Intragastric ph With Oral vs Intravenous Bolus Plus Infusion Proton- Pump Inhibitor Therapy in Patients With Bleeding Ulcers LOREN LAINE, ABBID SHAH, and SHAHROOZ BEMANIAN Division of Gastrointestinal and Liver Diseases, Keck School of Medicine, University of Southern California, Los Angeles, California Background & Aims: Intravenous bolus plus infusion proton pump inhibitor (PPI) therapy is recommended for patients with bleeding ulcers and higher risk stigmata. If frequent oral dosing of PPIs provided similar antisecretory effect, this might be preferred based on ease and cost. Methods: Patients presenting with overt bleeding due to ulcers had intragastric ph probes placed after endoscopy and baseline ph recorded. They were randomly assigned to intravenous lansoprazole (90-mg bolus followed by 9-mg/h infusion) or oral lansoprazole (120-mg bolus followed by 30 mg every 3 hours). ph was recorded for 24 hours. Results: Intragastric ph was >6 for 67.8% of the study period with intravenous PPI (n 32) and 64.8% with oral PPI (n 34): difference, 3.0%; 95% confidence interval (CI): 9.2% to 15.2%. Intragastric ph was >6 for >60% of the study period in 22 (68.8%) patients receiving intravenous and 22 (64.7%) patients receiving oral PPI: difference, 4.0%; 95% CI: 18.7% to 26.8%. At 1 hour, mean phs for intravenous vs oral were 5.3 0.4 vs 3.3 0.4, respectively (difference, 2.0; 95% CI: 0.8 3.1; P.001). At >1.5 hours, 95% CIs of the differences for all hourly mean phs included zero. Mean ph rose above 6 after 2 3 hours of intravenous PPI and 3 4 hours of oral PPI. Conclusions: Frequent oral PPI may be able to replace the currently recommended intravenous bolus plus infusion PPI therapy in patients with bleeding ulcers, although the possibility that intravenous PPIs are superior cannot be definitively excluded given our relatively wide confidence intervals. Intravenous PPI provides more rapid increase in ph, reaching mean ph of 6 approximately 1 hour sooner than oral PPI. Ulcers are the most common source of bleeding in patients hospitalized with upper gastrointestinal hemorrhage. 1 Endoscopic therapy clearly improves outcomes in patients with ulcers and active bleeding or a nonbleeding visible vessel. 2 In addition, intravenous proton pump inhibitor (PPI) therapy, given as a bolus followed by continuous infusion, is recommended in patients with active bleeding, nonbleeding visible vessels, and adherent clots. 3,4 Randomized controlled trials indicate that this PPI regimen significantly decreases recurrent bleeding, even in patients who have received endoscopic therapy. 4,5 In vitro data suggest that acid plays an important role in impairing hemostasis and causing clot lysis and that a ph above 6 would be required to promote clot formation and stability. 6,7 To reach and sustain intragastric ph 6, intravenous PPIs, given as a bolus followed by a constant infusion, have been employed. Initial studies suggested that such infusions maintained intragastric ph 6 for a majority of a 24-hour period. 8 However, administration of intravenous PPIs requires additional expense, personnel, and equipment as compared with oral PPI therapy. If administration of intermittent oral doses of PPI could provide antisecretory effect comparable with continuous infusion of intravenous PPI, it might be preferred because of lower cost and greater ease of administration. Doses of intravenous PPIs do not have greater pharmacodynamic effects than equivalent doses of oral PPIs beyond 1 hour after administration. 9 However, the elimination half-lives of PPIs are approximately 1 hour. Therefore, after clearance of a single dose of a PPI, new proton pumps can produce acid. The reason that the constant infusion of intravenous PPIs should be effective at maintaining a ph above the target of 6 relates to the fact that a constant level of PPI is present throughout the day so that any newly activated proton pumps can be inhibited. Frequent oral dosing of PPIs, even when given as a bolus and then every 3 hours, achieves ph 6 for 40% of the study period in healthy volunteers, 10 suggesting that constant infusion of intravenous PPIs would still be preferred in patients with bleeding ulcers. However, recent North American studies in healthy volunteers demonstrated that standard regimens of bolus and constant infusion PPIs (pantoprazole and lansopra- Abbreviations used in this paper: PPI, proton pump inhibitor. 2008 by the AGA Institute 0016-5085/08/$34.00 doi:10.1053/j.gastro.2008.03.006 1836
