- Email: [email protected]
A prospective, randomized trial of endoscopic hemoclip placement and distilled water injection for treatment of high-risk bleeding ulcers
A prospective, randomized trial of endoscopic hemoclip placement and distilled water injection for treatment of high-risk bleeding ulcers Yuh-Chyi Chou, MD, Ping-I. Hsu, MD, Kwok-Hung Lai, MD, Ching-Chu Lo, MD, Hoi-Hung Chan, MD, Chi-Pin Lin, MD, Wen-Chi Chen, MD, Chang-Bih Shie, MD, E.-Ming Wang, MD, Nan-Hua Chou, MD, Wency Chen, PhD, Gin-Ho Lo, MD Kaohsiung, Taiwan
Background: Although endoscopic hemoclip therapy is widely used in the treatment of GI bleeding, there are few prospective trials that assess its efficacy. This study evaluated the efficacy and safety of hemoclip placement and distilled water injection for the treatment of high-risk bleeding ulcers. Methods: Seventy-nine patients with major stigmata of ulcer hemorrhage were randomly assigned to either endoscopic hemoclip placement (n = 39) or injection with distilled water (n = 40). Results: Initial hemostasis was achieved in all patients treated with hemoclips and 39 treated by distilled water injection (respectively, 100.0% vs. 97.5%; p = 1.00). Bleeding recurred in 4 and 11 of patients, respectively, in the hemoclip and water injection groups. It occurred significantly more frequently in the injection group (hemoclip, 10.3%; injection, 28.2%; p = 0.04). No major procedurerelated complication occurred in either group. Emergency operations were performed in 5.1% of patients treated with hemoclips versus 12.5% of those in the water injection group (p = 0.43). Hospital days and mortality rate were similar in both groups. Conclusion: Endoscopic hemoclip placement is a safe and effective hemostatic method that is superior to distilled water injection for treatment of bleeding peptic ulcer. (Gastrointest Endosc 2003;57:324-8.)
Bleeding is a common, potentially life-threatening complication of peptic ulcer disease. Endoscopic hemostatic techniques, the treatment of choice, provide better outcomes compared with medical and surgical therapies.1-4 The commonly used methods are local injection (alcohol, dilute epinephrine, hypertonic saline solution) and thermal coagulation (heat probe, monopolar or bipolar electrocoagulation, laser, argon plasma).5-7 Any of the hemostatic techniques are effective in high-risk patients for control of bleeding and reducing the need for surgery. Nonetheless, acute recurrence of bleeding occurs in 10% to 30% of cases.5-10 A disadvantage of Received June 11, 2002. For revision August 20, 2002. Accepted October 12, 2002. Current affiliations: Division of Gastroenterology, Department of Internal Medicine, Surgery, Kaohsiung Veterans General Hospital, Kaohsiung Military General Hospital, National YangMing University, Kaohsiung 813, Taiwan, Department of Medicine, Poo-Sheng Hospital, Ping-Tong 905, Taiwan. Presented in part as an abstract at the annual meeting of the Gastroenterological Society of Taiwan, October 27, 2001, Taipei, Taiwan. Supported by research grant No. VGHKS-90-15 from the Research Foundation of Kaohsiung Veterans General Hospital. Reprint requests: Ping-I. Hsu, MD, Division of Gastroenterology, Department of Internal Medicine, Kaohsiung Veterans General Hospital, 386, Ta-Chung 1st Rd., Kaohsiung 813, Taiwan. Copyright © 2003 by the American Society for Gastrointestinal Endoscopy 0016-5107/2003/$30.00 + 0 doi:10.1067/mge.2003.103 324
GASTROINTESTINAL ENDOSCOPY
sclerosant injection and thermal coagulation is the potential for excessive tissue injury leading to necrosis and perforation.11-14 With epinephrine injection, there is also the possibility of inducing a hypertensive crisis in patients with underlying cardiovascular disease.15 Endoscopic metallic clip application to control bleeding from peptic ulcer was introduced by Hayashi et al.16 in 1975. Initial results were discouraging because of the complexity of the application technique and low rates of clip retention. Technical improvements in both the clip and the clip delivery system have been introduced, the modified clip having better grasping capability while causing less tissue trauma.17,18 However, experience with the endoscopic use of hemoclips for bleeding peptic ulcers is still limited.19 Only 2 prospective randomized trials comparing hemoclip therapy with traditional endoscopic methods have been published.20,21 Chung et al.21 showed that the hemoclip method is an effective hemostatic procedure and is safer than injection of hypertonic saline-epinephrine (HSE) in the management of patients with ulcer bleeding. A large hematoma with a newly developed ulcer at the periphery of the original ulcer may arise after HSE injection. Our group demonstrated that local tamponade with distilled water is effective and safe for the management of bleeding peptic ulcer.8 Theoretically, local injection of distilled water to control bleeding VOLUME 57, NO. 3, 2003
Bleeding ulcers: endoscopic hemoclipping vs. distilled water injection
from peptic ulcers causes less tissue injury than injection of epinephrine and carries no risk of inducing a hypertensive crisis. The present randomized study prospectively compared the efficacy of hemoclip therapy and distilled water injection for control of bleeding from peptic ulcer with respect to the outcomes of initial hemostasis, recurrent bleeding, need for surgical intervention, and mortality. PATIENTS AND METHODS Patients All patients hospitalized from July 1999 to July 2001 because of hematemesis, melena, or hematochezia were considered for inclusion in the study. Patients were included if emergency endoscopy disclosed a peptic ulcer with an active bleeding visible vessel (BVV; spurting or oozing) or a nonbleeding visible vessel (NBVV). A NBVV was defined as a raised red or bluish-red hemispheric lesion protruding from the ulcer base, without actively bleeding. Exclusion criteria were (1) the presence of another possible bleeding site (e.g., esophageal varices, gastric cancer), (2) coexistence of an acute significant illness (e.g., sepsis, stroke, acute myocardial infarction, acute respiratory failure, acute surgical abdomen), and, (3) systemic bleeding tendency (e.g., platelet <50,000/mm3, prothrombin time >3 seconds, treatment with an anticoagulant). Methods Upper endoscopy was performed within 24 hours of hospital admission. No patients had undergone prior endoscopic treatment of any kind. Endoscopy was performed with various standard upper endoscopes. Gastric lavage was carried out before endoscopy. Ulcers with stigmata of bleeding were irrigated with water by means of the accessory channel of the endoscope. Patients were randomly assigned to either hemoclip therapy or distilled water injection. Randomization was carried out during endoscopy by an individual not directly involved in the study or care of the patient who opened sealed and numbered envelopes containing the treatment assignments, which had been randomized with a computer program. All patients or relatives gave written consent before treatment. The study protocol was approved by the ethics committee of our hospital. Endoscopic hemoclip therapy was performed with stainless steel hemoclips (MD 850, Olympus Optical Co., Ltd., Tokyo, Japan) with prongs that measured 6 mm in length and 12 mm in width. Clips were applied with a clip application device (HX-5LR, Olympus) passed through the 2.8-mm diameter accessory channel of a standard endoscope. Hemoclips were individually loaded and deployed as described previously.19 Distilled water injection therapy was performed with 0.5-mL to 2-mL boluses of distilled water to a maximum of 20 mL. Injections were placed in all 4 quadrants surrounding the bleeding point and then into the vessel directly. Once hemostasis was achieved, the bleeding site was observed for at least 5 minutes. Initial hemostasis was defined as no endoscopic evidence of bleeding during 5 VOLUME 57, NO. 3, 2003
Y-C Chou, P-I Hsu, K-H Lai, et al.
