High- Versus Low-dose Proton Pump Inhibitors After

American Journal of Gastroenterology ISSN 0002-9270C 2008 by Am. Coll. of Gastroenterology doi: 10.1111/j.1572-0241.2008.02149.xPublished by Blackwell Publishing

ORIGINAL CONTRIBUTIONS

Endoscopy

High- Versus Low-Dose Proton Pump Inhibitors AfterEndoscopic Hemostasis in Patients With Peptic UlcerBleeding: A Multicentre, Randomized StudyAngelo Andriulli, M.D.,1 Silvano Loperfido, M.D.,2 Rosaria Focareta, M.D.,3 Pietro Leo, M.D.,4

Fabio Fornari, M.D.,5 Antonietta Garripoli, M.D.,6 Paolo Tonti, M.D.,7 Sergio Peyre, M.D.,8

Antonio Spadaccini, M.D.,9 Riccardo Marmo, M.D.,10 Antonio Merla, M.D.,1 Alessandro Caroli, M.D.,2

Gian Battista Forte, M.D.,3 Angelo Belmonte, M.D.,4 Giovanni Aragona, M.D.,5 Gianni Imperiali, M.D.,11

Fabrizio Forte, M.D.,3 Fabio Monica, M.D.,2 Nazario Caruso, M.D.,1 and Francesco Perri, M.D.1

Divisions of Gastroenterology, 1Casa Sollievo della Sofferenza Hospital, IRCCS, San Giovanni, Rotondo,Italy, 2Regional Hospital, Treviso, Italy, 3 S. Sebastiano Hospital, Caserta, Italy, 4P.O. Annunziata Hospital,Cosenza, Italy, 5G. da Saliceto Hospital, Piacenza, Italy, 6 Maria Vittoria Hospital, Torino, Italy,7Ospedali Riuniti, Foggia, Italy, 8Ivrea Hospital, Ivrea, Italy, 9S. Pio da Pietrelcina Hospital, Vasto, Italy,10L. Curto Hospital, Polla, Italy, and 11Valduce Hospital, Como, Italy

BACKGROUND: The most effective schedule of proton pump inhibitor (PPI) administration following endoscopichemostasis of bleeding ulcers remains uncertain.

METHODS: Patients with actively bleeding ulcers and those with nonbleeding visible vessel or adherent clotwere treated with epinephrine injection and/or thermal coagulation, and randomized to receiveintravenous PPIs according to an intensive regimen (80 mg bolus followed by 8 mg/h ascontinuous infusion for 72 h) or a standard regimen (40 mg bolus daily followed by saline infusionfor 72 h). After the infusion, all patients were given 20 mg PPI twice daily orally. The primary endpoint was the in-hospital rebleeding rate, as ascertained at the repeat endoscopy.

RESULTS: Bleeding recurred in 28 of 238 patients (11.8%) receiving the intensive regimen, and in 19 of 236(8.1%) patients receiving the standard regimen (P = 0.18). Most rebleeding episodes occurredduring the initial 72-h infusion: 18 (7.6%) and 19 events (8.1%) in the intensive and standardgroups, respectively (P = 0.32). Mean units of blood transfused were 1.7 2.1 in the intensive and1.5 2.1 in the standard regimen group (P = 0.34). The duration of hospital stay was

3012 Andriulli et al.

endoscopic hemostasis was conducted in 11 Italian centreson a nonprofit, voluntary basis after the protocol receivedapproval by local ethics committees. All study subjects gavewritten informed consent and were enrolled between October2004 and March 2007. No industry support in the ideation,planning, designing, and conducting of the trial or in the anal-ysis of the experimental data was pursued.

Patient PopulationPatients who presented with overt gastrointestinal bleeding ora recent history (

Proton-Pump Inhibitors After Endoscopic Hemostasis 3013

243 Assigned toan intensive regimen

239 Assigned tostandard regimen

238 included in analyses

236 included in analyses

5 Excluded from

analysis

3 Withdrew

voluntarily

2 Had warfarin

anticoagulation

3 Excluded from analysis

1 Withdrew voluntarily

1 Had a final diagnosis of

pancreatic gastrinoma

1 Had warfarin anticoagulation

482 Randomly assigned

to a study group

943 Excluded:

125 declined participation

383 F2c / F3 ulcers

74 MalloryWeiss or gastric erosions

33 cancer ulcers

12 unstable condition or moribond

84 Had begun to receive PPI

27 unable to provide written informed

consent

43 on anticoagulation therapy

22 failed endoscopic therapy

140 Excluded for other reasons

1425 Evaluated

Figure 1. Flow chart of patients throughout the study.

