Addition of a Second Endoscopic Treatment Following Epinephrine Injection Improves Outcome in High-Risk Bleeding Ulcers

GASTROENTEROLOGY 2004;126:441 450 Addition of a Second Endoscopic Treatment Following Epinephrine Injection Improves Outcome in High-Risk Bleeding Ulcers XAVIER CALVET,* MERCEDES VERGARA,* ENRIC BRULLET,* JAVIER P. GISBERT, and RAFEL CAMPO* *Unitat de Malalties Digestives, Hospital de Sabadell/UDIAT, Institut Universitari Parc Taulí, Universitat Autónoma de Barcelona, Barcelona; and Servicio de Aparato Digestivo Hospital de la Princesa, Madrid, Spain Background & Aims: Endoscopic therapy reduces the rebleeding rate, the need for surgery, and the mortality in patients with peptic ulcer and active bleeding or visible vessel. Injection of epinephrine is the most popular therapeutic method. Guidelines disagree on the need for a second hemostatic procedure immediately after epinephrine; although it seems to reduce further bleeding, its effects on morbidity, surgery rates, and mortality remain unclear. The aim of this study was to perform a systematic review and meta-analysis to determine whether the addition of a second procedure improves hemostatic efficacy and/or patient outcomes after epinephrine injection. Methods: An extensive search for randomized trials comparing epinephrine alone vs. epinephrine plus a second method was performed in MED- LINE and EMBASE and in the abstracts of the AGA Congresses between 1990 and 2002. Selected articles were included in a meta-analysis. Results: Sixteen studies including 1673 patients met inclusion criteria. Adding a second procedure reduced the further bleeding rate from 18.4% to 10.6% (Peto odds ratio 0.53, 95% CI: 0.40 0.69) and emergency surgery from 11.3% to 7.6% (OR: 0.64, 95% CI: 0.46 0.90). Mortality fell from 5.1% to 2.6% (OR: 0.51, 95% CI: 0.31 0.84). Subanalysis showed that the risk of further bleeding decreased regardless of which second procedure was applied. In addition, the risk was reduced in all subgroups, although reduction was more evident in high-risk patients and when no scheduled follow-up endoscopies were performed. Conclusions: Additional endoscopic treatment after epinephrine injection reduces further bleeding, need for surgery, and mortality in patients with bleeding peptic ulcer. Endoscopic therapy reduces the rebleeding rate, the need for surgery, and the morbidity and mortality of patients bleeding from a peptic ulcer. 1 Many different endoscopic hemostatic techniques have been developed and studied over the last 25 years. Methods are based on the injection of vasoconstrictor substances (epinephrine), sclerosant substances (polidocanol, absolute alcohol), clotting factors (thrombin, fibrin glue), or adhesives (cyanoacrylate). Thermal therapies include laser, monopolar electrocoagulation, argon plasma coagulation, bipolar probes, and heater probe. More recently, the use of mechanical devices to clip the bleeding vessel (hemoclip) has been incorporated. Epinephrine injection alone or in combination with another technique has become the most popular endoscopic method for emergency endoscopic hemostasis because of its safety, low cost, and easy application. 2 Although some randomized studies have established that epinephrine in combination with a second hemostatic technique is better than epinephrine alone, the guidelines do not make clear recommendations on this point. 3,4 The addition of a second endoscopic technique can increase the cost and the risk of complication of the procedure, and it remains unclear whether the reduction of further bleeding actually offsets these drawbacks. The aim of the present study was to perform a metaanalysis of the studies comparing the safety and efficacy of endoscopic injection of epinephrine alone vs. epinephrine plus a second hemostatic method in the treatment of patients with a bleeding peptic ulcer. Materials and Methods Literature Search and Identification of Primary Studies A literature search was performed in September 2002 using the MEDLINE and EMBASE databases and the Cochrane Controlled Trials Register. The strategy included the words (peptic ulcer OR gastric ulcer OR duodenal ulcer) AND (bleeding OR haemorrhage OR hemorrhage) AND (sclerotherapy OR sclerosis OR injection). We also conducted Abbreviation used in this paper: OR, odds ratio. 2004 by the American Gastroenterological Association 0016-5085/04/$30.00 doi:10.1053/j.gastro.2003.11.006

442 CALVET ET AL. GASTROENTEROLOGY Vol. 126, No. 2 Table 1. Criteria for Further Bleeding and Mortality Study Criteria for clinical further bleeding Mortality Balanzo et al. 19 Not specified Not specified Choudari and Palmer 20 Fresh hematemesis or melena shock or fall in hemoglobin 2 g/dl 30-Day mortality Chung et al. 21 Fresh hematemesis or melena shock or fall in hemoglobin 2 g/dl Hospital mortality Chung et al. 22 Not specified Hospital mortality Chung et al. 23 Not specified Hospital mortality Chung et al. 24 Fresh hematemesis or melena shock or transfusion 8 U Hospital mortality Garrido et al. 25 Hematemesis or fresh melena Not specified Kubba et al. 26 Fresh hematemesis or melena shock or fall in hemoglobin 2 g/dl 30-Day mortality Lee et al. 27 Not specified Not specified Lin et al. 28 Unstable vital signs or continued bloody stools or hematemesis Hospital mortality Lin et al. 29 Unstable vital signs or continued bloody stools or hematemesis Hospital mortality Loizou and Brown 30 Fresh hematemesis or melena shock or fall in hemoglobin despite transfusion Hospital mortality Pescatore et al. 31 