Table 1. Clinical characteristics of study groups Hemoclip (n = 39) Age (y) (SD) 63.2 (16.9) Gender (M:F) 31:8 Smoking 10 (25.6%) Alcohol abuse 5 (12.8%) NSAID use 6 (15.4%) Hypovolemic shock 12 (30.8%) Hemoglobin (g/dL) (SD) 9.2 (2.3) Endoscopic findings Ulcer size (cm) (SD) 1.2 (0.7) Ulcer site Stomach 19 Duodenum 20 Bleeding type Spurting vessel 5 Oozing vessel 15 Nonbleeding visible 19 vessel
Distilled water injection (n = 40) p Value 64.8 (15.0) 31:9 7 (17.5%) 5 (12.5%) 3 (7.5%) 8 (20.0%) 9.8 (2.7) 1.0 (0.5)
0.65 0.83 0.38 0.97 0.31 0.24 0.23 0.28 0.91
19 21 0.94 5 14 21
NSAID, Nonsteroidal anti-inflammatory drugs.
minutes of observation after therapy. After therapeutic endoscopy patients were observed closely. While the patient was in the hospital, treatment included partial parenteral nutrition and intravenous administration of ranitidine (50 mg every 8 hours) for 2 days. After the 48hour observation period, patients were given a soft diet for 48 hours and then a regular diet. Orally administered ranitidine (150 mg twice daily) was prescribed upon resumption of oral intake. Hemoglobin level was checked daily for 3 days and blood transfusion(s) given if the hemoglobin concentration decreased below 8 gm/dL or there was deterioration in vital signs. A clinician, independent of the endoscopist, observed patients for evidence of recurrent bleeding. This was defined as recurrent hemorrhage during an 8-week observation period apart from persistent bleeding during the initial attempt to induce hemostasis. Evidence of recurrent bleeding included the following: (1) new hematemesis, (2) aspiration of fresh blood from a nasogastric tube, or (3) continuous melena with a pulse rate greater than 100 beats/minute, a decrease in systolic blood pressure exceeding 30 mm Hg, or a decrease in hemoglobin of at least 2 gm/dL. When recurrent bleeding was suspected, endoscopy was performed immediately and the same hemostatic method as at randomization was used. If hemostasis could not be achieved, surgery was performed. After discharge, patients were treated with ranitidine (150 mg twice daily) for up to 8 weeks and were requested to return to the outpatient clinic 14 days, 4 weeks, and 8 weeks after initial hemostasis. Statistics Sample size was calculated based on previous experiences.6,21 The rate of recurrent bleeding after hemoclip hemostasis was 5%21 and after distilled water injection 29%.6 A sample size of 38 was thus required for each group to achieve a statistical power of 80% at 10% type I error. GASTROINTESTINAL ENDOSCOPY
325
Y-C Chou, P-I Hsu, K-H Lai, et al.
Bleeding ulcers: endoscopic hemoclipping vs. distilled water injection
Table 2. Clinical outcomes for hemoclip and injection groups Hemoclipping (n = 39) Initial hemostasis Recurrent bleeding Emergency operation Hospital days (SD) Mortality
39 4 2 9.3 1
(100.0%) (10.3%) (5.1%) (8.1) (2.6%)
Distilled water injection (n = 40) p Value 39 12 5 8.0 2
(97.5%) (28.2%) (12.5%) (5.1) (5.0%)
1.00 0.04* 0.43 0.37 1.00
*p < 0.05.
The Student t test was used to compare the mean value of continuous variables. The chi-square test, with or without Yates correction for continuity, and the Fisher exact test were used when appropriate to compare the location of the bleeding lesions, initial hemostasis, recurrent bleeding, emergency surgery, and mortality between groups. A p value of <0.05 was considered significant.
