Search

Background Esophageal cancer in the United States carries a poor prognosis with overall 5 year survival rate of approximately 10%. Due to this dismal outcome, data analyzing factors predictive of survival for greater than 5 years are not available. Methods Single institution retrospective review of esophageal resection for curative intent from 1984 to 2004. We identified 50 actual 5 year survivors (long term survivors, LTS) out of 266 patients (19%) with invasive esophageal cancer and, using multivariate logistic regression, compared characteristics between the LTS, and short-term (

The role of operation in patients with Multiple Endocrine Neoplasia Type 1 (MEN-1) and Zollinger-Ellison Syndrome (ZES) is controversial. Our institutional bias for this disease has, in general, been towards aggressive imaging and operative removal of localized gastrinomas. Few studies have reported long-term outcomes in patients with MEN-1 and ZES. A single institution retrospective review of all patients with MEN-1 and ZES from 1970 to present was performed. Twelve patients were identified (median age = 37 years at diagnosis). The median follow-up was 18 years from diagnosis of ZES. Common symptoms associated with gastrinoma in these patients were diarrhea (n = 6), abdominal pain (n = 4), and nausea/vomiting (n = 4). Most commonly identified sites of gastrinoma were: pancreas (n = 10), duodenum (n = 4), lymph nodes (n = 3), and liver (n = 1). Fifteen celiotomies were performed in total (median = 1; range 0-3). Operative procedures performed included: distal pancreatectomy (n = 4), acid reducing procedure (n = 4), enucleation of pancreatic gastrinoma (n = 3), duodenal resection (n = 3), pancreaticoduodenectomy (n = 1), and other (n = 7). One patient had a transient biochemical cure after operation lasting 3 years. Only one patient in this series had documented liver metastases of gastrinoma and no patients expired of metastatic gastrinoma. There was one postoperative patient death, secondary to respiratory arrest thought to be a result of aspiration or pulmonary embolus. Three patients died of nondisease related causes, and seven patients were alive at the time of last follow-up. Operations rarely result in biochemical cures in patients with MEN-1 and ZES. In our experience, resection of localized gastrinomas often did not require extended surgical resection and were associated with excellent long-term outcomes.

Background Most patients with Zollinger-Ellison Syndrome (ZES), even those in whom gastrinoma is found and resected at initial operation, will suffer from persistent or recurrent disease in longterm followup. There is currently no consensus about managing patients with recurrent or persistent ZES. Our unit has historically maintained an aggressive approach toward monitoring and reoperation for patients with sporadic ZES. Study Design We performed a review of a consecutive series of patients evaluated and managed at our institution between 1970 and 2007 for ZES. "Biochemical cure" was defined as normal serum gastrin assays and negative imaging studies. Reoperations were performed for elevations in serum gastrin assays and positive findings on imaging studies. Results Fifty-two patients with sporadic ZES were analyzed. Median followup was 14 years. Among patients with sporadic ZES, 37 patients underwent operative management. The most common operations were resection of duodenal gastrinoma (n=8) and total gastrectomy (n=7). Nine patients underwent 15 reoperations for recurrent or persistent disease. "Biochemical cure" was obtained in four patients (44%) undergoing reoperation for ZES. Three of these patients remained without evidence of recurrence at 4, 9, and 12 years after their curative re-resection. Only one of nine patients who underwent reoperation died of metastatic gastrinoma. Conclusions Primary and reoperative surgery in patients with sporadic ZES results in a significant rate of "biochemical cure." In selected patients with recurrent or persistent disease, reoperation for resection of gastrinoma is associated with excellent longterm survival and is warranted.

Autoimmune pancreatocholangitis is characterized by sclerosing inflammation of the biliary tree or pancreatic duct and can mimic pancreaticobiliary malignancy. Serum immunoglobin (Ig) G4 values seem to be helpful in distinguishing autoimmune pancreatocholangitis from pancreatic malignancy in the Japanese population; however, its significance in the Western population has not been well studied.We report a retrospective analysis of 7 consecutive patients with autoimmune pancreatocholangitis and compare them to 23 patients with pancreatic malignancy. Clinical presentation, diagnostic tests, and preoperative IgG4 levels were reviewed in all patients. Presence of autoimmune pancreatocholangitis or pancreatic malignancy was determined by pathologic analysis in all patients and reviewed by a single pathologist.In all patients, autoimmune pancreatocholangitis manifested in a similar fashion to pancreatic malignancy. Median IgG4 levels were far lower in pancreatic cancer patients with localized, resectable disease (24 mg/dL), locally advanced disease (24 mg/dL), and metastatic disease (28 mg/dL) as compared with patients with autoimmune pancreatocholangitis (142 mg/dL, P.05). Only one patient with pancreatic cancer had an IgG4 level that was100 mg/dL. In contrast, all patients with autoimmune pancreatitis or cholangitis had levels100 mg/dL. However, in five of these seven patients, IgG4 levels were below the upper limits of normal.Autoimmune pancreatocholangitis mimics pancreatobiliary malignancy. Serum IgG4 values seem to be helpful in distinguishing autoimmune pancreatocholangitis from malignancy in the Western population. However, absolute values seem to be lower in the United States compared with Japan. The upper limit of normal as reported in laboratories in the United States may not be useful in identifying abnormally high IgG4 values. A new upper limit of normal may need to be defined because IgG subclass determinations are being used more frequently in Western patients with biliary obstruction.

Duodenal gastrointestinal stromal tumors (GIST) have been described primarily in isolated case reports. In order to learn more about duodenal GIST, a retrospective review of patients with GIST managed at a single institution between 2000 and 2005 was conducted. Thirty-eight GIST of the stomach and small bowel were analyzed. Eight (21%) were duodenal GIST. The median size of duodenal GIST (6.0 cm) and small bowel GIST (6.3 cm) was larger than the median size of gastric GIST (3.0 cm). The most common presentation of duodenal GIST was bleeding (50%) which was similar to other small bowel GIST (49%) but different from gastric GIST which were most commonly an incidental finding (62%). Two patients (25%) with duodenal GIST had a history of neurofibromatosis. The duodenal GIST were located in the 2nd (n = 5, 63%) and 3rd portion of duodenum (n = 3, 37%). Seven of 8 patients underwent complete resection of duodenal GIST. Pancreaticoduodenectomy was the most common operation performed (n = 5); 2 patients were treated with partial duodenal resection. No patients undergoing pancreaticoduodenectomy (n = 5) were found to have lymph node metastases. No patients received neo-adjuvant or adjuvant therapy with Imatinib. Following resection, 2 patients have recurred (12 and 48 mo.), 4 patients are without disease (1, 6, 6, and 24 mo.), 1 patient died postoperatively. Duodenal GIST are relatively rare tumors that present most commonly with gastrointestinal bleeding. Duodenal GIST are associated with neurofibromatosis. Many duodenal GIST require pancreaticoduodenectomy for complete removal.

Carcinoma of the esophagus and gastroesophageal junction (GEJ) is an aggressive disease with a poor prognosis.1–3 Because of poor outcomes with surgery alone, an increasing number of patients with locoregionally advanced esophageal cancer are treated with preoperative chemoradiation (CRT) and surgery.4,5 The impact of this treatment paradigm shift on the pathologic stage (pTNM) esophageal cancer staging system has not been clearly defined. When surgery alone is the primary therapy, the best predictor of survival and long-term survival is the surgical pathologic stage according to the American Joint Committee on Cancer (AJCC) staging criteria.6,7 Because of pathologic down-staging, the possibility exists that CRT-treated patients may have different survivals than patients treated with surgery alone. Additionally, it is well known that a complete pathologic response (pP) (path CR, 0% residual cancer) following CRT is associated with improved long-term survival.8–10 We have recently demonstrated that a partial pP is also prognostic following CRT, with improved outcomes in patients who have partial pPs (1%–50% residual cancer) compared with no pP (>50% residual cancer).11 This manuscript extends these initial observations and evaluates the impact of complete and partial pPs on the pTNM esophageal cancer staging system. Our study shows that pP is an independent predictor for long-term survival in patients treated with CRT. In this manuscript, we propose a modified pTNM esophageal cancer staging system to incorporate this prognostic factor (pP: P0, 0% residual; P1, 1%–50% residual; P2, >50% residual) into the pTNM staging system to better predict long-term outcome for esophageal cancer patients treated with CRT.