Fresh hematemesis or melena shock or fall in hemoglobin despite transfusion 30-Day mortality Sollano et al. 32 Recurrence of hematemesis, melena, or anemia Not specified Villanueva et al. 33 Fresh hematemesis or melena hypovolemia or fall of hemoglobin requiring transfusion Hospital mortality Villanueva et al. 34 Not specified Not specified a manual search of abstracts submitted to the Digestive Diseases Week between 1990 and 2002. The search was primarily planned to include articles in English, French, or Spanish. Articles selected in the search were reviewed separately by 2 of the authors, and those fulfilling the inclusion criteria were selected for further analysis. In addition, a fully recursive search of reference lists of the original studies was performed to find studies not identified by the previous searches. Papers recorded in the personal databases of the authors were also reviewed and included when appropriate. Studies Selection Studies designed to compare the efficacy of different endoscopic methods to achieve definitive hemostasis in peptic ulcer patients were evaluated separately by 2 of the authors. The inclusion criteria were as follows: (1) Articles or abstracts should report the results of comparative, randomized trials. (2) Studies must include patients with hemorrhage from peptic ulcer disease (gastric or duodenal) with major stigmata of bleeding as defined by groups Ia (spurting hemorrhage), Ib (oozing hemorrhage), IIa (nonbleeding visible vessel), and IIb (adherent clot) of the Forrest classification. 5 (3) Studies had to include at least 2 branches of endoscopic treatment: epinephrine alone vs. epinephrine associated with a second hemostatic method. (4) The data on the baseline characteristics of the patients (number, age, sex, and others), inclusion and exclusion criteria, and results should allow adequate evaluation. Further clinically significant bleeding was defined according to the criteria established in each study (Table 1). Bleeding had to be confirmed by endoscopy. Criteria Used for the End Points of the Study Primary end point: further bleeding. In most of the studies, the primary end point was defined as endoscopic therapy failure, i.e., a combination of persistent hemorrhage and recurrence during follow-up. Here, for the sake of simplicity, this primary end point is referred to throughout the text as further bleeding a more usual term and one that gives a clearer image. Clinical criteria used for presuming further bleeding differed between the studies and are shown in Table 1. Bleeding was confirmed by endoscopy in all the studies. Clinically silent bleeding observed at scheduled endoscopies was not considered for the analysis. A subanalysis comparing data separately for persistent hemorrhage and recurrence during follow-up was also performed. Surgery and mortality. Criteria for emergency surgery were not specified in the majority of the studies, and few gave the criteria for mortality; 30-day mortality or in-hospital mortality were the most often used (Table 1). Data Extraction Data were extracted separately by 2 of the authors and reviewed by a third. If results were discordant, papers were jointly reviewed until the differences were resolved. Quality Assessment The quality of the studies included was assessed using the criteria proposed by Chalmers et al. 6 This method evaluates the design, implementation, and analysis of randomized controlled trials. The overall index of trial quality was weighted as follows: trial design and protocol (0.60), statistical analysis (0.30), and presentation of results (0.10). Final quality score ranged from 0 to 1, with maximum quality studies rating 1. Statistical Analysis Main comparisons contrasted epinephrine injection vs. epinephrine injection plus another hemostatic method. The primary outcome variable was further bleeding, defined as persistence or recurrence of bleeding during follow-up. Emergency surgery during hospitalization, morbidity, and mortality rate were also analyzed. Subanalysis for further bleeding was performed to examine the efficacy of the different techniques (sclerosant agents, mechanical hemostasis, or thermal devices) associated to epinephrine injection vs. epinephrine alone. Subanalysis was also performed depending on the type of

February 2004 ENDOSCOPIC TREATMENT IN BLEEDING PEPTIC ULCERS 443 Table 2. Reasons for Excluding Articles Study Reasons for exclusion Pescatore et al. 8 Preliminary results of Pescatore et al. 31 Male et al. 9 Preliminary results of Lin et al. 29 Gevers et al. 10 Combines epinephrine plus polidocanol plus hemoclip De Goede et al. 11 Preliminary results of Givers et al. 10 Chung et al. 12 Preliminary results of Chung et al. 22 Chung et al. 13 Preliminary results of Chung et al. 23 Chung et al. 14 Preliminary results of Chung et al. 21 Chua et al. 15 Not randomized Buffoli et al. 16 Not randomized Dedeu et al. 17 Does not specify number of patients in each group Ell 18 Does not specify number of patients in each group peptic ulcer hemorrhage: active spurting or oozing (Forrest Ia or Ib), nonbleeding visible vessel (Forrest IIa), or adherent clot (Forrest IIb). Also, further bleeding rates were divided into failure to achieve initial hemostasis and recurrence during follow-up and analyzed separately. Finally, to ascertain the influence of second-look endoscopy on the results, studies that performed this procedure were also analyzed separately. Peto odds ratios and 95% confidence intervals were used for comparisons. Prior to meta-analysis, the heterogeneity of results was assessed by means of a Q test. Because of the low power of the test, a cutoff P value of 0.20 was established as a threshold for homogeneity. In case of lower values indicating heterogeneity, a random-effect model was used for the analysis. If no heterogeneity was observed, odds ratios for all studies were pooled in a global odds ratio by means of a fixed-effects model. All results were obtained using the freeware program Review Manager 4.1.3. The statistical tests and formula implemented in RevMan are described in the RevMan User Guide. 