FDA STANDARD OF +/-20% ADEQUATE TO ASSURE BIOEQUIVALENCE OF GENERICS, TASK FORCE MAINTAINS; +/-10% CRITERIA HAS "SOME MERIT" AS EXTRA ASSURANCE

FDA's current bioequivalence evaluation criteria, which allows generic drugs to be approved when the mean values are within +/-20% of a reference product, is an adequate standard, the agency's Bioavailability Task Force concluded in its report on bioequivalence issues. The task force supported the current criteria in response to the possibility raised at the September 1986 hearing that the +/-20% standard is too lax and could result in clinical differences between products. The report states that the task force "favors the use of a 90% confidence interval based on the two one-sided t-test approach as the best available method for evaluating bioequivalence." Although it concluded that the current criteria is adequate, the task force added, however, that "there may be merit" in a proposal to change the acceptance criteria to +/-10%. The task force suggested that using a nonstatistical +/-10% mean test as an additional criteria "would add significantly to the assurance of the bioequivalence of generic drugs." The proposal to change the acceptance criteria was made by one of two expert consultants who reviewed the task force report and made additional recommendations. The consultants also participated in the public hearing, serving as expert commentators on the public presentations. According to the report, "one consultant agreed with the agency; the other consultant believed that the agency should modify its position and require that mean AUC [area under the curve] values be within 10% rather than within 20%. After discussion, the task force concluded that the agency should consider the feasibility of adding an additional nonstatistical criteria for the mean difference of AUC to be +/-10%." The task force explained, however, that it "believes that [the +/-10% standard] is less stringent than the currently employed criterion based on confidence intervals. The task force feels that as an additional requirement -- over and above the currently employed criterion -- it would further reassure the public of the comparability of the innovator and generic products." However, the group stressed its view that such an additional requirement is unnecessary, pointing out that "it can be estimated that only about 1% of generic drugs approved with current procedures would fail to pass this additional criterion." The agency also noted that for post-1962 drugs approved over a two-year period under the Waxman-Hatch bill, "the mean bioavailability difference between the generic and innovator product is 3.5%. Additionally, 80% of the values for drugs approved since 1984 were within +/-5.0% of the reference drug value." If adopted, the task force said, the additional requirement "would be applied prospectively and would also apply to reformulations of innovator products." The task force noted that while the +/-20% standard is adequate in most cases, for some drugs or drug classes tighter limits are appropriate. "These situations must be identified on the basis of clinical evidence demonstrating a need to tighten the generally applied standard." Such evidence could consist of a prospective clinical study, the report states. The task force noted that the +/-20% rule "can and has been modified for drugs with a narrow therapeutic window, e.g. warfarin" and for drugs with a wider therapeutic window, such as some psychotropic agents. FDA's 75-75 rule, the task force noted, has been dropped as a basis for making bioequivalence decisions in favor of the more rigorous 90% confidence interval. The task force agreed with the consensus developed at the public hearing that the 75-75 rule should not be used. Commenting on the clinical significance of the +/-20% standard, the task force noted that "clinical studies of effectiveness have difficulty detecting differences in dose of even 50-100%. Few drugs are given on a mg per kg basis to account for weight differences and few drugs have their dosage adjusted in actual clinical practice for factors that may affect blood concentrations in individuals. Thus, the variability inherent in medical practice and biological variation may cause plasma levels to vary in individuals by much more than 20%." The agency's criteria for evaluating bioequivalence studies, in any case, should be made "more widely available," the task force recommended. The group suggested that FDA "publish procedures under which drugs or classes of drugs would be added to or deleted from the list of drugs subject to either more strict or less strict criteria than the general rule." Discussing other issues related to the analysis of bioequivalence studies, the report states that FDA's Division of Biometrics should undertake a "comprehensive evaluation" to determine how "outliers" should be analyzed. Outliers are study subjects who produce data that is inconsistent with data from other study subjects. The task force said that "the identification and treatment of outliers in bioequivalence studies deserves more attention than it has received to date. Removing certain subjects from consideration in a study because their data do not conform to the rest of the data may