Peptic Upper Gastrointestinal Bleeding: Diagnosis and Treatment. A Monocentric Experience on a 5 Years Period

Transcription

1 ORIGINAL ARTICLES Peptic Upper Gastrointestinal Bleeding: Diagnosis and Treatment. A Monocentric Experience on a 5 Years Period I. SPOREA 1, DANIELA LAZĂR 1, ALINA POPESCU 1, ROXANA ŞIRLI 1, A. GOLDIŞ 1, M. STRĂIN 1, ADRIANA TUDORA 1, MIRELA DĂNILĂ 1, V. DĂNILĂ 1, CORINA VERNIC 2 1 Department of Gastroenterology and Hepatology, 2 Department of Medical Biostatistics, University of Medicine and Pharmacy Timişoara, Romania Aim. We evaluated all the cases of upper gastrointestinal hemorrhages admitted in the Department of Endoscopy of our Clinic during a 5 years period. Material and method. 810 patients were included in our study, 64.3% males and 36.7% females, mean age 58.7±15.2 years (17 96 years). The main cause of the upper digestive hemorrhage was peptic ulcer (82.6%), equally divided in gastric and duodenal. All the ulcers were assessed according to the Forrest classification. Results. Endoscopic haemostasis was performed in 40% of all cases, only epinephrine injection (1/10,000) in 44.4% of cases; combined therapy (injection + clip or thermocoagulation) in 40.3% of the cases and clip or thermocoagulation alone in 15.3% of the cases. A marked reduction of haemostasis using epinephrine injection alone (as monotherapy) was observed during the period of study. Postendoscopic treatment rebleeding occurred in 19.8% of cases; 3.6% of the patients have had a fatal outcome and surgical treatment was needed in 2.7% of cases. Conclusion. In an experienced Department of Endoscopy, the majority of upper gastrointestinal hemorrhages can be endoscopically treated with good results. In the last years, endoscopical haemostatic bitherapy (adrenaline injection+clipping or bipolar coagulation) replaced injection of adrenaline like monotherapy for ulcer hemostasis. Key words: upper gastrointestinal hemorrhage, etiology, endoscopic haemostasis. Despite the progresses made in the last decade in the technique of endoscopic haemostasis, the non-variceal upper digestive bleeding (UDB) still remains a very important cause of death. The fatality rate in these cases is 5 10%, linked either to the severity of bleeding, or to the co-morbidities, more frequent in elder patients [1 3]. The role of emergency endoscopy is well defined in UDB: it must be performed in a hemodynamic stable patient, aimed at finding the source of bleeding and performing endoscopic haemostasis [4]. The main causes of non-variceal UDB are duodenal and gastric ulcers, but there are other conditions that may generate important blood loss (e.g. Dieulafoy lesion). The ever growing consumption of non-steroidal anti-inflammatory drugs (NSAID s) and aspirin, especially popular among the elderly, is the main cause of ulcers [5]. Also, Helicobacter pylori (Hp) related ulcers can sometimes lead to UDB in areas in which the prevalence of this infection is still rather high. Lately, therapeutic strategies began to be standardized for bleeding ulcers. Since Calvet s meta-analysis has been published in 2004 [6], the combined treatment, epinephrine injection (EI) + clip placement or thermal coagulation, is the standard of care regarding endoscopic treatment in bleeding ulcers (Forrest I or II), opposed to EI alone. The choice of the endoscopic technique that should be associated to EI, thermal coagulation or clip placement, is still subject of debate [7]. The aim of our study was to assess the main causes of non-variceal UDB cases addressed to the Department of Endoscopy from the Gastroenterology Clinic in Timişoara, as well as the haemostasis techniques that were used for haemorrhagic ulcers and their results, during a 5 years period. Most of the UDB cases from the South-Western part of Romania are addressed to the Department of Gastroenterology from the County Emergency Hospital Timişoara, where a total of 36,360 emergency case presentations were recorded in 2003; 40,794 in 2004; 49,800 in 2005; 51,650 in 2006 and 51,164 in ROM. J. INTERN. MED., 2009, 47, 4,

2 348 I. Sporea et al. 2 MATERIAL AND METHOD We retrospectively analyzed all the non-variceal UDB cases hospitalized in the Department of Gastroenterology of the County Emergency Hospital Timişoara, between January 1 st 2003 and December 31 st 2007, in order to analyze the frequency of bleeding from ulcers, the therapy employed for haemostasis for bleeding ulcers and the posttherapeutic results. On our batch we have assessed the following parameters: age, gender, type of lesion, number of ulcers, type of endoscopic haemostasis, signs of shock (systolic blood pressure-sbp, pulse rate-pr), hemoglobin value, number of administered blood units, use of NSAIDs or anticoagulation treatment, endoscopic second look, rate and moment of rebleeding, endoscopic treatment in case of rebleeding, number of hospitalization days, need of