zole) maintained intragastric ph 6 only 25% 45% of a 24-hour study period. 11,12 These studies raise the possibility that frequent oral dosing of PPIs may not be inferior to constant infusion of PPIs. In addition, they raise the question of whether PPIs have less antisecretory effect in healthy volunteers than in patients presenting with bleeding ulcers, the intended patient population for constant infusion PPI therapy. To determine whether frequent oral dosing of PPI achieves an antisecretory effect similar to intravenous PPI infusion in the relevant patient population, we conducted a randomized controlled trial of intravenous bolus plus constant infusion PPI vs frequent oral dosing of a PPI in US patients presenting with bleeding ulcers. Patients and Methods Patients presenting to Los Angeles County University of Southern California Medical Center with overt evidence of upper gastrointestinal bleeding (melena, hematemesis, hematochezia) in whom endoscopy performed within the first day of admission revealed an ulcer in the stomach or duodenum were eligible for inclusion in the study. Endoscopic therapy was performed in patients with active bleeding or nonbleeding visible vessels; adherent clots resistant to vigorous irrigation were not treated with endoscopic therapy. Exclusion criteria included prior gastric surgery, recent use of PPIs (2 weeks) or H2-receptor antagonists (1 week), and active bleeding after endoscopic hemostasis. Intragastric ph was monitored with a single-channel antimony ph probe and an external recording device (Digitrapper; Medtronic, Minneapolis, MN). At the end of the endoscopic procedure, the ph probe was inserted via the nares and placed into the proximal stomach 10 cm below the gastroesophageal junction as determined by endoscopy (assuming a 5-cm correction for transnasal passage). Baseline ph was recorded for 1 hour, and then the study medication was given. Continuous intragastric ph was then recorded for 24 hours of PPI therapy. The 1-hour delay in PPI administration after endoscopy in our study was necessary to measure ph but theoretically could have somewhat decreased PPI efficacy if the lack of bleeding during that hour led to fewer activated proton pumps. Patients were randomly assigned using a computergenerated randomization sequence with concealed allocation by study investigators to receive intravenous lansoprazole (Prevacid; TAP Pharmaceuticals Products Inc, Lake Forest, IL) 90-mg bolus followed by 9-mg/h infusion or oral lansoprazole (Prevacid Solutab; TAP Pharmaceuticals Products Inc) 120-mg bolus orally followed by 30 mg orally every 3 hours from 3 hours to 21 hours. Therapy was not blinded (ie, no double-dummy technique was employed). The intravenous dosing was chosen based on the study of Metz et al, 11 which showed no evidence of increased antisecretory effect with a 120-mg bolus as compared with a 90-mg bolus. In addition, the 9-mg/h infusion studied by Metz et al provides the closest match to the oral regimen we used. The primary end-points were time intragastric ph was 6 over a 24-hour period and the proportion of patients who maintain intragastric ph 6 for 60% of the 24-hour period. We chose 60% as a threshold because clinical trials have demonstrated the efficacy of bolus followed by constant infusion intravenous PPI therapy in patients with bleeding ulcers, 3 5 and the pharmacodynamic studies performed in patients with bleeding ulcers from outside the United States revealed that mean times ph was 6 with intravenous bolus and constant infusion PPIs were as low as 60%. 