Effects of 6 months supervised physical training on muscle strength and aerobic capacity in patients undergoing Roux-en-Y gastric bypass surgery: a randomized controlled trial.

Obesity and physical inactivity are major health problems. Roux-en-Y gastric bypass (RYGB) surgery results in significant weight loss and reduces obesity-related morbidity and mortality. Physical activity lowers the risk of cardiovascular disease and premature death. The aims of this study were to elucidate the effects of RYGB followed by 6 months of supervised physical training on physical capacity. In a randomized controlled trial, 60 participants eligible for RYGB were randomized 6 months post-surgery to either two weekly physical training sessions for 26 weeks (INT) or a control group (CON). Aerobic capacity (VO2 max), muscle strength (MS) of the shoulder and hip and physical function were measured pre-surgery and 6, 12 and 24 months post-surgery. RYGB per se decreased MS in all tested muscle groups, had no effects on VO2 max but improved physical function. After the intervention, INT had a significant 0.33 L min-1 increase in VO2 max compared to CON (95% CI: 0.07-0.57, P = 0.013). Furthermore, MS in the hip adductor increased significantly with 13 N (95% CI: 3.6-22.4, P = 0.007) and a between-group difference was found in the Stair Climb Test (0.46 repetitions [95% CI: 0.02-0.91, P = 0.042]). The effects were not maintained at follow-up. Supervised physical training following RYGB improved VO2 max, hip MS and physical function, but the positive effects were not maintained at follow-up. While activities of daily life may become easier as a result of RYGB, the observed extensive post-operative loss of MS requires more attention to increase the patient's physical capacity prospectively.

Systematic review of risk prediction models for diabetes after bariatric surgery.

BACKGROUND: Diabetes remission is an important outcome after bariatric surgery. The purpose of this study was to identify risk prediction models of diabetes remission after bariatric surgery. METHODS: A systematic literature review was performed in MEDLINE, MEDLINE-In-Process, Embase and the Cochrane Central Register of Controlled Trials databases in April 2015. All English-language full-text published derivation and validation studies for risk prediction models on diabetic outcomes after bariatric surgery were included. Data extraction included population, outcomes, variables, intervention, model discrimination and calibration. RESULTS: Of 2330 studies retrieved, eight met the inclusion criteria. Of these, six presented development of risk prediction models and two reported validation of existing models. All included models were developed to predict diabetes remission. Internal validation using tenfold validation was reported for one model. Two models (ABCD score and DiaRem score) had external validation using independent patient cohorts with diabetes remission assessed at 12 and 14 months respectively. Of the 11 cohorts included in the eight studies, calibration was not reported in any cohort, and discrimination was reported in two. CONCLUSION: A variety of models are available for predicting risk of diabetes following bariatric surgery, but only two have undergone external validation.

Esophagogastric junction distensibility in hiatus hernia.

Hiatus hernia is known to be an important risk factor for developing gastroesophageal reflux disease. We aimed to use the endoscopic functional lumen imaging probe (EndoFLIP) to evaluate the functional properties of the esophagogastric junction. EndoFLIP assessments were made in 30 patients with hiatus hernia and Barrett's esophagus, and in 14 healthy controls. The EndoFLIP was placed straddling the esophagogastric junction and the bag distended stepwise to 50 mL. Cross-sectional areas of the bag and intra-bag pressures were recorded continuously. Measurements were made in the separate sphincter components and hiatus hernia cavity. EndoFLIP measured functional aspects such as sphincter distensibility and pressure of all esophagogastric junction components and visualized all hiatus hernia present at endoscopy. The lower esophageal sphincter in hiatus hernia patients had a lower pressure (e.g. 47.7 ± 13.0 vs. 61.4 ± 19.2 mm Hg at 50-mL distension volume) and was more distensible (all P < 0.001) than the common esophagogastric junction in controls. In hiatus hernia patients, the crural diaphragm had a lower pressure (e.g. 29.6 ± 10.1 vs. 47.7 ± 13.0 mm Hg at 50-mL distension volume) and was more distensible (all P < 0.001) than the lower esophageal sphincter. There was a significant association between symptom scores in patients and EndoFLIP assessment. Conclusively, EndoFLIP was a useful tool. To evaluate the presence of a hiatus hernia and to measure the functional properties of the esophagogastric junction. Furthermore, EndoFLIP distinguished the separate esophagogastric junction components in hiatus hernia patients, and may help us understand the biomechanics of the esophagogastric junction and the mechanisms behind hiatal herniation.

