sub-hepatic fluid collections, and distant infections) and length of hospital stay. Separate analyses were performed for each outcome by using odds ratio (OR) or weighted mean difference (WMD). Both random and fixed effect models were used. Publication bias was accessed by funnel plot. All studies were graded by Jadad scores. Heterogeneity among studies was assessed by calculating I2 measure of inconsistency. Results: Eight RCTs (N= 1,361) met the inclusion criteria. The antibiotics used in the trials were: cefazolin (3 trials), cefotaxim (2 trials), cefotetan (1 trial), cefuroxime (1 trial), & cefotetan/cefazolin (1 trial). Antibiotics were administered preoperatively in all studies. 3 trials used multiple doses; first dose pre-operatively and other doses post-operatively. There was no significant heterogeneity among the studies (I2=0%). Quality score ranged from 2-5. Prophylactic antibiotics did not decrease the odds of post-operative infective complications (including superficial, deep and distant infections) (OR 0.64, 95% CI: 0.31-1.30, p=0.22), superficial wound infections (OR 0.74, 95% CI 0.30-1.82, p=0.52), sub-hepatic fluid collections (OR 1.03, 95% CI: 0.21- 5.13, p=0.98), or distant infections (OR 0.50, 95% CI: 0.13-1.97, p=0.32). Prophylactic antibiotics also did not alter the length of hospital stay (WMD 0.02, 95%CI -0.10-0.14, p= 0.77). Funnel plot did not reveal the presence of publication bias. Pooling of data from high quality studies (Jadad score > 3) also did not reveal a reduction in the odds for total infection (OR 0.75, 95% CI 0.35-1.63, p= 0.72). Conclusions: Prophylactic antibiotics do not prevent infections in low risk patients undergoing laparoscopic cholecystectomy. 250 Comparison of Surgically Resected Polypoid Lesions of the Gallbladder to Their Pre-Operative Ultrasound Characteristics Martin D. Zielinski, Peyton W. Davis, Florencia G. Que, Michael L. Kendrick, Thomas D. Atwell Background: Polypoid lesions of the gallbladder have been a common finding on ultrasound examinations of the abdomen and are more prevalent since our use of equipment incorporat- ing pulse shaping, increased bandwidth and better phase use for image reconstruction began in 1996. Our study correlates the pre-operative ultrasonographic findings of these lesions to the surgically resected specimen with specific regard to identifying neoplastic polyps. Methods: A retrospective review was performed of 137 patients who had a pre-operative ultrasonographic diagnosis of a polypoid lesion of the gallbladder and subsequently under- went cholecystectomy between August 1996 and July 2007 at the Mayo Clinic Rochester. Results: 114 pseudopolyps (cholesterol polyps, inflammatory polyps and adenomyomas) and 23 true polyps (83.2% and 16.8% respectively) were identified on histopathologic analysis. 30 polyps had suspicious ultrasonographic characteristics for neoplastic changes. 26 were ≥ 10mm, 3 had vascularity and 1 demonstrated invasion. Of these, there were 21 pseudopolyps, 2 benign adenomas and 7 with neoplastic changes on final pathology (2 low grade dysplasia, 2 high grade dysplasia and 3 adenocarcinomas). 2 asymptomatic polyps, sized 6 mm and 7 mm by ultrasound, were identified pre-operatively and not regarded as suspicious but had neoplastic changes at pathology (one low grade and one high grade dysplasia). 25 patients were followed with at least two serial ultrasound examinations. Of these, 6 demonstrated polyp growth of at least 3 mm. None of these specimens demonstrated neoplastic changes. The positive predictive value and negative predictive value for ultrasound diagnosing neoplasia based on current criteria was 23% and 98% respectively with a false negative rate of 7%. Conclusion: Histopathologic analysis of polypoid lesions of the gallbladder continues to be the gold standard to identify neoplasia. Ultrasound has been used extensively in the pre-operative management of these lesions but modern ultrasound techniques are unable to differentiate between pseudopolyps and true polyps with any certainty. We identified 2 polyps with neoplastic changes that were less than 10 mm. Therefore, we recommend decreasing the current threshold for surgical resection to 5 mm while continuing to offer cholecystectomy for lesions that demonstrate vascularity, show invasion or are symptomatic. 