affect the validity of the study. In most cases, one cannot determine whether the apparently nonconforming data is due to a laboratory error, data transcription error, or other causes unrelated to bioequivalence. This argues against removing the data." In responding to the question of superior bioavailability in a generic compared to a reference drug, the task force noted that FDA will approve such products. "FDA is prepared to approve generic products where the innovator is poorly available and the generic product matches the bioavailability of a more fully bioavailable formulation and produces blood levels equivalent to those of the innovator product when given at a lower dose." The group added that such products are approvable, although they would not be considered bioequivalent to the innovator drug and could not be handled under the ANDA suitability petition process of the Waxman/Hatch law. There are "only a handful of approved products on the market that are considered to be poorly bioavailable," the task force pointed out, adding that "this does not appear to be a major and immediate public health problem" because those poorly bioavailable products have each been tested clinically for safety and effectiveness and they have each been on the market without problem for some time. An example of a poorly bioavailable innovator product is SmithKline's Dyazide; FDA last year approved Bolar's ANDA for a generic version. SmithKline is developing a formulation with better bioavailability. The agency's Biometrics Division should also conduct a study to determine if current bioequivalence study protocols can adequately account for variability within subjects, the task force recommended. The task force declared: "Ordinarily, a two-period crossover design is an adequate design for comparing the bioavailability of two products. If however, variability within subjects in such a study is large, it may be helpful to measure the within-subject variability directly by using a replicate design in which patients are given the same treatment more than once. Three- or four-period crossover studies with two treatments could be employed for a short time to gather data about whether such designs might prove superior. The agency is considering this concept and will ask for public comments if it decides to implement it. Designs using stable isotopes can also be used to assess intrasubject variability." Increasing the precision of the assessment of variability may not be particularly useful, the task force commented. "If appears that any residual error in the statistical analysis of the study is more likely to be related to product variability rather than to subject variability," the group said. Bioequivalence studies should continue to use healthy volunteers as study subjects, the task force agreed, explaining that "the important question is not whether patients are different from volunteers, but whether, and when, these differences could cause two products that seem bioequivalent in normals to be bio-inequivalent in a clinical setting." The agency should "pursue . . . any creditable leads suggesting patient factors not tested for currently which might lead to differences in bioequivalence." The agency "would welcome the conduct of such studies by the industry," the task force suggested. FDA BIOEQUIVALENCE TASK FORCE REPORT CONCLUSIONS The following five and one-half pages are a compliation of the Conclusions and Recommendations from FDA's report of the Bioequivalence Task Force. The report, released during the week of Feb. 8-12, addresses issues raised at the three-day opening hearing on bioequivalence held in the autumn of 1986. The compilation by "The Pink Sheet" puts conclusions and recommendations together for easier reading and understanding. I. DESIGN OF BIOEQUIVALENCE STUDIES q. Are single-dose studies adequate? CONCLUSION: The Task Force believes that as a general rule a single-dose study is adequate. Multiple-dose studies should be performed only when a single-dose study is not a reliable indicator of bioavailability, e.g., because of the kinetics of the drug. 2. Should a three-period bioequivalence study with both a solution and reference product as reference standards be required? CONCLUSION: The Task Force believes that a change to a three-period bioequivalence study using a solution is not warranted for most cases. The usefulness of a solution as an anchor is warranted only when information about the relative bioavailability of a product is unknown or the bioavailability is known to be poor. A more clearly defined benefit should be shown to justify the increased costs associated with the three-period design before it is required in every case. 3. Does the use of normal volunteers adequately account for the potentially altered absorption capacity and metabolism of special populations? CONCLUSION: The important question is not whether patients are different from volunteers, but whether, and when, these differences could cause two products that seem bioequivalent in normals to be bioinequivalent in a clinical setting. A search of the literature to identify these factors in patients revealed very few relevant publications. The Task Force believes that it is preferable to subject healthy