Module based training improves and sustains surgical skills: a randomised controlled trial.

PURPOSE: Traditional surgical training is challenged by factors such as patient safety issues, economic considerations and lack of exposure to surgical procedures due to short working hours. A module-based clinical training model promotes rapidly acquired and persistent surgical skills. METHODS: A randomised controlled trial concerning supervised hernia repair in eight training hospitals in Denmark was performed. The participants were 18 registrars [Post graduate year (PGY) 3 or more] in their first year of surgical specialist training. The intervention consisted of different modules with a skills-lab course followed by 20 supervised Lichtenstein hernia repairs. Operative performance was video recorded and blindly rated by two consultants using a previously validated skills rating scale (8-40 points). Outcome measures were change in the ratings of operative skills and operative time. RESULTS: In the intervention group (n = 10) the average rating of operative skills before intervention was 22.5 (20.6-24.3) and after 26.2 (23.5-28.8), p = 0.044. At follow-up after 1 year, rating was 26.9 (23.4-30.4), p = 0.019. In the conventionally trained group average rating was 23.4 (19.4-27.3) at start and 21.7 (17.3-26.1) at end, p = 0.51. At start no difference was detected between the two groups, p = 0.59; by 1 year the difference was statistically significant favouring intervention, p = 0.044. Operative time showed similar results in favour of the intervention. CONCLUSIONS: A module-based training model in Lichtenstein hernia repair was preferable in both short and long-term compared with standard clinical training. The model will probably be applicable to other surgical training procedures.

Proton pump inhibitor use may not prevent high-grade dysplasia and oesophageal adenocarcinoma in Barrett's oesophagus: a nationwide study of 9883 patients.

BACKGROUND: Proton pump inhibitors (PPI) may potentially modify and decrease the risk for development of oesophageal adenocarcinoma in Barrett's oesophagus (BO). AIM: To investigate if the intensity and adherence of PPI use among all patients with BO in Denmark affected the risk of oesophageal adenocarcinoma. METHODS: We performed a nationwide case-control study in Denmark among 9883 patients with a new diagnosis of BO. All incident oesophageal adenocarcinomas and high-grade dysplasias were identified, and risk ratios were estimated on the basis of prior use of PPIs. Sex- and age-matched BO patients without dysplasia or malignancies in a 10:1 ratio were used for comparison. Conditional logistic regression was used for analysis, adjusting for low-grade dysplasia, gender and medication. RESULTS: We identified 140 cases with incident oesophageal adenocarcinomas and/or high-grade dysplasia, with a median follow-up time of 10.2 years. The relative risk of oesophageal adenocarcinoma or high-grade dysplasia was 2.2 (0.7-6.7) and 3.4 (95% CI: 1.1-10.5) in long-term low- and high-adherence PPI users respectively. CONCLUSIONS: No cancer-protective effects from PPI's were seen. In fact, high-adherence and long-term use of PPI were associated with a significantly increased risk of adenocarcinoma or high-grade dysplasia. This could partly be due to confounding by indication or a true negative effect from PPIs. Until the results from future studies hopefully can elucidate the association further, continuous PPI therapy should be directed at symptom control and additional modalities considered as aid or replacement.

Reliable and valid assessment of Lichtenstein hernia repair skills.

PURPOSE: Lichtenstein hernia repair is a common surgical procedure and one of the first procedures performed by a surgical trainee. However, formal assessment tools developed for this procedure are few and sparsely validated. The aim of this study was to determine the reliability and validity of an assessment tool designed to measure surgical skills in Lichtenstein hernia repair. METHODS: Key issues were identified through a focus group interview. On this basis, an assessment tool with eight items was designed. Ten surgeons and surgical trainees were video recorded while performing Lichtenstein hernia repair, (four experts, three intermediates, and three novices). The videos were blindly and individually assessed by three raters (surgical consultants) using the assessment tool. Based on these assessments, validity and reliability were explored. RESULTS: The internal consistency of the items was high (Cronbach's alpha = 0.97). The inter-rater reliability was very good with an intra-class correlation coefficient (ICC) = 0.93. Generalizability analysis showed a coefficient above 0.8 even with one rater. The coefficient improved to 0.92 if three raters were used. One-way analysis of variance found a significant difference between the three groups which indicates construct validity, p < 0.001. CONCLUSIONS: Lichtenstein hernia repair skills can be assessed blindly by a single rater in a reliable and valid fashion with the new procedure-specific assessment tool. We recommend this tool for future assessment of trainees performing Lichtenstein hernia repair to ensure that the objectives of competency-based surgical training are met.

