1.27 anticipate recurrence after surgical resection. C-Reactive Ascites 0.21 1.23 Protein (CRP) is a valuable marker for predicting the out- come of several diseases including CD. The exact role of COPD 0.18 1.19 CRP as a prognostic factor for future recurrence in CD is not Weight loss > 10% 0.08 1.09 yet determined. Moreover no data are available investigat- Age (absolute number 0.04 1.04 ing specifi c CRP modifi cations in these patients following multiplied by coeffi cient) surgery. Pneumonia 0.03 1.03 Creatinine > 1.2 0.02 1.02 OBJECTIVE: of present study was to determine the peri- Hematocrit > 38 0.02 1.02 operative behaviour of the CRP in CD patients undergoing Operative Factors: Bowel 0.51 1.66 elective ileo-cecal resection. Our hypothesis is that peri- Gangrene operative CRP changes are disease-specifi c and therefore Bowel Resection 0.15 1.16 could detect subset of patient with more aggressive disease. Contaminated Case 0.07 1.07 Secondary objective was to investigate the role of CRP as a Postoperative Factors: Shock 1.64 5.18 potential early prognostic marker for future recurrence. CVA 1.05 2.87 METHODS: 52 patients undergoing IC resection for CD Acute Renal Failure 0.90 2.47 were prospectively enrolled. Serial CRP levels were assessed Re-Intubation 0.71 2.03 perioperatively: time 0, postoperative day (POD) 1 and POD Myocardial Infarction 0.64 1.89 6. CD patients’ perioperative CRP fi ndings were compared Poster Abstracts Bleeding 0.17 1.18 against same interval assessments of two control groups Return to OR 0.11 1.12 undergoing right colectomy and appendicectomy. Crohn’s Sunday Failure to wean 0.05 1.06 Disease Activity Index (CDAI) and Rutgeerts’ score (RS) were evaluated for recurrence during 3 year follow-up protocol. Coeffi cients can be added together RESULTS: As expected, in all 3 groups CRP signifi cantly CONCLUSION: This is to date the most powerful and the increased 24 hours after surgery vs baseline but the increase only validated nomogram to predict 30-day mortality fol- was signifi cantly higher in CD patients than in controls lowing surgery for SBO. This model represents an easy-to- (p < 0.001). Comparing to control groups CRP remained use tool for surgeons to risk-stratify and counsel patients remarkably high in CD (mean 32.2 mg/L) at POD 6. Dif- and can be used as a quality outcome measure. Implement- ference between groups was statistical signifi cant (p 0.03). ing strategies to modify certain risk factors may lower mor- All CD patients evaluated at 3 year follow up were in clini- tality in surgical SBO cases. cal remission. Endoscopic recurrence (RS > 2) was found in 51% at 1 year and in 42% at 3 years. Possible relation between endoscopic recurrence rate or severity and peri- operative CRP levels was investigated: multivariate ordinal regression showed that postoperative increment of CRP is a prognostic factor of recurrence at 3 years. CONCLUSION: Present preliminary data show disease- specifi c perioperative CRP levels for CD patients that refl ect immunomodulation impairment involved in disease eti- ology. The degree of such immunitary change and con- sequent severity of disease might be explored early after surgery by determining CRP alterations. Data from larger series can confi rm that perioperative CRP levels might be considered a novel prognostic factors of surgical recurrence.