RESULTS Seventy-nine patients were included in the study and were randomly assigned to undergo either endoscopic hemoclip application (39 patients) or distilled water injection (40 patients). Data on clinical characteristics of patients at entry are summarized in Table 1. The 2 groups were comparable with regard to clinical features, site and size of ulcers, and bleeding severity. At index endoscopy, 49% of randomized patients had BVVs (spurting, 12%; oozing, 37%) and 51% had NBVVs. All patients were followed through the 8-week period after initial endoscopy. Treatment results are shown in Table 2. A total of 82 clips were applied in the hemoclip group (mean 2.1 clips/patient, range 1-4). Initial hemostasis was achieved in all patients in the hemoclip group and 39 patients treated by distilled water injection (respectively, 100% vs. 97.5%; p = 1.00). Bleeding recurred in 4 patients in the hemoclip group and 11 in the injection group, occurring significantly more frequently in the injection group (hemoclip, 10.3%; injection, 28.2%; p = 0.04). All of the 16 patients in whom initial treatment failed had some form of subsequent intervention, either endoscopic alone (hemoclip group, 2; injection group, 7) or surgical (hemoclip group, 2; injection group, 5). Two of 3 deaths occurred in patients with uncontrollable bleeding. One in the injection group underwent surgery but died of septic shock. Another (hemoclip group) developed recurrent bleeding and died in hypovolemic shock and respiratory failure. The third patient (injection group) did not have recurrent bleeding, but died because of a combination of respiratory failure and myocardial infarction. There were no significant differences between the 326
GASTROINTESTINAL ENDOSCOPY
groups in length of hospital stay, rate of emergency surgery or mortality. No major procedure-related complications (e.g., perforation or worsening of bleeding as a result of the hemostatic procedure) were observed in either group. DISCUSSION Ulcer bleeding is a potentially life threatening condition. Injection techniques are commonly employed for endoscopic treatment, with dilute epinephrine being among the most frequently used solutions. Pinkas et al.22 evaluated injection of normal saline solution versus a 1:10,000 solution of epinephrine in a canine model of gastric ulcers with bleeding arteries. The rate of bleeding from 11 to 30 minutes was reduced to a greater degree with epinephrine than with normal saline solution, but not significantly so. The results of injection therapy in animal models, however, have not been consistent with clinical outcomes for commonly used agents. The hemostatic effects of normal saline solution or distilled water with dilute epinephrine have not been shown to be significantly different.5,8 Additionally, Randall et al.23 demonstrated that normal saline solution caused less local tissue injury than epinephrine. If effective, injection of distilled water or saline would be an ideal therapeutic option because of wide availability, low cost, reduced tissue injury, and absence of systemic side effects. Endoscopic hemostasis by application of hemoclips was originally described by Hayashi et al.16 in 1975, but abandoned because of its technical complexity. Hachisu17 and Hachisu et al.18 improved the mechanics of the clip applicator with further modifications being made in 1996.24 The modified clipdevice has the advantages of being rotatable and it can withstand autoclaving.24 Theoretically, endoscopic hemoclip application is the optimal method when a visible vessel is present. Hemostasis may be definitive when the vessel is properly clamped, and the targeted application of the clip, which has no thermal or chemical effect, nearly eliminates the risk of tissue injury. In an animal study comparing mechanical, injection, and thermal methods of hemostasis, only mechanical methods were effective for control of bleeding from vessels greater than 2 mm in diameter.25 The current study prospectively compared the hemostatic effects of endoscopic hemoclip application and distilled water injection for bleeding peptic ulcers with BVVs or NBVVs. There were no significant differences in initial hemostasis rates between the two patient groups, but bleeding recurred significantly less frequently in the hemoclip group (respectively, 10.3% vs. 28.2%; p = 0.04). The results of the VOLUME 57, NO. 3, 2003
Bleeding ulcers: endoscopic hemoclipping vs. distilled water injection
present study are consistent with those from a retrospective study by Nagayama et al.26 that demonstrated comparable results for initial hemostasis rates between endoscopic hemoclip application and ethanol injection (96% vs. 96%). However, the rate of recurrent bleeding was lower in the hemoclip group compared with the injection group (15% vs. 29%). The results of Nagayama et al.26 and those of the current investigation suggest that hemoclip application successfully controls bleeding from an ulcer vessel for a longer time than local injection. The tamponade effect of distilled water or other sclerosants on a bleeding vessel may decrease over time as the injected solution is gradually absorbed. Chung et al.21 conducted a randomized trial of endoscopic hemostasis involving 124 patients with active bleeding or visible vessels at endoscopy. Permanent hemostasis was achieved in 95% of patients in the hemoclip group, 85% in HSE injection group, and 95% in combined therapy group. These small differences were not significant. However, all of 3 procedure-related complications (exaggerated bleeding, submucosal hematoma) occurred in the patients treated by HSE injection. Cipolletta et al.20 randomly assigned 112 patients with a BVV or NBVV to either endoscopic hemoclip application or heat probe thermocoagulation, and found the 2 techniques to be equally effective in achieving initial hemostasis. The frequency of recurrent bleeding and the need for emergency surgery were, however, significantly lower for the hemoclip technique compared with the thermal method. No complications occurred in either treatment group in the present study. Thermal coagulation and epinephrine or ethanol injection have been reported to induce complications such as exacerbation of ulcers and even perforation in variable numbers of patients.11,12 Endoscopic hemoclip application induces less tissue injury, and thus far no complications have been reported as associated with this endoscopic treatment method. Endoscopic hemoclip application does have some drawbacks. Loading the clip on the application device was time consuming and application was technically difficult when the angle of approach was tangential to a lesion, as in the cardia of the stomach, the posterior wall of the proximal body, and the duodenal bulb.19,20 It is therefore advantageous to have at least 2 application devices available to avoid delays in the application of successive hemoclips.19 A preloaded hemoclip device has been developed (Olympus) that is rotatable and eliminates the time consumed in loading the clips. To overcome problems that arise when the approach to the lesion is tangential, cap-fitted endoscopeassisted hemoclip application has been described.27 VOLUME 57, NO. 3, 2003
Y-C Chou, P-I Hsu, K-H Lai, et al.