251 The Incidental Asymptomatic Pancreatic Lesion: Nuisance or Threat? Teviah E. Sachs, Wande B. Pratt, Mark P. Callery, Charles M. Vollmer Introduction: Although asymptomatic pancreatic lesions (APLs) are being discovered incid- entally with increasing frequency, their true significance remains obscure. Treatment decisions pivot off concerns for malignancy, but at times might be excessive. To better understand the role of surgery, we scrutinized a spectrum of APLs as they presented to our surgical practice over defined periods. Methods: All incidentally identified APLs that were operated upon during the past 5 years were clinically and pathologically annotated. Among features evaluated were method/reason for detection, location, morphology, interventions and pathology. For the past 2 years, since our adoption of the Sendai guidelines for cystic lesions, we scrutinized our approach to all patients presenting with APLs, operated upon or not. Results: Over 5 yrs, APLs were identified during evaluation of: GU/Renal (14%), Asymptomatic rise in LFTs (13%), Screening/Surveillance (7%) and Chest Pain (6%). APLs occurred throughout the pancreas (body/tail-63%; head/UP-37%) with 51% being solid. 110 operations were performed with no operative mortality including 89 resections (Distal- 57; Whipple- 32) and 21 other procedures. During these 5 years, APLs accounted for 27% of all pancreatic resections we performed. In all, 21 different diagnoses emerged including IPMN (19%), serous cystadenoma (13%), and neuroendocrine tumors (13%), while 8% of pts had >1 distinct pathology and 12% had no actual pancreatic lesion at all. Invasive malignancy was present 17% of the time and these pts were older (67.3 to 60.9 yrs; p<.05). Adenocarcinoma predominated (16%). An asymptomatic rise in LFTs correlated significantly (p<.01) with malignancy. Seven pts ultimately opted for operation over continued observation (mean 2.7 yrs), but none had cancer. In the last 2 years, we have evaluated 132 new patients with APLs, representing 47% of total referrals for pancreatic conditions. Nearly half were operated upon, with a 3:2 ratio of solid to cystic lesions. This differs significantly (p<.05) from the previous 3 years (2:3 ratio), reflecting tolerance for cysts <3 cm and side-branch IPMN. Surgery was undertaken more often for solid APLs (74%) than cysts (32%). Some solid APLs were actually unresectable cancers. Due to anxiety, 2 pts requested operation over continued observation, and neither had cancer. Conclusions: APLs occur commonly, are often solid and reflect a spectrum of diagnoses. Sendai guidelines are not transferable A-845 SSAT Abstracts to solid masses, but have safely refined management of cysts. An asymptomatic rise in LFTs cannot be overlooked, nor should a patient or doctor's anxiety, given the prevalence of cancer in APLs. 338 Cervical Nodal Metastasis from Intrathoracic Esophageal Squamous Cell Carcinoma Is Not Necessarily An Incurable Disease Daniel K. Tong, Simon Law, Kam Ho Wong, John Wong Background It remains controversial whether metastatic cervical lymph nodes in patients with intrathoracic esophageal cancer signify distant metastases, and therefore incurable; or they should be regarded as regional spread with still a potential for cure. The aim of this study is to review the treatment results in this group of patients. Patients and Methods Patients with intrathoracic esophageal squamous cell carcinoma from 1995-2007 were included. Only those who had cervical lymph node spread confirmed by fine needle aspiration cytology were studied. Patients with cancers in the cervical esophagus or gastric cardia, or histology other than that of squamous cell were excluded. Results There were 115 patients who satisfied the inclusion criteria, of whom 98 (89.9%) were men. The median age was 62 years (range: 34-88). Treatment methods included: chemoradiation plus salvage surgery [n= 22 (20.2%)]; chemoradiation [n=51 (46.8%)]; primary surgical resection [n=4 (3.7%)]; chemotherapy [n=2 (1.8%)]; radiotherapy [n=7 (6.4%)]; chemoradiation with bypass surgery [n=2 (1.8%)]; prosthetic intubation [n=12 (11%)]; and no intervention [n=9(8.2%)]. The median survival of patients with chemoradiation plus salvage surgery was significantly longer than those with surgery alone [34.8months vs 5.9months; p=0.029] or patients with chemoradiation [34.8 months vs 8.9 months; p=0.0003]. There was no hospital mortality in patients who had chemoradiation plus salvage surgery. Twelve out of the 22 patients who had chemoradiation and resection were downstaged from stage IV to stage 0-II. Conclu- sions Despite staged as stage IV disease; prognosis of patients with metastatic cervical nodes was not uniformly poor. Up to 20% could derive good survival after chemoradiation and surgical resection. 