[Endoscopic treatment of bleeding ulcers: has everything been said and done?]. / Tratamiento endoscópico de la úlcera hemorrágica: 'todo dicho, todo hecho?

Endoscopic treatment reduces bleeding recurrence, the need for surgery and mortality in patients with bleeding ulcers. However endoscopic treatment fails in 10-15% of patients, leading to high morbidity and mortality. The therapeutic measures with demonstrated effectiveness in reducing the risk of hemorrhagic recurrence and its complications are combined endoscopic treatment (adrenaline plus a second hemostatic intervention) and proton pump inhibitors. Also useful, although there is less evidence, are immediate resuscitation and <> endoscopy. Some studies suggest that activated recombinant factor VII infusion or supra-selective arterial embolization can be useful in severe hemorrhage. Further studies are required to determine optimal treatment according to the characteristics of each patient.

[Cecal cytomegalovirus infection following appendicectomy in a patient with ulcerative colitis]. / Infección cecal por citomegalovirus en paciente apendicectomizado con colitis ulcerosa.

We report a patient who, 3 months after being diagnosed with ulcerative colitis, was admitted to hospital because of malaise and right lower abdominal pain. An open appendectomy was performed. Histological study showed ulcerative colitis affecting the appendix. After surgery, the patient presented a refractory outbreak of ulcerative colitis requiring treatment with steroids and cyclosporin A. Despite this treatment, the patient continued to pass abundant fresh blood associated with severe anemia. Colonoscopy showed only granular and congestive cecal mucosa. Biopsies showed intracytoplasmic inclusion bodies with immunohistochemical stains positive for cytomegalovirus (CMV) infection. Rectorrhagia and anemia quickly disappeared after beginning treatment with ganciclovir. Appendicular ulcerative colitis is not uncommonly associated with distal colitis. In addition, diffuse CMV infection complicating ulcerative colitis treatment is not unusual. By contrast, isolated, segmentary infection by CMV in the proximal colon is extremely rare. Until now, only three patients with localized CMV infection have been described, and all three cases occurred in the context of ileoanal anastomosis.

Modified seven-day, quadruple therapy as a first line Helicobacter pylori treatment.

BACKGROUND: Cure rates of 7-day triple therapy seem to be decreasing. Quadruple therapies may be an alternative, although their complex administration makes patient acceptance difficult. OBJECTIVE: To test the usefulness of a thrice a day, quadruple therapy to cure Helicobacter pylori infection. PATIENTS AND METHODS: A total of 122 consecutive patients with peptic ulcer and Helicobacter pylori infection were treated with omeprazole 20 mg b.d., tetracycline chlorhydrate 500 mg t.d.s., metronidazole 500 mg t.d.s., and bismuth subcitrate 120 mg t.d.s. administered with meals for 7 days. Cure was tested by either endoscopy or breath test after 2 months, and by urea breath test 6 months after therapy. RESULTS: Seven patients were lost to follow-up. Of the remaining 115, 110 were cured at the first control, giving an intention-to-treat cure rate of 90.2% (95% CI: 83-95%) and a per protocol cure rate of 95.7% (95% CI: 90-98%). One hundred three patients returned for a 6-month breath test; all but one were cured. Side-effects were minimal or minor in 47 patients (40.8%) and moderate in four (3.4%). Compliance was good, 95% of patients taking more than 90% of the pills. Six (5%) patients stopped treatment after 1, 2, 4 (two patients) and 6 (two patients) days. CONCLUSION: Thrice a day quadruple therapy shows excellent cure rates, far above 90%, is well-tolerated and compliance is easy. Head-to-head comparison with triple therapies as first line Helicobacter pylori treatment seems warranted.

Evaluation of Helicobacter pylori diagnostic methods in patients with liver cirrhosis.