surgery and rate of deaths. All the ulcer lesions were classified using Forrest classification [8] and Rockall score was used [9]. RESULTS A number of 810 successive patients with non-variceal UDB were included in our study. Most of the patients were male 513 (64.3%) and 297 (36.7%) were female, with a mean age of 58.7±15.2 years (ranging from 17 to 96 years old). The main cause of UDB was ulcer, found in 82.6% of the cases (669 patients). In 17.4% cases (141 patients) other types of lesions were the cause of bleeding. THE GROUP OF PATIENTS WITH BLEEDING PEPTIC ULCERS From the 810 patients with UDB, in 669 (82.6%) cases gastric or duodenal ulcers were the cause of hemorrhage. From the 669 ulcers, 45% (301 patients) were gastric ones, 44.1% (295 patients) were duodenal and in 9.7% of the cases (65 patients) both gastric and duodenal ulcers were present. In a smaller number of cases 1.2% (8 patients) the ulcer occurred on an operated stomach. Out of the total of 669 patients with ulcer, 432 (64.6%) were male and 237 (35.4%) were female. In patients with UDB caused by ulcers, the mean age was 59.0±15.2 years (ranging from 17 to 96 years). In 47.5% (317 patients) of the cases, bleeding occurred in patients older than 60 years, a category of patients in which co-morbidities are often present. According to the age group, the bleeding ulcers occurred in: 20 years 7 cases (1%); years 30 cases (4.5%); years 42 cases (6.3%); years 101 cases (15%); years 172 cases (25.7%); years 147 cases (22%); years 128 cases (19.2%); years 40 cases (6.0%); > 90 years 2 cases (0.3%). (Fig. 1) Fig. 1. Distribution on age groups of patients with bleeding ulcer.

3 3 Peptic upper gastrointestinal bleeding 349 According to the Forrest classification, the hemorrhagic ulcers were (Fig.2): type I: 127/669 cases (19%) I a: 32/669 cases (4.8%); I b: 95/669 cases (14.2%); type II: 275/669 cases (41.1%) II a: 125/669 cases (18.7%); II b: 99/669 cases (14.8%); II c: 51/669 cases (7.6%); type III: 267/669 cases (39.9%). The mean Rockall score for these patients was 5.04±2.38 (ranging from 1 to 16 points). Regarding the drug-related etiology of ulcer, aspirin intake was blamed in 180 of the cases (26.9%), NSAID s in 35 of the cases (5.2%), 34 patients (5%) followed a chronic anticoagulant treatment, and 4 (0.6%) patients were treated with prednisone. In patients with ulcerous UDB, 2.7% of the cases (18/669 patients) had to be referred to surgery because the hemorrhage could not be managed by endoscopic therapy. Unfortunately 3.6% of the cases with ulcerous UDB died (24 of the 669 patients). THE ENDOSCOPIC THERAPY After the initial endoscopic assessment of the UDB, in patients with Forrest type Ia, Ib, IIa, and IIb endoscopic lesions, endoscopic haemostasis was performed. Either monotherapy, or combined treatment were performed. Fig. 2. Ulcers repartition according to the Forrest classification. In 268 patients (40.0% of cases) with bleeding ulcer, endoscopic therapy was performed, either monotherapy (EI) in 160 of the cases, or combined in 108 of the cases. Second look endoscopy was performed in 409 cases (61.1%), most frequently in the first 24 hours after the first endoscopy. Posttherapeutic rebleeding occurred in 53/268 cases (19.8%), mostly in the first 24 hours (21 cases), in 31 cases in the first 8 days after admission, only in one case it occurred 17 days after the first bleeding. Concerning the hemodynamic status of the patients, signs of hypovolemic shock were also searched for. SBP < 100 mmhg was observed in 99 patients (14.8%), and PR>100 bpm in 138 patients (20.6%). The mean level of Hb was of 10.60±30.8 g% (with a minimum of 3 g%), and the average number of blood units that were administered to these patients was of 2.35±1.43/patient (ranging from 1 to 9 units/patient). Blood was administered in 285 patients (42.6%). In the group of patients with ulcerous UDB, a mean number of 5.94±3.63 hospitalization days were needed (ranging from 1 to 40 days). Mono-therapy (n=160 patients) The following techniques were used as monotherapy during the first therapeutic endoscopy (Fig. 3): Epinephrine injection (EI) 1/10,000 in 119/160 of the cases (74.4%); Thermal coagulation (TC) 29/160 of the cases (18.1%); Clip placement (CP) 10/160 of the cases (6.2%); Banding ligation (B) 1/160 of the cases; Argon plasma coagulation (APC) 1/160 of the cases. In the majority of cases (74.4%), epinephrine injection 1/10,000 was used as mono-therapy. The mean volume of injected epinephrine solution was 8.52±4.43 ml/patient, ranging from 1 to 30 ml. Combined therapy (n=108 patients) The following therapeutic combinations were used (Fig. 4): EI + TC 73 cases (67.6%); EI + CP 22 cases (20.4%); Triple therapy: EI + TC + CP 9 cases (8.3%); TC + CP 3 cases (2.8%); EI + TC + APC 1 case (0.9%). Therefore, as a combined therapy, in approximately 2/3 of the cases association between EI and the TC was used.

