|
DIVISION OF ANTI-INFECTIVE DRUG PRODUCTS
POINTS TO CONSIDER
CLINICAL DEVELOPMENT AND LABELING OF
ANTI-INFECTIVE DRUG PRODUCTS
March 1995 Addendum
**DRAFT**
Helicobacter pylori-Associated Peptic Ulcer Disease
Indication # 25
(Revised: June 9, 1995)
The enclosed addendum represents the current major issues related to clinical trials supporting the labeling indication of reduced H. pylori associated peptic ulcer recurrence. This document is meant to guide sponsors during the process of clinical protocol development. Since the field of H. pylori is rapidly expanding, we are soliciting ideas, recommendations, and comments regarding the attached H. pylori Points to Consider Document. Please address comments to the following address:
Maureen Dillon-Parker
Consumer Safety Officer
Division of Anti-Infective Drug Products
HFD-520
5600 Fishers Lane
Rockville, Maryland 20957
H. pylori Eradication as a Surrogate for Reduction of Peptic Ulcer Recurrence
 The agency has concluded that H. pylori eradication can be used as a surrogate for reduction of peptic ulcer recurrence for the purpose of clinical trials. Members of the Division of Anti-Infective Drug Products and the Division of Biometrics recently completed a literature based meta-analysis of studies that evaluated both peptic ulcer recurrence and H. pylori eradication in patients who received various anti-H. pylori therapeutic regimens. These data were presented and discussed at FDA (internal) Scientific Rounds on January 11, 1995, and the following policy decisions were made.
The FDA will accept H. pylori eradication at 4 weeks after completion of therapy as a surrogate for reductions in the rate of gastric and duodenal ulcer recurrence. These policy changes will be applicable for the design of treatment trials which seek to demonstrate reduction of ulcer recurrence in patients with active peptic ulcer disease who are infected with H. pylori and are not taking non-steroidal anti-inflammatory agents.
By definition, ulcer recurrence assumes past healing. Hence, clinical trials that use H. pylori eradication as a surrogate for reduced ulcer recurrence must analyze H. pylori eradication only in patients who achieved ulcer healing.
H. pylori Resistance, Compliance, and Side Effects
A legitimate concern is that approval of H. pylori regimens may lead to the development of widespread resistance among strains in the United States. There is evidence to suggest that this is possible. Over 80 percent of strains in some developing countries are resistant to metronidazole. Some antimicrobial agents are not associated with the development of resistance in vivo while other agents have been shown to be associated with the development of resistance.
If the development of resistance to a specific antimicrobial agent quickly develops, this should become apparent in phase 1 and phase 2 clinical trials when both eradication rates and resistance rates (before and after therapy) should be determined. However, because of the problem of primary resistance in certain geographic areas and the possibility of unrecognized resistance in Phase 1 and 2 trials, we also recommend assessing for resistance during Phase 3 trials.
The following points should be considered during clinical trials:
- The incidence of side effects with multi-agent regimens (double and triple therapy)
- The relationship between the incidence of side effects and compliance
- The relationship between compliance and eradication failure
- The possibility that H. pylori resistance will persist even after stopping the drug regimen
- The possibility that resistant strains can be spread among the community
- The availability of alternate therapeutic options given the possibility that resistant strains may develop on therapy or spread throughout the community.
Since anti-H. pylori regimens can cause gastrointestinal side effects, it is recommended that in phase 3 clinical trials both potential side effects from medication (especially gastrointestinal side effects) and gastrointestinal symptoms be monitored before and during medication administration. These assessments may help distinguish between ulcer symptoms and medication side effects.
Randomization, Blinding, and Controls
All patients who are included in H. pylori eradication treatment trials should be randomized by treatment group. Blinding to treatment group and the results of H. pylori diagnostic tests is recommended if possible. There are four persons which should be blinded: the patient, the investigator/endoscopist, the microbiologist, and the pathologist. Whereas both the patient and the investigator/endoscopist should be blinded to all H. pylori diagnostic tests at enrollment, they do not need to be blinded to urease testing at the eradication time point. On the other hand, the pathologist should be blinded to both urease testing and culture results and the microbiologist should be blinded to both histology and urease testing results regardless of when these tests are performed (at study inclusion or at the eradication time point). The pathologist and microbiologist should also be blinded to all clinical information including treatment group.