Biliary diversion procedures are performed during gastric surgery to decrease bile reflux. A 1-day dual-radionuclide examination was studied to determine its potential in the evaluation of the effectiveness of the Braun enteroenterostomy in reducing bile reflux and its effects on gastric emptying. Orally ingested gallium 67–labeled egg and intravenously administered technetium 99m diisopropyl-imino-diacetic acid were imaged simultaneously. This provided a way to depict both bile reflux and gastric emptying on the same day in patients who underwent gastric surgery. Overall, the Braun enteroenterostomy trades bile reflux, a symptomatic and premalignant disease, for gastroparesis, a less severe and often treatable disease. © RSNA, 2003

Purpose To evaluate parameters that may influence prognosis in patients treated with preoperative radiotherapy (RT) or chemoradiation. Methods and Materials One hundred seventy-six patients with esophageal carcinoma received preoperative radiotherapy (45 patients) or chemoradiation (131 patients). Forty-three received no surgery (NS), 32 had exploratory surgery (ES), and 101 received definitive surgery (DS). Results Five-year cause-specific survival and absolute survival rates were overall, 19% and 16%; NS group, 0% and 0%; ES group, 3% and 3%; DS group, 30% and 26%. On univariate analysis, definitive surgery (P

Introduction and aims Cystic neoplasms of the pancreas may be inadvertently treated as benign pseudocysts in clinical practice, often without the use of cytology, cyst tumor markers, or histopathology. We assessed the utility of EUS-guided fine-needle aspiration (EUS-FNA) to assist in the diagnosis and management of pancreatic cysts. Methodology All patients who had pancreatic cysts detected by EUS over a 24-month period were analyzed. Preoperative diagnosis was derived from an algorithm combining clinical history and endosonographic features. In selected cases, EUS-FNA was performed and cyst fluid aspirates were analyzed. Surgical specimens served as diagnostic standard. Results A total of 43 patients with pancreatic cysts underwent 45 EUS examinations. Surgical specimens were obtained from 9 patients (mucinous cystadenocarcinoma, 3; adenocarcinoma, 3; pancreatic endocrine tumor, 2; and benign cyst, 1); diagnostic EUS correctly predicted malignant cysts in 8/9 (88.9%). One case inaccurately interpreted by EUS as cystic neoplasm turned out to be a benign cyst on resection. Twenty-one patients underwent EUS-FNA. The cytologic interpretation was adenocarcinoma in 9.5% (2/21); suspicious for malignancy or atypical cells in 19.0% (4/21); benign in 66.6% (14/21); and insufficient cells in 4.8% (1/21). Conclusion The information gathered from clinical history and EUS, complemented by fluid analysis after EUS-guided FNA, predicts neoplastic pancreatic cysts and assists in decision-making for medical or surgical approach.

A retrospective evaluation was performed of the use of gallium imaging in patients with known severe pancreatitis to detect infection in pancreatic and peripancreatic fluid collections.Gallium-67 single-photon emission computed tomography (SPECT) studies were retrospectively reviewed in patients with complicated pancreatitis. Only patients who had undergone interventional procedures within 10 days of the scanning were included in our analysis. A total of 23 scans from 20 patients were reviewed. SPECT imaging was typically performed 48-72 hr after injection of the gallium. All studies were correlated with conventional CT findings. Findings from subsequent interventions (results of aspiration, Gram stains, or cultures) were used as evidence of infection.Twenty patients underwent either percutaneous or surgical drainage within 10 days of their gallium scanning. One patient underwent gallium scanning on three different occasions and underwent three different interventional procedures after each of the gallium scans, bringing the total number of cases in our study to 23. Of these 23 cases, 18 patients (78%) with gallium scans showing positive findings for infection had infected fluid; five patients (22%) with negative findings for infection on gallium scans had sterile fluid (p0.00001). No false-positive scans were found among our study cases, and we found no correlation between the uptake of gallium and the presence or absence of pancreatic necrosis.Gallium does not actively accumulate in all patients with severe pancreatitis, and gallium uptake does not correlate with the presence or absence of necrosis. In patients with severe pancreatitis complicated by fluid collections or inflammatory masses, gallium SPECT is a useful predictor of infection and can be used to help guide subsequent intervention. Gallium SPECT allows targeting sites of infected fluid in patients with multiple fluid collections and potentially obviates intervention in patients with sterile fluid collections.

Background The management of small pancreatic cystic lesions presents a clinical challenge. Methods We reviewed our experience with 78 patients who presented with a cystic pancreatic lesion who underwent operative management between 1995 and 2005. Data on cyst characteristics were analyzed in the context of pathologic findings following resection. Results Among 78 patients, there were 55 (71%) females; median age 63 years. Patients presented with: an incidental finding (48%), pain (40%), acute pancreatitis (4%), other (8%). Operations were distal pancreatectomy (n = 47), pancreaticoduodenectomy (n = 16), and other (n = 15). Most patients had a non-malignant lesion (n = 65, 83%) (mucinous cystadenoma (n = 29), serous cystadenoma (n = 15), IPMN without invasion (n = 8), pseudocyst (n = 8), other benign (n = 5)). Malignant lesions (adenocarcinoma, neuroendocrine tumor, and other) were found in 13 patients (17%). The risk of malignancy increased with size: 5 cm (n = 30), 30%. Pre-operative cyst fluid cytology was performed in 41 patients. The negative predictive value (NPV) of cytology for malignancy was 88% and the positive predictive value (PPV) was 80%. The NPV of CA 19-9 for malignancy was 90%; the PPV was 50%. Conclusions Initial conservative management of small cystic pancreatic lesions may be indicated in selected patients. J. Surg. Oncol. 2009;100:372–374. © 2009 Wiley-Liss, Inc.

ObjectiveIn a randomized, double-blind, multicenter trial, ciprofloxacin/metronidazole was compared with imipenem/cilastatin for treatment of complicated intra-abdominal infections. A secondary objective was to demonstrate the ability to switch responding patients from intravenous (IV) to oral (PO)

OBJECTIVE: This retrospective, nonrandomized review evaluates 125 patients with esophageal carcinoma (adenocarcinoma and squamous cell) who underwent either surgery only or preoperative chemotherapy and/or radiation therapy followed by surgery. Major end points were survival and postchemoradiation downstaging. METHODS: Forty-four patients underwent radiation therapy of 4500 cGy over 5 weeks. Fluorouracil and cisplatin were administered on the first and fifth week of radiotherapy. Ninety-eight patients underwent "potentially curative" resections-transhiatal esophagectomy (70), Lewis esophagogastrectomy (25), and left esophagogastrectomy (3). All patients with preoperative adjuvant therapy underwent endoscopy and biopsy before surgery. RESULTS: There were no differences in overall mortality (5%) or surgical complications in either group. Fourteen of 44 patients (32%) downstaged to complete pathologic response, with 5-year survival of 57%. Fifteen of 44 patients (34%) downstaged to microscopic residual tumor, with 1- and 3-year survival of 77% and 31%, respectively. Twenty-eight of 29 patients in the two downstaged groups were lymph node negative. Overall, 5-year survival in the adjuvant therapy plus surgery group versus surgery only was 36% and 11% (p = 0.04). Five-year survival in lymph node-negative adjuvant therapy and surgery patients was 49% (p = 0.005). Positive nodes in the surgery only group was 48% versus 23% in the adjuvant therapy and surgery group (p = 0.02). CONCLUSION: Although retrospective and nonrandomized, these results suggest that preoperative chemoradiation results in significant clinical and pathologic downstaging, increases survival, and may sterilize local and regional lymph nodes, accounting for both downstaging and survival statistics.