Esophageal distension parameters as potential biomarkers of impaired gastrointestinal function in diabetes patients.

BACKGROUND: Gastrointestinal (GI) symptoms, such as nausea, vomiting, bloating, postprandial fullness, and abdominal pain, are frequent in patients with diabetes mellitus (DM). The pathogenesis is complex and multi-factorial. To determine easy accessible and valid biomarkers for disordered GI function in DM patients, we aimed to study esophageal mechanical parameters and their relation to symptoms typically arising from the digestive tract. METHODS: Seventeen patients with longstanding DM and GI symptoms and 13 healthy controls were studied using ultrasound monitored esophageal distension. The sensory response was recorded and their symptoms registered. Biomechanical parameters, such as compliance and stiffness were computed from luminal diameters during distension based on the ultrasound images and from pressure data. Biomechanical and sensory parameters were correlated with the clinical data. KEY RESULTS: Diabetes patients had reduced esophageal sensitivity compared with controls (P = 0.046). The esophageal compliance was reduced (P = 0.004) and the esophageal stiffness was increased (P = 0.004) in the diabetes patients. Among patients, both postprandial fullness/early satiety and bloating correlated negatively to the esophageal compliance parameters (all P < 0.05). CONCLUSIONS & INFERENCES: Patients with long-standing DM and GI symptoms had reduced esophageal sensitivity together with reduced compliance and increased stiffness, which were correlated to the patients' GI symptoms. Biomechanical parameters obtained during distension may serve as biomarker for similar pathophysiologic effects of diabetes in the stomach and small bowel. They may contribute to our understanding of the pathophysiology underlying GI dysfunction and symptoms in patients with longstanding DM.

Randomised clinical trial: the efficacy of a transient receptor potential vanilloid 1 antagonist AZD1386 in human oesophageal pain.

BACKGROUND: Many patients with gastro-oesophageal reflux disease (GERD) are hypersensitive to heat and acid and may respond insufficiently to standard treatment. Antagonists of the heat and acid receptor 'transient receptor potential vanilloid 1'(TRPV1) are a potential drug class for GERD treatment. AIM: To investigate the effect of a TRPV1 antagonist (AZD1386) on experimentally induced oesophageal pain. METHODS: Twenty-two healthy men (20-31 years) participated in this randomised, placebo-controlled, double-blinded, crossover study examining the effects of a single-dose oral AZD1386 (30 and 95 mg). Subjects were block-randomised. On treatment days, participants were stimulated with painful heat, distension, electrical current and acid in the oesophagus. Heat and pressure pain on the forearm were somatic control stimuli. DATA ANALYSIS: intention-to-treat. RESULTS: A total of 21 participants completed the protocol and 1 voluntarily discontinued. In the oesophagus, both 30 and 95 mg of AZD1386 increased pain thresholds to heat stimuli 23% [95% confidence interval (CI): 10-38%] and 28%, respectively (CI: 14-43%). The skin heat tolerance was increased 2.1 °C (CI: 1.1-3.2 °C) after 30 mg AZD1386 and 4.0 °C (CI: 3.0-5.0 °C) after 95 mg. Heat analgesia persisted for 2.5 h. Pain thresholds to the other stimuli were unaffected by AZD1386. 50% reported 'feeling cold' and body temperature increased in all subjects exposed to 30 and 95 mg AZD1386 (mean increase 0.4±0.3 °C and 0.7±0.3 °C, respectively, P<0.05). CONCLUSIONS: AZD1386 increased oesophageal and skin heat pain thresholds and had a safe adverse-event profile. This drug class may have a potential for treatment of GERD.

Partial liver ischemia is followed by metabolic changes in the normally perfused part of the liver during reperfusion.