BACKGROUND: Helicobacter pylori-associated peptic ulcer is a frequent complication in cirrhotic patients and its morbidity rate is high. In spite of this, diagnostic methods for H. pylori infection have not been fully evaluated in these patients. AIM: To evaluate H. pylori diagnostic methods in patients with liver cirrhosis. METHODS: One hundred and one cirrhotic patients were included in the study. Three antral and two corpus biopsies were obtained for rapid urease test of the antral mucosa, and Giemsa stain and immunohistochemistry were performed for both the corpus and antrum. Serology, 13C-urea breath test and faecal H. pylori antigen determination were also carried out. RESULTS: Sixty-two patients were positive and 35 were negative for H. pylori infection; four were indeterminate. The sensitivity and specificity were 90.4% and 100%, respectively, for antral histology, 100% and 100% for gastric body histology, 90.4% and 100% for antral immunohistochemistry, 96.2% and 96.7% for body immunochemistry, 85.7% and 97% for rapid urease test, 83.6% and 55.9% for serology, 96.4% and 97.1% for 13C-urea breath test and 75.4% and 94.1% for faecal antigen. CONCLUSION: The most reliable tests for H. pylori infection in cirrhotic patients were the 13C-urea breath test and gastric body histology.

Limited usefulness of a seven-day twice-a-day quadruple therapy.

OBJECTIVE: To test the usefulness of a twice-a-day, simplified quadruple therapy to cure Helicobacter pylori infection. DESIGN: Helicobacter pylori-positive ulcer patients were treated with omeprazole 20 mg twice a day (b.d.), amoxicillin 1 g b.d., tinidazole 500 mg b.d. and bismuth subcitrate 240 mg b.d. for 7 days in an experimental, noncomparative pilot study. SETTING: The gastroenterology unit of a county hospital. PARTICIPANTS: Forty-four consecutive patients with peptic ulcer disease and H. pylori infection. MEASUREMENTS: Cure was tested by either endoscopy or breath test after 2 months, and by urea breath test 6 months after therapy. RESULTS: One patient was lost to follow-up. Of the remaining 43, 37 were cured at the first control, giving an intention-to-treat cure rate of 84.1% (95% CI 69-93%) and a per protocol cure rate of 86% (95% CI 71-94%). Thirty-three cured patients agreed to return for a six-month breath test. All but one were cured (long-term per protocol cure rate 82.1%; 95% CI 66-92%). CONCLUSIONS: This particular quadruple therapy is well tolerated and easy to comply with. However, cure rates did not reach 90%.

Efficacy of low and standard midazolam doses for gastroscopy. A randomized, double-blind study.

AIM: To evaluate the efficacy and safety of two different doses of intravenous midazolam (35 and 70 microg/kg) compared to placebo in patients undergoing gastroscopy. PATIENTS AND METHODS: Sixty patients scheduled for diagnostic gastroscopy were selected according to factors previously reported to affect tolerance (Eur J Gastroenterol Hepatol 1999; 11:201-204) and were randomly assigned to receive premedication with midazolam 35 microg/kg iv, midazolam 70 microg/kg iv or placebo iv. Oxygen saturation was continuously monitored during the procedure. Patients' tolerance, time to discharge and post-sedative inconvenience were evaluated using visual analogue scales and a questionnaire. RESULTS: Patients receiving either dose of midazolam showed better tolerance of gastroscopy than those receiving the placebo. Fewer patients receiving 70 or 35 microg/kg of midazolam were reluctant to undergo a further gastroscopy compared to those receiving the placebo (2, 1 and 9 patients respectively, P = 0.01). Compared to patients receiving midazolam 70 microg/kg, those receiving midazolam 35 microg/kg were discharged earlier (29.3+/-14.4 versus 43.1+/-12.4 min respectively, P < 0.001), experienced less post-sedative inconvenience (8 versus 15 patients slept for > 1 h at home respectively, P = 0.02), and suffered fewer clinically relevant desaturation episodes (< 90%) (0 versus 5 patients respectively, P = 0.04). CONCLUSIONS: Low doses of intravenous midazolam (35 microg/kg) are adequate and safe when sedation is indicated for gastroscopy.

Validation of phone interview for follow-up in clinical trials on dyspepsia: evaluation of the Glasgow Dyspepsia Severity Score and a Likert-scale symptoms test.