101 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Su1664 Su1665 Predictors of Recurrence and Post Recurrence Survival Adenomas of the Ampulla of Vater: A Comparison of in Patients with Resected Ampullary Adenocarcinoma Outcomes of Operative and Endoscopic Resections Irene Epelboym1, Susan Hsiao2, James A. Lee1, Beth Schrope1, Edwin O. Onkendi1, Jordan Rosedahl2, William S. Harmsen2, John A. Chabot1, Helen Remotti2, John D. Allendorf1 Florencia G. Que1 1Surgery, Columbia University Medical Center, New York, NY; 1Surgery, Mayo Clinic, Rochester, MN, Rochester, MN; 2Biomedical 2Pathology, Columbia University Medical Center, New York, NY Statistics and Informatics, Mayo Clinic, Rochester, MN, Rochester, MN BACKGROUND: Ampullary neoplasms are a rare subset of BACKGROUND: Data comparing operative and endo- intestinal cancers, the only treatment for which is complete scopic resection of adenomas of the ampulla of Vater is surgical resection. Controversy exists, however, with regard limited. We reviewed our experience in the treatment of to need for and type of adjuvant therapy. The management adenomas of the ampulla of Vater and compared the opera- approach is even less clear for those patients in whom the dis- tive and endoscopic approaches. ease recurs. In this report, we aim to determine patient and METHODS: Retrospective review of all patients in the gas- histological factors predictive of recurrence, and to describe trointestinal endoscopy and surgical databases treated for the survival experience of those with recurrent disease. adenomas of ampulla of Vater at our institution from 1992 METHODS: Patients who underwent surgical resection for to 2009. Clinicopathologic factors, morbidity, mortality, ampullary adenocarcinoma at our institution were identi- recurrence and survival of patients treated by endoscopic fi ed, and histological diagnosis was confi rmed by inde- and surgical resection were comparatively analyzed. pendent pathologist review. Presenting features, operative RESULTS: A total of 137 patients (mean age 59.3 yrs), characteristics, postoperative outcomes, and overall and were treated for adenomas of the ampulla of Vater; 75 disease free survival were evaluated. Selected resection spec- (55%) males, follow up 91% (mean 4.6 years). The adeno- imens were stained for presence of CK7, CK20, and CDX2 mas were tubular in 55 (40%) patients, tubulovillous in 62 using standard methods. (45%) and villous in 20 (15%). Obstructive jaundice was Continuous variables were compared using Student’s t-test. more common in the operative resection group (p < 0.01)). Categorical variables were compared using chi-square or Endoscopic resection was performed in 100 (73%) patients; Fisher’s exact test. Predictors of recurrence were analyzed operative resection was performed in 37 (27%). Sixty seven using logistic regression. Survival was evaluated using percent of patients required only 1 endoscopic resection Kaplan-Meier method, and differences among groups were [piecemeal resection in 24 (36%)], while 33 (24%) required assessed by log-rank test. 2 or more resections (range 2–5). Patients who underwent RESULTS: Between 1990 and 2011, 79 patients underwent operative resection often had larger tumors >3.6 cm (p < pancreaticoduodenectomy for ampullary adenocarcinoma. 0.001) or intraductal extension (p = 0.04). Intraductal Thirty patients received adjuvant chemotherapy, which was extension and ulceration had no effect on recurrence (p gemcitabine based for 29 (96.6%). Among 74 R0 resections, values = 0.62, 1.0) in both groups. Postoperative compli- there were 24 cases of recurrence over 273 person-years cations occurred in 48% of patients; post-endoscopic com- (median follow-up 28.7 months, median time to recurrence plications in 30% of patients (p = .09). Post endoscopic 8.7 months). Four (16.7%) were in the surgical bed and 20 resection complications included bleeding in 18 (7 required (83.3%) distantly, predominantly in the liver. In univariable transfusion or endoscopic or angiographic intervention); analysis, no single demographic or clinical characteristic, pancreatitis (mild in 11; severe necrotizing in 1); ampul- nor histologic staining pattern, was a statistically signifi - lary obstruction from edema or blood clot in 2 and duode- cant predictor of recurrence. Lymph node positivity was nal perforation in 1. Postoperative complications included signifi cant in univariable but not in multivariable analysis, pancreatic leak (9), surgical site infection (4), anastomotic and pathologic T stage was unassociated with recurrence. leak (3), delayed gastric emptying (2), myocardial ischemia Recurrent disease was managed by surgical resection in 2 or dysrhythmia (2), and renal failure (1). One patient died cases, one local and one metastatic, after which the patients of pancreatic leak with MOSF following operative resection survived 15.8 and 3.4 months, respectively. Fifteen patients of a 6 cm sessile adenoma (mortality of 2%). Endoscopic received chemotherapy (either 5FU or gemcitabine based) resection was associated with a 3-fold higher risk of recur- only. Systemic therapy was not offered to 2 patients. Post- rences than operative resection, 5% of which were invasive recurrence survival was not signifi cantly different among cancers in both groups. Performing 2 or more endoscopic those who had surgery, chemotherapy, or no treatment (8.8 resections for complete tumor removal relative to 1 com- vs 8.0 vs 3.9 months, p = 0.39). Additionally, among those plete initial resection was associated with 5 times higher who received chemotherapy, difference in median post- risk of recurrence (p < 0.001). recurrence survival was not statistically signifi cant in 5FU CONCLUSION: Endoscopic resection of adenomas of compared with gemcitabine groups (16 vs 3.5 months, p = ampulla of Vater is associated with a 3-fold higher recur- 0.107). rence rate than operative resection; recurrences may be CONCLUSIONS: Optimal treatment approach for recur- invasive. There is a 5-fold higher risk of recurrence if 2 or rent ampullary adenocarcinoma remains unclear. Survival more endoscopic resections are needed for complete tumor is equivalent whether surgical resection or systemic chemo- removal as compared to one complete initial resection. therapy is employed, and no single cytotoxic protocol is Operative resection is associated with lower recurrence rates associated with improved outcome. for larger tumors and tumors with intraductal extension.