Delayed gastric emptying and altered upper gut transit or both are common following Roux-en-Y gastrojejunostomy and are thought to be due to altered efferent limb transit secondary to isolation of the Roux limb from the duodenal pacemaker. We postulated that preservation of myoneural continuity of the Roux limb with the duodenal pacemaker would enhance solid gastric emptying, while division of the afferent limb of a Billroth II gastrojejunostomy (B-II), isolating the efferent jejunal limb from the duodenal pacemaker, would slow gastric emptying. Solid gastric emptying was measured in 14 dogs, who then underwent gastric vagotomy and antrectomy. Eight animals were reconstructed with a Roux-en-Y gastrojejunostomy, preserving myoneural but not luminal continuity of the Roux limb with the afferent limb via a muscularis bridge, while six dogs underwent standard B-II gastrojejunostomy. Serosal electrodes were placed on the afferent and efferent jejunal limbs. Gastric emptying was restudied, with fed and fasted myoelectric recordings. The bridge was then divided to create a standard Roux, while the afferent limb was transected and reanastomosed just proximal to the gastrojejunostomy in the B-II dogs to isolate the efferent limb from the duodenal pacemaker, with repeat studies. Bridge dogs had delayed solid gastric emptying compared to their preoperative state, despite normal efferent limb motility. Gastric emptying was not further altered by division of the bridge, despite altered efferent limb motility, including a decreased migrating myoelectric complex (MMC) cycle length (62 ± 1 vs 128 ± 18 min, P < 0.001), slow wave frequency (SWF) (14.1 ± 0.6 vs 17.6 ± 0.6 cpm, P < 0.01), postptrandial suppression of MMCs (60 ± 18 vs 303 ± 28 min, P < 0.01), and an inverted SWF gradient, indicating ectopic pacemakers. Solid gastric emptying was unchanged after standard B-II compared to the preoperative state, even following transection and reanastomosis of the afferent limb. Solid gastric emptying was faster in the latter group than in Roux animals, despite similar efferent limb motility. We conclude that myoneural continuity of the efferent limb of a gastrojejunostomy with the duodenal pacemaker is less important than luminal continuity of the afferent and efferent limbs in determining postgastrectomy solid gastric emptying.

Surgery and radiotherapy yield equivalent results for patients with carcinoma of the cervical esophagus, with long-term survival of less than one in four patients. The advantages of radiotherapy are lower rates of acute morbidity and mortality compared with surgery, and potential for larynx preservation. The advantage of surgery is that the transposed stomach may function better over the long term than an irradiated esophagus, which tends to become stenotic over time. Patients with resectable cancers who are in good general medical condition may be treated with preoperative irradiation and surgery in an effort to improve the likelihood of local control and obtain a good functional result. Patients with relatively early lesions who are not good surgical candidates can be treated with high-dose radiation therapy. Patients with very advanced local disease and those with distant metastases are treated with palliative irradiation. Concurrent chemotherapy and radiotherapy has been shown to be superior to radiotherapy alone in managing esophageal cancer. Although relatively few patients with cervical esophageal cancer have been treated with this combination, some studies suggest it may be curative for patients with early lesions of the cervical esophagus, without the need for esopha-gectomy. Adjuvant chemotherapy is usually not used outside of a study setting.

OBJECTIVE AND SUMMARY BACKGROUND: Symptomatic, medically resistant postgastrectomy patients with alkaline reflux gastritis (ARG) have increased enterogastric reflux (EGR) documented by quantitative radionuclide biliary scanning. Even asymptomatic patients after gastrectomy have increased EGR compared with nonoperated control patients. Roux-en-Y biliary diversion, although successfully treats the clinical syndrome of ARG, has a high incidence of early and late postoperative severe gastroparesis, Roux limb retention (the Roux syndrome), or both, which often requires further remedial surgery. As an alternative to Roux-en-Y diversion, this review evaluates the efficacy of the Braun enteroenterostomy (BEE) in diverting bile away from the stomach in patients having gastric operations. Based on previous pilot studies, the BEE is positioned 30 cm from the gastroenterostomy. METHODS: Thirty patients had the following operations and were evaluated: standard pancreatoduodenectomy (8), vagotomy and Billroth II (BII) gastrectomy (6), BII gastrectomy only (10), and palliative gastroenterostomy to an intact stomach (6). All anastomoses were antecolic BII with a long afferent limb and a 30-cm BEE. Four symptomatic patients with medically intractable ARG and chronic gastroparesis had subtotal BII gastric resection with BEE rather than Roux-en-Y diversion. Eight control symptomatic patients and six asymptomatic patients with previous BII gastrectomy and no BEE were evaluated. Radionuclide biliary scanning was performed within 30 days in all patients and at 4 to 6 months in 14 patients. Bile reflux was expressed as an EGR index (%). RESULTS: After operation, 18 of 34 patients (53%) had no demonstrable EGR while in the fasting state for as long as 90 minutes. The range of demonstrable bile reflux (EGR) in the remaining 16 patients was from 2% to 17% (mean, 4.5%). Enterogastric reflux in the 14 control patients (with no BEE) ranged from 5% to 82% (mean, 42%). The four patients with ARG and chronic gastroparesis treated by subtotal gastrectomy and BEE had postoperative EGR of 0%, 2%, 2%, and 4%, respectively. They are asymptomatic with no evidence of bile reflux gastritis. In the 14 patients who had late evaluation, EGR ranged from 0% to 16% (mean, 5.5%). No patient had signs or symptoms of ARG after operation. CONCLUSIONS: Braun enteroenterostomy successfully diverts a substantial amount of bile from the stomach. The ARG syndrome might be prevented by performing BEE during gastric resection or bypass in a variety of operations. Conversion to a BII with BEE may be an alternative to Roux-en-Y diversion in treating medically resistant ARG and subsequent may avoid the Roux syndrome.

OBJECTIVE: To evaluate the efficacy of two distinct imaging techniques to predict, before operation, unresectability compared with standard computed tomographic scan (CT). SUMMARY BACKGROUND: Accurate preoperative identification of the number, size, and location of hepatic lesions is crucial in planning hepatic resection for colorectal hepatic metastases. Although infusion-enhanced CT is the standard, its limitations are the imaging of relatively isodense and/or small (< 1 cm) lesions. The increased sensitivity of CT arterial portography (CTAP) may be offset by false-positive results caused by benign lesions and flow artifacts. METHODS: Fifty-eight selected patients considered to be eligible for resection by standard CT had laparotomy. Before operation and in addition to CT, all patients had CT arterial portography and hepatic artery perfusion scintigraphy (HAPS) using radiolabeled macroaggregated albumin. Early studies showed an increased sensitivity for detecting small lesions using the invasive CTAP. Similarly, the HAPS study has detected malignant lesions not observed by standard CT. RESULTS: Of 58 patients having laparotomy, 40 were resectable by either lobectomy (22) or trisegmentectomy (1) and the rest by single or multiple wedge resections. Eighteen patients could not be resected because of combined intra- and extrahepatic disease or the number and location of metastases. Standard CT detected 64% of all lesions (12% of lesions less than 1 cm). Unresectability was accurately predicted by CTAP and HAPS in 16 (88%) and 15 (83%), respectively, of the 18 patients considered ineligible for resection at laparotomy. Of the 40 patients who had resection for possible cure, CTAP and HAPS falsely predicted unresectability in 6 of 40 patients (15%) and in 10 of 40 patients (25%), respectively. The positive predictive value for unresectability of CTAP and HAPS was 73% and 60%, respectively. False-positive lesions after CTAP included hemangiomas, cysts, granulomas, and flow artifacts. False-positive HAPS lesions included patients in whom no tumor was found at surgery but with some identified by intraoperative ultrasound, blind biopsy, and blind resection. CONCLUSIONS: False-positive results by HAPS and CTAP may limit the ability of these tests to accurately predict unresectability before operation and may deny patients the chance for surgical resection. The HAPS study does, however, detect small lesions not seen by CT or CTAP. Standard CT, although less sensitive, followed by surgery and intraoperative ultrasound, does not necessarily preclude patients who could be resected.