By pushing against the lesion with the edge of the transparent cap, a more en-face approach is possible. To our knowledge, the present study is the first prospective randomized trial comparing the hemostatic effects of endoscopic hemoclip application and distilled water injection in patients with bleeding ulcers. The results confirm that hemoclip application is an effective and safe hemostatic method that is superior to distilled water injection for the management of bleeding peptic ulcers. ACKNOWLEDGMENTS The authors express their deep appreciation to Miss Yu-Shan Chen, and Min-Rong Huang for their assistance in the clinical follow-up of the patients. REFERENCES 1. Laine L. Multipolar electrocoagulation vs. injection therapy in the treatment of bleeding peptic ulcers. Gastroenterology 1990;99:1303-6. 2. Cook DJ, Guyatt GH, Salena BJ, Laine LA. Endoscopic therapy for acute nonvariceal upper gastrointestinal hemorrhage. A meta-analysis. Gastroenterology 1992;102:139-48. 3. Gralnek IM, Jensen DM, Gornbein J, Kovacs TO, Juthaba R, Freeman ML, et al. Clinical and economic outcomes of individuals with sever peptic ulcer hemorrhage and nonbleeding visible vessel: an analysis of two prospective clinical trials. Am J Gastroenterol 1998;93:2047-56. 4. Laine L, Peterson WL. Bleeding peptic ulcer. New Engl J Med 1994;331:717-27. 5. Lin HJ, Perng CL, Lee FY, Chan CY, Huang ZC, Lee SD, et al. Endoscopic injection for arrest of peptic ulcer hemorrhage: final results of a prospective, randomized comparative trial. Gastrointest Endosc 1993;39:15-9. 6. Laine L, Estrada R. Randomized trial of normal saline solution injection versus bipolar electrocoagulation for treatment of patients with high-risk bleeding ulcers: is local tamponade enough? Gastrointest Endosc 2001;55:6-10. 7. Cipolletta L, Bianco MA, Rotondano G, Piscopo R, Prisco A, Garofano ML. Prospective comparison of argon plasma coagulator and heater probe in the endoscopic treatment of major peptic ulcer bleeding. Gastrointest Endosc 1998;48:191-5. 8. Lai KH, Peng SN, Guo WS, Lee FY, Chang FY, Malik U, et al. Endoscopic injection for the treatment of bleeding ulcers: local tamponade or drug effect? Endoscopy 1994;26:338-41. 9. Jaramillo JL, Carmona C, Galvez C, de la Mata M, Mino G. Efficacy of the heater probe in peptic ulcer with a non-bleeding visible vessel. A controlled randomized study. Gut 1993;34:1502-6. 10. Llach J, Bordas JM, Salmeron JM, Panes J, Garcia-Pagan JC, Feu F, et al. A prospective randomized trial of heater probe thermocoagulation versus injection therapy in peptic ulcer hemorrhage. Gastrointest Endosc 1996;43:117-20. 11. Rutgeerts P, Geboes K, Vantrappen G. Tissue damage produced by hemostatic injections [abstract]. Gastrointest Endosc 1986;32:AB179. 12. Bedford RA, van Stolk R, Sivak MV Jr, Chung RS, Van Dam J. Gastric perforation after endoscopic treatment of a Dieulafoy’s lesion. Am J Gastroenterol 1992;87:244-7. 13. Dell’Abate P, Spaggiari L, Carboynani P, Soliani P, Karake I, Foggi E. An unusual complication of sclerotherapy. Endoscopy 1991;23:352-3. 14. Loperfido S, Patelli G, La Torre L. Extensive necrosis of gasGASTROINTESTINAL ENDOSCOPY
327
Y-C Chou, P-I Hsu, K-H Lai, et al.
15.
16.
17. 18.
19.
20.
21.
tric mucosa following injection therapy of bleeding peptic ulcer. Endoscopy 1990;22:285-6. Rutgeers P, Geboes K, Vantrappen G. Experimental studies of injection therapy for severe nonvariceal bleeding in dogs. Gastroenterology 1989;97:610-21. Hayashi T, Yonezawa M, Kawabara T. The study on staunch clip for the treatment by endoscopy. Gastroenterol Endosc 1975;17:92-101. Hachisu T. Evaluation of endoscopic hemostasis using an improved clipping apparatus. Surg Endosc 1988;2:13-7. Hachisu T, Miyazaki S, Hamaguchi K. Endoscopic clipping marking of lesions using the newly developed HX-3L clip. Surg Endosc 1989;3:142-7. Soehendra N, Sriram PVJ, Ponchon T, Chung SCS. Hemostatic clip in gastrointestinal bleeding. Endoscopy 2001;33: 172-80. Cipolletta L, Bianco MA, Marmo R, Rotondano G, Piscopo R, Vinginani AM, et al. Endoclips versus heater probe in preventing early recurrent bleeding from peptic ulcer: a prospective and randomized trial. Gastrointest Endosc 2001;53:147-51. 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 injec-
Bleeding ulcers: endoscopic hemoclipping vs. distilled water injection
22.
23.
24.
25.
26.
27.
tion and a combination of the two for the management of bleeding peptic ulcers. Gastrointest Endosc 1999;49:13-8. Pinkas H, McAllister E, Norman J, Robinson B, Brady PG, Dawson PJ. Prolonged evaluation of epinephrine and normal saline solution injections in an acute ulcer model with a single bleeding artery. Gastrointest Endosc 1995;42:51-5. Randall GM, Jensen DM, Hirabayashi K, Machicado GA. Controlled study of different sclerosing agents for coagulation of canine gut arteries. Gastroenterology 1989;96:1274-81. Hachisu T, Yamada H, Satoh SI, Kouzu T. Endoscopic clipping with a new rotatable clip-device and a long clip. Dig Endosc 1996;8:127-33. Hepworth CC, Kadirkamanathan SS, Gong G, Swain CP. A randomized controlled comparison of injection, thermal and mechanical endoscopic methods of hemostasis on mesenteric vessels. Gut 1998;42:462-9. Nagayama K, Tazawa J, Sakai Y, Miyasaka Y, Yu SH, Sakuma I, et al. Efficacy of endoscopic clipping for bleeding gastroduodenal ulcer: comparison with topical ethanol injection. Am J Gastroenterol 1999;94:2897-901. Lin LF, Siauw CP, Ho KS, Tung JC. Endoscopic hemoclip treatment of gastrointestinal bleeding. Chang Gung Med J 2001;24:307-12.