4 350 I. Sporea et al. 4 Fig. 3. Techniques used as mono-therapy in UDB. Fig. 4. Combined techniques used for haemostasis in UDB. If further therapeutic endoscopy was needed (62 cases), the haemostatic technique most frequently used was EI (16 cases 25.7%), followed by the EI + TC combination (14 cases 22.6%) and TC alone (14 cases 22.6%), and CP combined or not with EI (6 cases 9.7% mono-therapy and 6 cases 9.7% of combined therapy. Sometimes, all three techniques were used (6 cases 9.7%). Concerning the use of adrenaline injection (119 cases = 44.7%) vs. definitive therapy (147 cases= 55.3%) we obtained the following results: Epinephrine injection (1/10,000) in 119 cases; Definitive therapy (combined therapy or directly TC or CP) in 147 cases from which: EI + TC: 73 cases (49.6%) EI + CP: 22 cases (14.9%) CP: 10 cases (6.8%) TC: 29 cases (19.7%) EI + TC + CP: 9 cases (6.2%) TC + CP: 3 cases (2.1%) EI + TC + APC: 1 case (0.7%) Regarding the type of endoscopic treatment, we noticed a statistically significant decrease in the number of epinephrine injections between 2003 and % vs. 21.5% (p< ES) (Table I, Fig. 5). Table I Types of endoscopic treatment during the analyzed period Year Epinephrine injection alone Number of patients (%) Combined treatment (two or more techniques) + Definitive treatment (89.2%) 6 (10.8%) (51.1%) 22 (48.9%) (42.6%) 35 (57.4%) (14.6%) 35 (85.4%) (22.2%) 49 (77.8%) Fig. 5. The trend of the endoscopic treatment during time.

5 5 Peptic upper gastrointestinal bleeding 351 DISCUSSION The Endoscopy Department of the Gastroenterology Clinic in Timişoara receives a great number of the UDBs from Timiş County, but also from the neighboring areas. The emergency endoscopy service is manned by 5 up to 12 senior and junior endoscopists (fellows of gastroenterology in the last two years), depending on the analyzed period. The endoscopies were performed during working hours, or on call, out of working hours. The patients have been haemodynamically stabilized before the endoscopy. Also, after the endoscopic haemostatic procedures, all patients received high doses of i.v. PPI s: omeprazole, esomeprazole or pantoprazole, 80 mg bolus followed by 8 mg/hour infusion for 72 hours, and then an oral PPI, 40 mg once daily for the remaining hospitalization days. This standard procedure of PPI administration after therapeutic endoscopy, in continuous infusion after a previous bolus dose, is used by most groups (10, 11). It was demonstrated that high doses of PPI following haemostasis significantly reduce the rebleeding rate (OR 0.49, 95%CI ) and the need for surgical intervention, as compared to placebo or H2 RA [11]. Also, in patients with highrisk endoscopic lesions (i.e. spurting and oozing bleeding, or non-bleeding visible vessel), postendoscopic high dose PPI infusion reduces the mortality in treated patients (OR 0.53, 95% CI ) [11][12]. It was also demonstrated [13] that PPI administration in bolus, followed by continuous infusion in the pre-endoscopic period (especially endoscopy was delayed for up to 12 hours), reduces the number of active bleedings and the number of endoscopic haemostatic procedures. In our group of 810 patients with nonvariceal UDB, the main causes of hemorrhage were, as expected, gastric and duodenal ulcers, in more than 80% of the cases. Other studies have also demonstrated ulcers to be the most frequent cause of bleeding [14], with an incidence almost equal for gastric and duodenal ulcers (29% vs. 30%), similar to our data (45% vs. 44.1%). The mean age in our patients with nonvariceal UDB was 58.7 years (ranging from 17 to 96 years), lower than in Imperiale s study (63.4 years) (15), Romagnuolo s study (66 years) (16) or than in Parente s study (67.3 years) [10]. We compared our data to other studies regarding the mean number of blood units (in our group 2.3 units/patient, ranging from 1 to 9); the mean number of hospitalization days (in our group 5.7 days); the rate of transfers to surgery (in our group 2.7% of the cases); and the mortality 