7 Results Included and Excluded Studies The preliminary search identified 27 studies (26 in English, 1 in Spanish, and none in French). 8 34 A wider search including articles in other languages did not find any additional papers suitable for inclusion. Eleven studies were excluded from further analysis. 8 18 Reasons for exclusion are outlined in Table 2. Sixteen studies were included in the meta-analysis. 19 34 Characteristics of the studies are shown in Tables 1, 3, and 4. Quality Assessment The articles included ranged in quality from 0.38 to 0.88. Individual assessment of quality is shown in Table 3. Most of the studies were published as full papers and were rated high in the quality scores. Epinephrine vs. Epinephrine and a Second Endoscopic Method Sixteen articles compared epinephrine injection vs. epinephrine plus other endoscopic methods for the endoscopic treatment of peptic ulcers. A total of 1673 patients were included. The further bleeding rate in the epinephrine group was 155 of 840 (18.4%) vs. 88 of 833 (10.6%) patients in the combined therapy group. Peto odds ratio was 0.53 (95% CI: 0.40 0.69) (Figure 1). The need for emergency surgery was evaluated in 15 studies with a total of 1588 patients. In the epinephrine group, 90 of 795 (11.3%) patients needed surgical intervention and 60 of 793 (7.6%) in the combined therapy. Peto odds ratio was 0.64 (95% CI: 0.46 0.90) (Figure 2). The mortality rate was evaluated in 15 studies and 1588 Table 3. Characteristics of the Studies Included in the Meta-analysis Study Patients (N) Quality Additional treatment Second look Medical treatment Balanzo et al. 19 64 0.73 Thrombin injection Yes Not specified Choudari and Palmer 20 107 0.79 Ethanolamine injection Yes H2 antagonists Chung et al. 21 83 0.82 Hemoclip Yes Ranitidine 50 mg/6 h IV Chung et al. 22 196 0.77 Sodium tetradecyl sulfate Yes Anti-H2 Chung et al. 23 160 0.75 Ethanol injection Yes Ranitidine Chung et al. 24 270 0.88 Heat probe Yes Not specified Garrido et al. 25 85 0.59 Polidocanol injection No Not specified Kubba et al. 26 140 0.83 Thrombin injection No Not specified Lee et al. 27 60 Abstract Ethanol injection No Not specified Lin et al. 28 64 0.74 Ethanol injection No Ranitidine or cimetidine Lin et al. 29 64 0.64 Bipolar electrocoagulation Yes Omeprazole 40 mg/6 h IV Loizou and Brown 30 42 0.70 NYAG laser No Ranitidine 300 mg/12 h PO Pescatore et al. 31 135 0.76 Fibrin glue injection Yes Not specified Sollano et al. 32 61 0.38 Polidocanol injection Yes Not specified Villanueva et al. 33 63 0.86 Polidocanol injection Yes Ranitidine Villanueva et al. 34 79 Abstract Hemoclip No Not specified IV, intravenously; PO, per os.

444 CALVET ET AL. GASTROENTEROLOGY Vol. 126, No. 2 Table 4. Volume of Injected Epinephrine and Complications and Blood Transfusion Requirements Study Epinephrine volume Complications Transfusion Balanzo et al. 19 Not specified Not specified A: 3.9 U, B: 3.14 U Choudari and Palmer 20 7.5 ml in both groups Not specified Not specified Chung et al. 21 Not specified A: 1 bleeding, 2 sub mucosal hematoma. B: None A: 9.1 U, B: 8.9 U Chung et al. 22 Not specified A: None, B: Necrosis of lesser curvature A: 3U(0 41), B: 3U(0 19) Chung et al. 23 A: 9.5 5.5 ml, B: 10.1 6.2 ml Not specified A:3U(0 20), B: 2U(0 23) Chung et al. 24 Not specified A: None, B: 2 perforations A: 2U(0 18), B: 3U(0 29) Garrido et al. 25 Not specified Not specified Not specified Kubba et al. 26 A: 10 ml, B: 7 ml No complications A: 297 (total U), B: 219 (total U) Lee et al. 27 A: 17.6 ml, B: 14.5 ml Not specified Not specified Lin et al. 28 A: 6 ml, B: 5.5 ml No complications A: 1.2 L 1.1 L, B: 1.2 L 1.6 L Lin et al. 29 A: 7.1 ml, B: 6.8 ml No complications A: 0.7 (0.5 1) L, B: 0.6 (0.3 0.8) L Loizou and Brown 30 10 30 ml (both groups) A: Spurting hemorrhage, B: None A: 5.8 4.3 U, B: 5.1 3.2 U Pescatore et al. 31 A: 10.3 1.8 ml, B: 7.5 3.5 ml A: 1 hemorrhage, B: 1 perforation A: 2U(2 4),B:2U(2 4) Sollano et al. 32 A: 8.6 ml, B: 11 ml A: 1 mucosal necrosis, B: 1 mucosal necrosis Not specified Villanueva et al. 33 A: 12 3.5 ml, B: 15.2 6.2 ml A: None, B: Pneumoperitoneum A: 2.03 2.04 U, B: 1.9 2.5 U Villanueva et al. 34 Not specified Not specified A: 2.6 2.1 U, B: 1.6 2.4 U A, Epinephrine alone; B, combined therapy; N (n-n), median (range); N n, mean SD; U, blood units. patients, with a total of 41 of 795 (5.1%) patients in the epinephrine group and 21 of 793 (2.6%) in the combined therapy group. Peto odds ratio was 0.51 (95% CI: 0.31 0.84) (Figure 3). In addition, 10 studies, including 1118 patients, reported data about complications of endoscopic therapy. The number and type of complications in each group are detailed in Table 4. Significant complications were found in 6 of 560 (1.1%) patients in the epinephrine group and in 6 of 558 patients (1.1%) in the combined therapy group. Peto odds ratio was 1.01 (95% Figure 1. Further bleeding rates for endoscopic treatment with epinephrine injection vs. combined therapy.