BACKGROUND: Temporary vascular in- and outflow occlusion is an effective technique for bleeding control during liver resection. However, occlusion can result in ischemia/reperfusion (I/R) injury to the liver. The aim of this study in a porcine model was to investigate the effect of in- and outflow occlusion of part of the liver on the metabolism of the normally perfused parenchyma of the same liver measured by microdialysis. METHODS: Eight pigs underwent laparotomy. A microdialysis catheter was inserted into in the left and right part of the liver, respectively. Microdialysis samples were collected every 30 min. Occlusion of the left part of the liver was achieved for 60 min, followed by 5 h of reperfusion. Samples were analyzed for glucose, lactate, pyruvate and glycerol. Blood samples were drawn to determine standard liver and biochemical parameters. RESULTS: Comparing the ischemic part of the liver with the normally perfused part, significant differences in the levels of lactate, pyruvate and glycerol were found. During reperfusion, similar and continuous decreases below baseline levels were observed for lactate and pyruvate in both the ischemic and normally perfused part of the liver. No significant changes in liver parameters or blood glucose levels were seen. CONCLUSIONS: Partial ischemia of the liver is without effects on metabolism in the normally perfused part. Metabolic changes in the ischemic part of the liver were reversible. However, partial liver ischemia was followed by similar continuous decreases in lactate and pyruvate levels in the whole liver, even though the ischemic insult was not detectable in transaminase levels.

A new distensibility technique to measure sphincter of Oddi function.

BACKGROUND: Evaluation of the biliary tract is important in physiological, pathophysiological, and clinical studies. Although the sphincter of Oddi (SO) can be evaluated with manometry, this technique has several limitations. This may explain the difficulties in identifying pathophysiological mechanisms for dysfunction of the SO and in identifying patients who may benefit from certain therapies. To encompass problems with manometry, methods such as the functional lumen imaging probe (FLIP) technique have been developed to study GI sphincters. This study set about miniaturising the FLIP probe and validating it for measurements in the SO. In order to get a better physiological understanding of the SO the aims were to show the sphincter profile in vivo and motility patterns of SO in pilot studies using volunteers that were experiencing biliary type pain but had normal SO manometry. METHODS: The SO probe was constructed to measure eight cross-sectional areas (CSA) along a length of 25 mm inside a saline-filled bag. To validate the technique for profiling the SO, six perspex cylinders with different CSAs were measured nine times to assess reproducibility and accuracy. KEY RESULTS: Reproducibility and accuracy for these measurements were good. The probe performed well in bench tests and was therefore tested in four humans. The data indicated that it was possible to make distensions in the human SO and that a geometric sphincter profile could be obtained. CONCLUSIONS & INFERENCES: The probe will in future studies be tested for diagnostic purposes related to sphincter of Oddi diseases.

An endoscopic method for thermal and chemical stimulation of the human oesophagus.

Previous methods for visceral thermal stimulation have lacked control of the temperature rate and visual inspection of the organ. The aims of this study was to develop a method for linear control of heat stimulation in the human oesophagus combined with endoscopy, to assess the reproducibility of this method and to investigate sensitivity to thermal stimulation of the distal oesophagus before and after acid perfusion. A probe with a 2.8 mm endoscope inside was constructed permitting heat and chemical stimulation. Three different temperature ramps were applied in the distal oesophagus in 12 healthy subjects by recirculation of heated water in a bag. Endoscopy of the oesophageal mucosa was performed prior to experimental stimulation. The temperature, the time of stimulation and the area under the temperature curve (AUC) were measured at the pain detection threshold. Thermal stimulation was repeated after perfusion of the oesophagus with acid. The method was tested on two subsequent days to assess reproducibility. All subjects had a normal endoscopic examination. Day-to-day reproducibility was good for the three temperature ramps (intra-class correlations >0.6). The subjects tolerated less heat stimulation, a decrease in AUC (P = 0.0003), a decrease in time to pain detection threshold (P = 0.005) and decreased temperature at pain detection threshold (P = 0.0001) after acid perfusion. The slow ramp was the most sensitive, showing a decrease in AUC of 29%. The present method was easily implemented and showed good reproducibility. It can potentially be used in basic experiments, drug and clinical studies as it provides a controllable thermal stimulus.

Intraductal papillary mucinous tumor of the pancreas (IPMT): follow-up of twelve cases.

INTRODUCTION: Intraductal papillary mucinous tumours (IPMT) were described as a distinct entity in I982. The extent of surgical resection for this disease remains controversial. METHODS: Twelve patients with a diagnosis of IPMT were included in the present retrospective study. RESULTS: Ten out of twelve patients had symptoms suggesting chronic pancreatitis. Two patients were not operated on due to biopsy-verified metastases in the liver. Nine patients were treated with a total pancreatectomy and one with a pancreaticoduodenectomy. In the ten patients operated on for IPMT, histological examination showed eight non-invasive- and two invasive carcinomas. In six cases, multifocal extensive intraductal changes were found, affecting either most of or the whole pancreas. There was no perioperative mortality. Six patients were alive at follow-up without recurrence and four patients were dead, two of them with recurrence. CONCLUSION: IPMTs represent a subgroup of pancreatic neoplasms with a favourable prognosis, and the resection should aim at removing all dysplastic foci. In cases with diffuse dilatation of the main pancreatic duct, widespread tumour involvement of the duct system can be expected and total pancreatectomy should be the operation of choice.