Background and Objectives Morbidity rates following pancreaticoduodenectomy (PD) remain high with delayed gastric emptying (DGE) and slow resumption of oral diet contributing to increased postoperative length of stay. A Braun enteroenterostomy has been shown to decrease bile reflux following gastric resection. We hypothesize that addition of Braun enteroenterostomy during PD would reduce the sequelae of DGE. Methods From our PD database, patients were identified that underwent classic PD with partial gastrectomy from 2001 to 2006. All patients with reconstruction utilizing a single loop of jejunum at the University of Florida Shands Hospital were reviewed. Demographics, presenting signs and symptoms, pathologic diagnoses, and postoperative morbidity were compared in those patients undergoing reconstruction with an additional Braun enteroenterostomy (n = 70) to those not undergoing a Braun enteroenterostomy (n = 35). Results Patients undergoing a Braun had NG tubes removed earlier (Braun: 2 days, no Braun: 3 days, P = 0.002) and no significant change in postoperative vomiting (Braun: 27%, no Braun: 37%, P = 0.37) or NG tube reinsertion rates (Braun: 17%, no Braun: 29%, P = 0.21). Median postoperative day with tolerance of oral liquids (Braun: 5, no Braun: 6, P = 0.01) and solid diets (Braun: 7, no Braun: 9, P = 0.01) were significantly sooner in the Braun group. DGE defined by two criteria including the inability to have oral intake by postoperative day 10 (Braun: 10%, no Braun: 26%, P

Postoperative gastric myoelectric activity, gastric emptying, and clinical course were correlated in 17 patients at high risk of developing gastroparesis after gastric surgery. In addition, an attempt was made to pace the stomach with an electrical stimulus and determine the effect of pacing on early postoperative gastric emptying. Gastric dysrhythmias (bradygastria, slow wave frequency2 cycles/min; tachygastria, slow wave frequency4 cycles/min) persisted beyond the first postoperative day in 6 patients (35%). Delayed gastric emptying was identified by a radionuclide meal in 15 patients (88%), but symptoms of gastroparesis developed in only 6 of 15 (40%). Patients with postoperative gastroparesis had more frequent dysrhythmias than asymptomatic patients (67% vs. 18%), but these differences were not significant, although we cannot exclude a type II statistical error. Gastric rhythm was entrained in 10 of 16 patients (63%). Pacing increased the gastric slow wave frequency (3.1 vs. 4.1 cycles/min; P0.01) but did not improve gastric emptying (gastric retention at 60 minutes, 86% +/- 6% for control and 90% +/- 2% for paced). In conclusion, gastric dysrhythmias do not appear to play a major role in the development of postsurgical gastroparesis. Although gastric rhythm could be entrained in the majority of patients, pacing did not improve gastric emptying overall.

Delayed gastric emptying occurs frequently following Roux-en-Y gastrojejunostomy. The role of vagal denervation in the etiology of this "Roux-stasis syndrome" is controversial. This study evaluates the effect of selective vagotomy on gastric emptying and motility following Roux-en-Y. Four dogs underwent control gastric emptying studies. The animals then underwent selective vagotomy, antrectomy, and Billroth II gastrojejunostomy, with placement of serosal electrodes. Gastric emptying was assessed with simultaneous myoelectric recordings, and the animals were converted to Roux-en-Y, followed by repeat studies. Gastric emptying was unchanged following selective vagotomy, antrectomy, and Billroth II gastrojejunostomy (T 1/2: 132 +/- 18 min [SEM] versus 118 +/- 14 min control) but was markedly delayed following Roux-en-Y diversion (T 1/2: 286 +/- 44 min; p less than 0.01). All animals went into the fed pattern following Billroth II gastrojejunostomy (migrating myoelectric complex [MMC] interval: 326 +/- 6 min postprandial versus 92 +/- 5 min fasting; p less than 0.01), but no fed pattern was recognized in three of four animals following Roux-en-Y diversion (MMC interval: 68 +/- 12 min postprandial versus 62 +/- 1.5 min fasting; p = NS). In a canine model, selective vagotomy does not prevent delayed gastric emptying or myoelectric alterations following Roux-en-Y.

Endoscopic therapy of acute and chronic pancreatitis has decreased the need for operative intervention. However, a significant proportion of patients treated endoscopically require definitive surgical management for persistent symptoms. Our aim was to determine which patients are likely to fail with endoscopic therapy, and to assess the clinical outcome of surgical management. Patients were identified using ICD-9 codes for pancreatic disease as well as CPT codes for endoscopic therapy followed by surgery. Patients with documented acute or chronic pancreatitis treated endoscopically prior to surgical therapy were included (N = 88). The majority of patients (65%) exhibited chronic pancreatitis due to alcohol abuse. Common indicators for surgery were: persistent symptoms, anatomy not amenable to endoscopic treatment and unresolved common bile duct or pancreatic duct strictures. Surgical salvage procedures included internal drainage of a pseudocyst or an obstructed pancreatic duct (46%), debridement of peripancreatic fluid collections (25%), and pancreatic resection (31%). Death occurred in 3% of patients. The most common complications were hemorrhage (16%), wound infection (13%), and pulmonary complications (11%). Chronic pancreatitis with persistent symptoms is the most common reason for pancreatic surgery following endoscopic therapy. Surgical salvage therapy can largely be accomplished by drainage procedures, but pancreatic resection is common. These complex procedures can be performed with acceptable mortality but also with significant risk for morbidity.

The purpose of this study was to compare gastric emptying and Roux myoelectric activity in a canine model. Four dogs underwent truncal vagotomy, antrectomy, and 40 cm Roux-en-Y gastrojejunostomy, with placement of serosal electrodes. Following recovery, gastric emptying was determined scintigraphically with a radiolabeled solid meal, and fasting and fed small-bowel myoelectric activity was obtained. Gastric emptying was markedly slowed compared to control unoperated animals (202 +/- 91 versus 46 +/- 12 min; P less than 0.05). Slow wave frequency declined in the Roux limb compared to the duodenum (14.2 +/- 0.4 versus 18.0 +/- .06 counts per minute; P less than 0.01). No gradient in slow wave frequency was observed in the Roux limb, although one animal was noted to have reversed propagation of slow waves in the proximal Roux limb. Migrating myoelectric complexes (MMCs) were coordinated between the Roux limb and jejunum distal to the enteroenterostomy, but not with the duodenum. Periodicity of the MMCs was different in the Roux limb and duodenum (98.6 +/- 6.3 versus 138 +/- 17.5 min; P less than 0.05). None of the animals converted to the fed myoelectric pattern with a 272 kcal meal (MMC periodicity in the Roux limb = 99 +/- 10 min postprandially, P = N.S.). These quantitative and qualitative alterations in myoelectric activity may contribute to the observed delay in gastric emptying following Roux-en-Y gastrojejunostomy.

Background Ampullectomy may be an appropriate oncologic procedure in selected patients. Sparse data exist on procedure-related complications and the relationship between histologic analysis and outcomes. Study Design We retrospectively reviewed our experience with ampullectomy in 29 patients with a preoperative benign histologic diagnosis over 15 years (1991 to 2006). Presenting signs, symptoms, and preoperative diagnostic studies were reviewed. Postoperative complications and followup for recurrence were recorded. The abilities of preoperative histologic biopsy, intraoperative frozen section, and final histologic analysis to guide management and predict outcomes were determined. Results Median age was 63 years. Jaundice was present in 30% of patients. Median length of hospital stay was 9 days. Forty-five percent of patients had a complication, and there was one postoperative mortality (3%). Ampullary adenomatous neoplasms were present in 89% of patients. Preoperative biopsy had complete concordance with final pathology in 76% of patients. Preoperative biopsy and intraoperative frozen section failed to identify carcinoma in four patients. Pancreaticoduodenectomy was performed within 7 days in the postoperative period in three of these patients. After ampullectomy (median followup=16 months), recurrences were identified in two patients (8%) with benign tumors. No patients with high-grade dysplasia (n=4) have had recurrence. Conclusions Preoperative biopsy and intraoperative frozen section analysis have limitations in the management of patients undergoing ampullectomy. High-grade dysplasia on preoperative biopsy is not an absolute contraindication to ampullectomy. Morbidity of ampullectomy is significant, but longterm outcomes of this procedure, in patients without invasive malignancy, are acceptable.