Moving? To ensure continued service please notify us of a change of address at least 6 weeks before your move. Phone Subscription Services at 800-654-2452 (outside the U.S. call 407-345-4000) or fax your information to 407-363-9661.
328
GASTROINTESTINAL ENDOSCOPY
VOLUME 57, NO. 3, 2003
Background: Although endoscopic hemoclip therapy is widely used in the treatment of GI bleeding, there are few prospective trials that assess its efficacy. This study evaluated the efficacy and safety of hemoclip placement and distilled water injection for the treatment of high-risk bleeding ulcers. Methods: Seventy-nine patients with major stigmata of ulcer hemorrhage were randomly assigned to either endoscopic hemoclip placement (n = 39) or injection with distilled water (n = 40). Results: Initial hemostasis was achieved in all patients treated with hemoclips and 39 treated by distilled water injection (respectively, 100.0% vs. 97.5%; p = 1.00). Bleeding recurred in 4 and 11 of patients, respectively, in the hemoclip and water injection groups. It occurred significantly more frequently in the injection group (hemoclip, 10.3%; injection, 28.2%; p = 0.04). No major procedurerelated complication occurred in either group. Emergency operations were performed in 5.1% of patients treated with hemoclips versus 12.5% of those in the water injection group (p = 0.43). Hospital days and mortality rate were similar in both groups. Conclusion: Endoscopic hemoclip placement is a safe and effective hemostatic method that is superior to distilled water injection for treatment of bleeding peptic ulcer. (Gastrointest Endosc 2003;57:324-8.)
Bleeding is a common, potentially life-threatening complication of peptic ulcer disease. Endoscopic hemostatic techniques, the treatment of choice, provide better outcomes compared with medical and surgical therapies.1-4 The commonly used methods are local injection (alcohol, dilute epinephrine, hypertonic saline solution) and thermal coagulation (heat probe, monopolar or bipolar electrocoagulation, laser, argon plasma).5-7 Any of the hemostatic techniques are effective in high-risk patients for control of bleeding and reducing the need for surgery. Nonetheless, acute recurrence of bleeding occurs in 10% to 30% of cases.5-10 A disadvantage of Received June 11, 2002. For revision August 20, 2002. Accepted October 12, 2002. Current affiliations: Division of Gastroenterology, Department of Internal Medicine, Surgery, Kaohsiung Veterans General Hospital, Kaohsiung Military General Hospital, National YangMing University, Kaohsiung 813, Taiwan, Department of Medicine, Poo-Sheng Hospital, Ping-Tong 905, Taiwan. Presented in part as an abstract at the annual meeting of the Gastroenterological Society of Taiwan, October 27, 2001, Taipei, Taiwan. Supported by research grant No. VGHKS-90-15 from the Research Foundation of Kaohsiung Veterans General Hospital. Reprint requests: Ping-I. Hsu, MD, Division of Gastroenterology, Department of Internal Medicine, Kaohsiung Veterans General Hospital, 386, Ta-Chung 1st Rd., Kaohsiung 813, Taiwan. Copyright © 2003 by the American Society for Gastrointestinal Endoscopy 0016-5107/2003/$30.00 + 0 doi:10.1067/mge.2003.103 324
GASTROINTESTINAL ENDOSCOPY
sclerosant injection and thermal coagulation is the potential for excessive tissue injury leading to necrosis and perforation.11-14 With epinephrine injection, there is also the possibility of inducing a hypertensive crisis in patients with underlying cardiovascular disease.15 Endoscopic metallic clip application to control bleeding from peptic ulcer was introduced by Hayashi et al.16 in 1975. Initial results were discouraging because of the complexity of the application technique and low rates of clip retention. Technical improvements in both the clip and the clip delivery system have been introduced, the modified clip having better grasping capability while causing less tissue trauma.17,18 However, experience with the endoscopic use of hemoclips for bleeding peptic ulcers is still limited.19 Only 2 prospective randomized trials comparing hemoclip therapy with traditional endoscopic methods have been published.20,21 Chung et al.21 showed that the hemoclip method is an effective hemostatic procedure and is safer than injection of hypertonic saline-epinephrine (HSE) in the management of patients with ulcer bleeding. A large hematoma with a newly developed ulcer at the periphery of the original ulcer may arise after HSE injection. Our group demonstrated that local tamponade with distilled water is effective and safe for the management of bleeding peptic ulcer.8 Theoretically, local injection of distilled water to control bleeding VOLUME 57, NO. 3, 2003
Bleeding ulcers: endoscopic hemoclipping vs. distilled water injection
from peptic ulcers causes less tissue injury than injection of epinephrine and carries no risk of inducing a hypertensive crisis. The present randomized study prospectively compared the efficacy of hemoclip therapy and distilled water injection for control of bleeding from peptic ulcer with respect to the outcomes of initial hemostasis, recurrent bleeding, need for surgical intervention, and mortality. PATIENTS AND METHODS Patients All patients hospitalized from July 1999 to July 2001 because of hematemesis, melena, or hematochezia were considered for inclusion in the study. Patients were included if emergency endoscopy disclosed a peptic ulcer with an active bleeding visible vessel (BVV; spurting or oozing) or a nonbleeding visible vessel (NBVV). A NBVV was defined as a raised red or bluish-red hemispheric lesion protruding from the ulcer base, without actively bleeding. Exclusion criteria were (1) the presence of another possible bleeding site (e.g., esophageal varices, gastric cancer), (2) coexistence of an acute significant illness (e.g., sepsis, stroke, acute myocardial infarction, acute respiratory failure, acute surgical abdomen), and, (3) systemic bleeding tendency (e.g., platelet <50,000/mm3, prothrombin time >3 seconds, treatment with an anticoagulant). Methods Upper endoscopy was performed within 24 hours of hospital admission. No patients had undergone prior endoscopic treatment of any kind. Endoscopy was performed with various standard upper endoscopes. Gastric lavage was carried out before endoscopy. Ulcers with stigmata of bleeding were irrigated with water by means of the accessory channel of the endoscope. Patients were randomly assigned to either hemoclip therapy or distilled water injection. Randomization was carried out during endoscopy by an individual not directly involved in the study or care of the patient who opened sealed and numbered envelopes containing the treatment assignments, which had been randomized with a computer program. All patients or relatives gave written consent before treatment. The study protocol was approved by the ethics committee of our hospital. Endoscopic hemoclip therapy was performed with stainless steel hemoclips (MD 850, Olympus Optical Co., Ltd., Tokyo, Japan) with prongs that measured 6 mm in length and 12 mm in width. Clips were applied with a clip application device (HX-5LR, Olympus) passed through the 2.8-mm diameter accessory channel of a standard endoscope. Hemoclips were individually loaded and deployed as described previously.19 Distilled water injection therapy was performed with 0.5-mL to 2-mL boluses of distilled water to a maximum of 20 mL. Injections were placed in all 4 quadrants surrounding the bleeding point and then into the vessel directly. Once hemostasis was achieved, the bleeding site was observed for at least 5 minutes. Initial hemostasis was defined as no endoscopic evidence of bleeding during 5 VOLUME 57, NO. 3, 2003
Y-C Chou, P-I Hsu, K-H Lai, et al.