OBJECTIVE: To validate two widely used dyspepsia scores performed by phone interview. DESIGN: Spanish translations of the Glasgow Dyspepsia Severity Score and a Likert-scale symptomatic test were evaluated. Responsiveness to the treatment, validity of the tests, and reproducibility of tests performed by phone interview were assessed. SETTING: Gastroenterology and endoscopy unit of a county hospital. PARTICIPANTS: Group I consisted of 16 ulcer patients undergoing Helicobacter pylori eradication; Group II consisted of 29 healthy volunteers; and Group III consisted of 95 patients undergoing upper endoscopy. MEASUREMENTS: Glasgow Severity Dyspepsia Score and Likert test. RESULTS: Both tests showed adequate improvement (responsiveness) after H. pylori eradication. With regard to validity, the Glasgow and Likert test were significantly higher in 95 patients undergoing endoscopy than in 29 healthy controls. Analysis of reproducibility showed that intraobserver variation was low on both the Glasgow and Likert scores. No differences were found between consecutive tests regardless of whether both were performed by phone (24 patients) or one by phone and the other by clinical interview (40 patients). Interobserver variation analysis showed that the Glasgow test remained highly reproducible even when performed by different observers using different methods (clinical interview 8.83, phone 8.44, P = 0.12). By contrast, Likert-scale tests showed significant differences between observers for all symptoms except abdominal pain. CONCLUSIONS: (1) The Glasgow score is highly reproducible even when performed by different observers and using different methods. (2) By contrast, Likert tests show greater variability. To be reproducible in different conditions, they need to be performed by the same observer.

Identification of factors that influence tolerance of upper gastrointestinal endoscopy.

OBJECTIVE: Unsedated gastroscopy is unpleasant for some patients. The identification of factors related to tolerance would permit the selection of patients for sedation. The aim of the present study was to identify these factors. METHODS: Five hundred and nine patients underwent diagnostic gastroscopy after the administration of topical pharyngeal anaesthesia, without sedation. Patients were grouped as to whether they had undergone prior examinations or not. Tolerance was assessed with a visual analogue scale and a questionnaire. RESULTS: Two hundred and seventy-three (54%) patients underwent gastroscopy for the first time, and 236 (46%) patients had prior experience. Patient tolerance was poor in 84 of 273 (31%) patients undergoing gastroscopy for the first time, and in 61 of 236 (26%) patients with prior experience. Logistic regression analysis identified the following variables related to poor tolerance: (a) in patients undergoing gastroscopy for the first time: presence of gag reflex (odds ratio (OR) = 3.42, 95% confidence interval (CI) 1.90-6.17), apprehension (OR = 2.57, CI 1.33-4.95), young age (OR = 0.95, CI 0.93-0.98) and high level of anxiety (OR = 1.91, CI 0.96-3.89); (b) in patients with prior experience: apprehension (OR = 4.21, CI 1.93-9.20), poor tolerance of prior examinations (OR = 4.92, CI 1.93-12.5) and female (OR = 2.23, CI 1.09-4.57). CONCLUSIONS: The above-mentioned factors are predictive of poor tolerance, and may enable the identification of those patients who might benefit more from sedation for gastroscopy.

Diagnosis of Helicobacter pylori infection in dyspeptic patients by stool antigen detection usefulness of a new monoclonal enzyme immunoassay test.

BACKGROUND: Helicobacter pylori antigens can be measured in human stools with an enzyme immunoassay, which may prove to be a valuable non-invasive diagnostic tool. Aim. To evaluate the usefulness of a new monoclonal enzyme immunoassay for detecting H. pylori antigens in dyspeptic patients' faeces (FemtoLab H. pylori, Connex, Martinsried, Germany). PATIENTS: H. pylori infection was determined in 75 patients (49 men, 26 women, mean age 52 + 16.5) for histology and rapid urease test. METHODS: H. pylori status was established by concordance of the reference tests. FemtoLab H. pylori was measured in triplicate. In addition, two determinations of a polyclonal faecal antigen test (HpSA, Platinum Premier HpSA, Meridian Diagnostic Inc., Cincinnati, USA) were also performed. Sensitivity, specificity, positive and negative predictive values were calculated. Concordance between different measurements was estimated by Kappa statistics. RESULTS: The sensitivity of the FemtoLab H. pylori immunoassay ranged from 98 to 100% and its specificity was 76%. Positive and negative predictive values were 91 and 94-100%, respectively. Concordance coefficients ranged from 0.81 to 0.92. Corresponding HpSA values were 69, 86, 92 and 53%, respectively. Concordance coefficient was 0.61. CONCLUSIONS: FemtoLab H. pylori is a very sensitive, specific, highly reproducible and easy-to-perform tool for diagnosis of H. pylori infection.

Occupation and gastric cancer in Spain.

