Evidence for Endoscopic Ulcers as Meaningful Surrogate Endpoint for Clinically Significant Upper Gastrointestinal Harm

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2009;7:1156 –1163

REVIEW Evidence for Endoscopic Ulcers as Meaningful Surrogate Endpoint for Clinically Significant Upper Gastrointestinal Harm ANDREW MOORE,* INGVAR BJARNASON,‡ BYRON CRYER,§ LUIS GARCIA–RODRIGUEZ,储 LARRY GOLDKIND,¶ ANGEL LANAS,# and LEE SIMON** *Pain Research and Nuffield Department of Anaesthetics, University of Oxford, John Radcliffe Hospital, Oxford, United Kingdom; ‡Department of Medicine, King’s College Hospital, Denmark Hill, London, United Kingdom; §Medical Service, Department of Veterans Affairs Medical Center, Dallas, Texas; 储Spanish Centre for Pharmacoepidemiological Research (CEIFE), Madrid, Spain; ¶Uniformed Services University of Health Sciences, School of Medicine, Bethesda, Maryland; # Department of Gastroenterology, University Hospital, Institute of Health Sciences of Aragon, Zaragoza, Spain; **Harvard Medical School, Beth Israel Deaconess Medical Center, Boston, Massachusetts

mal biologic processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention.

See Editorial on page 1147. BACKGROUND & AIMS: Surrogate endpoints are biomarkers intended to substitute for a clinical endpoint. Are endoscopic ulcers a useful surrogate endpoint for a biological progression to clinical endpoints of ulcer complications (perforation, ulcers, and bleeds), hospital admission, or death? METHODS: Review of randomized trials, meta-analyses, clinical outcomes trials, and observational studies. RESULTS: No large study examined both endoscopic and clinical endpoints. Endoscopic ulcers and clinically significant ulcer complications were affected in the same direction and to about the same extent in 4 distinct circumstances: (1) by risk factors—age, previous history of symptomatic ulcer or bleeding, Helicobacter pylori, aspirin; (2) in studies of antiulcer treatments with differing modes of action, especially in relation to nonsteroidal antiinflammatory drug toxicity, and Helicobacter pylori infection; (3) in studies evaluating ulcer complications with Cox-2 selective drugs and nonsteroidal anti-inflammatory drugs; and (4) in studies of interventions in patients with high risk of recurrent ulcer bleed needing nonsteroidal anti-inflammatory drug therapy. All study designs showed consistent and reproducible effects on gastrointestinal ulcer complications paralleling endoscopy. CONCLUSIONS: Consistent and plausible findings from disparate populations and designs make endoscopic ulcers a strong candidate for surrogacy, though direct progression from endoscopic ulcers to ulcer complications cannot be demonstrated. Large outcome studies are needed to establish the power of the surrogacy, absolute risk of clinical outcomes, and to identify the totality of risks and benefits of new pharmacologic therapies.

A

surrogate is something that acts for or takes the place of another. A National Institutes of Health working group defined surrogacy based on prediction of more serious outcome from epidemiologic, therapeutic, pathophysiologic, or other scientific evidence.1 ●

A biological marker (biomarker) is a characteristic that is objectively measured and evaluated as an indicator of nor-

●

A clinical endpoint is a characteristic or variable that reflects how a patient feels, functions, or survives.

●

A surrogate endpoint is a biomarker that is intended to substitute for a clinical endpoint. A surrogate endpoint is expected to predict clinical benefit (or harm, or lack of benefit or harm) based on epidemiologic, therapeutic, pathophysiologic, or other scientific evidence.

There are other, generally similar, definitions.2,3 Surrogate endpoints are widely used: LDL cholesterol and increased blood pressure as indicators of more important outcomes of stroke and heart attack,4 CD4 cell count and viral load for risk of rapid progression to AIDS,5 and eradication of Helicobacter pylori for peptic ulcer recurrence.6 Surrogate endpoint presence does not imply the occurrence of the outcome, but rather the representation of a process leading to serious disease or death. Predictive markers, by contrast, are characteristics associated with response or lack of response to a particular therapy. Markers can be both surrogate (LDL cholesterol and cardiovascular event reduction in trials of cholesterol reduction therapy), and predictive (increased LDL cholesterol associated with increased risk of adverse cardiovascular outcomes). Gastroenterologists use surrogate endpoints. For example, while adenomatous colorectal polyps have a small probability of developing into a metastatic tumor, it is believed that most colon cancers develop from such polyps. Colorectal polyps are recognized as a surrogate endpoint for colon cancer, despite the fact that colon cancer does not occur in most patients with polyps; standard clinical practice is to remove all detectable adenomatous colorectal polyps on discovery at colonoscopy. Abbreviations used in this paper: CI, confidence interval; COX-1, cyclooxygenase-1; COX-2, cyclooxygenase-2; GI, gastrointestinal; H2RAs, histamine-2 receptor antagonists; NSAIDs, traditional nonsteroidal anti-inflammatory drugs; OR, odds ratio; PPI, proton pump inhibitor; RCT, randomized controlled trial; RR, relative risk. © 2009 by the AGA Institute 1542-3565/09/$36.00 doi:10.1016/j.cgh.2009.03.032

November 2009

The US Food and Drug Administration recognizes colon polyps as surrogate endpoints for colon cancer.7 Upper gastroduodenal endoscopy investigates signs and symptoms of clinically significant upper gastrointestinal (GI) tract damage, and for surveillance for esophageal and gastric cancer in high-risk populations. Endoscopy is also used in clinical trials, regardless of signs or symptoms, to determine the occurrence of damage in the upper GI tract. Endoscopy has become the gold standard for assessing the potential upper GI benefits of antiulcer agents and damage caused by drugs like aspirin, traditional nonsteroidal anti-inflammatory drugs (NSAIDs), and cyclooxygenase-2 (COX-2)-selective NSAIDs (coxibs). It has been argued that there is a biological progression from lesser to more severe disease with NSAID adverse events, from endoscopic erosions and asymptomatic ulcers, to symptomatic ulcers, to ulcer complications (bleeding and perforation), and death,8,9 though asymptomatic ulcers can bleed. Endoscopic ulcers may represent an early step in the biological progression from mucosal injury to symptomatic ulcer and ulcer complication. These complications include: ●

Obstruction complicating peptic ulcers: this is a function of the ulcer’s anatomical location. Lesions involving the pylorus are more likely to present with obstruction than those in the gastric corpus.