3.6% in our group. In Parente s study [10], in a group of 272 patients, the mean number of blood units administered was of 1.8 and 3 units/patient (senior vs. junior endoscopists, respectively), and the transfer to the surgical department was 4% vs. 10% (senior vs. junior endoscopists), yielding an average of 5.5%. The mortality in this study was 13.2%. In Schemmer s study [14], in a group of 121 patients, the mortality rate was 14%; in Romagnuolo s study (16) in a group of 1860 patients, mortality was 5.3%. In Lau s study [13], performed on 638 patients, the mortality rate was 2.5%, the mean number of transfused blood units was 1.5, and 1.6% of the cases needed surgery. In Imperiale s study [15] performed on 391 patients, death occurred in 3.1% of the cases. As one can observe, the mortality rate and the necessity for surgery varies very much from one study to another. The mortality rate varies between 2.5% [13] and % [10][14]. This wide variation is influenced by multiple factors: the mean age of the analyzed group; the incidence of severe co-morbidities [15][17]18]; the ulcer site (high risk in bulbar ulcers of the posterior wall) [14]; the experience of the endoscopists team, [10], etc. The mortality rate in our group 3.6% of the cases in a large group of patients is a relatively low one. We think that it was positively influenced by the rather low mean age of the group. Considering the ageing population in which UDB occurs more and more often, (a fact that we obviously acknowledged in the last decade), we expect that the mortality rate will probably increase. The number of cases sent to surgery, approximately 4.7% of the cases diagnosed with non-variceal UDB, also fits within the limits of the published data. The second issue that we wish to analyze in this paper is the endoscopic haemostasis. Out of the 669 cases of UDB on ulcers, in 268 patients (approximately 40%), an endoscopic therapy was needed. The remaining ulcers did not need any endoscopic therapy, since they were type II c or III lesions according to the Forrest classification. Regarding the choice of the haemostasis technique, the decision regarding injecting epinephrine 1/10,000 alone, or using a combination of EI with a second technique (bipolar coagulation or clip

6 352 I. Sporea et al. 6 placing), each endoscopy unit has its own strategy. Calvet s meta-analysis [6] demonstrated the superiority of the combined therapy versus simple EI, so the haemostasis strategy changed in most centers (including our own). The current standard of care regarding haemostasis always includes the CP or TC. A recent meta-analysis published by Malmo in 2007, including 2472 patients [19], compared EI to EI + a second endoscopic technique. It demonstrated that combined endoscopic therapy reduces the risk of recurrent bleeding (OD=0.59 [ ], p=0.0001) and the risk of emergency surgery (OR=0.6 6[ ], p=0.03) and showed a trend toward a reduction of the death risk (OR=0.68 [ ], p=0.06), but failed to demonstrate that any combination of treatments is better than either mechanical therapy alone (CP) or thermal therapy (TC) alone. Since our study is a retrospective one, initiated on January 1 st 2003, it includes cases of EI used as monotherapy (119 cases). Regarding the rebleeding rate, from the 268 ulcers that were treated in our department, rebleeding occurred in 19.8% of the cases. These results are similar to the published ones: 18.0% in Romagnuolo s study [16]; 21% in Wolf s study, who also demonstrated that mild to moderate anticoagulation does not increase the risk of re-bleeding following endoscopic therapy [20]; 4.6% in Imperiale s study [15]; 14% 37% in Parente s study, depending on the endoscopist s experience [10]; and 2.8% 20.3% in Liou s study, depending on the volume of injected epinephrine solution [21]. These large studies demonstrated a wide rebleeding rate range, from 2.8% to 37%, influenced by many factors, out of which the experience of the endoscopist and volume of injected epinephrine seem to have