people, rather than patients, to the rigors of blood sampling and other discpomforts of bioequivalence testing. Moreover, use of patients would invariably increase intersubject variability and possibly intrasubject variability as well. Thus far there have been few, if any documented examples of problems associated with the use of normals to predict bioequivalence, although there have been relatively few rigorous attempts to document problems. The Task Force believes that at this time it remains appropriate to determine bioequivalence based on testing in healthy volunteers. The Agency recognizes the possibility that some conditions could affect bioavailability and is prepared to modify its position regarding the use of normal subjects if such a situation is adequately documented for a given drug. RECOMMENDATION: The Task Force recommends that the Agency be prepared to pursue, in house or extramurally, any credible leads suggesting patient factors not tested for currently which might lead to differences in bioequivalence. The Agency would welcome the conduct of such studies by the industry. 4. Should bioequivalence studies include measuring of clinically active metabolites? CONCLUSION: The Task Force believes that clinically active metabolites should be measured when they have significant pharmacologic activity. 5. Should FDA develop individual criteria for each drug or class of drugs? CONCLUSION: The Task Force endorses the current system of informal guidances on how to design and conduct studies prepared by the Division of Bioequivalence. This system of informal guidances promotes cooperation and consultation between FDA and industry, fosters scientific discussion and investigation, and permits the flexibility necessary to ensure that the guidances contain up-to-date scientific information and testing approaches. The general decisional criteria discussed elsewhere in this report (see II-1) could be better articulated and more widely publicized. Specific statistical methodology and review criteria could be incorporated into the guidance system or in another forum when appropriate. 6. Can dissolution testing assure bioequivalence? Should it be employed as a substitute for in vivo study in humans? Does adequate information exist to justify a waiver of in vivo studies based on dissolution alone? Should drugs be approved based on dissolution only without a relationship of in vitro data to in vivo performance? CONCLUSIONS: The Task Force believes there is not yet evidence to show that any particular dissolution pattern alone will assure bioequivalence. Dissolution testing can be used for drugs where there is a known in vivo/in vitro relationship, and is used for pre-1962 drugs not suspected of having, or not likely to have, a bioavailability problem. For all other solid oral drugs, an in vivo bioequivalence study on the drug product is required to support at least one strength of the product. The Task Force believes that dissolution testing is important in assuring lot-to-lot uniformity, and in supporting minor alterations to drug products (see 21 CFR 320.22(d)). Also, it is FDA policy that if a product meets in vivo bioequivalence study requirements at one strength, and the formulations of additional strengths are proportional to the strength tested in the in vivo bioequivalence study, and the additional strengths meet dissolution requirements, then further in vivo bioequivalence studies are not required for the additional strengths unless there is evidence of safety or efficacy problems. This policy applies to generic and innovator products. The Task Force believes these policies are sound, but does not recommend expanding the use of in vitro testing beyond these limits. 1. Do or should bioequivalence studies consider the effect of excipients on bioavailability of drug products? CONCLUSION: The Task Force agrees that the rare incidence of allergies and toxicity to excipients may pose a problem for a few patients. Information on excipients for all drug products is currently being addressed by the PMA and the PA with their voluntary labeling guidelines and this information will help enable patients to be alerted to an allergenic potential. The effect of excipients on bioavailability is assessed by current bioequivalence studies. 8. How should FDA assure lot-to-lot uniformity? Are in vivo bioequivalence studies necessary to approve all formulation changes? CONCLUSION: The Task Force believes that dissolution testing is appropriate for assuring lot-to-lot uniformity. The more difficult question is the extent to which particular changes in formulation may affect the bioavailability of a drug product. The FDA may waive the requirement for submission of evidence of in vivo bioavailability under certain conditions for solid oral dosage forms. Generally, a waiver is granted for a 'minor' change in formulation (e.g., change in color). The dividing line between a minor and major reformulation is not always clear, however, and even a series of minor reformulations could have the same implication as a major reformulation. RECOMMENDATION: The Task Force recommends that a public meeting be held or some other forum be used, e.g., a Federal Register notice soliciting information and comment to allow discussion of such questions regarding reformulation changes. 