Cytokine changes during warm ischemia and reperfusion of the pig liver with or without preconditioning.

BACKGROUND: Hepatic inflow occlusion results in ischemia-reperfusion injury. The aim of the present porcine study was to investigate whether the pro- and anti-inflammatory cytokine response is involved in mediating the protective effect of ischemic preconditioning (IPC) during, and after warm liver ischemia. METHODS: Fifteen randomized pigs--7 non-IPC and 8 (IPC)--underwent laparotomy followed by 60 min of total ischemia with or without IPC continued by 3 h of reperfusion. Plasma cytokines (IL-6, IL-8, IL-10, and TNF-alpha) were measured during the study period as well as liver parameters (alanine-aminotransferase, alkaline phosphatase, bilirubin, and prothrombin time). RESULTS: In the IPC group, IL-6 increased significantly during reperfusion compared to baseline and the non-IPC group. TNF-alpha increased nonsignificantly in the non-IPC group, while the levels remained stable in the IPC group. IL-8 and IL-10 increased in both groups after reperfusion. Only minor differences were observed in liver parameters. CONCLUSIONS: Warm liver ischemia with or without IPC activates inflammatory cytokines. IL-6 increased significantly in the IPC group compared to the non-IPC group, while the opposite was observed for TNF-alpha. These cytokine changes may be involved in the hepatoprotective mechanism induced by IPC.

Pain in chronic pancreatitis: the role of neuropathic pain mechanisms.

Pain mechanisms in patients with chronic pancreatitis are incompletely understood and probably multifactorial. Recently, evidence from experimental human pain research has indicated that in many of these patients pain processing in the central nervous system is abnormal and mimics that seen in neuropathic pain disorders. The current review focuses on several lines of evidence supporting this hypothesis. Hence, the spontaneous and postprandial pain in chronic pancreatitis may reflect the characteristic pain features seen in patients with neuropathic pain. Biochemical and histopathological findings in tissues from patients with chronic pancreatitis are similar to those observed in patients with other nerve fibre lesions. Experimental studies have shown that patients with chronic pancreatitis show signs of spinal hyper-excitability counter-balanced by segmental and descending inhibition. Changes in the brain with cortical reorganisation to gut stimulation and increased activity in specific electroencephalographic features characteristic for neuropathic pain are also seen in patients with chronic pancreatitis. Finally, principles involved in the treatment of pancreatic pain have many similarities with those recommended in neuropathic pain disorders. In conclusion, a mechanism-based understanding of pain in chronic pancreatitis may have important implications for the treatment.

Endoscopic gallbladder drainage of patients with acute cholecystitis.

BACKGROUND AND STUDY AIM: The standard treatment for acute cholecystitis is early laparoscopic cholecystectomy. In cases of increased operative risk surgery may be postponed or rejected, and instead alternative methods, such as percutaneous or endoscopic drainage, may be attempted. This paper is a retrospective assessment of our results with endoscopic gallbladder drainage (EGBD) by means of endoscopic retrograde cholangiopancreatography (ERCP). PATIENTS AND METHODS: Over a 9-year period, data from all patients who underwent attempted EGBD, primarily referred for diagnosis and treatment of cholestasis by ERCP, were analyzed. EGBD was attempted in those patients who had concomitant acute cholecystitis. RESULTS: EGBD was successful in 24 of 34 patients with acute cholecystitis (70.6%). The success rate rose from 50% during the first 4-year period to 89% during the subsequent 5-year period. Of the 24 patients in whom EBGD had been successful, 21 showed clinical improvement and 10 were finally managed nonoperatively. Of the 21 clinically improved patients, 14 underwent elective surgery a median of 24 days later. Of the 10 patients in whom EGBD failed nine underwent surgery four of whom required surgery within one week. CONCLUSION: Treatment of acute cholecystitis in patients could be done successfully by EGBD, and in 70% of cases cholecystectomy could then be carried out on an elective basis rather than as emergency surgery. Some patients could be treated with EGBD alone. Future prospective trials will clarify the role of EGBD in patients with acute cholecystitis.