The Division does not advocate the use of placebo controls for peptic ulcer studies designed to eradicate H. pylori in patients with active peptic ulcer disease. The choice of appropriate comparator regimens is difficult since there is no currently approved regimen for H. pylori eradication. Only maintenance H2blockers are currently approved for patients with healed ulcers to reduce ulcer recurrence, and placebo controlled studies using these agents have not been carried out for periods longer than 1 year. Sponsors are encouraged to discuss with the Division the need for comparator regimens during the development of clinical protocols.
Factorial Design
Section 21, Part 300.50 of the Code of Federal Regulations (CFR) titled Fixed-combination prescription drugs for humans states that "two or more drugs may be combined in a single dosage form when each component makes a contribution to the claimed effects and the dosage of each component (amount, frequency, duration) is such that the combination is safe and effective for a significant patient population requiring such concurrent therapy as defined in the labeling for the drug. Special cases of this general rule are where a component is added: (1) "To enhance the safety or effectiveness of the principal active component..." This regulation suggests that each of the components of a fixed combination regimen needs to contribute to the overall efficacy of the combination. Although drug combinations for H. pylori are not "fixed", factorial designs are recommended by the Division to ensure that each component contributes to overall efficacy. Since factorial designs may induce some complexity to H. pylori study designs, "modified factorial designs" may be considered in certain situations: (1) when H. pylori resistance has been shown to develop in association with antibiotic monotherapy (making monotherapy unethical) and, (2) when eradication rates for single agents have been well documented in the literature. It is suggested that sponsors discuss appropriate factorial designs with the Division during protocol development.
Inactive (Healed) Peptic Ulcer Disease
To be completed
Inclusion and Exclusion Criteria
To be completed
Diagnosis of Peptic Ulcer Disease
Endoscopy is now considered the gold standard for the diagnosis of peptic ulcer disease. However, if inactive ulcers are studied, an alternative diagnostic technique is the upper GI series. Since the upper GI series is highly specific for peptic ulcer disease, the results of this test may be used to diagnose peptic ulcers prior to study inclusion (as a second line diagnostic modality) if the size and location of the these ulcers are clearly documented. In contrast, because of the poor sensitivity of this test, it should not replace endoscopy to confirm ulcer healing after the completion of therapy.
Diagnosis of H. pylori Infection
The Division currently requires endoscopic diagnostic tests to confirm and exclude the diagnosis of H. pylori infection. The number of positive or negative tests required depends on whether these tests are to be used for the initial diagnosis (before treatment, to confirm infection) or used to document eradication (after the completion of therapy). The definition of baseline infection and eradication following therapy need to be considered separately (in terms of the number and type of endoscopic tests used) since at baseline it is important to achieve high specificity (low false positives) to confirm infection while at the test-of-cure time point sensitivity is more important to exclude infection (low false negatives). The following definitions of infection (or no infection) have been developed to assist sponsors in deciding which patients should be considered infected, not infected, or not evaluable based on endoscopic tests. It is important to note that there is no single correct definition of these terms. True definitions are based on the quality and quantity of endoscopic H. pylori diagnostic tests. The Division's current opinion is that more tests are better than fewer tests. However, the Division also recognizes that the methods used for sample collection and performance of each of these tests may be more important than the actual number of tests used to diagnose infection.
The Division maintains that inclusion into a clinical study only requires a single positive CLOtest, but evaluability for analysis requires either a positive culture or both a positive CLOtest and positive histology. If the culture is negative or missing and the histology/CLOtest results are incongruent, these patients should be considered non-evaluable.
The Division maintains that only one endoscopic test needs to be positive for the patient to be considered an eradication failure. Patients will be considered non-evaluable at the test-of-cure time point if only one test is available and that test is negative. If only two tests are available at the test-of-cure time point and these two tests are negative, infection in these patients will be considered eradicated. It is important to point out that culture is not required in the study design as long as both the CLOtest and histology are included. Table 1 summarizes the Division's current recommendation concerning the use of endoscopic diagnostic tests to diagnose H. pylori infection.
Many factors have been reported to affect the ability to identify H. pylori with histologic techniques. These include the type of stain used, the experience and interest of the pathologist, blinding of the pathologist, the magnification used, the number of biopsy samples taken (and the location from which they were taken), and the number of slides examined from each biopsy. In 1990 the World Congress of Gastroenterology recommended using 4 biopsies (2 from the corpus and 2 from the antrum) and the use of either the Warthin-Starry or a Giemsa stain to better visualize organisms. Since multiple biopsies are needed (for study purposes and for patient care), it is suggested that only 2 antral biopsies be used to identify H. pylori with histologic means. The pathologist should have extensive experience in the identification of this organism and should have demonstrated a high degree of intra-observer agreement. Multicentered trials should use an experienced pathologist(s) to read all slides. Whereas reading of histology slides are likely to be needed for patient care, centralized reading of histology slides should be made available for analysis. Both H & E stains (to determine extent of gastritis) and either Giemsa or Warthin-Starry stain (to identify organisms) should be used at $ 100X magnification.