●

Perforation complicating peptic ulcers: like obstruction, this also depends on anatomical location. Most perforating ulcers occur in the duodenum.

●

Bleeding: this is less predictable, but is increasingly seen in association with antithrombotic agents.

●

Death: with improved resuscitation and endoscopic therapy techniques, this complication largely depends on comorbidity.

Endoscopic ulcers expose GI toxicity, but are they a good surrogate endpoint for the development of serious upper GI harm? This review examines the evidence for endoscopic ulcers as surrogate endpoints for serious upper GI complications and symptomatic ulcers, especially their association with use of NSAIDs and coxibs.

Evidence Ideally, evidence for similar effects on surrogate and clinical endpoints would come from the same trial, or set of trials. Because evidence like this is unavailable, 4 other sources of evidence have been used to assess the relationship between endoscopic ulcers and ulcer complications, where factors affecting clinical outcomes affect endoscopic outcomes, in the same direction, and to much the same extent. These are: ●

Effects of risk factors: age, previous history of symptomatic ulcer or bleeding, presence of Helicobacter pylori, and aspirin or NSAID use.

●

Clinical trials of antiulcer treatments.

●

Clinical trials of coxibs and NSAIDs.

●

Interventions in patients with high risk of recurrent GI bleed who need NSAID therapy.

ENDOSCOPIC ULCERS AS A SURROGATE ENDPOINT

1157

Risk Factors Age Pooled analysis of 12 RCTs showed the incidence of endoscopic ulcers to increase linearly with age with placebo and NSAIDs.10 NSAID-induced endoscopic gastroduodenal ulcer prevalence increased from 10% in patients aged 20 –29 years to 34% in those aged 70 –79 years.11 Patients aged ⱖ75 years had increased risk of upper GI complications (odds ratio [OR] 2.5; 95% CI, 1.5– 4.1).12 There was a trend toward more endoscopic ulcers and a significant increase in perforations, ulcers, and bleeds in those aged ⱖ65 years in a posthoc analysis of rofecoxib or ibuprofen use.13 These findings from clinical trials concur with those of observational studies.14 –17 Meta-analysis of good quality observational studies showed increased risk of gastrointestinal bleeding of more than 4-fold in patients over 70 years.15 The increase in hospital admissions for upper GI bleeding in current NSAID users and nonusers was dramatic after the seventh decade.17

History of Previous Ulcer or Bleeding History of upper GI disease and presence of gastroduodenal erosions at baseline were associated with increased risk of endoscopic ulcers in analysis of 2 endoscopy studies of patients taking placebo, rofecoxib, or ibuprofen.13 Previous upper GI disease had a relative risk (RR) of 4.2 (2.5–7.1) for endoscopic ulcers and 3.8 (1.4 –11) for ulcer bleeds. Baseline gastroduodenal ulcers had a RR of 4.4 (2.6 –7.5) for endoscopic ulcers and 5.0 (1.9 –14) for bleeds. Observational studies also show a history of peptic ulcer disease and ulcer complications to be important risk factors.15,18 History of ulcer disease increased the risk of a bleed 6-fold, and previous ulcer bleed by almost 15-fold in a metaanalysis of observational studies.15 Relative risks of upper GI bleeding were 4.0 (3.2– 4.9) in patients with a history of uncomplicated ulcer and 8.7 (7.8 –11) in patients with complicated ulcers.18

Helicobacter pylori Helicobacter pylori eradication significantly reduced endoscopic ulcers in the short term compared with no treatment (RR 0.4, 0.3– 0.5), and reduced long term recurrence, including ulcer bleeding (RR 0.2, 0.1– 0.3).19 Randomized trials evaluating Helicobacter pylori eradication in NSAID-naïve users show consistent reduction in the incidence of endoscopic ulcers;20 –23 in 1 study the RR of endoscopic ulcer in the eradication group compared with placebo was 0.35 (0.15– 0.81).21 Two long-term randomized trials have shown reduced ulcer bleeding after Helicobacter pylori eradication. In one, rebleeding occurred in 0% over 7 months with eradication compared with 29% with maintenance ranitidine.24 In another, recurrent bleeding with eradication (4.2%) was less than with maintenance ranitidine (8.3%) over 1.6 years.25

Aspirin Alone Aspirin inhibits cyclooxygenase-1 (COX-1) activity in the gastric mucosa, with increased risk of ulceration and risk of hospitalization due to GI bleeding.26 –29 Endoscopic studies show dose-related increase in ulcer incidence from 1.6% with 81 mg aspirin to 7.6% with 325 mg;29 the majority developed ulcers on 3.9 g aspirin/day.30 GI bleeding is dose-related, with number

1158

MOORE ET AL

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 11

needed to harm of 120 for doses ⬍162 mg daily, but with number needed to harm of 60 for higher doses.28 Compared with those not taking low dose aspirin, the OR for upper GI bleed is 2.4 (1.8 –3.3).31 For serious bleeding, event rates increase from about 0.3% at doses ⬍162 mg daily to almost 1% at 1500 mg.29