February 2004 ENDOSCOPIC TREATMENT IN BLEEDING PEPTIC ULCERS 445 Figure 2. Surgical rates for endoscopic treatment with epinephrine injection vs. combined therapy. CI: 0.32 3.16). No significant heterogeneity was found in any of the analyses. Epinephrine vs. Epinephrine and a Second Injected Agent All articles that compared substances administered by endoscopic injection (sclerosants such as ethanol, polidocanol, ethanolamine, or tetradecyl sulfate; adhesive agents such as cyanoacrylate, and thrombotic substances such as a fibrin glue or thrombin) were analyzed together. Eleven studies with 1135 patients were included. 19,20,22,23,25 28,31 33 The further bleeding rate was 108 of 575 (18.8%) patients in the epinephrine group and 68 of 560 (12.1%) in the combined therapy group. Peto odds ratio was 0.60 (95% CI: 0.43 0.82). The need for surgery was evaluated in 10 studies with a total of 1050 patients. Fifty-nine of 530 (11.1%) patients in the epinephrine group and 49 of 520 (9.4%) in the combined therapy group needed surgical intervention; Peto odds ratio was 0.82 (95% CI: 0.55 1.22). Mortality rate was evaluated in 10 studies with a total of 1050 patients. In the epinephrine group, 28 of 530 patients died (5.3%) vs. 11 of 520 (2.1%) in the combined group, Peto odds ratio was 0.41 (95% CI: 0.21 0.77). Epinephrine vs. Epinephrine and Mechanical Endoscopic Methods Two studies analyzed the efficacy of adding a mechanical hemoclip to achieve hemostasis in bleeding peptic ulcers. 21,34 A total of 162 patients were evaluated. The further bleeding rate was 15 of 78 (19.2%) in the epinephrine group and 7 of 84 (8.3%) in the combined therapy group. The Peto odds ratio was 0.40 (95% CI 0.16 0.98). Both studies evaluated the need for surgery. Nine of 78 (11.5%) patients in the epinephrine group and 2 of 84 (2.4%) in the combined therapy group required surgical intervention. Peto odds ratio was 0.24 (95% CI: 0.07 0.82). Mortality rate was 3 of 78 (3.8%) patients in the epinephrine group compared with 1 of 84 (1.2%) in the combined therapy group. Peto odds ratio was 0.34 (95% CI: 0.05 2.42). Epinephrine vs. Epinephrine and Thermal Methods Three studies compared epinephrine alone with epinephrine combined with thermal hemostatic methods (contact heat probe, N-YAG laser or bipolar electrocoagulation). 24,29,30 A total of 376 patients were evaluated.

446 CALVET ET AL. GASTROENTEROLOGY Vol. 126, No. 2 Figure 3. Mortality rates for endoscopic treatment with epinephrine injection vs. combined therapy. The further bleeding rate was 32 of 187 (17.1%) patients in the epinephrine group vs. 13 of 189 (6.9%) in the combined therapy group. Peto odds ratio was 0.37 (95% CI: 0.19 0.69). The surgery rate was analyzed in 376 patients, being required in 22 of 187 (11.8%) patients in the epinephrine group and in 9 of 189 (4.8%) in the combined therapy group. Peto odds ratio was 0.40 (95% CI: 0.19 0.83). The mortality rate was also analyzed in 376 patients. It was 10 of 187 (5.3%) patients in the epinephrine group and 9 of 189 (4.8%) in the combined therapy group. Peto odds ratio was 0.88 (95% CI: 0.35 2.23). Further Bleeding Rates in Patients With Active Hemorrhage (Forrest Ia or Ib) or Nonbleeding Visible Vessel (Forrest IIa) or Adherent Clot (Forrest IIb) Eleven studies with a total of 913 patients analyzed further bleeding rates when peptic ulcers were actively bleeding (spurting or oozing, Forrest Ia or Ib). 19 26,28,30,33 Eighty-nine of 453 (19.6%) patients further bled in the epinephrine alone group vs. 54 of 460 (11.7%) in the combined therapy group. Peto odds ratio was 0.55 (95% CI: 0.38 0.79). Seven studies with a total of 353 patients gave data on peptic ulcers with nonbleeding visible vessel (Forrest IIa). 19 21,25,26,30,33 Twenty-three of 182 (12.6%) patients in the epinephrine alone group presented further bleeding vs. 14 of 171 (8.2%) in the combined therapy group. Peto odds ratio was 0.62 (95% CI: 0.31 1.22). Only 1 study 25 included patients with adherent clot (Forrest IIb), and, therefore, a meta-analysis could not be performed. Initial Failure of Hemostasis vs. Recurrence During Follow-up Thirteen studies gave separate data on the patients who did not achieve initial hemostasis. 19,21 24,26 33 A total of 1402 patients were evaluated. Of these, 20 of 703 (2.8%) patients in the epinephrine group and 17 of 699 (2.4%) in the combined therapy group presented initial failure of hemostasis, Peto odds ratio 0.84 (95% CI: 0.44 1.62). In the 1365 patients with initial hemostasis, the rate of recurrence of bleeding during follow-up was 108 of 683 (15.8%) in the epinephrine group and 61 of 682 (8.9%) in the combined therapy group. Peto odds ratio was 0.53 (95% CI: 0.38 0.73).