Statistical AnalysisAll patients who took at least one dose of drugs were includedinto the analysis. Categorical data were expressed as propor-tions (%), and continuous data as means standard deviation(SD). P values for the primary end point were obtained fromthe two-sided 2 Pearsons test. For other P values, the two-sample t-test was used for continuous variables and the Fisherexact test used for discrete variables.

A prediction model for rebleeding was developed includ-ing the following covariates: Rockall score ( or


Table 1. Characteristics of the 474 Patients Included in the Analysis

Intensive Regimen Standard RegimenCharacteristic (N = 238) (N = 236)Ageyr 66.3 15.6 66.8 16.7Age > 70 yrno. (%) 116 (48.7) 121 (51.3)Menno. (%) 151 (63.4) 156 (66.1)Hemoglobing/dL 9.1 2.4 9.5 2.2Patients with hemoglobin 10 g/dLno. (%) 155 (65.1) 146 (61.9)

Shock at presentationno. (%) 20 (8.4) 35 (14.8)Rockall score

Mean SD 4.7 1.3 5.0 1.56 pointsno. (%) 64 (26.9) 81 (34.3)

Severe comorbidityno. (%) 19 (8.0) 30 (12.7)Bleeding during hospitalization

no. (%) 10 (4.2) 15 (6.4)Risk factors for bleeding peptic ulcerno. (%)

Use of NSAID 75 (31.5) 74 (31.4)Use of aspirin 72 (30.3) 82 (34.7)

Time from bleeding to endoscopy6 h 77 (32.4) 76 (32.1)Between 6 and 12 h 74 (31.1) 78 (32.2)Between 12 and 24 h 55 (23.1) 49 (20.8)>24 h 32 (13.4) 33 (13.9)

Previous ulcer diseaseno. (%) 32 (13.4) 36 (15.3)Previous bleedingno. (%) 25 (10.5) 14 (5.9)Peptic ulcer as source of bleedingno. (%)

Duodenal ulcer 150 (63.0) 143 (60.6)Gastric ulcer 88 (37.0) 93 (39.4)

Ulcer size 20 mmno. (%) 21 (8.8) 15 (6.4)Forrest classification of endoscopic stigmata of hemorrhageno. (%)

1 aspurting 25 (10.5) 25 (10.6)1 b oozing 82 (34.5) 73 (30.9)2 anonbleeding visible vessel 81 (34.0) 85 (36.0)2 badherent clot 50 (21.0) 53 (22.5)

Endoscopic hemostasis modalityno. (%)Unimodal 119 (50.0) 136 (57.6)Multimodal 119 (50.0) 100 (42.4)

Type of PPI administeredno. (%)Omeprazole 167 (70.2) 163 (69.1)Pantoprazole 71 (29.8) 73 (30.9)

NSAID = nonsteroidal anti-inflammatory drug; PPI = proton pump inhibitor; SD = standard deviation.

therapy (Table 1): two-thirds of patients were men, abouthalf were older than 70 yr, and a majority had taken aspirinor nonsteroidal anti-inflammatory drugs before hospitaliza-tion. Nineteen and 30 patients in the intensive and standardregimen groups, respectively, presented with a severe comor-bidity (P = 0.09). Equal proportions of patients bled duringtheir hospital stay for unrelated illnesses. Signs of hemody-namic instability were more frequent in the standard regimen:shock was more common, mean Rockall score higher, andmore patients had a Rockall score 6. An equal number ofpatients had duodenal and gastric ulcers, and prevalence ofactive arterial bleeding, oozing, nonbleeding visible vessel, oradherent clot was not significantly different. (Table 1). Abouthalf of ulcers were injected with epinephrine solution, and theremaining received epinephrine in association with thermaltherapy (bipolar electrocoagulation or argon plasma coagula-tion) or apposition of clips. In the two treatment groups, 70%of patients received omeprazole and the remaining patients

received pantoprazole. All patients completed the assignedschedule of PPIs administration. No drug-related side effectswere reported in either group.