Low Dose Aspirin in Combination With NSAIDs Low dose aspirin used together with NSAIDs is a potent risk factor for upper GI ulcers and ulcer complications in endoscopic studies and clinical outcomes studies, though the magnitude of the effect depends on the comparator—low dose aspirin alone, NSAID alone, or nonuse of low dose aspirin/ NSAID.32 Compared with aspirin alone, increased endoscopic ulcer incidence occurs with aspirin plus NSAIDs. In one trial, the RR for endoscopic ulcers was 16 (4.1– 66) for naproxen plus aspirin 81 mg compared with aspirin 81 mg alone.33 In another the RR was 3.6 (1.7–7.6) for naproxen plus aspirin 325 mg compared with aspirin 325 mg alone.34 A subgroup analysis of a large randomized trial showed no significant increased risk of ulcer complications with aspirin (ⱕ325 mg) in combination with NSAIDs (RR 1.7, 0.7– 4.3).35 NSAIDs plus aspirin had higher rates of bleeding and hospital admission (hazard ratio 1.6; 95% CI, 1.3–2.1) than either celecoxib plus aspirin (hazard ratio 1.0; 95% CI, 0.8 –1.3) or NSAIDs alone (referent) in an observational study of elderly patients.36 NSAID plus aspirin had increased risk of upper GI bleeding and perforation compared with nonusers (RR 13, 8.5–21), aspirin alone (RR 2.0, 1.7–2.3), or aspirin plus low/ medium-dose NSAIDs (RR 2.2; 1.0 – 4.6).37 Combination of NSAID or coxib plus low dose aspirin was associated with major increases in RR for upper GI bleeding of 13 (7–23) and 15 (3.3– 64), respectively, after adjustment for confounders.18

Ulcer Prevention Strategies Misoprostol Misoprostol is effective in reducing endoscopic gastroduodenal ulcers in patients taking NSAIDs. In a meta-analysis, ORs for an endoscopic gastric or duodenal ulcer were 0.3 (0.2– 0.4) and 0.5 (0.3– 0.7) at 3–24 months with misoprostol compared with control.38 Misoprostol compared with placebo in NSAID users had a RR of 0.3 (0.27– 0.41) for endoscopic ulcers and 0.6 (0.36 – 0.91) for serious GI events (defined as hemorrhage, recurrent upper GI bleeds, perforation, obstruc-

tion, or melena, including death from any of these) in a metaanalysis.39 Two 1-year studies of arthritis patients on chronic NSAID therapy showed that misoprostol reduced the risk of gastric ulcers.40,41

Histamine-2 Receptor Antagonists In a systematic review of RCTs comparing histamine-2 receptor antagonists (H2RAs) with placebo in NSAID users, the RR for all endoscopic ulcers was 0.4 (0.2– 0.7),38 but without a significant reduction in serious GI events (RR 0.3, 0.01– 8.1), because of small numbers of events.37 Two observational studies showed that H2RAs had reduced the risk of upper GI peptic ulcer bleeding. In 1 study42 the RR was 0.7 (0.5– 0.9), and in the other31 the OR was 0.6 (0.4 – 0.8).

Proton Pump Inhibitors Proton pump inhibitors (PPIs) reduce endoscopic ulcers compared with placebo in patients receiving NSAIDs (RR 0.4, 0.3– 0.5), but with a nonsignificant reduction of GI complications (including hemorrhage, recurrent upper GI bleeds, perforation, obstruction, or melena; RR 0.5, 0.1–2.9), because of small numbers of events.38 Combined analysis of 2 RCTs showed a reduction in serious GI adverse events from 2.7% to 1.0% with esomeprazole 20 mg or 40 mg, but with only 4 upper GI bleeds, all with placebo in patients taking NSAID, coxib, or coxib with low dose aspirin.43 Esomeprazole plus celecoxib reduced rebleeding rates to 0 over 13 months in high risk patients with previous upper GI bleeding, compared to 9% with celecoxib alone.44 In 1 observational study42 the RR for upper GI bleeding was 0.3 (0.27– 0.39); in another31 the OR was 0.6 (95% CI, 0.4 – 0.9); a third reported an OR of 0.8 (0.6 –1.0).45 Two observational studies showed that PPI use is associated with a reduced risk of recurrent upper GI bleeding. In 1 study46 the RR was 0.2 (0.02–1.0); in the other47 the OR was 0.51 (0.26 – 0.99).

COX-2 Selective Agents Versus NSAIDs Randomized Trials Table 1 summarizes consistently lower RR rates for coxibs than NSAID comparators for endoscopic gastroduodenal, gastric, and duodenal ulcers.48 For all COX-2 selective drugs compared with all NSAIDs, the RR for gastroduodenal ulcers was 0.26 (0.23– 0.30). Absolute rates of endoscopic gastric ulcers were 3.8% and 19% for coxib and NSAID, and for endo-

Table 1. Relative Risk (95% CI) of Endoscopic Gastroduodenal, Gastric, and Duodenal Ulcers, Ulcer Complications, and Ulcer Complications/Symptomatic Ulcers for Coxibs Versus NSAIDs in RCTs48 COX-2 compared with NSAIDs

Gastroduodenal ulcer

Gastric ulcer

Duodenal ulcer

Ulcer complications (POBs)

Ulcer complications and symptomatic ulcers (PUBs)