102 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1666 Clinical Outcomes for Neuroendocrine Tumors of the Duodenum and Ampulla of Vater: A Population-Based Study Reese W. Randle, Shuja Ahmed, Naeem A. Newman, Clancy J. Clark General Surgery, Wake Forest Baptist Health, Winston Salem, NC BACKGROUND: Neuroendocrine tumors (NETs) of the duodenum are quite rare representing only 4% of all car- cinoid tumors. Limited single-institution case series indi- cated that ampullary NETs have worse survival than NETs located in the duodenum. The aim of the current study was to evaluate the overall survival (OS) of patients with ampul- lary NETs compared to patients with duodenal NETs using a population based registry. METHODS: We conducted a retrospective comparative cohort study using the Surveillance, Epidemiology, and End Results (SEER) registry from 1988 to 2009. Patients with Su1667 pathology confi rmed NETs of the duodenum and ampulla of Vater were identifi ed, and overall survival was evaluated Laparoscopic vs. Open Bilateral Inguinal Hernia Repair: using Kaplan-Meier estimates and Cox proportional hazard A NSQIP Analysis regression. Multi-variable survival analyses included covari- Muhammad Asad Khan, Roman Grinberg, John Afthinos, ates with p < 0.1 and less than 10% of data missing. Karen E. Gibbs RESULTS: The study cohort included 1360 (92%) patients Staten Island University Hospital, Staten Island, NY with duodenal NETs and 120 (8%) with ampullary NETs. Ampullary NETs were larger (median tumor size 18 vs. 10 OBJECTIVES: Laparoscopic inguinal herniorrhaphy was mm, p < 0.001), higher grade (poorly and undifferentiated introduced into surgical practice in 1990. It has shown a tumor 42% vs. 12%, p < 0.001), and higher SEER historic great deal of promise and was shown to allow quicker and stage (distant metastasis 18% vs. 9%, p < 0.001) than duo- more thorough assessment and repair of bilateral groin denal NETs. Ampullary NETs were also more likely to be hernias. However the evolution of tension-free open repair Poster Abstracts resected (78% vs. 60%, p < 0.001). OS was signifi cantly with mesh allows use of local or regional anesthesia and is worse for patients with ampullary NETs than for patients associated with rapid recovery and a low recurrence rate. We Sunday with duodenal NETs (median OS 98 vs. 143 months; HR sought to compare a large number of patients and compare 1.38, 95% CI 1.02–1.86, p = 0.037). For resected patients national trends and outcomes between these approaches. (n = 878), OS was similar between ampullary and duode- METHODS: The NSQIP database was queried for laparo- nal NETs (median OS 182 vs. 164 months; HR 1.42, 95% scopic or open bilateral inguinal hernia repair. Age, gender CI 0.96–2.09, p = 0.078). Using univariate survival analy- and comorbidities were quantifi ed and outcomes data col- ses, signifi cant predictors for worse OS in resected patients lected. Specifi cally, morbidity, mortality, length of stay and included older age (p < 0.001), larger tumor size (p = 0.035), operative times were examined. Statistical analysis was then higher grade (p < 0.001), higher SEER historic stage (p < performed. A p-value of <0.05 was considered signifi cant. 0.001), and radiation treatment (p = 0.003). After adjust- RESULTS: A total of 4985 patients were identifi ed, of ing for signifi cant predictors of OS, ampullary NETs had sig- which 2025 patients underwent open repair of bilateral nifi cantly worse OS than duodenal NETs (HR 1.63, 95% CI inguinal hernia and 2960 patients underwent laparoscopic 1.05–2.53, p = 0.031). repair. CONCLUSIONS: NETs of the ampulla of Vater are more CONCLUSIONS: Nationally, 59.4% of bilateral hernias advanced at presentation and have worse OS than duode- were repaired laparoscopically. More patients with diabetes, nal NETs. After controlling for signifi cant predictors of OS, HTN and history of CABG underwent open repair. Laparo- tumor location remained an independent predictor of OS scopic and open approaches have a similar complication in resected patients. profi le. Operative time was shorter in the laparoscopic group (75.9 ± 35 vs. 85.2 ± 38.3 min, p < 0.001) as was length of stay (0.18 ± 1.1 vs. 0.409 ± 3.0 days, p < 0.001).