statistical significance. Regarding the combined therapy, in our study the endoscopists preferred the association of EI with TC (in 67.6% of the cases), rather than EI + CP (in 20.4% of the cases), in the rest of the cases various combinations of haemostasis techniques being used. These differences are probably due to personal preferences, but we must emphasize that there are some localizations where CP is more difficult to perform (posterior wall of the duodenal bulb, posterior wall of the gastric body and lesser curve of the stomach) [7]. Regarding the superiority of the TC over CP or vice-versa, Sung s analysis [7], performed on 1,156 patients, showed no improvement in definitive haemostasis with CP as compared with TC (81.5% vs 81.2%; RR 1.00, 95% CI ). There were also no differences between CP and TC regarding rebleeding, the need for surgery and mortality. The same study revealed that the rate of definitive haemostasis was higher with CP (86.5%) than with EI (75.4%), RR 1.14, (95% CI ), as well as in CP+EI (88.5%) as compared with EI alone (78.1%), RR 1.13, (95% CI ). The conclusion of this study was that successful application of CP is superior to EI alone, but comparable to TC in obtaining definitive haemostasis. Based on the most recent scientific studies, the endoscopist can decide to perform CP, or TC, depending on his/her preferences and experience and on the localization of the endoscopic lesions. Most often, he/she will inject epinephrine 1/10,000, especially in the case of active bleeding lesions (type Ia or I b according to the Forrest classification). In conclusion, the analysis of the 810 cases of non-variceal UDB, hospitalized in an experienced endoscopy center, has shown that more than 80% of these hemorrhages were caused by gastric or duodenal ulcers (with approximately equal incidence). Ulcerous UDB had in our department a mortality rate of 3.6%, and the rate of haemostatic surgical interventions was 2.7%. Regarding hemorrhagic ulcers, endoscopic haemostasis was needed in 40% of the cases. In endoscopically treated ulcers, rebleeding occurred in 19.8% of the cases. During the time of this study, we observed a decreasing number of epinephrine injections in favor of definitive treatment (bipolar coagulation or clipping). In an experienced endoscopic center, the endoscopic treatment of ulcerous UDB was successful in the overwhelming majority of cases, the mortality rate and the transfer rate to surgery can be maintained at relatively low levels. Scop. Am evaluat toate cazurile de hemoragie digestivă superioară investigate în cadrul Departamentului de Endoscopie al clinicii noastre pe o perioadă de 5 ani.

7 7 Peptic upper gastrointestinal bleeding 353 Material şi metodă. Am inclus în studiul nostru 810 pacienţi, 64.3% bărbaţi şi 36.7% femei, cu vârsta medie 58.7±15.2 ani (17 96 ani). Principala cauză de hemoragie digestivă superioară a fost ulcerul peptic (82.6%), cu distribuţie egală pentru ulcerele gastrice şi duodenale. Toate ulcerele au fost evaluate după clasificarea Forrest. Rezultate. La 40% din cazuri s-a efectuat hemostază endoscopică, în 44.4% din cazuri aceasta constând doar din injectare de adrenalină (1/10.000); terapie combinată (injectare + clipare sau termocoagulare) a fost folosită în 40.3% din cazuri, iar cliparea sau termocoagularea ca monoterapii au fost folosite în 15.3% din cazuri. Pe perioada studiului s-a observat o reducere importantă a proporţiei folosirii injectării de adrenalină ca monoterapie hemostatică. Resângerarea post-terapie endoscopică a apărut la 19.8% din cazuri; 3.6% din pacienţi au decedat iar 2.7% din cazuri au necesitat intervenţie chirurgicală. Concluzie. Într-un Departament de Endoscopie cu experienţă majoritatea hemoragiilor digestive superioare pot fi tratate endoscopic cu rezultate bune. În ultimii ani biterapia (injectarea de adrenalină + clipare sau coagulare bipolară) a înlocuit monoterapia prin injectare de adrenalină în terapia hemostatică a ulcerelor. Corresponding author: I. Sporea Snagov Str