Risk of peptic ulcer, oesophagitis, pancreatitis or gallstone in patients with unexplained chest/epigastric pain and normal upper endoscopy: a 10-year Danish cohort study.

BACKGROUND: No studies have examined the risk of upper gastrointestinal diseases among patients with unexplained chest/epigastric pain (UCEP) and a normal upper endoscopy. AIM: To examine the relative risk of peptic ulcer, oesophagitis, pancreatitis or gallstone in UCEP patients. METHODS: This Danish 10-year cohort study focused on UCEP patients (n = 386), diagnosed in 1992-93. Ten age- and gender-matched controls were selected per patient from Denmark's Civil Registration System (n = 3860). Kaplan-Meier analysis and Cox's regression analysis was used to calculate the risk of hospitalization for peptic ulcer, oesophagitis, pancreatitis or gallstone. RESULTS: Compared with controls, the adjusted relative risks among UCEP patients <1 and > or = 1 year after upper endoscopy were for peptic ulcer 2.0 [95% confidence interval (CI) 0.2-18.4] and 1.7 (95% CI 0.9-3.4), for oesophagitis 8.2 (95% CI 1.2-59.2) and 1.9 (95% CI 0.7-5.0), for pancreatitis 9.2 (95% CI 2.0-41.8) and 3.9 (95% CI 1.4-10.5), and for gallstone 14.1 (95% CI 5.4-37.2) and 3.3 (95% CI 1.9-5.8). CONCLUSIONS: UCEP is positively associated with all study outcomes especially in the first year after upper endoscopy, indicating that peptic ulcer, oesophagitis, pancreatitis or gallstone could be underlying early UCEP symptoms. However, the long-term association remained strong for pancreatitis and gallstone, suggesting a genuine excess risk.

Cortical changes to experimental sensitization of the human esophagus.

Topographical organization in the neocortex shows experience-dependent plasticity. We hypothesized that experimental sensitization of the esophagus results in changes of the topographical distribution of the evoked potentials and the corresponding dipole source activities to painful stimulation. An endoscopic method was used to deliver 35 electrical stimuli at the pain threshold to a fixed area of the mucosa in 10 healthy volunteer men and women. The stimulations were repeated after 30 min (reproducibility experiment), and after 60 min following perfusion of 200 ml 0.1 N hydrochloric acid (sensitization experiment). During stimulation the electroencephalogram was recorded from 64 surface electrodes. The sensitization resulted in a decrease in the pain threshold (F=6.2; P=0.004). The topographic distribution of the evoked potentials showed reproducible negative (N1, N2) and positive (P1, P2) components. After acid perfusion a reduced latency and a change in localization was seen for the P1 subdivided into frontal and occipital components (F=29.5, P<0.001; F=53.7, P<0.001). Furthermore the sensitization resulted in a reduction of the latency for P2 (F=6.2, P=0.009). The source analysis showed consistent dipolar activity in the bilateral opercular-insular cortex before and after acid perfusion. For the anterior cingulate dipole there was a reduction in latency (P=0.03) and a posterior shift (P=0.0002) following acid perfusion. The findings indicate that short-term sensitization of the esophagus results in central neuroplastic changes involving the cingulate gyrus, which also showed pathological activation in functional diseases of the gut, thus reflecting the importance of this region in visceral pain and hyperalgesia.

Transcatheter arterial embolization in patients with bleeding duodenal ulcer: an alternative to surgery.

PURPOSE: To evaluate the efficacy and safety of transcatheter arterial embolization (TAE) in patients with bleeding/rebleeding duodenal ulcers. MATERIAL AND METHODS: Over a 6-year-period, 40 consecutive patients with bleeding/ rebleeding after endoscopic therapy and/or surgery for duodenal ulcer were included in the study. Superselective angiographic catheterization and coil embolization were performed by the same interventional radiologist. RESULTS: Lasting hemostasis was achieved in 26 of 40 patients (65%). Transfusion requirement was reduced from median 14 (range 3-35) units of blood before TAE to 2 (range 0-53) units after TAE. Ten patients died, five because of continuous bleeding. No adverse effects as a result of TAE were seen. CONCLUSION: TAE is an effective and safe treatment in a significant proportion of patients with bleeding duodenal/rebleeding ulcers after therapeutic endoscopy and/or surgery.