103 THE SOCIETY FOR SURGERY OF THE ALIMENTARY TRACT
Table 1: Patient Comorbidites neoadjuvant therapeutic concepts (QCGC’07/09) and com- pared to former data obtained in a study with similar design Open Repair Laparoscopic (EGGCS’02) but no neoadjuvant treatment arm. N = 2025 Repair N = 2960 P-Value Male gender 1877 (92.7%) 2825 (95.4%) <.001 RESULTS: From 01/01/2007–12/31/2009, 2,897 patients Diabetes on oral 96 (4.7%) 101 (3.4%) 0.013 from 141 hospitals were enrolled in the study with the Diabetes on Insulin 23 (1.1%) 20 (0.7%) 0.013 following rates (QCGC’07/09 [EGGCS’02: n = 1,139 patients]): Resection (91.2 [87.1]%), gastrectomy (74.5 HTN 742 (36.6) 834 (28.2%) <0.001 [79.8]%), R0-resection (82.8 [82.3]%) explorative lapa- CHF in 30 days 2 (0.1%) 2 (0.1%) 1 rotomy (4.9 [6.3]%), UICC-III/IV (45.2 [41.8]%), hospital History of MI in 6 months 1 (0%) 2 (0.1%) 1 lethality (6.0 [8.3]%) & esophagojejunal anastomotic insuf- Prior PCI 117 (5.8%) 109 (3.7%) .001 fi ciency (6.0 [5.8]%). Prior CABG 123 (6.1%) 105 (3.5%) <.001 PAD 13 (0.6%) 12 (0.4%) 0.308 • After inauguration of multimodal procedures (n = 498; 18%) ESRD 10 (0.5%) 9 (0.3%) 0.350 Smoker 409 (20.2%) 566 (19.1%) 0.363 — the proportion of patients with no or only pallia- Bleeding disorder 46 (2.3%) 34 (1.1%) .003 tive surgical intervention decreased considerably— ASA III or above 25 (1.2%) 14 (0.5%) .005 palliative rate (no R0-resection, i.e., palliative or no operation: In 2002 [40%] vs. 2007–2009 [24.5%]); Table 2: Patient Outcome and Complications — hospital lethality (overall, 6%/with neoadjuvant chemotherapy, 3.4%) and peri- & postoperative Laparoscopic morbidity did not increase; Open Repair Repair N = 2025 N = 2960 P-Value — 4.4% (n = 23/521) of patients with neoadjuvant Superfi cial SSI 8 (0.4%) 8 (0.4%) 1 treatment could not be resected. Deep incisional SSI 1 (0%) 0 0.406 • There were no changes of the distribution of tumor sites Pneumonia 4 (0.2%) 3 (0.1%) 0.452 and stages (according to UICC classifi cation), in particu- Unplanned Intubation 1 (0%) 2 (0.1%) 1 lar, no reduction of advanced tumor stages. PE 1 (0%) 2 (0.1%) 1 • Gastrectomy rate decreased from 79.8 to 74.5%—the Return to OR 19 (0.9%) 24 (0.8%) 0.643 supposed reduction of radicality at the primary tumor ARF 0 (0%) 2 (0.1%) 0.517 lesion was associated with an extension of lymphade- UTI 6 (0.3%) 9 (0.3%) 1 nectomy compared with 2002–D1: 11.9% (n = 245/2,052 MI 0 1 1 resected patients with curative intention); D2: 79.5%; DVT 3 (0.1%) 3 (0.1%) 0.135 D3/4: 5.7/2.8%. Operative time (min) 85.2 ± 38.3 75.9 ± 35 <0.001 • Two trends continued: Predominating esophagojejunal Length of stay (days) 0.409 ± 3.0 0.18 ± 1.1 <0.001 stapler anastomosis, 96% (pouch: Approx. 20%). Mortality 2 (0.1%) 0 (0%) 0.165 • A hospital-volume effect could not be observed. • A postoperative adjuvant chemotherapy (only patients Clinical: Stomach without neoadjuvant treatment) received 15.8% of the patients (n = 327). • The 5-year survival rate of the whole patient group Su1668 (including curative & palliative intention) increased Outcome of the Surgical Treatment of Gastric Cancer from 40.0% up to 48.5% but, in particular, in stage II–in stage IV: Increase from 3.5 to 11.3%). After Inauguration of the Neoadjuvant Concept Using a Systematic Multicenter Prospective Clinical CONCLUSION: After inauguration of multimodal con- cepts, there was an effective improval of the overall survival Observational Study without increased perioperative risk in the daily clinical care Frank Meyer1, Karsten Ridwelski2, Lutz Meyer3, Henry Ptok4, of gastric cancer; however, patients undergoing endoscopic Ingo Gastinger4, Hans Lippert1 tumor ablation & exclusively palliative chemotherapy were 1Department of Surgery, University Hospital, Magdeburg, Germany; not included. 2Department of Surgery, Municipal Hospital, Magdeburg, Germany; 3Department of Surgery, Municipal Hospital, Plauen, Germany; 4Department of Surgery, Municipal Hospital, Cottbus, Germany AIM & Methods: By means of a systematic multicenter prospective observational study, quality of surgical care for a representative group of patients with gastric cancer in daily clinical practice was investigated after inauguration of