The association between occupational exposure and stomach cancer was investigated in a multicenter case-referent study conducted in Spain on 354 histologically confirmed cases and 354 hospital referents, matched by age, gender, and residence. An increased risk of gastric cancer was found for coal mining workers [odds ratio (OR) 11.8], but the number of workers was small, and therefore the 95% confidence interval (95% CI) was wide (95% CI 1.36-103). An increased risk was observed for wood and furniture workers (OR 1.76), construction workers (OR 1.68), and glass and ceramic workers (OR 2.18), but none of these risks were statistically significant. According to an occupation-exposure linkage system an increased risk was found for occupations associated with exposure to silica and mineral dust (OR 1.80, 95% CI 0.90-3.59). All of the OR estimates were adjusted for the confounding factors socioprofessional status and dietary habits. The possibility of a causal association between stomach cancer and coal and mineral dust is supported by the results.

[A descriptive epidemiological study of ulcerative colitis in a community hospital (1985-1989)]. / Estudio epidemiológico descriptivo de la colitis ulcerosa en un hospital comunitario (1985-1989).

BACKGROUND: The incidence of ulcerative colitis (UC) shows a marked geographical variation. It is high in Anglo-Saxon and Scandinavian countries, while in Mediterranean countries the available data on its epidemiological features are few. Spain is not an exception. METHODS: The new cases of UC diagnosed during the period 1985-1989 in the Hospital de Sabadell (a reference center of an area with a markedly industrial population of 343924) were retrospectively evaluated. The definitive diagnosis of UC required a negative fecal microbiological study and findings in colonoscopy and biopsy consistent with UC. RESULTS: During the study period 68 new cases of UC were diagnosed, representing a mean yearly incidence of 3.95/10(5) population, and 5.26/10(5) considering only the adult population (age greater than 14 years). No significant differences were found (p greater than 0.05) in the sex and age distributions, although the male/female ratio was 1.06. In 84% of cases UC was limited to the rectum and/or distal colon, with a low frequency of pancolonic forms (5%). In 70% of patients UC was classified as a low severity disease. CONCLUSIONS: Although the estimation of incidence only from hospital cases underestimates the true incidence, and also considering the limitation of comparing results of studies from several time periods, the incidence of UC in our area is the highest one reported to the present time in Spain and Southern Europe. Nevertheless, it is lower than that in most Northern European countries.

[Epidemiological study on the incidence of inflammatory bowel disease in 4 Spanish areas. Spanish Group on the Epidemiological Study of Inflammatory Bowel Disease]. / Estudio epidemiológico sobre la incidencia de enfermedad inflamatoria intestinal en cuatro áreas españolas. Grupo Español para el Estudio Epidemiológico de la Enfermedad Inflamatoria Intestinal.

BACKGROUND: The incidence of inflammatory bowel disease (IBD) shows marked geographical variations. The aim of this study was to determine and compare the incidence of IBD in four Spanish areas: Sabadell (Northeast), Vigo (Northwest), Mallorca island and Motril (South). PATIENTS AND METHODS: Prospective survey based on inception cohorts over a two-years period (1 October 1991 to September 1993). Subjects were the patients resident in these areas and diagnosed of IBD according to a standard protocol for case ascertainment and definition. RESULTS: Altogether 328 cases were identified, of whom 191 were diagnosed as ulcerative colitis (UC), 135 as Crohn's disease (CD) and 2 as indeterminate colitis. The overall adjusted incidence rate per 100,000 persons between ages 15-64 years of UC and CD were respectively 9.8 and 5.2 in Sabadell, 7.7 and 5.0 in Vigo, 7.8 and 5.8 in Mallorca and 4.3 and 6.5 in Motril. The Incidence rate ratio showed no significant differences for either conditions among these areas. The global adjusted incidence rate of UC in Spanish areas (8.0; IC 95%: 6.3-9.7) was significantly lower to that of Northern European countries while that of CD (5.5; IC 95%: 4.1-6.9) was between that of Northern and Southern Europe with no significant differences. CONCLUSIONS: The incidence of IBD did not show differences among the Spanish areas, and rates are between 2 and 6 times higher than those previously reported. The incidence of UC is significantly lower than that observed in the North of Europe, while for CD the incidence is between that of Northern and Southern Europe.

[Prognostic factors of early mortality in the 1st episode of hemorrhage caused by esophageal varices]. / Factores predictivos de mortalidad precoz en el primer episodio hemorrágico por varices esofágicas.