Three CLOtests have been approved to diagnose H. pylori infection. However, none of these are approved for therapy monitoring purposes. Nevertheless, the Division currently suggests that the use of this test should be considered at both baseline and test-of-cure time points. To maximize sensitivity at the eradication time point, 2 antral biopsies should be used. Whereas a 3 hour reading is appropriate for the baseline pre-therapy CLOtest, it is best to read the CLOtest after 24 hours to increase sensitivity at the test-of-cure time point.
The most important factor to consider in performing H. pylori culture is the experience of the microbiologist(s). It is recommended that H. pylori microbiology only be performed in centers who have a good track record. Whereas the specificity approximates 100 percent regardless of methodology, sensitivity is known to vary markedly among published studies. It is recommended that 2 antral biopsy specimens be frozen locally and shipped to a center experienced in the recovery of H. pylori through culture. Since the methodologies of gastric biopsy freezing and subsequent recovery of H. pylori have not been standardized, H. pylori culture performed locally is acceptable. Microbiologic identification from pure culture should be confirmed with oxidase, urease, and catalase testing as well as gram stain.
In addition to endoscopic diagnostic tests, multiple various non-invasive tests have been developed to confirm and exclude the diagnosis of H. pylori. Although urea breath tests have not yet been approved by the FDA, they have been shown to be highly sensitive and specific for the diagnosis of H. pylori. In the future, they may be especially useful in the stablishment of eradication post-therapy. Currently, the Division maintains that urea breath tests should not be used to define infection or eradication for the purpose of clinical trials. However, the Division does not object to their use in clinical trials for the purpose of generating data which standardizes their use for initial diagnosis of infection and for the purpose of monitoring therapy.
Multiple serologic tests are approved for the detection of H. pylori antibodies. They may be used at the pre-therapy time point for the purpose of study inclusion, but should not be used to determine evaluability because these tests do not detect active infection. It follows that these tests should not be used to confirm eradication. Since all patients enrolled in clinical studies will require a baseline endoscopy (to document the presence or absence of ulcer), the Division also recommends using endoscopic (direct) tests rather than serology to define evaluability for the purpose of efficacy and safety pretherapy.
Statistical Considerations in H. pylori Trials
Expeditious patient accrual and comparison of outcomes across centers, types of institutions, geographic regions, etc., are two primary advantages of multicenter trials over single center trials. Since it is important to discern whether inconsistencies in treatment effect across centers are real, or merely artifacts, adequate numbers of patients should be enrolled in each center.
The Division suggests enrolling sufficient numbers of evaluable patients per arm per center to adequately evaluate the potential treatment by center interaction. Ideally, a minimum of 10 evaluable patients per treatment arm should be an enrollment goal for centers participating in these trials.
Currently, the Division recommends a factorial or "modified factorial" trial design to establish the effectiveness of combination products for H. pylori. To be considered a successful trial under this design, the results must demonstrate that the combination product is statistically and clinically superior in efficacy to each of its component parts. The suggested analytical approach is classical hypothesis testing.
In the future, if it is felt to be ethically unacceptable to conduct studies that withhold treatment from infected patients when an approved effective therapy is available, an active-controlled study design may be used to establish the effectiveness of a new H. pylori regimen. To be considered a successful trial under this design, the results must demonstrate that the new H. pylori regimen is statistically and clinically equivalent in efficacy to an approved H. pylori regimen. The suggested analytical approach is estimation via two-sided 95% confidence intervals of the treatment difference (test minus control) in success rate. In this approach, the confidence interval must include zero, and the lower limit of the confidence interval must not exceed the clinically specified boundary for establishing efficacy equivalence. These boundaries vary depending on the treatment success rates observed in the trial as follows:
Table 2
| If the observed success rate for the better of the two agents is: |
The lower bound of the C.I. (in absolute value) must not exceed: |
| 90% |
10% |
| 80-89% |
15% |
| <80% |
20% |
DAIDP's Points to Consider
1992 Addendum: Dose-Response Testing
Date created: November 26, 1996; last update: July 6, 2005 |