13 Secondary end-points included the proportion of time ph was 5 and 4 and the mean hourly phs (including mean ph at 0.5 and 1.5 hours) during the study period after drug administration. Our hypothesis was that ph control with oral PPI would be noninferior to intravenous PPI. We prespecified that a difference between treatments of 20% for intravenous vs oral PPI in our primary end-points would be considered noninferior. The sample size of 66 patients was originally chosen based on a 1-sided 95% confidence interval (CI), but our present analyses are all 2-sided. A sample size of 66 patients will produce an upper bound of the 2-sided 95% CI of the difference for intravenous vs oral PPI 20% if the 2 treatments achieve a 60% mean proportion of time ph 6 (with a variance of 1000) or if they achieve a rate of 80% control of ph 6 for 60% of the 24-hour period. We planned a modified intent-totreat analysis to include all patients who were randomized and had any intragastric ph data recorded. We present absolute risk differences with 95% CIs of the differences for comparisons of the end-points. We also used a Mann Whitney U test for comparison of ph values in the first 2 hours because a bolus of intravenous PPI has been reported to raise ph significantly more than an oral PPI in the first hour after administration. 9 Multiple logistic regression analysis was performed with the dichotomous dependent outcome variable of our coprimary end-point (ph 6 for 60% of the time) and the independent variables of race/ethnicity (Hispanic vs non- Hispanic), Helicobacter pylori status (positive vs negative), age ( 60 vs 60 years), ulcer location (duodenal vs gastric), and therapy (intravenous vs oral PPI). The relatively small sample size of the study limits the power of this regression analysis. The study was approved by the University of Southern California Health Sciences Institutional Review Board, and all patients provided informed consent, signing study consent forms. 1837
Figure 1. Flow diagram of progress of patients through study from initial assessment for eligibility through randomization, drug allocation, and analysis. Results Patients were enrolled between April 2006 and December 2007. A flow diagram of the progress of the patients in the trial is shown in Figure 1. Selected characteristics of the 66 patients in the 2 study groups are shown in Table 1. The age range of patients was 26 to 82 years in each group. A majority of patients had H pylori infection. Higher-risk stigmata that would currently lead to bolus plus constant infusion intravenous PPI therapy (active bleeding, nonbleeding visible vessel, or adherent clot resistant to irrigation) were seen in 7 patients in the intravenous group and 6 in the oral group. The mean baseline ph was 2.0 0.3 in the intravenous PPI group and 1.9 0.3 in the oral PPI group. Gastric biopsies in 2 patients who were randomized (1 in the intravenous group and 1 in the oral group) subsequently revealed gastric carcinoma; both patients had baseline achlorhydria. Based on our prespecified modified intent-to-treat analysis, these patients were included in all analyses presented. Exclusion of these patients results in a 1% absolute decrease in each of the coprimary end-points for each study group. The results for the primary and secondary end-points related to overall ph control during the study period are shown in Table 2. Intragastric ph was maintained at 6 for 67.8% of the study period with intravenous PPI and 64.8% with oral PPI. The upper bound of the 95% CI of this 3.0% difference was 15.2% (consistent with intravenous PPI maintaining ph 6 for as much as 15.2% more of the study period than oral PPI). Intragastric ph was maintained at 6 for 60% of the study period in 68.8% of the intravenous PPI group and 64.7% of the oral PPI group. The upper bound of the 95% CI of this 4.0% difference was 26.8%. In a post hoc assessment, we also determined the proportion of patients who achieved the threshold ph of 6 for 80% of the study period to explore a more rigorous end-point for ph control. Results for intravenous vs oral PPI were 13 (40.6%) vs 13 (38.2%) patients, respectively (difference, 2.4%; 95% CI: 21.2% 26.0%), for ph 6; 23 (71.9%) vs 23 (67.6%) patients, respectively (difference, 4.2%; 17.9% 26.4%) for ph 5; and 27 Table 1. Selected Characteristics of the Study Groups Intravenous PPI, n 32 (%) Oral PPI, n 34 (%) Mean ( SE) age 50 2 54 2 Male/female 22 (69)/10 (31) 20 (59)/14 (41) H pylori infection a 21 (66) 20 (59) Race/ethnicity White 1 (3) 1 (3) African American 5 (16) 4 (12) Hispanic 20 (63) 20 (59) Asian 6 (19) 9 (26) Ulcer location Gastric 15 (47) 18 (53) Duodenal 17 (53) 16 (47) a H pylori status not available for 5 patients. PPI, proton pump inhibitor. 1838