To evaluate the outcome of aggressive conservative therapy in patients with esophageal perforation.The treatment of esophageal perforation remains controversial with a bias toward early primary repair, resection, and/or proximal diversion. This review evaluates an alternate approach with a bias toward aggressive drainage of fluid collections and frequent CT and gastographin UGI examinations to evaluate progress.From 1992 to 2004, 47 patients with esophageal perforation (10 proximal, 37 thoracic) were treated (18 patients early [24 hours], 29 late). There were 31 male and 16 females (ages 18-90 years). The etiology was iatrogenic (25), spontaneous (14), trauma (3), dissecting thoracic aneurysm (3), and 1 each following a Stretta procedure and Blakemore tube placement.Six of 10 cervical perforations underwent surgery (3 primary repair, 3 abscess drainage). Nine of 10 perforations healed at discharge. In 37 thoracic perforations, 2 underwent primary repair (1 iatrogenic, 1 spontaneous) and 4 underwent limited thoracotomy. Thirty-4 patients (4 cervical, 28 thoracic) underwent nonoperative treatment. Thirteen of the 14 patients with spontaneous perforation (thoracic) underwent initial nonoperative care. Overall mortality was 4.2% (2 of 47 patients). These deaths represent 2 of 37 thoracic perforations (5.4%). There were no deaths in the 34 patients treated nonoperatively. Esophageal healing occurred in 43 of 45 surviving patients (96%). Subsequent operations included colon interposition in 2, esophagectomy for malignancy in 3, and esophagectomy for benign stricture in 2.Aggressive treatment of sepsis and control of esophageal leaks leak lowers mortality and morbidity, allow esophageal healing, and avoid major surgery in most patients.

Previously, we demonstrated that ketorolac, a nonsteroidal antiinflammatory drug (NSAID), prevented postoperative small bowel ileus in a rodent model. The aim of this study was to evaluate the effect of salsalate, an NSAID without antiplatelet effect, on postoperative ileus alone or in combination with morphine.Forty-eight rats underwent placement of duodenal catheters and were then randomly assigned to one of eight groups (n = 6). Four groups had standardized laparotomy following drug administration, whereas 4 groups underwent the same treatment without laparotomy: control and morphine animals received 0.1 mL alcohol via the catheter, whereas salsalate and salsalate-plus-morphine animals received salsalate (15 mg/kg) dissolved in 0.1 mL alcohol. The animals also received 0.5 mg/kg morphine (morphine and salsalate plus morphine) or the same volume of saline (control and salsalate) subcutaneously. Transit was measured following the injection of a nonabsorbed marker via the duodenal catheter and is defined as the geometric center (GC) of distribution. An additional 20 rats had serosal electrodes placed on the jejunum, and were assigned to one of four treatment groups (control, salsalate, morphine, and salsalate plus morphine; n = 5 each group). Myoelectric activity was recorded until the reappearance of the migrating myoelectric complex (MMC) following laparotomy.Laparotomy and morphine independently reduced small bowel transit (P = 0.0006 and 0.006, respectively, by three-way analysis of variance [ANOVA]; GC 4.3 +/- 0.2 control versus 2.2 +/- 0.3 laparotomy versus 3.6 +/- 0.4 morphine), but morphine did not further worsen postoperative transit (GC 2.4 +/- 0.4; P = 0.42). Although salsalate did not alter baseline transit, pretreatment improved postoperative transit (P = 0.0002; GC 3.6 +/- 0.4). This effect was lost with the addition of morphine (GC 2.7 +/- 0.2; P = 0.21). The MMCs returned earlier after laparotomy in salsalate-pretreated rats (63 +/- 18 minutes salsalate versus 160 +/- 12 minutes laparotomy; P0.01, one-way ANOVA). However, this effect was also lost in animals receiving morphine (106 +/- 16 min; P0.05).Salsalate improves postoperative small bowel motility in a rodent model; however, this effect is masked by morphine.

The role of enterococcus in intraabdominal infection is controversial. This study examines the contribution of enterococcus to adverse outcome in a large intraabdominal infection trial.A randomized prospective double-blind trial was performed to compare two different antimicrobial regimens in combination with surgical or percutaneous drainage in the treatment of complicated intraabdominal infections. A total of 330 valid patients was enrolled from 22 centers in North America.In 330 valid patients, 71 had enterococcus isolated from the initial drainage of an intraabdominal focus of infection. This finding was associated with a significantly higher treatment failure rate than that of patients without enterococcus (28% versus 14%, p0.01). In addition, only Acute Physiology and Chronic Health Evaluation II score and presence of enterococcus were significant independent predictors of treatment failure when stepwise logistic regression was performed (p0.01 and0.03). Risk factors for the presence of enterococcus include age, Acute Physiology and Chronic Health Evaluation II, preinfection hospital length of stay, postoperative infections, and anatomic source of infection. There was no difference between the clinical trial treatment regimens with regard to overall failure, failure associated with enterococcus, or frequency of enterococcal isolation.This study is the first to report enterococcus as a predictor of treatment failure in complicated intraabdominal infections. This trial also identifies several significant risk factors for the presence of enterococcus in such infections.

Octreotide, a somatostatin analogue that inhibits the release of most gut peptides, hastens the resolution of experimental postoperative ileus, suggesting that gut peptides mediate this process. We studied the role of two gut peptides involved in the control of normal gut motility, vasoactive intestinal peptide (VIP), and substance P (SP), in the initiation and maintenance of postoperative small bowel ileus in rats by preoperative administration of VIP and SP receptor antagonists, (VIP-ra and SP-ra). Thirty male Sprague-Dawley rats (300-350 g) underwent laparotomy. One half underwent placement of a duodenal catheter for transit studies while the other half had serosal electrodes placed on the proximal jejunum for myoelectric recordings. Six days later, animals were separated into three treatment groups of five each. Control animals were pretreated with ip saline, while the others received either VIP-ra or SP-ra prior to standardized laparotomy. Following abdominal closure, [Na51]CrO4 was injected into the duodenum and the animals were sacrificed 25 min later. The small bowel was then excised and divided into 10 equal segments. Small bowel transit was calculated as the geometric center of [Na51]CrO4 distribution. The interval until the return of migrating myoelectric complexes (MMCs) was determined in animals with intestinal electrodes. VIP-ra-treated rats demonstrated a 67% improvement in the geometric center of radiolabel relative to controls and SP-ra-treated rats had a 23% improvement (3.67 +/- 0.06 VIP-ra vs 2.69 +/- 0.09 SP-ra vs 2.20 +/- 0.09 control, P < 0.01). MMCs returned 180 +/- 17 min in controls vs 99 +/- 14 min in VIP-ra-treated rats (P < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Background The "Roux stasis syndrome" is characterized by symptoms of upper gut stasis following Roux-en-Y gastrojejunostomy (RG). Whether symptoms result from delayed gastric emptying, altered Roux-limb transit, or both has never been settled, partly because of the difficulty of measuring Roux-limb transit. The aim of this study was to develop a model to simultaneously quantitate Roux-limb transit and gastric emptying. Methods Rats underwent vagotomy and antrectomy with RG or Billroth II reconstruction (B-II). Gastrointestinal transit of a solid meal (Technetium-99m sulfur colloid-labelled egg white) was assessed 0.5, 1, and 1.5 hours postprandial (pp). Transit of a liquid marker (Na 51 -CrO 4 injected through an efferent-limb catheter) was measured at 25 minutes pp. Results Solid gastric emptying was slower in RG than in B-II rats at 60 and 90 minutes pp. More of the solid meal and of the liquid marker was retained in the Roux limb than the efferent limb of the B-II at all time points ( P Conclusions In a rodent model, Roux-en-Y gastrojejunostomy is associated with delayed gastric emptying and slowed efferent-limb transit of solids and liquids.