104 54TH ANNUAL MEETING • MAY 17–21, 2013 • ORLANDO, FL
Su1669 10, 15, 20, 30, 40, 50, 60, 75, 90, 105, 120, 135, 150, 165, 180 minutes after meal. 13CO2 content was measured by The Tolerance to Volume Load, But Not the Reservoir or infrared spectrophotometry. Wagner-Nelson analysis was Emptying Capacity, Refl ects the Patients’ Living States performed on measured data. The reservoir and emptying After Gastrectomy capacity of the remnant stomach were calculated as reten- Koji Nakada, Masahiko Kawamura, Hideo Konishi, Taizo Iwasaki, tion rate at 5 minutes after meal (RR5) and half emptying Keishiro Murakami, Fumiaki Yano, Kazuto Tsuboi, time (T1/2), respectively. For drink test, mineral water (10 ml/kg) was ingested for 5 min at equal rate. The severity Yoshio Ishibashi, Norio Mitsumori, Nobuyoshi N. Hanyu, and the duration of abdominal symptoms caused by DT Hideyuki Kashiwagi, Noburo Omura, Katsuhiko Yanaga was scored 0 (none) to 3 (severe) and 0 (none) to 4 (more Surgery, The Jikei University School of Medicine, Tokyo, Japan than 30 minutes), respectively. The tolerance to volume Upper gastrointestinal tract has various physiological roles, load was assessed by DT total score (the sum of both sever- such as, the capacity to store or empty the ingested food by ity and duration scores). The questionnaire was performed the stomach, and the tolerance to volume load to receive a to examine the ingested amount of food per meal, the fre- certain amount of food by the upper gastrointestinal tract quency of daily meals, the change in body weight and the as a whole. The gastric surgery may alter these physiologi- restriction to daily life. cal properties, and then, could impair the patient’s living The multivariable analysis was performed to explore the states. However, the information about this concern is effect of physiological properties on patients’ living states limited. after gastrectomy. AIM: To study the effect of physiological properties of the RESULTS: (Table 1) The reservoir and emptying capac- upper gastrointestinal tract on patients’ living states after ity of the remnant stomach had no effect on patients’ gastrectomy. living states. The impairment in the tolerance to volume METHODS: 13C-acetate breath test (13-BT), drink test load (higher scores) resulted in reduced meal amount and (DT) and the questionnaire had performed in the patients restricted daily life. The frequency of daily meal was higher [n = 53] who received conventional gastrectomy (total in the patients with total gastrectomy. with Roux-en-Y [TG; n = 17], distal with Billroth-I [n = CONCLUSION: The tolerance to volume load by DT, but 17], distal with Roux-en-Y [n = 19]). For 13-BT, liquid meal not the reservoir or emptying capacity by 13C-BT, refl ected (200kcal/200 ml) mixed with 100 mg of 13C-acetate sodium the patients’ living states after gastrectomy. salt was given. Breath samples were collected before and 5,
Table 1: The Effect of Physiological Properties on Patients’ Living States After Gastrectomy Poster Abstracts Ingested Amount of Food Frequency of Daily Meals Change in Body Weight Restriction to Daily Life