Table 2. Intragastric ph Results for 24-Hour Period of PPI Administration Intravenous PPI (n 32) Oral PPI (n 34) Difference (95% CI) Mean ( SE) percent time ph 6 67.8 4.5 64.8 4.2 3.0 ( 9.2 15.2) Mean ( SE) percent time ph 5 81.4 3.4 79.4 3.7 2.0 ( 8.1 12.0) Mean ( SE) percent time ph 4 90.5 2.0 87.6 2.9 2.9 ( 4.3 10.0) Number of patients with ph 6 for 60% of study period 22 (68.8%) 22 (64.7%) 4.0% ( 18.7% 26.8%) Number of patients with ph 5 for 60% of study period 28 (87.5%) 29 (85.3%) 2.2% ( 14.3% 18.7%) Number of patients with ph 4 for 60% of study period 30 (93.8%) 31 (91.2%) 2.6% ( 10.1% 15.3%) PPI, proton pump inhibitor. (84.4%) vs 28 (82.4%) patients, respectively (difference, 2.0%; 15.9% 20.0%), for ph 4. The effect of the PPI therapies over time is shown in Figure 2. At 0.5 hours after initiation of therapy, the mean phs were 3.9 0.4 with intravenous PPI vs 2.6 0.4 with oral PPI (difference, 1.3; 95% CI of difference: 0.2 2.4; P.04 by Mann Whitney U test); at 1 hour, mean phs were 5.3 0.4 vs 3.3 0.4, respectively (difference, 2.0; 95% CI: 0.8 3.1; P.001); at 1.5 hours, mean phs were 5.5 0.4 vs 4.4 0.5, respectively (difference, 1.1; 95% CI: 0.1 2.3; P.09); and, at 2 hours, mean phs were 5.6 0.4 vs 4.9 0.4, respectively (difference, 0.7; 0.4 1.8; P.20). Beyond 1 hour of therapy, the 95% CIs of the differences for all hourly mean phs included zero. The mean ph rose above 6 after between 2 and 3 hours of intravenous PPI and between 3 and 4 hours of oral PPI. Multiple logistic regression analysis showed no evidence of a significant difference between intravenous and oral PPI when adjusted for age, H pylori status, race/ethnicity, and ulcer location: the odds ratio (OR) was 1.2 (95% CI: 0.4 3.6; P.80), which was identical to the OR when only treatment was included as an independent variable. H pylori infection was associated with an increased response to PPI (OR, 5.3; 95% CI: 1.4 2.1; P.02), and duodenal ulcer was associated with a decreased response (OR, 0.2; 95% CI: 0.4 0.7; P.01). Age and race/ethnicity were not associated with ph response to PPI. Figure 2. Hourly mean intragastric ph over 24-hour period of PPI administration. H pylori-positive patients had greater mean time ph 6 than H pylori-negative patients with oral PPI (71.5% 3.8% vs 45.1% 9.7%, respectively; 95% CI of difference: 3.6% 49.3%) but not with intravenous PPI (67.9% 6.0% vs 66.8% 7.5%, respectively). Patients with duodenal ulcers tended to have lower response to PPIs than those with gastric ulcers for oral (56.6% 7.1% vs 72.0% 4.1%, respectively; 95% CI of difference: 32.4% 1.6%) and intravenous PPI (64.6% 6.5% vs 71.4% 6.3%, respectively; 95% CI of difference: 25.3% 11.7%). Results in the 13 patients with higher risk stigmata showed mean percent time ph 6 was 72.3 8.5 for oral PPI and 76.1 12.6 for intravenous PPI (95% CI of difference: 38.6 to 30.9). No episodes of recurrent bleeding or death occurred in either group, and no adverse effects with PPI therapy were noted. Discussion This randomized trial compared the current recommended medical therapy for patients presenting with bleeding ulcers and high-risk features, bolus plus continuous infusion intravenous PPI, to frequent oral PPI therapy. Maintenance of intragastric ph 6 with the PPI regimen has been the goal for treatment of patients with ulcer bleeding, although this is based on experimental data and has not been verified clinically. An alternate hypothesis might be that pepsin-induced clot lysis is the most important factor to target to decrease recurrent ulcer bleeding and that maintaining the intragastric ph above 4 5 may be adequate for this purpose. 7,14 We used measures of the proportion of time ph was maintained above 6 as our primary end-points, although we also prespecified assessment of the proportion of time ph was maintained above 5 and above 4. We found that the proportion of time ph was maintained above 6 was similar with intravenous (68%) and oral (65%) PPI administration. The proportions of time the ph was above 5 and above 4 also were similar, with differences between intravenous and oral therapies of 2% to 3%. Results were also similar when assessing the proportion of patients who were able to sustain ph above 6 (as well as above 5 and above 4) for 60% of the study period and for the more rigorous end-point of 80% of the study period: differences ranged from 2% to 4% in favor of intravenous 1839