February 2004 ENDOSCOPIC TREATMENT IN BLEEDING PEPTIC ULCERS 447 Role of Second Look Endoscopy Ten studies performed one or more second look endoscopies 24 to 72 hours after the initial technique. 19 24,29,31 33 When active bleeding or persistent high-risk stigmata were observed, a second therapeutic procedure was performed. A total of 1202 patients were included in these studies. The further bleeding rate was 95 of 605 (15.7%) patients in the epinephrine group and 68 of 597 (11.4%) in the combined therapy group. Peto odds ratio was 0.69 (95% CI: 0.49 0.96). Six studies with a total of 470 patients did not schedule a second-look endoscopy. 25 28,30,34 Sixty of 235 (25.5%) patients in the epinephrine group further bled after the first endoscopic procedure compared with 20 of 235 (8.5%) in the combined therapy group. Peto odds ratio was 0.27 (95% CI: 0.15 0.46). Discussion Guidelines agree that there is no clear evidence that any technique is superior to injection of epinephrine alone for the endoscopic treatment of high-risk bleeding peptic ulcers. 3,4 However, a few individual studies have shown a significant reduction in the further bleeding rates with the addition of a second endoscopic treatment 24 26,29 (Figure 1). Advancing further in this direction, the results of the present meta-analysis clearly suggest that combined therapy is the treatment of choice for high-risk bleeding peptic ulcers. Although the absolute improvements in hemostatic efficacy are relatively small (from 5% to 10%), they represent a 30% to 50% reduction in the relative risk of a recurrent hemorrhage. Failure of endoscopic therapy is the main predictor of the need for surgery and morbidity and mortality in bleeding peptic ulcer patients. 35 Therefore, it is no surprise that reduction of further bleeding rates decreased the need for surgery and improved survival. One of the major fears about using combined therapy is the possible risk of gastric wall necrosis and/or perforation. The present meta-analysis shows that the risk of significant complications is very low (1.1% in the epinephrine group and 1.1% in the combined therapy group) and that there were no differences between groups. Looking at the complications in detail, induction of massive bleeding requiring surgery was the most frequent in the group of adrenaline alone; gastric wall necrosis appeared in 3 patients (2 in the combined therapy group and 1 in the adrenaline alone group), and perforation (4 patients) was observed only in the combined therapy group. These complications appeared with both further injection and thermal methods. Thus, it would seem that perforation or necrosis was slightly more frequent in the combined therapy group (6/558 patients) than in the adrenaline alone group (1/560 patients), although this difference did not achieve statistical significance. In any case, this possible small increase in the risk of perforation or necrosis is clearly compensated for by the benefits in reducing further bleeding, which result in a significant decrease in the need for surgery and mortality; it is therefore not a reason for avoiding combined therapy. The subanalysis shows many interesting data: first, that the improvement in prognosis seems to be more evident in higher risk patients that is, those with Forrest I bleeding ulcers or when no second-look endoscopy is performed. However, because the basal risk of further bleeding is much lower in Forrest II patients or when repeated endoscopy is performed, the sample size needed to demonstrate a significant reduction in further bleeding may be larger than those available for the subanalysis. Second, that beneficial effects of combined therapy seem to derive from the reduction of recurrent bleeding during follow-up rates, and the ability to achieve initial hemostasis seems very similar in the 2 groups. The subanalysis also raises many unanswered questions. First, although it could be taken to suggest that the efficacy of the injection of a second agent is similar to that achieved with thermal and mechanical methods, this interpretation should be treated with extreme care. The only conclusion that can be drawn from the study is that, whichever second treatment is used, combined therapy seems to work better than adrenaline alone. By contrast, it cannot be concluded that a particular form of treatment is equal to another. There are 2 main reasons for this. First, the subgroups (further injection, thermal, or mechanical methods) include different procedures, which may present heterogeneous activity. In fact, each endoscopic treatment had different characteristics: epinephrine produces vasoconstriction, vessel compression, and platelet aggregation but seems not to induce permanent thrombosis in blood vessels. 36 Sclerosant agents such as polidocanol or ethanol can produce thrombosis of vessels favoring hemostasis, although they may also induce significant tissue injury. 37,38 Whether human thrombin injection could reduce the risk of tissue damage remains unclear, and thrombin is more expensive than other additional treatments. 39 Thermal agents also produce thrombosis of vessels and also risk, therefore, of damaging tissue. Among them, laser photocoagulation seems to be associated with higher risks of perforation, optical hazard, high cost, and imperfect hemostatic effect. Mul-