Bleeding recurred in 28 patients (11.8%, CI 7.616.0) inthe intensive regimen group and in 19 (8.1%, CI 4.511.5)in the standard regimen group; a difference of 3.8% (CI 1.7to 9.1%) (P = 0.18) (Table 2). At the repeat endoscopy, 9lesions were spurting ulcers, 20 oozing, 13 had a nonbleed-ing visible vessel, 3 an adherent clot, all but 2 of previouslesions were successfully controlled by a second attempt atendoscoscopic hemostasis; the remaining 3 lesions were clas-sified as Forrest 2c or 3 and were not treated endoscopically.Recurrent bleeding was most common during the period ofintravenous administration: within the first 3 days, 16 pa-tients (6.7%, CI 3.59.9) in the intensive regimen group and22 (9.3%, CI 5.513.1) in the standard regimen experiencedrebleeding (P = 0.32). Between the two treatment groups,rebleeding rates did not differ for gastric and duodenal ulcers


Table 2. Clinical Outcomes for the 47 Patients Included in the Analysis

Intensive Regimen Standard RegimenCharacteristic (N = 238) (N = 236) P ValueRebleedingno. (%) 28 (11.8) 19 (8.1) 0.18Rebleeding according to Forrest classification of endoscopic stigmata of hemorrhageno. (%)

1 aspurting 5 (2.1) 5 (2.1)1 boozing 11 (4.6) 8 (3.4)2 anonbleeding visible vessel 7 (2.9) 5 (2.1)2 badherent clot 5 (2.1) 1 (0.4) 0.61

Rebleeding according to time from index endoscopyno. (%)24 h 13 (5.5) 10 (4.2)>24 h to 7 days 15 (6.3) 9 (3.8) 0.68

Rebleeding according to Rockall score: no. (%)


11.8% and 8.0% of patients, respectively, with a correspond-ing 3.8% (CI 1.7 to 9.1%) difference in proportions. The sta-tistically equal rebleeding rate between the two intravenousPPI strategies is of relevance considering that patients treatedwith bolus PPIs presented more frequently with shock andsigns of hemodynamic instability, two predictors of worseoutcome (12, 14). Our negative results are in keeping withnegative data from four previous trials that compared thetwo PPI strategies (710), but are at odds with the Lin re-port (15). Therefore, we cannot endorse the reccommen-dation issued by two consensus statements on the routineuse of the intensive PPI regimen for peptic ulcer bleeding(2, 3).

The 11.8% and 8.1% rebleeding rates with intensive orstandard regimens, reported in this study, are in keeping withthe 13.2% probability of rebleeding reported in a pooled anal-ysis of data from eight trials on high-dose PPI infusion (13).Consequently, the divergent results between our study andprevious trials are unlikely to reflect major differences in pa-tients populations or endoscopic interventions. Reasons forthe discrepancy might be revealed by carefully inspectingthe comparator used in previous trials: whenever endoscopichemostasis was not applied, intravenous PPI therapy reducedthe rate of rebleeding when compared with placebo (16, 17)or H2-receptor antagonists (18). However, in patients whounderwent simultaneous endoscopic hemostasis, two trialsreported opposite results with the intensive PPI dosing ver-sus placebo (5, 6), a single report showed a benefit of high-dose PPI therapy in comparison to H2R antagonists (1921),and four studies did not show a significant difference be-tween an intensive regimen and standard bolus PPI injections(710).

As intravenous PPI therapy is expensive, to be cost-effective the extra cost of the medication must be offset bya reduction in the occurrence of important adverse clinicaloutcomes. In our study, patients who received the high-dosePPI regimen had no advantage with respect to transfusion re-quirement, need for surgery, length of hospital stay, or deathrate. Moreover, cost-effectiveness analyses have yielded con-trasting results regarding which of the two PPIs administra-tion strategies, oral (22) or intravenous (23), is more effectivein patients who have received concomitant endoscopic treat-ment. Standard PPI dosing has the obvious advantage of re-duced cost and was not shown to increase the risk of nonulcerdeaths (15, 17).

The rationale for the need for profound gastric acid sup-pression in peptic ulcer bleeding is extrapolated largely fromin vitro evidences that an acidic environment impairs plateletfunction and clot stabilization (3, 4). These in vitro data havebeen taken as supportive evidences for the in vivo clinicalbenefit. However, in healthy non-Asian volunteers, who aretypically Helicoacter pylori negative and rapid PPI metabo-lizers, the currently accepted high-dose PPI regimen main-tained intragastric pH above 6 for only 30% of the time (24,25), thus raising the possibility that we may not be achiev-ing the theoretically desirable pH even with intensive therapy.

Findings from this investigation indicate that we may not needintragastric pH maintained at


What Is New Here

We report the results of a multicentre randomizedcontrolled trial comparing an intensive high-doseacid-inhibiting regimen versus a low-dose standardregimen in the management of high-risk peptic ulcer-related bleeding after performance of appropriate en-doscopic hemostasis: no significant difference in out-comes was registered between the two regimens. Thesedata demonstrate that 3-day PPI infusion can be re-placed with a once-daily intravenous bolus, support-ing the use of a low-dose regimen for its cost-savingpotential.