All COX-2 selective Celecoxib Valdecoxib Rofecoxib Etoricoxib Lumiracoxib

0.26 (0.23–0.30) 0.21 (0.16–0.28) 0.29 (0.21–0.39) 0.26 (0.21–0.32) 0.37 (0.18–0.77) 0.26 (0.18–0.39)

0.21 (0.18–0.25) 0.20 (0.14–0.28) 0.24 (0.18–0.37) 0.20 (0.15–0.26) — 0.25 (0.16–0.40)

0.34 (0.25–0.45) 0.29 (0.18–0.47) 0.39 (0.21–0.70) 0.36 (0.14–0.93) — 0.20 (0.09–0.43)

0.39 (0.31–0.50) 0.23 (0.07–0.76) 0.35 (0.14–0.87) 0.42 (0.24–0.73) 0.57 (0.31–1.04) 0.36 (0.24–0.55)

0.41 (0.33–0.52) 0.39 (0.21–0.73) 0.23 (0.15–0.36) 0.44 (0.34–0.58) 0.50 (0.33–0.75) 0.56 (0.41–0.78)

POBs, perforations, obstruction, and bleeds; PUBs, perforation, obstruction, bleed, or symptomatic ulcers.

November 2009

ENDOSCOPIC ULCERS AS A SURROGATE ENDPOINT

scopic duodenal ulcers 1.6% and 5.2%, respectively. For coxibs, gastroduodenal ulcers were no different from placebo. Table 1 also summarizes RR for clinical outcomes of ulcer complications and ulcer complications plus symptomatic ulcers, with consistently lower rates for individual coxibs than NSAID comparators.48 For all coxibs compared with all NSAIDs, the RR for ulcer complications was 0.39 (0.31– 0.50), and for ulcer complications plus symptomatic ulcers 0.41 (0.33– 0.52). Absolute rates of ulcer complications were 0.2% for coxibs and 0.6% for NSAIDs, and ulcer complications plus symptomatic ulcers 0.6% and 1.1%, respectively.

Observational Studies Observational studies compare rates of upper GI events, usually hospital admission, for people who use coxib or NSAID with those who use neither, making adjustments for differences in risk factors.18,45,49 –51 They show consistently lower risk for coxib than for NSAID, comparable in size to those for endoscopic ulcers and clinical outcomes from RCTs (Table 2).

Very High GI Risk Populations Requiring NSAID Therapy There is consistent effect of protective therapies for endoscopic and bleeding ulcers in patients with a very high risk of bleeding, such as a previous GI bleed. US and UK guidelines recommend patients at risk for ulcer disease requiring treatment for arthritis receive coxib or NSAID in combination with PPI.52,53 A single trial compared celecoxib with diclofenac plus omeprazole in similar patients with previous NSAID-induced GI bleeding.54,55 Recurrent endoscopic ulcer incidence at 6 months was comparable with celecoxib (19%) and diclofenac plus omeprazole (26%).54 Recurrent bleeding during the 6-month study was comparable for celecoxib (4.9%) and diclofenac plus omeprazole (6.4%).55 With diclofenac plus omeprazole endoscopic ulcer incidence was 1.4 times higher than for celecoxib; the incidence of recurrent bleeds was 1.3 times higher. This argues for consistent

1159

effects of different therapies on both putative surrogate endpoint and clinical outcome.

Evidence for Endoscopic Ulcers as a Surrogate Endpoint Formal assessments of surrogacy have been made previously using systematic reviews or meta-analyses, in which surrogate and clinical endpoints have been collected in the same trials.56-58 These approaches are not applicable to endoscopic ulcers, as endoscopic ulcers and clinically significant events are not available in the same trial in large enough numbers to compare directly. An indirect approach must be used that seeks evidence that both outcomes change in the same way to the same extent, but in different studies. There appears to be no universal approach validated across a number of different circumstances. Indeed, a universal approach may be inappropriate because of the individual circumstances of each surrogate endpoint. Reference to general principles is the only practical method presently available, but seems to work for the case of endoscopic ulcers. We examined consistency of effect of a range of factors and interventions involving endoscopic ulcers and serious clinical upper GI events. These included effects of risk factors, antiulcer treatments, coxibs, and NSAIDs, and interventions in patients with high risk of recurrent GI bleed but needing NSAID therapy. (1)

(2) (3)

There was a consistent effect: risk factors and interventions that affected serious upper GI bleeding events generally had the same direction and comparable magnitude of effect on endoscopic ulcers. Endoscopic ulcers are prognostic of more serious upper GI bleeding events. There is biological plausibility for linking the 2 outcomes, with a significant body of evidence suggesting that endoscopic ulcers represent an early step in the biological progression from mucosal injury to symptomatic ulcer and ulcer complications.

Table 2. Risk of Events From Endoscopic, Clinical Outcomes, and Retrospective Population-Based Studies, Compared With Nonusers RR (95% CI) Endoscopy

COX-2 selective NSAID vs COX-2 selective NSAID vs COX-2 selective NSAID vs COX-2 selective NSAID vs Celecoxib vs NSAIDs Celecoxib vs diclofenac Valdecoxib vs NSAIDs Rofecoxib vs NSAIDs Lumiracoxib vs NSAIDs Etoricoxib vs NSAIDs

NSAIDs diclofenac ibuprofen naproxen

studiesa

Clinical outcomes

Ulcers

POBS

0.26 (0.23–0.30) 0.25 (0.18–0.35) 0.27 (0.23–0.32) 0.25 (0.20–0.32) 0.21 (0.16–0.28)

0.39 (0.31–0.50)

0.29 (0.21–0.39) 0.26 (0.21–0.32) 0.26 (0.18–0.39) 0.37 (0.18–0.77)

Adjusted RR (95% CI)