448 CALVET ET AL. GASTROENTEROLOGY Vol. 126, No. 2 tipolar electrocoagulation and heater probe thermocoagulation have been reported to produce excellent results; they are also less expensive and more easily portable than a laser. 40 Mechanical methods also closed the vessel. They were also associated with few complications, but there was a certain amount of technical difficulty in applying the hemoclip on the posterior wall of the proximal body and cardia of the stomach and on the posterior wall of the duodenum because of the requirement that the hemoclip meet the lesion at a right angle. 21,41 Even more important, the statistical treatment of the study is not designed to compare additional treatments head to head. In fact, we lack randomized trials comparing different additional treatments after epinephrine injection and must await future comparative studies to establish which is the best therapy to add to adrenaline. Nor did our study determine whether combined therapy is better than sclerosants or thermal or mechanical methods alone. The evidence in the literature to suggest so is scarce. In a small study Lin et al. found that combined therapy seems better than bipolar electrocoagulation alone. 29 Also, a recent meta-analysis concluded that combined injection is superior to a sclerosant alone. 42 In any event, the clinical relevance of this question may be minor. Adrenaline injection is cheap, easy to perform, and safe. In addition, according to our analysis, epinephrine is as good as combined therapy for achieving initial hemostasis. Therefore, by controlling active bleeding, it could allow a better endoscopic view and a more accurate targeting of the additional therapy. Therefore, there are few medical or economic arguments against adrenaline injection. An important methodologic point about the study was the definition of the primary end point. The marked heterogeneity in defining further bleeding among the studies precluded the definition of a homogeneous, predetermined criteria for the primary end point. We therefore accepted the definition established for each study. Further bleeding definitions are shown in Table 1. Because the analysis includes only randomized, comparative studies, the criteria were similar for the 2 groups (adrenaline alone and combined therapy) in each study, thus allowing further comparison. Furthermore, endoscopic confirmation of bleeding was required, thus reducing heterogeneity. Finally, results on further bleeding are consistent with those of variables with a lower degree of subjectivity in their evaluation, such as the need for surgery or mortality. This indirectly confirms that the assessment of further bleeding was adequate. An interesting question is how other therapeutic approaches such as the use of high-dose PPI or secondlook endoscopy could influence the efficacy of combined therapy. The use of high-dose proton pump inhibitors in bleeding peptic ulcer patients is gaining acceptance. The evidence clearly suggests that these drugs reduce the risk of rebleeding. 43 However, many points remain unclear, such as the cost effectiveness of this approach, the ideal drug dosage, and whether this strategy should be reserved for patients at high risk of rebleeding. Both combined therapy and proton pump inhibitor infusion are safe and comfortable for the patient. Furthermore, there is no reason to suspect that associating high-dose proton pump inhibitors and combined endoscopic therapy will show ill effects or decrease efficacy. Therefore, although the extent of the benefit of combining the 2 approaches remains uncertain, it seems reasonable to do so until more evidence becomes available. As far as second-look endoscopy is concerned, a recent meta-analysis showed that, although scheduled secondlook endoscopies reduced the rebleeding rate, they did not decrease the need for surgery or mortality. 44 Recently, it has been suggested that selective second-look endoscopy for selected high-risk patients could be a cost-effective approach. 45 However, this strategy exposes patients to uncomfortable and somewhat risky procedures and increases the workload of the endoscopy unit. The results of the present study indirectly suggest that the effect of associating a second hemostatic procedure after epinephrine injection is similar to that obtained using second-look endoscopy. For all these reasons, it remains unclear whether second-look endoscopy offers any added benefit to combined therapy associated with proton pump inhibitor infusion. In conclusion, the present study shows that adding a second endoscopic procedure after adrenaline injection reduces the rates of recurrence in high-risk bleeding peptic ulcer patients, thus reducing the need for surgery and mortality. Currently, there is no justification for preferring a specific second hemostatic method after adrenaline injection. In view of the evidence and while we await the results of further studies combined therapy should be considered as the standard procedure. References 1. Cook DJ, Guyatt GH, Salena BJ, Laine LA. Endoscopic therapy for acute nonvariceal upper gastrointestinal hemorrhage: a metaanalysis. Gastroenterology 1992;102:139 148. 2. Savides TJ, Jensen DM. Therapeutic endoscopy for nonvariceal gastrointestinal bleeding. Gastroenterol Clin North Am 2000;29: 465 487. 3. British Society of Gastroenterology Endoscopy Committee. Nonvariceal upper gastrointestinal hemorrhage: guidelines. Gut 2002;51(suppl 4):iv1 iv6.