AIM: To assess the prognostic factors of early mortality in cirrhotic patients with the first variceal bleeding episode. PATIENTS AND METHODS: Fifty-five cirrhotic patients with endoscopic evidence of variceal bleeding treated with sclerotherapy during emergent endoscopy were included. RESULTS: Permanent hemostasis was obtained in 36 patients (65.5%). Fourteen (25.4%) patients died within six weeks of the bleeding episode. Twenty-four variables obtained at admission and in the following days were compared between patients who survived (n = 41), and died (n = 14). In the univariate analysis the following variables were related to early mortality: prothrombin ratio (p = 0.04), the presence of ascites (p = 0.004) and encephalopathy (p = 0.06), albumin (p = 0.01), Child-Pugh score (p = 0.0003), hemostasis during endoscopy (p = 0.002), absence of rebleeding at 24 hours (p = 0.01) and early rebleeding (within five days after the bleeding episode) (p = 0.006). Multiple logistic regression identified the Child-Pugh score (OR 11.86, CI 95% 2.54-55.48; p = 0.001) and early rebleeding (OR 6.27, CI 95% 1.29-30.44; p = 0.02) as prognostic independent factors of early mortality. CONCLUSIONS: The degree of hepatic failure and early rebleeding are prognostic independent factors of early mortality in cirrhotic patients after the first variceal bleeding episode.

[Evaluation of a rapid test (ImmunoCard STAT! HpSA) for Helicobacter pylori detection in stools]. / Evaluación de un test rápido (ImmunoCard STAT! HpSA) para la detección de Helicobacter pylori en heces.

INTRODUCTION: The use of a rapid diagnostic technique may be extremely useful for the management of Helicobacter pylori infection. A new immunochromatographic in-office test (ImmunoCard STAT! HpSA, Meridian Diagnosis Inc, Cincinnati, Ohio, USA) for the detection of H. pylori in feces has recently become available. The aim of the present study was to evaluate the diagnostic reliability and reproducibility of the ImmunoCard STAT! HpSA test in patients with dyspepsia. PATIENTS AND METHODS: Sixty-three dyspeptic patients were enrolled. H. pylori status was determined by CLO-test and Giemsa staining of antral biopsy. Patients with a positive result for both tests were considered infected and those with a negative result for both tests were considered not infected. Fecal H. pylori antigen was tested twice by ImmunoCard STAT! HpSA. The sensitivity, specificity, positive predictive value and negative predictive value of each determination were calculated. The concordance between the two determinations was evaluated using Kappa statistics. RESULTS: Of the 63 patients, 46 were infected by H. pylori. Sensitivity, specificity and positive and negative predictive values were 89-91%, 86-93%, 96-98% and 72-75%, respectively. The correlation coefficient between determinations was 0.845. CONCLUSION: The new ImmunoCard STAT! HpSA test shows good sensitivity and reproducibility. Therefore, it could be highly useful in the management of H. pylori infection.

[Endoscopic ligation with elastic bands in the prevention of hemorrhage recurrence caused by esophageal varices. Study of 45 patients]. / Ligadura endoscópica con bandas elásticas en la prevención de la recidiva hemorrágica por varices esofágicas. Estudio en 45 pacientes.

The aim of the present study was to determine the usefulness of elastic band ligation in the prevention of hemorrhage recurrence by esophageal varices. Forty-five patients without known hepatocarcinoma who had survived a hemorrhagic variceal episode were included in the study. Seventeen patients (38%) were Child-Pugh A, 22 (49%) B, and 6 (13%) C, with the hepatitis C virus and alcohol being the etiology of cirrosis in 55 and 20% of the cases, respectively. The first ligation session was performed between the third and fifth days after the hemorrhagic episode and the posterior sessions were carried out at intervals of 2-4 weeks. The ligation sessions were performed without antibiotic prophylaxis and with placement of an overtube. A mean of 4 +/- 2 bands were placed per session (range, 1-8) and the mean number of sessions required per patient to achieve erradication of the varices was 3.5 +/- 1.5 (range, 2-8). The rate of bleeding recurrence was 17.7% (9 episodes, five by variceal rupture and four by ulcer secondary to ligation). All the episodes of bleeding recurrence occurred between the sessions, with the mortality being 11% (5/45 patients). In the 40 remaining patients the varices were erradicated although 19 (47.5%) required one or two additional sessions of sclerotherapy. The accumulated percentage of patients free of bleeding recurrence was 82% during a mean follow-up of 10.2 +/- 6.7 months. Ten lesions of dislaceration of the esophageal mucosa caused by placement of the were observed overtube. In conclusion, endoscopic elastic band ligation is a useful technique for the erradication of esophageal varices an in the prevention of bleeding recurrence.