0.46 (0.30–0.71) 0.34 (0.24–0.48) 0.23 (0.07–0.76) 0.31 (0.06–1.61)

studiesa

Retrospective population-based studiesb

PUBs

Upper GI bleed/hemorrhage

0.39 (0.21–0.73) 0.41 (0.30–0.55) 0.23 (0.15–0.36) 0.44 (0.34–0.58) 0.56 (0.41–0.78) 0.50 (0.33–0.75)c

0.23 (0.12–0.42) 0.31 (0.15–0.63)

POBs, perforations, obstruction, and bleeds; PUBs, perforation, obstruction, bleed, or symptomatic ulcers. aFrom Rostom et al.48 bCelecoxib vs NSAIDs from Lanas et al18 and coxibs vs NSAIDs from Mamdani et al.49 cNot including data from the Multinational Etoricoxib and Diclofenac Arthritis Long-term (MEDAL) programme.

0.53 (0.29–1.0)

1160

MOORE ET AL

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 11

(4) There is copious evidence from randomized trials, observational studies, and meta-analyses of RCTs and observational studies comparing coxibs with NSAIDs, with a large number of events.59 (5) Both outcomes are influenced by a wide variety of different drugs of different classes, with numerous different mechanisms of action. The evidence we have of endoscopic ulcers is that there is a prima facie case for regarding them as a useful surrogate endpoint for more serious upper GI bleeding events. What represents an ulcer for use in the clinical trial setting of ulcer prevention has been debated. A systematic review of the literature of ulcer prevention in NSAID users reported that a gastric or duodenal lesion ⱖ3 mm in diameter with a significant depth is the definition of ulcer used by a substantial majority of investigators.60

Limitations of the Evidence for Endoscopic Ulcers as a Surrogate The single largest limitation for considering endoscopic ulcers as a surrogate endpoint is the lack of evidence showing a direct relationship in the same trial between endoscopic gastroduodenal ulcers and clinically significant upper GI harm. Such evidence is available for blood pressure and cholesterol, but for endoscopy we only have information from a single trial, of limited size, and in patients at particularly high risk.54 Consequently, indirect evidence from a range of interventions and risk factors in direction and magnitude on endoscopic ulcers and clinical outcomes has been used. Trials conducted to assess GI damage from NSAIDs or coxibs exclude patients with gastroduodenal lesions from further participation in the studies once a lesion is detected. Susceptible patients are removed; those who remain are less

susceptible to GI damage, implying fewer symptomatic ulcers and GI complications later. Removing patients with endoscopic ulcers or high numbers of erosions makes it impossible to observe progression of surrogate endoscopic lesions to more serious clinical events. Another limitation is the evolution of outcomes in clinical outcome trials, with different definitions of GI events (Figure 1). Definitions have moved from assessment of serious upper GI ulcer complications in the Misoprostal Ulcer Complications Outcome Safety Assessment (MUCOSA) trial,12 to evaluation of visible upper GI complications (Celecoxib Long-term Arthritis Safety Study [CLASS]),35 to inclusion of symptomatic ulcers (VIGOR and Multinational Etoricoxib and Diclofenac Arthritis Long-term [MEDAL]),61,62 with addition of acute and chronic anemia attributed to bleeding (Therapeutic Arthritis Research and Gastrointestinal Event Trial [TARGET]).63 A composite measure of clinically significant upper and lower GI events (CSULGIES)64 may be 1 way of standardizing the impact of NSAIDs or other drugs throughout the GI tract, similar to the Anti-platelet Trialists’ Collaboration (APTC) end point for cardiovascular end points.65 Rules about use of gastroprotection and low dose aspirin have also evolved. Initial trials were conducted without gastroprotection, but MEDAL made strenuous efforts to ensure proper use of both. Only 7% of patients in MUCOSA used aspirin, no aspirin use was allowed in VIGOR, but it increased to 21% in CLASS, 24% in TARGET, and 33% in the MEDAL program. The MEDAL program also involved more extensive use of PPIs, in almost 40% of patients enrolled. Not all analyses were in agreement. Observational studies can also give conflicting results, with rofecoxib being associated with an increased risk of upper GI bleeding compared with nonusers, while celecoxib was not.18,49,66 Few studies take into account the effects of dose, which can be large.45 Small numbers

Figure 1. Changing definitions of gastrointestinal events.

November 2009

of events increase the risk of chance or misleading findings,67 and it has been suggested that 200 events are needed to be confident of a result.68