Table 1. Clinical characteristics of study groups Hemoclip (n = 39) Age (y) (SD) 63.2 (16.9) Gender (M:F) 31:8 Smoking 10 (25.6%) Alcohol abuse 5 (12.8%) NSAID use 6 (15.4%) Hypovolemic shock 12 (30.8%) Hemoglobin (g/dL) (SD) 9.2 (2.3) Endoscopic findings Ulcer size (cm) (SD) 1.2 (0.7) Ulcer site Stomach 19 Duodenum 20 Bleeding type Spurting vessel 5 Oozing vessel 15 Nonbleeding visible 19 vessel
Distilled water injection (n = 40) p Value 64.8 (15.0) 31:9 7 (17.5%) 5 (12.5%) 3 (7.5%) 8 (20.0%) 9.8 (2.7) 1.0 (0.5)
0.65 0.83 0.38 0.97 0.31 0.24 0.23 0.28 0.91
19 21 0.94 5 14 21
NSAID, Nonsteroidal anti-inflammatory drugs.
minutes of observation after therapy. After therapeutic endoscopy patients were observed closely. While the patient was in the hospital, treatment included partial parenteral nutrition and intravenous administration of ranitidine (50 mg every 8 hours) for 2 days. After the 48hour observation period, patients were given a soft diet for 48 hours and then a regular diet. Orally administered ranitidine (150 mg twice daily) was prescribed upon resumption of oral intake. Hemoglobin level was checked daily for 3 days and blood transfusion(s) given if the hemoglobin concentration decreased below 8 gm/dL or there was deterioration in vital signs. A clinician, independent of the endoscopist, observed patients for evidence of recurrent bleeding. This was defined as recurrent hemorrhage during an 8-week observation period apart from persistent bleeding during the initial attempt to induce hemostasis. Evidence of recurrent bleeding included the following: (1) new hematemesis, (2) aspiration of fresh blood from a nasogastric tube, or (3) continuous melena with a pulse rate greater than 100 beats/minute, a decrease in systolic blood pressure exceeding 30 mm Hg, or a decrease in hemoglobin of at least 2 gm/dL. When recurrent bleeding was suspected, endoscopy was performed immediately and the same hemostatic method as at randomization was used. If hemostasis could not be achieved, surgery was performed. After discharge, patients were treated with ranitidine (150 mg twice daily) for up to 8 weeks and were requested to return to the outpatient clinic 14 days, 4 weeks, and 8 weeks after initial hemostasis. Statistics Sample size was calculated based on previous experiences.6,21 The rate of recurrent bleeding after hemoclip hemostasis was 5%21 and after distilled water injection 29%.6 A sample size of 38 was thus required for each group to achieve a statistical power of 80% at 10% type I error. GASTROINTESTINAL ENDOSCOPY
325
Y-C Chou, P-I Hsu, K-H Lai, et al.
Bleeding ulcers: endoscopic hemoclipping vs. distilled water injection
Table 2. Clinical outcomes for hemoclip and injection groups Hemoclipping (n = 39) Initial hemostasis Recurrent bleeding Emergency operation Hospital days (SD) Mortality
39 4 2 9.3 1
(100.0%) (10.3%) (5.1%) (8.1) (2.6%)
Distilled water injection (n = 40) p Value 39 12 5 8.0 2
(97.5%) (28.2%) (12.5%) (5.1) (5.0%)
1.00 0.04* 0.43 0.37 1.00
*p < 0.05.
The Student t test was used to compare the mean value of continuous variables. The chi-square test, with or without Yates correction for continuity, and the Fisher exact test were used when appropriate to compare the location of the bleeding lesions, initial hemostasis, recurrent bleeding, emergency surgery, and mortality between groups. A p value of <0.05 was considered significant.