We have demonstrated that erythromycin improves gastric emptying in dogs following truncal vagotomy and Roux-en-Y antrectomy (VRYA). To explore its mechanism of action we studied gastric emptying and myoelectric activity in a canine Roux model and administered atropine simultaneously with erythromycin. Tachyphylaxis was evaluated following short-term administration. Four dogs with delayed gastric emptying following VRYA were studied. Radionuclide solid gastric emptying was measured, with simultaneous myoelectric recordings obtained from the duodenum and Roux limb. Study groups were: (1) saline control (VRYA dogs); (2) erythromycin 1 mg/kg iv over 1 hr; (3) erythromycin 3 mg/kg po tid for 1 week, with repeat studies using erythromycin 1 mg/kg iv over 1 hr; and (4) atropine 0.5 mg/kg iv bolus, followed by a 1-hr infusion of atropine 0.05 mg/kg and erythromycin 1 mg/kg. Control Roux animals had severe gastric retention (73 +/- 5% at 2 hr, compared to 27 +/- 6% following iv erythromycin (P less than 0.01). Clustered spike bursts were observed in the Roux limb following erythromycin. Atropine abolished the gastrokinetic response and suppressed the myoelectric response to erythromycin (81 +/- 3% retention at 2 hr, P less than 0.01 compared to erythromycin alone). The response to erythromycin was unchanged after 1 week of tid administration (40 +/- 14% retention at 2 hr postprandial, P = NS). Erythromycin improves gastric emptying in VRYA dogs via a cholinergic pathway and does not exhibit tachyphylaxis following short-term administration.

Delayed gastric emptying occurs in up to 50% of patients after truncal vagotomy and Roux-Y antrectomy and is often resistant to nonsurgical therapy. This study evaluates the effect of erythromycin, metoclopramide, and motilin on delayed gastric emptying in four dogs after Roux-Y antrectomy. Solid food gastric emptying was measured using a radionuclide technique. Study groups were: (1) saline control; (2) erythromycin 1 mg/kg intravenously over 1 hour; (3) erythromycin 3 mg/kg by mouth 45 minutes prior to feeding; (4) metoclopramide 0.6 mg/kg intravenously over 1 hour; and (5) motilin 500 ng/kg intravenously over 1 hour. After Roux-Y antrectomy, saline control dogs had 73% +/- 5% (SEM) gastric retention at 2 hours. After intravenous and oral erythromycin, gastric emptying improved at 2 hours to 27% +/- 6% and 39% +/- 5% (p less than 0.01 compared with control). Erythromycin intravenously and by mouth improved gastric emptying compared with metoclopramide (64% +/- 8%, p less than 0.05). Motilin enhanced gastric emptying to a similar degree as erythromycin, with a 2-hour gastric retention of 37% +/- 4% (NS). Erythromycin improved gastric emptying in dogs with severe Roux-Y gastroparesis and may have clinical application.

A Novel Gene Transfer Therapy Against Pancreatic Cancer (TNFerade) Delivered by Endoscopic Ultrasound (EUS) and Percutaneous Guided Fine Needle Injection (FNI) Kenneth J. Chang, Neil Senzer, Theodore Chung, J. Randolph Hecht, Stephen Vogel, Alexander Rosemurgy, John Nemunaitis, John Gibbs, Milind Javle, Tony Reid, Jennifer Macko, Paul Kessler, Mitchell C. Posner, James Farrell, L. Shane Grundy, Roy Soetikno, Irving Waxman, Nader Hanna We published two trials utilizing EUS-guided FNI for the delivery of immune therapy (Cancer 2000) and oncolytic viral therapy (Clin Cancer Res 2003) against pancreatic cancer. We now present a novel gene transfer therapy, TNFerade, delivered by EUS and percutaneous approaches (PTA) guided by ultrasound or computerized tomography (CT), in the treatment of unresectable, locally advanced adenocarcinoma of the pancreas (LAPC). TNFerade is a replicationdeficient adenovector containing human TNFa gene, regulated by a radiationinducible promoter Egr-1. Purpose: 1) Assess safety and efficacy of TNFerade plus chemoradiation in pts with LAPC. 2) Compare EUS and PTA as delivery modalities. Methods: Dose-escalating run-in design. Five-week treatment of weekly intratumoral injections of TNFerade (43 10 particle units (pu) in 2mL) via EUS or PTA (each institution allowed one preferred technique), continuous iv 5-FU (200 mg/m/d 3 5 days/wk) and radiation (50.4 Gy). TNFerade was delivered with a single needle pass at a single site in the tumor for PTA, while up to 4 injections were given by EUS. Endpoints: Safety and tumor response on spiral CT (core lab). Results: Of 37 pts, 17 had EUS and 20 had PTA (similar TNFerade doses). EUS vs PTA: T4 staging = 88% vs. 75%; N1 staging = 29% vs. 40%, mean tumor area (cm)= 18 vs.19; mean CA19-9= 4,396 vs. 1,707, meanKPS= 88 vs. 90. (p=NS for all). 1 dose limiting toxicity (gr 3 hypotension) in a PTApt; all other adverse events (AEs) potentially related to TNFerade were grade 1-2. Procedure related AEs were all grade 1-2 and were similar, except for injection site pain: 35% PTA vs 0% EUS (p = .01). Tumor responses and disease control were similar (Table 1). 4 patients underwent resection; one, an EUS patient, had a complete pathologic response. Conclusions: 1) TNFerade plus chemoradiation was well tolerated with maximum tolerated dose not reached; 2) indications of clinical activity were demonstrated by local tumor control and progression-free survival at up to 3 months; 3) EUS and PTA appear to be similarly safe and effective for delivery of TNFerade.

Background: EUS has established itself as the most accurate pre-operative staging modality for EC. Its utility in assessing response after neo-adjuvant therapy has been less promising. Aims: To determine whether the initial EUS T-stage predicts response to neo-adjuvant therapy and if the endoscopic appearance at the time of a re-staging EUS correlated with pathology from the resected specimen (pTNM). Methods: 32 patients (27M, 5F, av age 63.9±10.3) with EC underwent an esophagectomy between 7/98 and 11/99 after neo-adjuvant therapy. A staging EUS was performed followed by a repeat EGD and EUS prior to surgery. Results: 19 patients (16 adeno, 3 SCCA) had an EUS prior to neo-adjuvant therapy for EC. 17 patients underwent an EGD with 15 undergoing concomitant EUS after receiving neo-adjuvant therapy. 14 patients underwent an esophagectomy. 50% (5/10) of T3 lesions and 57% (4/7) of T2 lesions had no mass by EGD after neo-adjuvant therapy (NS). The presence of a mass at EGD after neo-adjuvant therapy had the following sensitivity, specificity, positive predictive value, and negative predictive value for detecting tumor in resected specimens: 80%,100%,100%, and 90%. A re-staging EUS was unable to correctly downstage 54.5 % (6/11) when compared to pTNM. 86% (6/7) of patients downstaged to at least T2 by a re-staging EUS had no tumor in the resected specimen. Conclusions: Response rates to neo-adjuvant therapy appear to be equal in patients staged either T2 or T3 by EUS. The absence of a visible mass by EGD after neo-adjuvant therapy has a NPV of 90% for detecting tumor in the resected specimen. Although the ability of EUS to assess response to neo-adjuvant therapy remains poor, EUS may help in identifying a subset of patients who have a low likelihood of residual EC if downstaged to T2.

Background Postoperative infections remain common after high-risk gastrointestinal procedures. PGG-glucan (Betafectin; Alpha Beta Technology Inc, Worcester, Mass), derived from yeast cell walls, promotes phagocytosis and intracellular killing of bacterial pathogens by leukocytes, prevents infection in an animal model of wound infection, and acts synergistically with antibiotics to reduce mortality in rat peritonitis. Hypothesis We hypothesized that infectious complications in these patients might be reduced by the administration of a nonspecific immune-enhancing agent. Design Multicenter, prospective, randomized, double-blind, placebo-controlled trial of 1249 patients prospectively stratified into colorectal or noncolorectal strata. Setting Thirty-nine medical centers throughout the United States. Patients Aged 18 years or older, scheduled for gastrointestinal procedure lasting 2 to 8 hours, with 2 or more defined risk factors. Interventions PGG-glucan, 0.5 mg/kg or 1.0 mg/kg, or placebo once preoperatively and 3 times postoperatively. All patients received standardized antibiotic prophylaxis. Main Outcome Measures Serious infection or death within 30 days. Results All randomized patients revealed no difference in serious infections and deaths in the treated groups compared with placebo groups (15% vs 14%, P >.90). In the prospectively defined noncolorectal stratum (n=391), PGG-glucan administration was associated with a statistically significant relative reduction (39%) in serious infections and death (placebo, 46 [36%] of 129 vs either PGG-glucan group, 29 [21%] of 132 and 28 [22%] of 130, P P P Conclusion Perioperative administration of PGG-glucan reduced serious postoperative infections or death by 39% after high-risk noncolorectal operations.