[Sedation in digestive endoscopy. Results of a hospital survey in Catalonia (Spain)]. / Sedación en la endoscopia digestiva. Resultados de una encuesta hospitalaria en Cataluña.

INTRODUCTION: The need for sedation is increasing in digestive endoscopy units (DEU). There are no data on the use of sedation in DEU in Catalonia (Spain). OBJECTIVE: To evaluate the use of sedation in DEU in Catalonia. MATERIAL AND METHOD: A questionnaire on the practice of sedation was designed and sent to the heads of medical and nursing staff of the DEU of 63 public and private hospitals in Catalonia. Two mailings were sent with an interval of three months between each. The questionnaire included 62 items on the characteristics of the hospital and the DEU, number of explorations, frequency of sedation use, drugs employed, participation of an anesthesiologist, use of monitoring, and complications. RESULTS: Forty-four DEU (70%) corresponding to 31 public hospitals and 13 private hospitals completed the questionnaire. Evaluation of sedation patterns was based on 105,904 explorations performed in the various DEU (56,453 gastroscopies, 47,278 colonoscopies and 2,173 endoscopic retrograde cholangiopancreatographies (ERCP) in 2001. Sedation, sedation-analgesia or anesthesia was used in 17% of gastroscopies, 61% of colonoscopies and 100% of ERCP. Sedation was administered by an anesthesiologist in 7% of gastroscopies, 25% of colonoscopies and 38% of ERCP. Anesthesiologist administration was more frequent in private than in public centers (gastroscopies: 25% vs. 2%; colonoscopies: 57% vs. 9%, p < 0.001). No deaths associated with the use of sedation were reported. Eighty-nine percent of the DEU complied with standard recommendations for the practice of sedation. CONCLUSIONS: In Catalonia, the use of sedation is highly variable, depending on the endoscopic procedure and the DEU. Use of sedation in infrequent in gastroscopy, fairly widespread in colonoscopy and routine in ERCP. Anesthesiologist administration is significantly more frequent in private hospitals. Most DEU follow standard sedation practices.

Meta-analysis: predictors of rebleeding after endoscopic treatment for bleeding peptic ulcer.

BACKGROUND: Determining the risk of rebleeding after endoscopic therapy for peptic ulcer bleeding (PUB) may be useful for establishing additional haemostatic measures in very high-risk patients. AIM: To identify predictors of rebleeding after endoscopic therapy. METHODS: Bibliographic database searches were performed to identify studies assessing rebleeding after endoscopic therapy for PUB. All searches and data abstraction were performed in duplicate. A parameter was considered to be an independent predictor of rebleeding when it was detected as prognostic by multivariate analyses in ≥2 studies. Pooled odds ratios (pOR) were calculated for prognostic variables. RESULTS: Fourteen studies met the prespecified inclusion criteria. Pre-endoscopic predictors of rebleeding were: (i) Haemodynamic instability: significant in 9 of 13 studies evaluating the variable (pOR: 3.30, 95% CI: 2.57-4.24); (ii) Haemoglobin value: significant in 2 of 10 (pOR: 1.73, 95% CI: 1.14-2.62) and (iii) Transfusion: significant in two of six (pOR not calculable). Endoscopic predictors of rebleeding were: (i) Active bleeding: significant in 6 of 12 studies (pOR: 1.70, 95% CI: 1.31-2.22); (ii) Large ulcer size: significant in 8 of 12 studies (pOR: 2.81, 95% CI: 1.98-4.00); (iii) Posterior duodenal ulcer location: significant in four of eight studies (pOR: 3.83, 95% CI: 1.38-10.66) and (iv) High lesser gastric curvature ulcer location: significant in three of eight studies (pOR: 2.86; 95% CI: 1.69-4.86). CONCLUSIONS: Major predictors for rebleeding in patients receiving endoscopic therapy are haemodynamic instability, active bleeding at endoscopy, large ulcer size, ulcer location, haemoglobin value and the need for transfusion. These risk factors may be useful for guiding clinical management in patients with PUB.