Reprint requests and correspondence: Angelo Andriulli, M.D.,Division of Gastroenterology, Casa Sollievo della Sofferenza Hos-pital, IRCCS, 71013, San Giovanni, Rotondo, Italy.

Received November 9, 2007; accepted May 5, 2008.

REFERENCES

1. British Society of Gastroenterology Endoscopy Commit-tee. Non-variceal upper gastrointestinal hemorrhage: Guide-lines. Gut 2002;51(Suppl 4):iv16.

2. Barkun A, Bardou M, Marshall JK, for the NonvaricealUpper GI Bleeding Consensus Conference Group. Con-sensus recommendations for managing patients with non-variceal upper gastrointestinal bleeding. Am Intern Med2003;139:84357.

3. Green FW Jr., Kaplan MM, Curtis LE, et al. Effect of acidand pepsin on blood coagulation and platelet aggregation.A possible contributor prolonged gastroduodenal mucosalhemorrhage. Gastroenterology 1978;74:3843.

4. Patchett SE, Enright H, Afdhal N, et al. Clot lysis by gastricjuice: An in vitro study. Gut 1989;30:10747.

5. Schaffalitzky de Muckadell UB, Havelund T, Harling H,et al. Effects of omeprazole on the outcome of endo-scopically treated peptic ulcers. Scand J Gastroenterol1997;32:3207.

6. Lau JYW, Sung JJY, Lee KKC, et al. Effect of intravenousomeprazole on recurrent bleeding after endoscopic treat-ment of bleeding peptic ulcers. N Engl J Med 2000;343:3106.

7. Schonekas H, Ahrens H, Pannewick U, et al. Comparisonof two doses of intravenous pantoprazole in peptic ulcerbleeding. Gastroenterology 1999;116:A305.

8. Udd M, Miettinen P, Palmu A, et al. Regular-dose versushigh-dose omeprazole in peptic ulcer bleeding: A prospec-tive randomized double-blind study. Scand J Gastroenterol2001;36:13328.

9. Dokas SM, Lazaraki GI, Kontoninas Z, et al. Bolus intra-venous omeprazole b.i.d vs. continuous intravenous omepra-zole infusion combined with endoscopic hemostasis in thetreatment of peptic ulcer bleeding. Preliminary results. Gut2004;53(Suppl VI):A290.

10. Bajaj JS, Dua KS, Hanson K, et al. Prospective, randomizedtrial comparing effect of oral versus intravenous pantopra-zole on rebleeding after nonvariceal upper gastrointestinalbleeding: A pilot study. Dig Dis Sci 2007;52:21904.

11. Andriulli A, Annese V, Caruso N, et al. Proton-pump in-hibitors and outcome of endoscopic hemostasis in bleeding

peptic ulcer: A series of meta-analysis. Am J Gastroenterol2005;100:20719.

12. Rockall TA, Logan RF, Devlin HB, et al. Risk assess-ment after acute upper gastrointestinal haemorrhage. Gut1996;38:31621.

13. Leontiadis GI, Sharma VK, Howden CW. Proton pump in-hibitor treatment for acute peptic ulcer bleeding. CochraneDatabase Syst Rev 2006;1:CD0020947.

14. Saeed ZA, Ramirez FC, Hepps KS, et al. Prospective vali-dation of the Baylor bleeding score for predicting the like-lihood of rebleeding after endoscopic hemostasis of pepticulcers. Gastrointest Endosc 1995;41:5615.

15. Lin HJ, Lo WC, Chen YC, Perng CL. Role of intravenousomeprazole in patients with high-risk of peptic ulcer bleed-ing after successful endoscopic epinephrine injection: Aprospective randomized comparative trial. Am J Gastroen-terol 2006;101:5005.

16. Hasselgren G, Lind T, Lundell L, et al. Continuous intra-venous infusion of omeprazole in elderly patients with pep-tic ulcer bleeding: Results of a placebo controlled multicen-ter study. Scand J Gastroenterol 1997;32:32833.

17. Khuroo MS, Farahat KL, Kagevi IE. Treatment with pro-ton pump inhibitors in acute non-variceal upper gastroin-testinal bleeding: A meta-analysis. J Gastroenterol Hepatol2005;20:1125.