Objective: To evaluate the safety and efficacy of cefepime hydrochloride plus metronidazole vs the combination of imipenem and cilastatin sodium in the treatment of complicated intra-abdominal infections in adult patients. Design: Prospective, randomized, double-blind multicenter study. Setting: University-affiliated hospitals in the United States and Canada. Patients: Three hundred twenty-three patients with complicated intra-abdominal infections in whom an operative procedure or percutaneous drainage was required for diagnosis and management. Intervention: Cefepime, 2 g, was administered intravenously every 12 hours (n=164) in addition to metronidazole, 500 mg (or 7.5 mg/kg) intravenously every 6 hours. Imipenen—cilastatin sodium, 500 mg, was administered intravenously every 6 hours (n= 159). Surgical infection management was determined by the patients' surgeons. Main Outcome Assessments: Clinical cure, defined as elimination of all signs and symptoms relevant to the original infection; and treatment failure, defined as persistence, increase or worsening of signs and symptoms resulting in an antibiotic change, requirement of an additional surgical procedure to cure the infection, or a wound infection with fever. Results: Of the initial isolates, 84% were susceptible to cefepime and 92% were susceptible to imipenemcilastatin. Among the 217 protocol-valid patients, those treated with cefepime+metronidizole were deemed clinical cures (88%) more frequently than were imipenemcilastatin—treated patients (76%) (P=.02). Using multivariate analysis to adjust for identified clinical risk factors for an adverse outcome (severity of presenting illness, isolation of enterococcus, type of infection, and duration of prestudy hospitalization), there was a trend (P=.06) toward a higher cure rate favoring cefepime+metronidazole. Pathogens were eradicated in significantly (P=.01) more patients treated with combined cefepime and metronidazole (89%) than with imipenem-cilastatin (76%). Conclusion: The combination of cefepime plus metronidazole is safe and effective therapy for patients with severe intra-abdominal infections. Arch Surg. 1997;132:1294-1302

While computerized tomographic (CT) scanning and intraoperative exploration are both considered accurate measures of liver involvement with metastatic disease, 10% to 30% of colorectal liver metastases remain undetected. Attempting to improve current methods for detecting colorectal liver metastases, CEA levels in gallbladder bile and serum from patients with known liver metastases were determined. One hundred per cent of patients with single and multiple metastases of various dimensions were observed to have gallbladder bile CEA levels strikingly higher than serum values (4.7 to 259 times greater, p = 0.0009). Linear regression analysis of estimated tumor volume and surface area versus gallbladder bile CEA levels predicted that very small tumors (less than or equal to 1 cm3 in volume) might produce detectable levels (9 to 41 ng/mL) of biliary CEA. For this reason, patients who lack clinical and radiologic evidence of distant metastases at the time of primary colorectal resection but who do have elevated gallbladder bile CEA levels (greater than or equal to 10 ng/mL) are being followed for the appearance of occult hepatic metastases.

Five patients with hepatorenal syndrome were treated with a LeVeen peritoneovenous shunt and furosemide. Four of the five patients responded immediately with urinary volumes in excess of 2 liters on the day of surgery and the subsequent 3 postoperative days. Urinary sodium excretion increased from 5 to 122 mEq/liter after insertion of a peritoneovenous shunt. Body weight decreased by 7.7 kg in 1 week. Serum creatinine decreased from 4.0 to 1.8 mg/dl in 1 week. There were two long-term survivors. Peritoneovenous shunting is an effective method of reversing the renal component in the hepatorenal syndrome.

Symptoms of severe nausea, vomiting, abdominal pain, and frequent bezoars, as well as objective gastric retention, can occur following Roux-Y biliary diversion for alkaline reflux gastritis. Medical therapy and prokinetic drugs have proven ineffective. This review evaluates 37 patients who underwent further gastric resection from 1979 to 1987 to improve gastric emptying and resolve symptoms. Fifteen patients underwent perioperative radionuclide solid-food gastric emptying studies. Seventy-three per cent (27 of 37 patients) of the patients who underwent further gastric resection (70% to 95%) had a satisfactory postoperative response. Twenty patients were graded Visick 1 or 2 and 7 Visick-3 patients, although much improved, still had some symptoms of gastroparesis. Twenty-seven per cent (10 of 37 patients) failed to improve and underwent completion total gastrectomy. Overall, 70% of this group had almost complete resolution of their symptoms. Three of 10 patients were considered "failures" due to postprandial pain in 1 and early vasomotor dumping in 2. Of the 10 patients who failed initial revisional surgery, 7 underwent a 70% to 80% subtotal gastric resection (STG) and 3 patients underwent 85% to 95% extensive resection (EXT.G.). Of the 15 patients who underwent perioperative radionuclide evaluation, a mean two-hour gastric retention of 61.4% +/- 4% (SEM) decreased to 25% +/- 4% following further gastric resection. Eight patients were in the STG group and seven patients were in the EXT.G group. Following STG, mean two-hour gastric retention of 58.2% +/- 3.5% decreased to 38% +/- 3% (p less than 0.05). In seven patients who underwent EXT.G, mean two-hour retention of 65% +/- 4% decreased to 10% +/- 2.5% (p less than 0.005). EXT.G resulted in normal gastric emptying and few late failures. In post-Roux-Y patients with symptoms of gastroparesis and documented gastric retention, EXT.G normalizes gastric emptying and restores a better quality of life. Total gastrectomy should be reserved for those patients who are failed by more extensive resection.

From 1973 to 1986, 22 patients underwent Roux-Y gastrojejunostomy for the early postgastrectomy dumping syndrome. In the early years, five patients underwent Roux-Y conversion with the addition of a 10 cm antiperistaltic jejunal segment interposed between the Roux-Y limb and the stomach. Within 4 years, all five patients had the jejunal segment removed due to severe symptoms of gastric retention. These patients underwent reconstruction to create Roux-Y limb only and joined the pool of 17 patients who underwent Roux-Y diversion only for the dumping syndrome. Overall, 19 of 22 patients (86 percent) had almost complete resolution of their dumping symptoms on long-term follow-up. Three patients showed no improvement, two with severe gastric retention and one with recurrent dumping symptoms. Overall, 5 of 22 patients (23 percent) had moderate to severe early and late postoperative gastric retention necessitating medical treatment in three and subsequent near-total gastrectomy in two. Although other procedures such as pyloric reconstruction or the addition of isoperistaltic or antiperistaltic jejunal interpositions have been reported to be equally successful in delaying gastric emptying and resolving dumping symptoms, we have preferred Roux-Y diversion for the treatment of combined alkaline reflux gastritis and dumping or the pure early vasomotor postgastrectomy dumping syndrome. As reported, we have abandoned the use of an antiperistaltic jejunal segment interposed between the stomach and the Roux-Y limb due to the high rate of postoperative gastric retention.