RESULTS Seventy-nine patients were included in the study and were randomly assigned to undergo either endoscopic hemoclip application (39 patients) or distilled water injection (40 patients). Data on clinical characteristics of patients at entry are summarized in Table 1. The 2 groups were comparable with regard to clinical features, site and size of ulcers, and bleeding severity. At index endoscopy, 49% of randomized patients had BVVs (spurting, 12%; oozing, 37%) and 51% had NBVVs. All patients were followed through the 8-week period after initial endoscopy. Treatment results are shown in Table 2. A total of 82 clips were applied in the hemoclip group (mean 2.1 clips/patient, range 1-4). Initial hemostasis was achieved in all patients in the hemoclip group and 39 patients treated by distilled water injection (respectively, 100% vs. 97.5%; p = 1.00). Bleeding recurred in 4 patients in the hemoclip group and 11 in the injection group, occurring significantly more frequently in the injection group (hemoclip, 10.3%; injection, 28.2%; p = 0.04). All of the 16 patients in whom initial treatment failed had some form of subsequent intervention, either endoscopic alone (hemoclip group, 2; injection group, 7) or surgical (hemoclip group, 2; injection group, 5). Two of 3 deaths occurred in patients with uncontrollable bleeding. One in the injection group underwent surgery but died of septic shock. Another (hemoclip group) developed recurrent bleeding and died in hypovolemic shock and respiratory failure. The third patient (injection group) did not have recurrent bleeding, but died because of a combination of respiratory failure and myocardial infarction. There were no significant differences between the 326
GASTROINTESTINAL ENDOSCOPY
groups in length of hospital stay, rate of emergency surgery or mortality. No major procedure-related complications (e.g., perforation or worsening of bleeding as a result of the hemostatic procedure) were observed in either group. DISCUSSION Ulcer bleeding is a potentially life threatening condition. Injection techniques are commonly employed for endoscopic treatment, with dilute epinephrine being among the most frequently used solutions. Pinkas et al.22 evaluated injection of normal saline solution versus a 1:10,000 solution of epinephrine in a canine model of gastric ulcers with bleeding arteries. The rate of bleeding from 11 to 30 minutes was reduced to a greater degree with epinephrine than with normal saline solution, but not significantly so. The results of injection therapy in animal models, however, have not been consistent with clinical outcomes for commonly used agents. The hemostatic effects of normal saline solution or distilled water with dilute epinephrine have not been shown to be significantly different.5,8 Additionally, Randall et al.23 demonstrated that normal saline solution caused less local tissue injury than epinephrine. If effective, injection of distilled water or saline would be an ideal therapeutic option because of wide availability, low cost, reduced tissue injury, and absence of systemic side effects. Endoscopic hemostasis by application of hemoclips was originally described by Hayashi et al.16 in 1975, but abandoned because of its technical complexity. Hachisu17 and Hachisu et al.18 improved the mechanics of the clip applicator with further modifications being made in 1996.24 The modified clipdevice has the advantages of being rotatable and it can withstand autoclaving.24 Theoretically, endoscopic hemoclip application is the optimal method when a visible vessel is present. Hemostasis may be definitive when the vessel is properly clamped, and the targeted application of the clip, which has no thermal or chemical effect, nearly eliminates the risk of tissue injury. In an animal study comparing mechanical, injection, and thermal methods of hemostasis, only mechanical methods were effective for control of bleeding from vessels greater than 2 mm in diameter.25 The current study prospectively compared the hemostatic effects of endoscopic hemoclip application and distilled water injection for bleeding peptic ulcers with BVVs or NBVVs. There were no significant differences in initial hemostasis rates between the two patient groups, but bleeding recurred significantly less frequently in the hemoclip group (respectively, 10.3% vs. 28.2%; p = 0.04). The results of the VOLUME 57, NO. 3, 2003
Bleeding ulcers: endoscopic hemoclipping vs. distilled water injection
present study are consistent with those from a retrospective study by Nagayama et al.26 that demonstrated comparable results for initial hemostasis rates between endoscopic hemoclip application and ethanol injection (96% vs. 96%). However, the rate of recurrent bleeding was lower in the hemoclip group compared with the injection group (15% vs. 29%). The results of Nagayama et al.26 and those of the current investigation suggest that hemoclip application successfully controls bleeding from an ulcer vessel for a longer time than local injection. The tamponade effect of distilled water or other sclerosants on a bleeding vessel may decrease over time as the injected solution is gradually absorbed. Chung et al.21 conducted a randomized trial of endoscopic hemostasis involving 124 patients with active bleeding or visible vessels at endoscopy. Permanent hemostasis was achieved in 95% of patients in the hemoclip group, 85% in HSE injection group, and 95% in combined therapy group. These small differences were not significant. However, all of 3 procedure-related complications (exaggerated bleeding, submucosal hematoma) occurred in the patients treated by HSE injection. Cipolletta et al.20 randomly assigned 112 patients with a BVV or NBVV to either endoscopic hemoclip application or heat probe thermocoagulation, and found the 2 techniques to be equally effective in achieving initial hemostasis. The frequency of recurrent bleeding and the need for emergency surgery were, however, significantly lower for the hemoclip technique compared with the thermal method. No complications occurred in either treatment group in the present study. Thermal coagulation and epinephrine or ethanol injection have been reported to induce complications such as exacerbation of ulcers and even perforation in variable numbers of patients.11,12 Endoscopic hemoclip application induces less tissue injury, and thus far no complications have been reported as associated with this endoscopic treatment method. Endoscopic hemoclip application does have some drawbacks. Loading the clip on the application device was time consuming and application was technically difficult when the angle of approach was tangential to a lesion, as in the cardia of the stomach, the posterior wall of the proximal body, and the duodenal bulb.19,20 It is therefore advantageous to have at least 2 application devices available to avoid delays in the application of successive hemoclips.19 A preloaded hemoclip device has been developed (Olympus) that is rotatable and eliminates the time consumed in loading the clips. To overcome problems that arise when the approach to the lesion is tangential, cap-fitted endoscopeassisted hemoclip application has been described.27 VOLUME 57, NO. 3, 2003
Y-C Chou, P-I Hsu, K-H Lai, et al.