February 2004 ENDOSCOPIC TREATMENT IN BLEEDING PEPTIC ULCERS 449 4. Feu F, Brullet E, Calvet X, Fernandez-Llamazares J, Guardiola J, Moreno P, Panades A, Salo J, Saperas E, Villanueva C, Planas R. Guidelines for the diagnosis and treatment of acute non-variceal upper gastrointestinal bleeding. Gastroenterol Hepatol 2003;26: 70 85. 5. Forrest JA, Finlayson ND, Shearman DJ. Endoscopy in gastrointestinal bleeding. Lancet 1974;2:394 397. 6. Chalmers TC, Smith H Jr, Blackburn B, Silverman B, Schroeder B, Reitman D, Ambroz A. A method for assessing the quality of a randomized control trial. Control Clin Trials 1981;2:31 49. 7. Clarke M, Oxman AD. Cochrane reviewers handbook 4.0, Oxford, England: The Cochrane Collaboration, 2001. 8. Pescatore P, Jorned P, Borovicka J, Suter W, Delarive J, Wissel P, Meyenberger C, Blum AL, Dorta G. Randomized study of epinephrine vs epinephrine fibrin injection in peptic ulcer bleeding. Interim analysis. Gastrointest Endosc 1999;49:A448. 9. Male GY, Lin H, Pergn CH, Lee FY, Chang FY, Lee SD. A prospective randomized trial comparing injection of epinephrine alone, bipolar electrocoagulation alone and the combination of injection of epinephrine with bipolar electrocoagulation for the arrest of peptic ulcer bleeding. Gastrointest Endosc 1999;49:A452. 10. Gevers AM, De Goede E, Simoens M, Hiele M, Rutgeerts P. A randomized trial comparing injection therapy with hemoclip and with injection combined with hemoclip for bleeding ulcers. Gastrointest Endosc 2002;55:466 469. 11. De Goede E, Gevers AM, Simoens M, Hiele M, Rutgeerts P. Interim results of a randomized trial comparing injection therapy with epinephrine polidocanol 1% vs hemoclip vs. injection combined with hemoclip for bleeding ulcers. Gastrointest Endosc 1998;53:A234. 12. Chung SC, Leung JW, Leung HT, Lo KK, Griffin SM, Li AK. Does adding a sclerosant improve the results of endoscopic epinephrine injection in actively bleeding ulcers? An interim report of a randomized trial. Gastrointest Endosc 1990;36:A38. 13. Chung SC, Leong HT, Chan AC, Yung MY, Leung JW, Li AK. Epinephrine or epinephrine plus alcohol for injection of bleeding ulcers? Gastrointest Endosc 1992;38:A21. 14. Chung IK, Kim HS, Park SH, Lee MH, Kim SJ. Comparison of hemostatic efficacy of endoscopic hemoclip, hypertonic saline epinephrine injection and combined method in treatment for bleeding peptic ulcer. Gastrointest Endosc 1997;45:A256. 15. Chua TS, Fock KM, Ng TM, Khor JL, Teo EK. Treatment of bleeding peptic ulcers with combination of epinephrine injection and endoscopic hemoclip application. Gastroenterology 2001;120: A4370. 16. Buffoli F, Graffeo M, Nicosia F, Gentile C, Cesar P, Rolfi F, Paterlini A. Peptic ulcer bleeding: comparison of two hemostatic procedures. Am J Gastroenterol 2001;96:89 94. 17. Dedeu JM, Villanueva C, Gomez C, Minana JM, Sola-Vera J, Gallego A, Kolle L, Soriano G, Sainz S, Torras X, Balanzó J. Endoscopic injection alone or combined with bipolar electrocoagulation for bleeding peptic ulcer. Preliminary results of a prospective and randomized trial. Gastrointest Endosc 2003; 49:A439. 18. Ell C. Endoclip versus injection therapy in bleeding gastroduodenal ulcers: results of a prospective randomized multicenter trial. Gastroenterology 2002;122:A1542. 19. Balanzo J, Villanueva C, Sainz S, Espinos JC, Mendez C, Guarner C, Vilardell F. Injection therapy of bleeding peptic ulcer. A prospective, randomized trial using epinephrine and thrombin. Endoscopy 1990;22:157 159. 20. Choudari CP, Palmer KR. Endoscopic injection therapy for bleeding peptic ulcer; a comparison of adrenaline alone with adrenaline plus ethanolamine oleate. Gut 1994;35:608 610. 21. Chung IK, Ham JS, Kim HS, Park SH, Lee MH, Kim SJ. Comparison of the hemostatic efficacy of the endoscopic hemoclip method with hypertonic saline-epinephrine injection and a combination of the two for the management of bleeding peptic ulcers. Gastrointest Endosc 1999;49:13 18. 22. Chung SC, Leung JW, Leong HT, Lo KK, Li AK. Adding a sclerosant to endoscopic epinephrine injection in actively bleeding ulcers: a randomized trial. Gastrointest Endosc 1993;39:611 615. 23. Chung SC, Leong HT, Chan AC, Lau JY, Yung MY, Leung JW, Li AK. Epinephrine or epinephrine plus alcohol for injection of bleeding ulcers: a prospective randomized trial. Gastrointest Endosc 1996;43:591 595. 24. Chung SS, Lau JY, Sung JJ, Chan AC, Lai CW, Ng EK, Chan FKL, Yung MY, Li KC. Randomised comparison between adrenaline injection alone and adrenaline injection plus heat probe treatment for actively bleeding ulcers. BMJ 1997;314:1307 1311. 25. Garrido A, Guerrero FJ, Perianes C, Arenas FJ, Palomo S. Inyección local terapéutica en la úlcera gastroduodenal sangrante; estudio comparativo de adrenalina frente adrenalina más agente esclerosante. Rev Esp Enferm Dig 2002;94:395 400. 