The effect of various prokinetic drugs was assessed in animals with Roux-en-Y gastrojejunostomy. The agents tested were (1) bethanechol 2.5 mg subcutaneously at 0 min and 30 min postprandially (pp); (2) metoclopramide 20 mg intravenous bolus at 0 min pp; (3) a combination of 1 and 2; (4) oxytocin 5 mg intramuscularly at 0 min and 240 min pp; (5) motilin at 100 ng/kg/hr; or (6) 300 ng/kg/hr continuous intravenous infusion from 0 to 270 min pp. Only bethanechol administration resulted in significantly less gastric retention (65 +/- 6% vs 32 +/- 5% retention at 5 hr). (P less than 0.002). The animal results with parenteral bethanechol were confirmed in humans with chronic delayed gastric emptying following Roux-en-Y gastrojejunostomy, with a decrease in gastric retention on radionuclide scan from 78.5 +/- 5% to 26 +/- 12% at 2 hr pp (P less than 0.01). Initially all patients responded with symptomatic improvement. However, subsequently 3/6 (50%) of patients required total or near total gastrectomy for recurrent symptoms of gastric stasis. Nevertheless, 2/6 (33%) of patients have no further evidence of gastric stasis, and a trial of bethanechol is recommended prior to considering further gastrectomy in patients with the Roux-stasis syndrome.

Recent reports cite an increased incidence in delayed gastric emptying following Roux-en-Y biliary diversion for alkaline reflux gastritis. The effect of Roux-en-Y diversion on the gastric emptying of liquids and solids was evaluated following vagotomy and antrectomy and vagotomy and subtotal gastrectomy. Twenty dogs underwent placements of large Thomas cannula in the stomach. Four dogs with intact stomachs served a controls. Eight dogs each with vagotomy and antrectomy were subdivided into Roux-en-Y gastrojejunostomy (RYA) and a Billroth II (B-IIA) group. Eight dogs each with vagotomy and subtotal gastrectomy were subdivided into similar groups. Four dogs - Roux-en-Y (RSTG) and four dogs - Billroth II (B-IISTG). Gastric emptying of solid food, normal saline and 25% dextrose was evaluated. RYA dogs demonstrated a significant delay in gastric emptying of solids compared with corresponding B-IIA animals. RYA dogs had 76, 61 and 42% of solid food retained at three, five and eight hours while B-II animals retained 56, 41 and 20%, respectively. The results are highly significant at all time intervals (p less than 0.001 at five and eight hours). Control animals retained 34, 17 and 3% of solid food at three, five and eight hours. RSTG animals had 73, 52 and 28% retained solids at three, five and eight hours, while B-IISTG animals had 55, 42 and 13% retention, respectively (p less than 0.05 at eight hours). Normal saline was significantly delayed in both Roux-en-Y subgroups compared with B-II dogs (p less than 0.02 in V/A, p less than 0.05 in V/STG). There was a trend toward delayed emptying of 25% dextrose in the Roux-en-Y groups, but significance was achieved only in the RYA compared with B-IIA groups (p less than 0.02 at 30 minutes). Delayed gastric emptying following Roux-en-Y gastrojejunostomy is documented in the experimental animals which underwent vagotomy and appears greater in magnitude than that observed following vagotomy and B-II gastrectomy. These data corroborate the clinical observations of severe delayed gastric emptying in patients undergoing Roux-en-Y diversions for alkaline gastritis.

Forty-two patients with acute cholangitis, as evidenced by fever (95%), jaundice (86%), and right upper quadrant pain (67%), were treated with fluid and electrolyte resuscitation, broad spectrum antibiotic coverage, and initial percutaneous transhepatic biliary drainage (PTD). Despite a 17% incidence of nondilated ductal systems, drainage was established in all patients using a 22-gauge "skinny" needle and "accordion" catheter. No attempt was made at definitive cholangiogram; only 1-2 mL of contrast were injected to confirm placement of the catheter. Sepsis began to resolve in all patients within 24 hours of PTD, after which definitive cholangiogram was performed. PTD was accompanied by a 7% (3/42) complication rate, none of which contributed to subsequent morbidity and mortality. Two patients in severe septic shock had PTD but died within 8 hours of admission, constituting a 5% mortality rate. Definitive therapy after resolution of sepsis included: surgical (16 patients), internal/external drainage (14 patients), balloon dilatation (10 patients), mono-octanoin infusion (1 patient), and ampullary dilatation (1 patient). The surgical morbidity rate was 18%. There was no mortality. PTD is effective in providing decompression as initial therapy for acute cholangitis with minimal morbidity. Accurate diagnosis provided by the definitive cholangiogram obviates the need for multiple surgical procedures. PTD provides a portal to the biliary tract for alternative procedures (i.e., internal/external drainage, balloon dilatation), especially in patients with medical contraindications to surgery.

Central venous catheter (CVC) vascular erosions are difficult to diagnose, and they cause serious complications. From 1985 to 1987, ten patients receiving the surgical services at the University of Florida suffered CVC vascular erosions. By chest roentgenogram, nine CVC tips were in the superior vena cava (SVC), although three catheter tips abutted the lateral wall of the SVC. One catheter tip was in the right atrium. All patients had sudden onset of symptoms, the most common of which was shortness of breath. Initial diagnosis was respiratory insufficiency in five patients, cardiac failure in three patients, pulmonary embolism in one, and sepsis in one. Four patients required intensive care. Two patients suffered pericardial tamponade, and pleural effusions developed in eight patients. One patient died of cardiac arrest. The average time interval from CVC placement to onset of symptoms was 60.2 hours, and from the onset of symptoms to the time of diagnosis, the interval was 16.7 hours. The mean volume obtained at thoracentesis was 1324 ml and at pericardiocentesis was 250 ml.

This is an analysis of 34 patients with carcinoma of the cervical esophagus treated with radiation therapy with curative intent at the University of Florida between September 1966 and May 1985. All patients have a minimum 2-year follow-up and 28 (82%) have at least 5 years of follow-up. Patients were staged according to the recommendations of the AJCC. Patients who died within 2 years of treatment with the primary site continuously disease-free were excluded from the local control analysis; all patients were included in the analysis of complications and survival. Irradiation resulted in control of the primary lesion in 1 of 2 patients who presented with T1 lesions, in 4 of the 12 patients with T2 lesions, and 3 of 17 patients who presented with T3 lesions. One patient with a T3 lesion that recurred locally was successfully salvaged by an operation. The 5-year absolute survival rates by stage were as follows: no patients with stage I lesions survived; of 11 stage II patients, one survived; and of 16 stage III patients, three survived. Interestingly, all four of the 5-year survivors were women.

Evaluation of gastric acid secretion and serum gastrin response to proximal or distal intestinal bypass and subsequent resection was undertaken. Ten dogs with gastric fistulas underwent control secretory studies consisting of basal acid output (BAO), maximal vagal stimulation (2 deoxy-d-glucose (2DG) 150 mg/kg iv), maximal histamine stimulation (160 μg/kg/hr iv), and test meal stimuli. Serum gastrin was measured basally and following gastric stimulation. Dogs were then divided into two bypass groups: Group I, distal bypass, leaving 40 cm jejunum and 4 cm ileum; and Group II, proximal bypass, leaving 4 cm jejunum and 40 cm ileum. Postoperatively, repeat studies were performed. Both groups then underwent resection of the bypassed segments followed by repeat secretory and gastrin tests. BAO was significantly increased in Group I dogs. BAO increased from a control of 0.04 ± 0.02 to 0.51 ± 0.19 meq H+/60 min (±SEM). No significant changes in BAO were observed in Group II following bypass. Acid secretion to maximal vagal or maximal histamine stimulation was not significantly different in either group. Serum gastrin response was unchanged in either group following a test meal. In Group I dogs, resection of the previously bypassed bowel produced a further significant increase in BAO (1.44 ± 0.4 meq H+/60 min). After resection of the bypassed bowel, there was no significant change in basal or stimulated gastrin. Gastric fistula BAO increased following 90% distal but not proximal bypass; subsequent resection of the bypassed segment produced a further increase in BAO. Acid secretory responses to both histamine and 2DG were not influenced by either bypass or resection. An increase in BAO without a change in the background of serum gastrin may suggest the presence of a gastric secretory inhibitor in the distal small bowel.

Limiting blood loss is a primary concern in hepatic resection. Torrential hemorrhage occasionally occurs and is often related to loss of control or injury to the hepatic veins or vena cava. Stapling devices, carefully applied, are presented as a safe and expedient means for controlling hepatic veins within the hepatic parenchyma.