Conclusions One of the most challenging aspects in designing clinical outcome trials is the choice of primary outcome measure to assess benefit. Clinically meaningful outcomes (in this case serious GI complications) are most useful, but it is increasingly unethical to have designs in which patients are unnecessarily exposed to developing such an outcome. The MEDAL program62 emphasizes that it is now not acceptable to expose patients at higher risk of GI bleed to NSAIDs without gastroprotection. Though this guidance is usually ignored in clinical practice,59 conventional outcome trials have to use existing best practice. Alternatives are the use of surrogate outcomes, despite their limitations, or trials in patients at high risk, with sufficient sensitivity to detect potential benefits of new interventions because of high event rates. A number of drugs have been approved with labeling that presents a reduced incidence of endoscopic ulcers as a surrogate marker for serious GI toxicity, without definitive serious outcome studies, like coadministration of PPIs with NSAIDs. There is a considerable body of evidence showing that interventions that alter endoscopic ulcer incidence also alter more serious ulcer complications, in the same direction and to much the same extent. This is seen with different interventions, with different modes of action, in evidence of good quality and quantity. The consistency and plausibility of findings from disparate populations and study designs makes endoscopic ulcers a strong candidate. The extent of data supporting the surrogacy of asymptomatic endoscopically-identified ulcers is similar to other fields where surrogates have been validated. While endoscopic ulcers may reasonably be considered as surrogates for clinical outcomes of bleeding, and therefore clinically relevant, large outcome studies are still needed to establish the strength of the surrogacy, absolute associated risk of clinical outcomes, and to identify the totality of risks and benefits associated with a new pharmacologic therapy. References 1. Biomarkers Definitions Working Group. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharm Ther 2001;69:89 –95. 2. Prentice RL. Surrogate endpoints in clinical trials: definition and operational criteria. Stat Med 1989;8:431– 440. 3. Temple R. Are surrogate markers adequate to assess cardiovascular disease drugs? JAMA 1999;282:790 –795. 4. International Conference on Surrogate Endpoints and Biomarkers. Remarks by: Jane E. Henney, MD, Commissioner of Food and Drugs. Available at: http://www.fda.gov/oc/speeches/surrogates8. html. Accessed September 2008. 5. Bren L. The advancement of controlled clinical trials. FDA Consumer Magazine. Last updated March-April 2007. Available at: http://www.fda.gov/fdac/features/2007/207_trials.html. Accessed September 2008. 6. Points to consider: clinical development and labelling of antiinfective drug products. Helicobacter pylori-associated peptic ulcer disease. Last updated June 1995. Available at: http://www.fda. gov/cder/guidance/hpptc.htm. Accessed September 2008.

ENDOSCOPIC ULCERS AS A SURROGATE ENDPOINT

1161

7. FDA Gastrointestinal Advisory Committee. Chemoprevention of colorectal cancer. Last updated March 2002. Available at: http://www.fda.gov/ohrms/dockets/ac/02/transcripts/3845t1. pdf. Accessed September 2008. 8. Tramer MR, Moore RA, Reynolds DJ, et al. Quantitative estimation of rare adverse events which follow a biological progression: a new model applied to chronic NSAID use. Pain 2000;85:169 – 182. 9. Hawkey CJ. Non-steroidal anti-inflammatory drugs and peptic ulcers. BMJ 1990;300:278 –284. 10. Boers M, Tangelder MJ, van Ingen H, et al. The rate of NSAIDinduced endoscopic ulcers increases linearly, but not exponentially with age: a pooled analysis of 12 randomized trials. Ann Rheum Dis 2006;66:417– 418. 11. Cheatum DE, Arvanitakis C, Gumpel M, et al. An endoscopic study of gastroduodenal lesions induced by nonsteroidal antiinflammatory drugs. Clin Ther 1999;21:992–1003. 12. Silverstein FE, Graham DY, Senior JR, et al. Misoprostol reduces serious gastrointestinal complications in patients with rheumatoid arthritis receiving nonsteroidal anti-inflammatory drugs. A randomized, double-blind, placebo-controlled trial. Ann Intern Med 1995;123:241–249. 13. Hawkey CJ, Laine L, Harper SE, et al. Influence of risk factors on endoscopic and clinical ulcers in patients taking rofecoxib or ibuprofen in two randomized controlled trials. Aliment Pharmacol Ther 2001;15:1593–1601. 14. MacDonald TM, Morant SV, Robinson GC, et al. Association of upper gastrointestinal toxicity of non-steroidal anti-inflammatory drugs with continued exposure: cohort study. BMJ 1997;315: 1333–1337. 15. Hernandez-Diaz S, Rodriguez LA. Association between nonsteroidal anti-inflammatory drugs and upper gastrointestinal tract bleeding/perforation: an overview of epidemiologic studies published in the 1990s. Arch Intern Med 2000;160:2093–2099. 16. Hernandez-Diaz S, Rodriguez LA. Incidence of serious upper gastrointestinal bleeding/perforation in the general population: review of epidemiologic studies. J Clin Epidemiol 2002;55:157– 163. 17. Perez-Gutthann S, Garcia-Rodriguez LA, Raiford DS. Individual nonsteroidal antiinflammatory drugs and other risk factors for upper gastrointestinal bleeding and perforation. Epidemiology 1997;8:18 –24. 18. Lanas A, Garcia-Rodriguez LA, Arroyo MT, et al. Risk of upper gastrointestinal ulcer bleeding associated with selective COX-2 inhibitors, traditional non-aspirin NSAIDs, aspirin, and combinations. Gut 2006;55:1731–1738. 19. Moayyedi P, Soo S, Deeks J, et al. Eradication of Helicobacter pylori for non-ulcer dyspepsia. Cochrane Database Syst Rev 2006;2:CD002096. 20. Graham DY, Lew GM, Klein PD, et al. Effect of treatment of Helicobacter pylori infection on the long-term recurrence of gastric or duodenal ulcer. A randomized, controlled study. Ann Intern Med 1992;116:705–708. 21. Chan FK, To KF, Wu JC, et al. Eradication of Helicobacter pylori and risk of peptic ulcers in patients starting long-term treatment with non-steroidal anti-inflammatory drugs: a randomised trial. Lancet 2002;539:9 –13. 22. Labenz J, Blum AL, Bolten WW, et al. Primary prevention of diclofenac associated ulcers and dyspepsia by omeprazole or triple therapy in Helicobacter pylori positive patients: a randomised, double blind, placebo controlled, clinical trial. Gut 2002; 51:329 –335. 23. Chan FK, Sung JJ, Chung SC, et al. Randomised trial of eradication of Helicobacter pylori before non-steroidal anti-inflammatory drug therapy to prevent peptic ulcers. Lancet 1997;350:975– 979.