18. Cardi M, Muttillo IA, Amadori L, et al. Omeprazole ver-sus ranitidine intraveineux dans le traitment de lulce`reduodenal hemorragique: Une etude randomisee prospective.Ann Chir 1997;51:1369.

19. Labenz J, Peitz U, Leusing C, et al. Efficacy of primed infu-sions with high dose ranitidine and omeprazole to maintainhigh intragastric pH in patients with peptic ulcer bleeding: Aprospective randomised controlled study. Gut 1997;40:3641.

20. Fried R, Beglinger C, Meier R, et al. Comparison of intra-venous pantoprazole with intravenous ranitidine in pepticulcer bleeding. Gut 1999;45(Suppl V):A100.

21. Lin HJ, Lo WC, Lee FY, et al. A prospective randomisedcomparative trial showing tat omeprazole prevents rebleed-ing in patients with bleeding peptic ulcer after endoscopictherapy. Arch Intern Med 1998;158:548.

22. Spiegel BM, Dulai GS, Lim BS, et al. The cost-effectivenessand budget impact of intravenous versus oral proton pumpinhibitors in peptic ulcer hemorrhage. Clin GastroenterolHepatol 2006;4:98897.

23. Barkun AN, Herba K, Adam V, et al. The cost-effectivenessof high-dose oral proton pump inhibition after endoscopy inthe acute treatment of peptic ulcer bleeding. Aliment Phar-macol Ther 2004;20:195202.

24. Netzer P, Gaia C, Sandoz M, et al. Effect of repeated injec-tion and continuous infusion of omeprazole and ranitidine onintragastric pH over 72 h. Am J Gastroenterol 1999;94:3517.

25. Metz DC, Amer F, Hunt B, et al. Lansoprazole regimens thatsustain intragastric pH >6: AN evaluation of intermittentoral and continuous intravenous lansoprazole or ranitidine.Aliment Pharmacol Ther 1999;23:98595.

26. Welage LS, Berardi RR. Evaluation of omeprazole, lan-soprazole, pantoprazole, and rabeprazole in the treatmentof acid-related diseases. J Am Pharm Assoc 2000;40:5262.

27. Chilovi F, Piazzi L, Zancanella L, et al. Intravenousomeprazole and pantoprazole after endoscopic treatmentof bleeding peptic ulcers. Gastrointest Endosc 2003;57:AB150.

28. Marmo R, Rotondano G, Piscopo R, et al. Dual ther-apy versus monotherapy in the endoscopic treatment of


high-risk bleeding ulcers: A meta-analysis of controlledtrial. Am J Gastroenterol 2007;102:27989.

29. ASGE guideline: The role of endoscopy in acute non-variceal upper-GI hemorrhage. Gastrointest Endosc 2004;60:497504.

30. Pilotto A, Franceschi M, Longoa MG, et al. Helicobacter py-lori infection and the prevention of peptic ulcer with protonpump inhibitors in elderly subjects taking low-dose aspirin.Dig Liver Dis 2004;36:66670.

31. Jensen DM, Pace SC, Soffer E, et al., and members of the 315Study Group. Continuous infusion of pantoprazole versusranitidine for prevention of ulcer rebleeding: A U.S. multi-center randomized, double-blind study. Am J Gastroenterol2006;101:19919.

CONFLICT OF INTEREST

Guarantor of the article: Angelo Andriulli, M.D.Specific author contributions: Angelo Andriulli designedthe study, wrote the original draft of the protocol, discussedit with the Ethics committee of the coordinating centre, acted

as the coordinator of the entire study, supervised the statisti-cal analysis of trial data, and wrote the manuscript. SilvanoLoperfido discussed and amended the original draft of theprotocol, presented and discussed the study with the localEthics committee, enrolled patients to the study, and reviewedthe final version of the manuscript. Rosaria Focareta, PietroLeo, Fabio Fornari, Antonietta Garritoli, Paolo Tonti, Ser-gio Peyre, Antonio Spadaccini, Riccardo Marmo, Alessan-dro Caroli, Gian Battista Forte, Angelo Belmonte, GiovanniAragona, Gianni Imperiali, Fabrizio Forte, Fabio Monica,Nazario Caruso agreed on the final version of the protocol ofthe study, discussed the protocol with their local Ethics com-mittees, enrolled patients to the study, and agreed on the finalversion of the manuscript. Antonio Merla enrolled patientsto the study, collected data from contributing investigators,and managed the electronic database. Francesco Perri ran thestatistical analysis of experimental data.Financial support: None.Potential competing interests: None.

Documents

High- Versus Low-dose Proton Pump Inhibitors After