PPI for all of these outcomes. In addition, adjustment for potential confounding variables such as H pylori, age, race/ethnicity, and ulcer location did not alter the results. Our study population may not be generalizable to the current US population given that 60% were Hispanic and 20% were Asian, but Hispanic status and Hispanic/ Asian status (data not shown) were not independent predictors of ph response to PPIs in multiple regression analysis. The results of our study do not allow us to conclude definitively that oral PPI is noninferior to intravenous. The difference of 3% in the proportion of the study period during which the mean ph was above 6 is consistent with a difference as high as 15% in favor of intravenous PPI, which is within our noninferior range. However, the difference of 4% more patients maintaining ph above 6 for 60% of the study period (our other coprimary end-point) was still consistent with a difference in favor of intravenous PPI as high as 27%. Furthermore, many gastroenterologists may view differences of 15% or 20% (our noninferiority level) as relatively large and potentially meaningful. The consistently small differences between intravenous and oral PPIs in the multiple analyses suggest that the difference between bolus plus constant infusion intravenous PPI and frequent oral PPI is small, but a larger sample size would be necessary to tighten the precision of the estimates and definitively confirm this finding. Use of oral PPIs in place of bolus plus constant infusion intravenous PPI should decrease costs and may simplify therapy. Given the short half-life of PPIs, we chose an oral dosing schedule of every 3 hours for proof of concept, realizing that in the hospital setting some might consider administration of a PPI orally every 3 hours as no less labor intensive than giving a bolus and constant infusion of intravenous PPI. In an attempt to improve ease of administration, future studies could assess the use of less frequent dosing of oral PPIs. The intravenous PPI regimen does appear to provide more rapid elevation of ph than the oral regimen, despite an initial 120-mg oral dose. In the first hour of PPI therapy, the ph was significantly higher with intravenous than with oral PPI (5.3 vs 3.3, respectively [95% CI of difference: 0.8 3.1]). Beyond 1 2 hours of treatment, the mean phs for the 2 regimens approached one another. The mean ph rose above 6 within 3 hours of intravenous PPI therapy and within 4 hours of oral PPI therapy. Reaching the threshold of efficacy as rapidly as possible after presentation should be desirable. However, the potential clinical benefit of reaching a target ph 1 hour earlier in preventing rebleeding is unknown and would be extremely difficult to identify in a randomized clinical trial. Our results suggest that intravenous and frequent oral PPIs are more effective in patients with bleeding ulcers than in healthy volunteers, at least within North America. Maintenance of intragastric ph above 6 in approximately two thirds of patients was markedly higher than the 25% 45% rates identified in recent North American studies of similar regimens of intravenous and oral PPI therapy in healthy volunteers. 10 12 One explanation may be a higher incidence of H pylori in ulcer patients than in healthy volunteers. Most ulcer patients have H pylori infection, and PPIs do have somewhat greater antisecretory effect in subjects with H pylori infection. Two of the cited volunteer studies did not include H pylori-positive subjects, whereas 50% of volunteers in the third study were H pylori positive. 10 12 In addition, blood theoretically may activate hydrogen-potassium AT- Pase pumps, increasing the efficacy of PPIs. Ulcer patients are also older than volunteers in these studies. Although PPIs are not noted to have an increased efficacy in older patients, pharmacokinetic studies have noted a decreased clearance and increased bioavailability of PPIs in the elderly population. 