26. Kubba AK, Murphy W, Palmer KR. Endoscopic injection for bleeding peptic ulcer: a comparison of adrenaline alone with adrenaline plus human thrombin. Gastroenterology 1996;111:623 628. 27. Lee MH, Jung HY, Yang SK, Kim HR, Homg WS, Min YI. Epinephrine alone versus epinephrine plus ethanol for injection therapy of bleeding peptic ulcers. Gastrointest Endosc 1997;45:A283. 28. Lin HJ, Perng CL, Lee SD. Is sclerosant injection mandatory after an epinephrine injection for arrest of peptic ulcer hemorrhage? A prospective, randomised, comparative study. Gut 1993;34:1182 1185. 29. Lin HJ, Tseng GY, Perng CL, Lee FY, Chang FY, Lee SD. Comparison of adrenaline injection and bipolar electrocoagulation for the arrest of peptic ulcer bleeding. Gut 1999;44:715 719. 30. Loizou LA, Bown SG. Endoscopic treatment for bleeding peptic ulcers: randomised comparison of adrenaline injection and adrenaline injection Nd:YAG laser photocoagulation. Gut 1991;32:1100 1103. 31. Pescatore P, Jornod P, Borovicka J, Pantoflickova D, Suter W, Meyenberger C, Blum AL, Dorta G. Epinephrine versus epinephrine plus fibrin glue injection in peptic ulcer bleeding: a prospective randomized trial. Gastrointest Endosc 2002;55:348 353. 32. Sollano JD, Ang VN, Moreno JA. Endoscopic hemostasis of bleeding peptic ulcers: 1:10000 adrenalin injection vs 1:10000 adrenalin 1% aethoxysclerol injection vs. heater probe. Gastroenterol Jpn 1991;26(suppl 3):83 85. 33. Villanueva C, Balanzo J, Espinos JC, Fabrega E, Sainz S, Gonzalez D, Vilardell F. Endoscopic injection therapy of bleeding ulcer: a prospective and randomized comparison of adrenaline alone or with polidocanol. J Clin Gastroenterol 1993;17:195 200. 34. Villanueva C, Balanzo J, Sabat M, Gallego A, Ortiz G, Soriano G, Sainz S, Torras X, Vilardell F. Injection therapy alone or with endoscopic hemoclip for bleeding peptic ulcer. Preliminary results of a randomized trial. Gastrointest Endosc 1996;43:A281. 35. Brullet E, Calvet X, Campo R, Rue M, Catot L, Donoso L. Factors predicting failure of endoscopic injection therapy in bleeding duodenal ulcer. Gastrointest Endosc 1996;43:111 116. 36. Lin HJ, Hsleh YH, Tseng GY, Perng CL, Chang FY, Lee SD. A prospective, randomized trial of large- versus small-volume endoscopic injection of epinephrine for peptic ulcer bleeding. Gastrointest Endosc 2002;55:615 619. 37. Randall GM, Jensen DM, Hirabayashi K, Machicado GA. Controlled study of different sclerosing agents for coagulation of canine gut arteries. Gastroenterology 1989;96:1274 1281. 38. Rutgeerts P, Geobes K, Vantrapen G. Experimental studies of injection therapy for severe nonvariceal bleeding in dogs. Gastroenterology 1989;97:610 621.

450 CALVET ET AL. GASTROENTEROLOGY Vol. 126, No. 2 39. Laine L. Endoscopic therapy for bleeding ulcers: room for improvement? Gastrointest Endosc 2003;57:557 560. 40. Llach J, Bordas JM, Salmeron JM, Panés J, García-Pagán JC, Feu F, Navasa M, Mondelo F, Pique JM, Mas A, Terés J, Rodés J.A prospective randomized trial of heater probe thermocoagulation versus injection therapy in peptic ulcer hemorrhage. Gastrointest Endosc 1996;43:117 120. 41. Simoens M, Rutgeerts P. Non-variceal upper gastrointestinal bleeding. Best Pract Res Clin Gastroenterol 2001;15:121 133. 42. Rollhauser C, Fleischer DE. Current status of endoscopic therapy for ulcer bleeding. Baillieres Best Pract Res Clin Gastroenterol 2000;14:391 410. 43. Gisbert JP, Gonzalez L, Calvet X, Roque M, Gabriel R, Pajares JM. Proton pump inhibitors versus H2-antagonists: a meta-analysis of their efficacy in treating bleeding peptic ulcer. Aliment Pharmacol Ther 2001;15:917 926. 44. Marmo R, Rotondano G, Bianco MA, Piscopo R, Prisco A, Cipolletta L. Outcome of endoscopic treatment for peptic ulcer bleeding: is a second look necessary? A meta-analysis. Gastrointest Endosc 2003;57:62 67. 45. Spiegel BM, Ofman JJ, Woods K, Vakil NB. Minimizing recurrent peptic ulcer hemorrhage after endoscopic hemostasis: the costeffectiveness of competing strategies. Am J Gastroenterol 2003; 98:86 97. Received May 29, 2003. Accepted October 23, 2003. Address requests for reprints to: Xavier Calvet, M.D., Unitat de Malalties Digestives, Hospital de Sabadell, Institut Universitari Parc Tauli, UAB, Parc Tauli, s/n, 08208 Sabadell (Barcelona), Spain. e-mail: xcalvet@cspt.es; fax: (34) 937160646. Supported in part by a grant from the Instituto de Salud Carlos III (C03/02). The authors thank the group that elaborated on the Catalan Guidelines for nonvariceal upper gastrointestinal bleeding, which were of invaluable help to determine the gaps in knowledge on the bleeding peptic ulcer treatment. We are indebted to all participants, especially Dr. Faust Feu as their coordinator. We thank Michael Maudsley for his help with the English.