339 The 3-Year Outcome of Optimal Medical or Surgical Management of GERD Patients with Barrett's Esophagus: the Lotus Trial Experience Stephen E. Attwood, Lars R. Lundell, Jan G. Hatlebakk, Stefan Eklund, Ola Junghard, Jean Paul Galmiche, Christian Ell, Roberto Fiocca, Tore Lind Introduction: The long-term management of gastroesophageal reflux disease (GERD) in patients with Barrett's esophagus (BE) is not well supported by an evidence-based consensus. In patients with BE, standard doses of acid-suppression therapy often result in incomplete reflux control, both symptomatically and as measured by pH monitoring. Similarly, the outcomes of anti-reflux surgery reported in the literature have been poor. This analysis aimed to compare the long-term outcome of optimal medical vs surgical anti-reflux treatment in patients with BE (biopsy-proven intestinal metaplasia). A secondary aim was to identify if there were any differences in treatment outcome comparing patients with and without BE. Methods: In the LOTUS trial (a European multicenter randomized study; ClinicalTrials.gov identifier: NCT00251927), standardized laparoscopic anti-reflux surgery (LARS) was com- pared with dose-adjusted medical therapy, starting with esomeprazole 20mg od (ESO), and increasing if needed to 20mg bd. Operative difficulty, complications, pre- and post-treatment symptoms (GSRS), time to treatment failure (ie, need for alternative therapy), pre- and post- treatment measurements of esophageal acid exposure (after 6 months) and endoscopic findings (at 3 years) are reported. Results: Out of 554 patients with chronic GERD, 60 patients had BE of whom 28 were randomized to medical treatment and 32 to LARS. The median acid exposure times before treatment were not significantly different in patients with or without BE. Before surgical treatment, patients with BE had acid with pH<4 in their esophagus for 13.2% of the time, which reduced to 0.4% 6 months after surgery. The corresponding figures for medically treated patients were 7.4% and 4.9%. Operative difficulty was greater in patients with BE (23% vs 13% in patients with no BE). This was based on a longer operating time (30% >2 hrs vs 23% in non-BE), and larger hiatus hernias (37% >5cm vs 15% in non-BE) requiring >3 crural sutures in 40% BE vs 28% non-BE. There was no apparent difference between the groups in post-operative complications. The GSRS scores of patients were similar for the medically and surgically treated groups, regardless of the presence or absence of BE, both at baseline and at 3 years. Three BE patients in the medically treated group and one in the surgically treated group were judged to be treatment failures at 3 years. Conclusion: There is a high degree of success at 3 years with optimal medical or surgical anti-reflux therapy. In a well-controlled surgical environment, the success of LARS in patients with BE is greater than expected and similar to that in patients without BE. 340 Zenker's Diverticula: Is a Tailored Approach Feasible? Christian Rizzetto, Mario Costantini, Raffaele Bottin, Elena Finotti, Lisa Zanatta, Martina Ceolin, Loredana Nicoletti, Giovanni Zaninotto, Ermanno Ancona Background: Zenker's diverticula (ZD) can be treated by transoral diverticulostomy or open surgery (upper esophageal sphincter myotomy and diverticulectomy or diverticulopexy). The aim of this study was to compare the effectiveness of minimally invasive (Group A) versus traditional open surgical approach (Group B) in the treatment of ZD. Methods: Between 1993 and September 2007, 128 patients underwent transoral stapling (n = 51) or cricopharyngeal myotomy and diverticulectomy or diverticulopexy (n = 77). All patients were evaluated for symptoms with a detailed questionnaire. Manometry recorded upper esophageal sphincter (UES) pressure, relaxations and intrabolus pharyngeal pressure. The dimension of the pouch was based on the barium swallow. The choice of the treatment was based on the operative risk, the size of the diverticulum and the patients' preference. Long- term follow-up data were available for 120/128 (94%) patients with a median follow-up of 40 months (IQR: 17-83). Results: Mortality was nil. Three patients (5.8%) in the group A SSAT Abstracts