By pushing against the lesion with the edge of the transparent cap, a more en-face approach is possible. To our knowledge, the present study is the first prospective randomized trial comparing the hemostatic effects of endoscopic hemoclip application and distilled water injection in patients with bleeding ulcers. The results confirm that hemoclip application is an effective and safe hemostatic method that is superior to distilled water injection for the management of bleeding peptic ulcers. ACKNOWLEDGMENTS The authors express their deep appreciation to Miss Yu-Shan Chen, and Min-Rong Huang for their assistance in the clinical follow-up of the patients. REFERENCES 1. Laine L. Multipolar electrocoagulation vs. injection therapy in the treatment of bleeding peptic ulcers. Gastroenterology 1990;99:1303-6. 2. Cook DJ, Guyatt GH, Salena BJ, Laine LA. Endoscopic therapy for acute nonvariceal upper gastrointestinal hemorrhage. A meta-analysis. Gastroenterology 1992;102:139-48. 3. Gralnek IM, Jensen DM, Gornbein J, Kovacs TO, Juthaba R, Freeman ML, et al. Clinical and economic outcomes of individuals with sever peptic ulcer hemorrhage and nonbleeding visible vessel: an analysis of two prospective clinical trials. Am J Gastroenterol 1998;93:2047-56. 4. Laine L, Peterson WL. Bleeding peptic ulcer. New Engl J Med 1994;331:717-27. 5. Lin HJ, Perng CL, Lee FY, Chan CY, Huang ZC, Lee SD, et al. Endoscopic injection for arrest of peptic ulcer hemorrhage: final results of a prospective, randomized comparative trial. Gastrointest Endosc 1993;39:15-9. 6. Laine L, Estrada R. Randomized trial of normal saline solution injection versus bipolar electrocoagulation for treatment of patients with high-risk bleeding ulcers: is local tamponade enough? Gastrointest Endosc 2001;55:6-10. 7. Cipolletta L, Bianco MA, Rotondano G, Piscopo R, Prisco A, Garofano ML. Prospective comparison of argon plasma coagulator and heater probe in the endoscopic treatment of major peptic ulcer bleeding. Gastrointest Endosc 1998;48:191-5. 8. Lai KH, Peng SN, Guo WS, Lee FY, Chang FY, Malik U, et al. Endoscopic injection for the treatment of bleeding ulcers: local tamponade or drug effect? Endoscopy 1994;26:338-41. 9. Jaramillo JL, Carmona C, Galvez C, de la Mata M, Mino G. Efficacy of the heater probe in peptic ulcer with a non-bleeding visible vessel. A controlled randomized study. Gut 1993;34:1502-6. 10. Llach J, Bordas JM, Salmeron JM, Panes J, Garcia-Pagan JC, Feu F, et al. A prospective randomized trial of heater probe thermocoagulation versus injection therapy in peptic ulcer hemorrhage. Gastrointest Endosc 1996;43:117-20. 11. Rutgeerts P, Geboes K, Vantrappen G. Tissue damage produced by hemostatic injections [abstract]. Gastrointest Endosc 1986;32:AB179. 12. Bedford RA, van Stolk R, Sivak MV Jr, Chung RS, Van Dam J. Gastric perforation after endoscopic treatment of a Dieulafoy’s lesion. Am J Gastroenterol 1992;87:244-7. 13. Dell’Abate P, Spaggiari L, Carboynani P, Soliani P, Karake I, Foggi E. An unusual complication of sclerotherapy. Endoscopy 1991;23:352-3. 14. Loperfido S, Patelli G, La Torre L. Extensive necrosis of gasGASTROINTESTINAL ENDOSCOPY
327
Y-C Chou, P-I Hsu, K-H Lai, et al.
15.
16.
17. 18.
19.
20.
21.
tric mucosa following injection therapy of bleeding peptic ulcer. Endoscopy 1990;22:285-6. Rutgeers P, Geboes K, Vantrappen G. Experimental studies of injection therapy for severe nonvariceal bleeding in dogs. Gastroenterology 1989;97:610-21. Hayashi T, Yonezawa M, Kawabara T. The study on staunch clip for the treatment by endoscopy. Gastroenterol Endosc 1975;17:92-101. Hachisu T. Evaluation of endoscopic hemostasis using an improved clipping apparatus. Surg Endosc 1988;2:13-7. Hachisu T, Miyazaki S, Hamaguchi K. Endoscopic clipping marking of lesions using the newly developed HX-3L clip. Surg Endosc 1989;3:142-7. Soehendra N, Sriram PVJ, Ponchon T, Chung SCS. Hemostatic clip in gastrointestinal bleeding. Endoscopy 2001;33: 172-80. Cipolletta L, Bianco MA, Marmo R, Rotondano G, Piscopo R, Vinginani AM, et al. Endoclips versus heater probe in preventing early recurrent bleeding from peptic ulcer: a prospective and randomized trial. Gastrointest Endosc 2001;53:147-51. 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 injec-
Bleeding ulcers: endoscopic hemoclipping vs. distilled water injection
22.
23.
24.
25.
26.
27.
tion and a combination of the two for the management of bleeding peptic ulcers. Gastrointest Endosc 1999;49:13-8. Pinkas H, McAllister E, Norman J, Robinson B, Brady PG, Dawson PJ. Prolonged evaluation of epinephrine and normal saline solution injections in an acute ulcer model with a single bleeding artery. Gastrointest Endosc 1995;42:51-5. Randall GM, Jensen DM, Hirabayashi K, Machicado GA. Controlled study of different sclerosing agents for coagulation of canine gut arteries. Gastroenterology 1989;96:1274-81. Hachisu T, Yamada H, Satoh SI, Kouzu T. Endoscopic clipping with a new rotatable clip-device and a long clip. Dig Endosc 1996;8:127-33. Hepworth CC, Kadirkamanathan SS, Gong G, Swain CP. A randomized controlled comparison of injection, thermal and mechanical endoscopic methods of hemostasis on mesenteric vessels. Gut 1998;42:462-9. Nagayama K, Tazawa J, Sakai Y, Miyasaka Y, Yu SH, Sakuma I, et al. Efficacy of endoscopic clipping for bleeding gastroduodenal ulcer: comparison with topical ethanol injection. Am J Gastroenterol 1999;94:2897-901. Lin LF, Siauw CP, Ho KS, Tung JC. Endoscopic hemoclip treatment of gastrointestinal bleeding. Chang Gung Med J 2001;24:307-12.
Moving? To ensure continued service please notify us of a change of address at least 6 weeks before your move. Phone Subscription Services at 800-654-2452 (outside the U.S. call 407-345-4000) or fax your information to 407-363-9661.
328
GASTROINTESTINAL ENDOSCOPY
VOLUME 57, NO. 3, 2003