15,16 In conclusion, frequent oral PPI therapy achieves 24-hour ph control that appears to be similar to intravenous bolus plus constant infusion PPI therapy, although the possibility that intravenous PPIs are superior cannot be definitively excluded given the relatively wide confidence intervals around our observed results. Because the only known mechanism by which PPI therapy reduces ulcer rebleeding is via its effect on acid secretion, our results suggest that frequent oral PPI therapy may be able to replace bolus plus constant infusion intravenous PPI for patients with ulcer bleeding. However, our end-points were surrogate pharmacodynamic end-points and not clinical end-points, which provide the most relevant information. The major pharmacodynamic benefit of the intravenous regimen was a more rapid elevation in intragastric ph. If reaching a target ph of 6 (or 5) an hour earlier is desired and felt to be clinically important, then intravenous PPI therapy would be preferred. References 1. Laine L, Peterson W. Bleeding peptic ulcers. N Engl J Med 1994; 331:717 727. 2. Cook DJ, Salena B, Guyatt GH, et al. Endoscopic therapy for acute non-variceal upper gastrointestinal hemorrhage a meta-analysis. Gastroenterology 1992;102:139 148. 3. Barkun A, Bardou M, Marshall JK. Nonvariceal Upper GI Bleeding Consensus Conference Group. Consensus recommendations for managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2003;139:843 857. 4. Sung JJY, Chan FKL, Lau JYW, et al. The effect of endoscopic therapy in patients receiving omeprazole for bleeding ulcers with nonbleeding visible vessels or adherent clots. A randomized comparison. Ann Intern Med 2003;139:237 243. 5. Lau JYW, Sung JJY, Lee KKC, et al. Effect of intravenous omeprazole on recurrent bleeding after endoscopic treatment of bleeding peptic ulcers. N Engl J Med 2000;343:310 316. 1840
6. Chaimhoff C, Creter D, Djaldetti M. The effect of ph on platelet and coagualtion factor activities. Am J Surg 1978;136:257 259. 7. Green FW, Kaplan MM, Curtis LE, et al. Effect of acid and pepsin on blood coagulation and platelet aggregation. Gastroenterology 1978;74:38 43. 8. Labenz J, Peitz U, Leusing C, et al. Efficacy of primed infusions with high-dose ranitidine and omeprazole to maintain high intragastric ph in patients with peptic ulcer bleeding: a prospective randomised controlled study. Gut 1997;40:36 41. 9. Freston JW, Pilmer BL, Chiu YL, et al. Evaluation of the pharmacokinetics and pharmacodynamics of intravenous lansoprazole. Aliment Pharmacol Ther 2004;19:1111 1122. 10. Pais SA, Nathwani R, Dhar V, et al. Effect of frequent dosing of an oral proton pump inhibitor on intragastric ph. Aliment Pharmacol Ther 2006;23:1607 1611. 11. Metz DC, Amer F, Hunt B, et al. Lansoprazole regimens that sustain intragastric ph 6.0: an evaluation of intermittent oral and continuous intravenous infusion dosages. Aliment Pharmacol Ther 2006;23:985 995. 12. Howden CW, Metz DC, Hunt B, et al. Dose-response evaluation of the antisecretory effect of continuous infusion intravenous lansoprazole regimens over 48 hours. Aliment Pharmacol Ther 2006;23:975 984. 13. Van Rensberg CJ, Hartmann M, Thorpe A, et al. Intragastric ph during continuous infusion with pantoprazole in patients with bleeding peptic ulcer. Am J Gastroenterol 2003;98:2685 2691. 14. Patchett SE, Euright H, Afdhal N, et al. Clot lysis by gastric juice: an in vitro study. Gut 1989;30:1704 1707. 15. Pilotto A, Franceschi M, Paris F. Recent advances in the treatment of GERD in the elderly: focus on proton pump inhibitors. Int J Clin Pract 2005;59:1204 1209. 16. Klotz U. Pharmacokinetic considerations in the eradication of Helicobacter pylori. Clin Pharmacokinetics 2000;38:243 270. Received January 4, 2008. Accepted March 6, 2008. Address requests for reprints to: Loren Laine, MD, Room 12-137, L. A. County and U. S. C. Medical Center, 1200 N. State St., Los Angeles, California 90033. e-mail: llaine@usc.edu; fax: (323) 226-7573. Supported by TAP Pharmaceutical Products, Inc, Lake Forest, IL. Conflicts of interest and financial disclosures: Loren Laine: research support for this study by TAP Pharmaceutical Products, Inc; consultant: AstraZeneca, Santarus, Horizon, Novartis, Merck. Abbid Shah and Shahrooz Bemanian: research support for this study by TAP Pharmaceutical Products, Inc. This trial was an investigator-initiated study, and the funding source had no input on the study design, study execution, data analysis, or drafting of the manuscript (ClinicalTrials.gov identifier: NCT00573924). 1841