1162

MOORE ET AL

24. Graham DY, Hepps KS, Ramirez FC, et al. Treatment of Helicobacter pylori reduces the rate of rebleeding in peptic ulcer disease. Scand J Gastroenterol 1993;28:939 –942. 25. Riemann JF, Schilling D, Schauwecker P, et al. Cure with omeprazole plus amoxicillin versus long-term ranitidine therapy in Helicobacter pylori-associated peptic ulcer bleeding. Gastrointest Endosc 1997;46:299 –304. 26. Yeomans ND, Lanas AI, Talley NJ, et al. Prevalence and incidence of gastroduodenal ulcers during treatment with vascular protective doses of aspirin. Aliment Pharmacol Ther 2005;22:795– 801. 27. Garcia Rodriguez LA, Hernandez-Diaz S, de Abajo FJ. Association between aspirin and upper gastrointestinal complications: systematic review of epidemiologic studies. Br J Clin Pharmacol 2001;52:563–571. 28. Derry S, Loke YK. Risk of gastrointestinal haemorrhage with long term use of aspirin: meta-analysis. BMJ 2000;321:1183–1187. 29. M Tramèr. Aspirin, like all other drugs, is a poison. BMJ 2000; 321:1170 –1171. 30. Lanza FL, Royer GL, Nelson RS. Endoscopic evaluation of the effects of aspirin, buffered aspirin, and enteric-coated aspirin on gastric and duodenal mucosa. N Eng J Med 1980;303:136 –138. 31. Lanas A, Bajador E, Serrano P, et al. Nitrovasodilators, low-dose aspirin, other nonsteroidal antiinflammatory drugs, and the risk of upper gastrointestinal bleeding. N Engl J Med 2000;343:834 – 839. 32. Hernández-Díaz S, García Rodríguez LA. Cardioprotective aspirin users and their excess risk of upper gastrointestinal complications. BMC Medicine 2006;4:22. 33. Goldstein JL, Aisenberg J, Zakko SF, et al. Endoscopic ulcer rates in healthy subjects associated with use of aspirin (81 mg q.d.) alone or coadministered with celecoxib or naproxen: a randomized, 1-week trial. Dig Dis Sci 2008;53:647– 656. 34. Goldstein JL, Lowry SC, Lanza FL, et al. The impact of low-dose aspirin on endoscopic gastric and duodenal ulcer rates in users of a non-selective non-steroidal anti-inflammatory drug or a cyclooxygenase-2-selective inhibitor. Aliment Pharmacol Ther 2006; 23:1489 –1498. 35. Silverstein FE, Faich G, Goldstein JL, et al. Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASS study: a randomized controlled trial. Celecoxib Long-term Arthritis Safety Study. JAMA 2000;284:1247–1255. 36. Rahme E, Bardou M, Dasgupta K, et al. Hospitalization for gastrointestinal bleeding associated with non-steroidal anti-inflammatory drugs among elderly patients using low-dose aspirin: a retrospective cohort study. Rheumatology (Oxford) 2007;46: 265–272. 37. de Abajo FJ, Garcia Rodriguez LA. Risk of upper gastrointestinal bleeding and perforation associated with low-dose aspirin as plain and enteric-coated formulations. BMC Clin Pharmacol 2001;1:1. 38. Rostom A, Dube C, Wells G, et al. Prevention of NSAID-induced gastroduodenal ulcers. Cochrane Database Syst Rev 2002;4: CD002296. 39. Brown TJ, Hooper L, Elliott RA, et al. A comparison of the costeffectiveness of five strategies for the prevention of non-steroidal anti-inflammatory drug-induced gastrointestinal toxicity: a systematic review with economic modelling. Health Technol Assess 2006;10:iii–xiii, 1. 40. Elliott SL, Yeomans ND, Buchanan RR, et al. Efficacy of 12 months’ misoprostol as prophylaxis against NSAID-induced gastric ulcers: a placebo-controlled trial. Scand J Rheumatol 1994; 23:171–176. 41. Agrawal NM, Van Kerckhove HE, Erhardt LJ, et al. Misoprostol coadministered with diclofenac for prevention of gastroduodenal ulcers. A one-year study. Dig Dis Sci 1995;40:1125–1131.