9. Should an alternate study designnn be considered as the standard for bioequivalence testing to determine intrasubject variability? CONCLUSION: Requiring a second study for bioequivalence determinations is not justified. The use of bioequivalence studies with an alternate study design is being considered. RECOMMENDATION: The Task Force recommends that the Division of Biometrics undertake a study of alternative study designs to address the need for any changes in current protocol design. II. DECISIONAL CRITERIA FOR BIOEQUIVALENCE 1. Should the current equivalence criteria be changed? CONCLUSION: The Task Force favors the use of a 90% confidence interval based on the two one-sided t-test approach as the best available method for evaluating bioequivalence. The Task Force concludes that some drugs or drug classes may require tighter limits than the generally applied +/- 20% rule. These situations must be identified on the basis of clinical evidence demonstrating a need to tighten the generally applied standard. Such evidence could include, for example, a prospective clinical study demonstrating that the usual criteria for bioequivalence measurements are not stringent enough. The Task Force also concludes that limits of +/- 20% effectively preclude true differences in means beyond those limits. The Task Force believes that there may be merit to the consultant's proposal for an additional criteria, because it would add significantly to the assurance of the bioequivalence of generic drugs, and would also preclude the unusual case of a real difference beyond +/- 10%. However, the Task Force does not believe it is necessary to require an additional criteria beyond the current requirements. RECOMMENDATION: The Task Force recommends that the Agency make more widely available the criteria it uses to make bioequivalence determinations, as well as any exceptions to those criteria, e.g., where the standards are more or less stringent for particular drugs for documented clinical reasons. The Task Force recommends that the Agency publish procedures under which drugs or classes of drugs would be added to or deleted from the list of drugs subject to either more strict or less strict criteria than the general rule. Also, the Task Force recommends that the Agency consider the feasibility of adding an additional nonstatistical criteria for the mean difference of AUC to be +/- 10%. However, the Task Force does not believe that such an additional criteria beyond the current requirements is necessary for assuring the bioequivalence of generic drugs. 2. Could the +/- 20% requirement lead to differences in products of 40-50%? CONCLUSON: The Task Force notes that for post-1962 drugs approved over a two-year period under the Waxman-Hatch bill, the mean bioavailability difference between the generic and innovator product is 3.5% (see the discussion under II-1). Additionally, 80% of the values for drugs approved since 1984 were within +/- 5.0% of the reference drug value. 3. Should the use of the 75/75 rule as a decisional tool be dropped by FDA? CONCLUSION: The Agency agrees with the consensus that the 75/75 rule should be dropped. The Task Force favors the use of a 90% confidence interval based on the two one-sided t-test approach to evaluate bioequivalence. This involves determining the confidence interval for the ratio of means using a modified t-test method. 4. Is product to product variability within an acceptable range? Should the Agency expend resources to answer this question, and if so, how should it be determined? CONCLUSION: The Task Force believes that current requirements are adequate to assure the quality and uniformity of all drug products. However, the variability among drug products deserves further study. RECOMMENDATION: The Task Force recommends that the Division of Biometrics gather data and develop statistical methodology to consider whether a problem exists regarding product variability. Appropriate action will be taken, should a problem be discovered. 5. How should outlying data be treated in bioequivalence analyses? CONCLUSION: The Task Force believes that neither testimony at the Hearing, nor any currently available document adequately addresses these issues. RECOMMENDATION: The Task Force recommends that the Division of Biometrics undertake a comprehensive evaluation of the treatment of outliers. III. AGENCY PROCEDURES AND REGULATORY ASPECTS OF BIOEQUIVALENCE 1. Should bioequivalence decisional criteria be published using notice and comment rulemaking? CONCLUSION: The Task Force believes the Agency can make clearer the decisional criteria it employs to determine bioequivalence. The specific guidances on how to design and conduct a bioequivalence study are good examples of the Agency's ability to convey this kind of information to the regulated industry. The Task Force does not agree, however, that notice and comment rulemaking is the appropriate mechanism for disseminating inforamtion about the decisional criteria. In this regard, we believe notice and comment rulemaking is too slow a process to accommodate new and evolving statistical and biopharmaceutical scientific methods and changes based on new information and experience. It