Timişoara isporea@umft.ro REFERENCES 1. LONGSTRETH G.F., Epidemiology of hospitalization for acute upper gastrointestinal hemorrhage: a population based study. Am. J. Gastroenterol., 1995; 90: TARGOWNIK L.W., NABALAMBA A., Trends in management and outcome of acute non-variceal upper gastrointestinal bleeding: Clin. Gastroenterol. and Hepatol., 2006; 4: AMERICAN SOCIETY OF GASTROINTESTINAL ENDOSCOPY. GUIDELINE, The role of endoscopy in acute nonvariceal upper GI hemorrhage. Gastrointest. Endosc., 2004; 60: COOK D.J., GAYATT G.H., SALENA B.J., LAINE L.A., Endoscopic therapy for acute non-variceal hemorrhage: a metaanalysis. Gastroenterology, 1992; 102: LANAS A., GARCIA-RODRIGUEZ L.A., ARROYO M.T. et al., Risk of upper gastrointestinal ulcer bleeding associated with selective cyclo-oxygenase-2 inhibitors, traditional non-aspirin non-steroidal anti-inflammatory drugs, aspirin and combinations. Gut, 2006; 55: CALVET X., VERGARA M., BRULLET E., GISBERT J.P., CAMPO R., Addition of a second endoscopic treatment following epinephrine injection improve outcome in high risk bleeding ulcers. Gastroenterology, 2004; 126: SUNG J.J.Y., TSOI K.T., LAI L.H., WU J.C., LAU J.Y., Endoscopic clipping versus injection and thermo-coagulation in the treatment of non-variceal upper gastrointestinal bleeding: a meta-analysis. Gut, 2007; 10: FORREST J.A., FINLAYSON N.D., SHEARMAN D.J., Endoscopy in gastrointestinal bleeding. Lancet, 1974; 2: ROCKALL T.A., LOGAN R.F., DEVLIN H.B., NORTHFIELD T.C., Risk assessment after acute upper gastrointestinal haemorrhage. Gut, 1996, 38 (3): PARENTE F., ANDERLONI A., BARGIGGIA S. et al., Outcome of non-variceal acute upper gastrointestinal bleeding in relation to the time of endoscopy and the experience of the endoscopist: A two-year survey. World J. Gastroenterol., 2005; 11, LEONTIADIS G.I., SREEDHARAN A., DORWARD S. et al., Systematic reviews of the clinical effectiveness and costeffectiveness of proton pump inhibitors in acute upper gastrointestinal bleeding. Health Technology Assessment, 2007; 51: LEONTIADIS G.I., SHARMA V.K., HOWDEN C.W., Proton pump inhibitor therapy for peptic ulcer bleeding: Cochrane Collaboration meta-analysis of randomized controlled trials. Mayo Clin. Proc., 2007; 82: LAU J.Y., LEUNG W.K., WU J.C.Y. et al., Omeprazole before endoscopy in patients with gastrointestinal bleeding. NEJM, 2007; 356: SCHEMMER P., DECKER F., DEI-ANANE G. et al., The vital threat of an upper gastrointestinal bleeding: Risk factors analysis of 121 consecutive patients. World J. Gastroenterol., 2006; 12:

8 354 I. Sporea et al IMPERIALE T.F., DOMINITZ J.A., PROVENZALE D.T. et al., Predicting poor outcome from acute upper gastrointestinal hemorrhage. Arch. Intern. Med., 2007; 167: ROMAGNUOLO J., BARKUN A.N., ENNS R., ARMSTRONG D., GREGOR J., Simple clinical predictors may obviate urgent endoscopy in selected patients with nonvariceal upper gastrointestinal tract bleeding. Arch. Intern. Med., 2007; 167: ENNS R.A., GAGNON Y.M., BARKUN A.N., ARMSTRONG D., GREGOR J.C., FEDORAK R.N., RUGBE INVESTIGATORS GROUP, Validation of the Rockall scoring system for outcomes from non-variceal upper gastrointestinal bleeding in a Canadian setting. World J. Gastroenterol., 2006; 12: THAM T.C., JAMES C., KELLY M., Predicting outcome of acute non-variceal upper gastrointestinal hemorrhage without endoscopy using the clinical Rockall Score. Postgrad. Med. J., 2006; 82: MARMO R., ROTONDANO G., PISCOPO R., BIANCO M., D ANGELLA R., CIPOLLETTA L., Dual therapy versus monotherapy in the endoscopic treatment of high-risk bleeding ulcers: A meta-analysis of controlled trials. Am. J. Gastroenterol., 2007; 102: WOLF A.T., WASAN S.K., SALZMAN J.R., Impact of anticoagulation on rebleeding following endoscopic therapy for nonvariceal upper gastrointestinal hemorrhage. Am. J. Gastroenterol., 2007; 102: LIOU T.C., LIN S.C., WANG H.Y., CHANG W.H., Optimal injection volume of epinephrine for endoscopic treatment of peptic ulcer bleeding. World J. Gastroenterol., 2006; 12: Received August 1, 2009