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 11

42. Lanas A, García-Rodríguez LA, Arroyo MT, et al. Effect of antisecretory drugs and nitrates on the risk of ulcer bleeding associated with nonsteroidal anti-inflammatory drugs, antiplatelet agents, and anticoagulants. Am J Gastroenterol 2007;102:507–515. 43. Scheiman JM, Yeomans ND, Talley NJ, et al. Prevention of ulcers by esomeprazole in at-risk patients using non-selective NSAIDs and COX-2 inhibitors. Am J Gastroenterol 2006;101:701–710. 44. Chan FK, Wong VW, Suen BY, et al. Combination of a cyclooxygenase-2 inhibitor and a proton-pump inhibitor for prevention of recurrent ulcer bleeding in patients at very high risk: a doubleblind, randomised trial. Lancet. 2007;369:1621–1626. 45. García-Rodríguez L, Barreales Tolosa L. Risk of upper gastrointestinal complications among users of traditional NSAIDs and COXIBs in the general population. Gastroenterology 2007;132: 498 –506. 46. García-Rodríguez LA, Ruigómez A. Secondary prevention of upper gastrointestinal bleeding associated with maintenance acid-suppressing treatment in patients with peptic ulcer bleed. Epidemiology 1999;10:228 –232. 47. Massó-Gónzalez EL, García-Rodríguez LA. PPIs reduce the longterm risk of recurrent UGIB: an observational study. Aliment Pharmacol Ther 2008;28:629 – 637. 48. Rostom A, Muir K, Dube C, et al. Gastrointestinal safety of cyclooxygenase-2 inhibitors: a Cochrane Collaboration systematic review. Clin Gastroenterol Hepatol 2007;5:818 – 828. 49. Mamdani M, Rochon PA, Juurlink DN, et al. Observational study of upper gastrointestinal haemorrhage in elderly patients given selective cyclo-oxygenase-2 inhibitors or conventional non-steroidal anti-inflammatory drugs. BMJ 2002;325:624. 50. MacDonald TM, Morant SV, Goldstein JL, et al. Channelling bias and the incidence of gastrointestinal haemorrhage in users of meloxicam, coxibs, and older, non-specific non-steroidal antiinflammatory drugs. Gut 2003;52:1265–1270. 51. Norgard B, Pedersen L, Johnsen SP, et al. COX-2-selective inhibitors and the risk of upper gastrointestinal bleeding in high-risk patients with previous gastrointestinal diseases: a populationbased case-control study. Aliment Pharmacol Ther 2004;19: 817– 825. 52. American College of Rheumatology Subcommittee on Osteoarthritis Guidelines: recommendations for the medical management of osteoarthritis of the hip and knee: 2000 update. Arthritis Rheum 2000;43:1905–1915. 53. National Collaborating Centre for Chronic Conditions. Osteoarthritis: national clinical guideline for care and management in adults. London: Royal College of Physicians, 2008. 54. Chan FK, Hung LC, Suen BY, et al. Celecoxib versus diclofenac plus omeprazole in high-risk arthritis patients: results of a randomized double-blind trial. Gastroenterology 2004;127:1038 – 1043. 55. Chan FK, Hung LC, Suen BY, et al. Celecoxib versus diclofenac and omeprazole in reducing the risk of recurrent ulcer bleeding in patients with arthritis. N Engl J Med 2002;347:2104 –2110. 56. Molenberghs G, Buyse M, Geys H, et al. Statistical challenges in the evaluation of surrogate endpoints in randomized trials. Control Clin Trials 2002;23:607– 625. 57. Hackshaw A, Knight A, Barrett-Lee P, et al. Surrogate markers and survival in women receiving first-line combination anthracycline chemotherapy for advanced breast cancer. Br J Cancer 2005;93:1215–1221. 58. Pocock SJ, Lansky AJ, Mehran R, et al. Angiographic surrogate end points in drug-eluting stent trials: a systematic evaluation based on individual patient data from 11 randomized, controlled trials. J Am Coll Cardiol 2008;51:23–32. 59. Moore RA, Derry S, Phillips CJ, et al. Nonsteroidal anti-inflammatory drugs (NSAIDs), cyxlooxygenase-2 selective inhibitors (coxibs) and gastrointestinal harm: review of clinical trials and clinical practice. BMC Musculoskelet Disord 2006;7:79.

November 2009

60. Yeomans ND, Naesdal J. Systematic review: ulcer definition in NSAID ulcer prevention trials. Aliment Pharmacol Ther 2008;27: 465– 472. 61. Bombardier C, Laine L, Reicin A, et al. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. VIGOR Study Group. N Engl J Med 2000;343:1520 –1528. 62. Laine L, Curtis SP, Cryer B, et al.; and the MEDAL Steering Committee: Assessment of upper gastrointestinal safety of etoricoxib and diclofenac in patients with osteoarthritis and rheumatoid arthritis in the Multinational Etoricoxib and Diclofenac Arthritis Long-term (MEDAL) programme: a randomised comparison. Lancet 2007;369:465– 473. 63. Schnitzer TJ, Burmester GR, Mysler E, et al. Comparison of lumiracoxib with naproxen and ibuprofen in the Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET), reduction in ulcer complications: randomised controlled trial. Lancet 2004;364:665– 674. 64. Schaefer M, DeLattre M, Gao X, et al. Assessing the costeffectiveness of COX-2 specific inhibitors for arthritis in the Veterans Health Administration. Curr Med Res Opin 2005;21: 47– 60. 65. Antiplatelet Trialists’ Collaboration. Collaborative overview of randomised trials of antiplatelet therapy-I: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ 1994;308:81–106. 66. Laine L, White WB, Rostom A, et al. COX-2 selective inhibitors in the treatment of osteoarthritis. Semin Arthritis Rheum 2008;38: 165–187. 67. Moore RA, Gavaghan D, Tramèr MR, et al. Size is everything– large amounts of information are needed to overcome random

ENDOSCOPIC ULCERS AS A SURROGATE ENDPOINT

1163

effects in estimating direction and magnitude of treatment effects. Pain 1998;78:209 –216. 68. Flather MD, Farkouh ME, Pogue JM, et al. Strengths and limitations of meta-analysis: larger studies may be more reliable. Control Clin Trials 1997;18:568 –579.

Reprint requests Address requests for reprints to: Dr Andrew Moore, Pain Research and Nuffield Department of Anaesthetics, University of Oxford, John Radcliffe Hospital, Oxford, United Kingdom, OX3 9DU. e-mail: [email protected]; fax: (44) 1865 234539. Acknowledgments Dr Julia Fawcett and colleagues at PAREXEL provided editorial support for the original review (white paper). For this paper they provided only formatting support for references. Conflicts of interest The authors disclose the following: All the authors have received research grants, consulting, or lecture fees from pharmaceutical companies, charities, and government sources at various times. The authors were involved in discussions concerning a review, supported by Pfizer Inc, for the creation of a white paper for presentation to regulatory authorities. The decision to write this paper was the decision of the authors. Pfizer and its agents had no role in the interpretation of data or writing of the paper, or in the decision to publish. The authors had the absolute right to publish the results of their research regardless of any conclusions reached. No payments were received for preparing this manuscript.