is also difficult to write a meaningful regulation in this area since bioequivalence is often a matter of judgment. To ensure the flexibility necessary to keep Agency criteria current, a method other than rulemaking is recommended. The Agency could publish a formal guideline specifying the decisional criteria used to evaluate bioequivalence studies. This guideline would be subject to public comment and would commit the Agency to follow the guideline. Under current administrative procedures, formal guidelines can be modified more quickly than can regulations. However, these formal guidelines still must be announced by Federal Register publication. Several speakers at the Hearing recommended including more information about the decisional criteria in the Orange Book. We suggest that a guideline on the bioequivalence decisional criteria could either be a part of the Orange Book, a supplement to the Orange Book, a separate publication, or perhaps some combination of those publications listed above. Others at the Hearing voiced concern regarding the cost of the Orange Book. Before including a guideline into the Orange Book, we must consider these concerns because the cost of the Orange Book is based on its total number of pages. RECOMMENDATION: The Task Force recommends that the Agency publish full information about the bioequivalence evaluation procedures and decisional criteria, either in the form of a formal guideline, or as a supplement or companion piece to the Orange Book. 2. Should FDA establish an advisory panel to advise the Agency on bioequivalence issues? CONCLUSION: The Task Force agrees with the principle of obtaining views about bioequivalence from outside the Agency especially concerning general bioequivalence issues. The Agency currently employs individual consultants on an ad hoc basis for guidance on a variety of issues including bioequivalence. The Agency also sponsors or cosponsors a number of meetings that address bioequivalence issues as well as others. The Task Force believes the Agency should continue these practices. In addition, the Agency should consider other ways to broaden outside input, e.g., augmenting existing standing advisory committees with biopharmaceutic experts. RECOMMENDATION: The Task Force recognizes that there is a significant interest in obtaining more outside input on bioequivalence issues and recommends that the Agency further this possibility. 3. Have there been therapeutic failures with approved generic products? Is the current adverse drug reaction monitoring program adequately detecting therapeutic failures? CONCLUSION: The Task Force concludes that FDA should enhance current procedures to better detect and evaluate reports of therapeutic failures that could be indicative of failure of a product. FDA should fully investigate possible inequivalence only when there is good evidence of a problem, and not on unsupported anecdotes. The medical community and the manufacturers should be encouraged to submit reports of therapeutic inequivalence with as much detail as possible, including blood level data. RECOMMENDATION: The Task Force recommends that detailed plans for identifying signals or clusters of possible important instances of product failure from the the ADR reporting system be developed. These plans should indicate how and when these signals will be communicated to the appropriate Agency components. The Task Force recommends that a policy be developed to outline how, when, and by whom signals should be investigated, the role of laboratory testing, requiring repeat bioequivalence testing, field work to investigate individual cases, and the responsibilities of the several offices involved. The Task Force recommends that the current ADR regulations be modified to require that reports be submitted to the Agency to aid in accomplishing these recommendations. 4. Can FDA's therapeutic equivalence list, (the Orange Book), be revised to show whether an AB rating is based on in vivo or in vitro testing? CONCLUSION: The Task Force believes the Orange Book should be modified so that it is possible to determine by the drug code the basis by which drugs were rated. RECOMMENDATIONS: The Task Force recommends that a method be devised to identify the drugs now rated AB on the basis of dissolution alone, and that the category AB be reserved for drugs that have been approved on the basis of an in vivo bioequivalence study. The Task Force recommends that the Orange Book be more widely advertised to pharmacists. This could be accomplished by working more closely with the states. Efforts to decrease the cost should be explored. This could be accomplished by publishing an abbreviated version of the list alone, by selling the Orange Book without the monthly supplements, or by enlisting the assistance of a private organization to make the book available at a lower cost. 5. Should the patient and the physician be notified by the pharmacist when generic substitution occurs? Is there a problem of indiscriminant substitution of products by pharmacists in states even when 'no substitution' is requested, and are products not rated as bioequivalent being substituted? Should FDA get involved in these issues? CONCLUSION: FDA does not regulate the practice of pharmacy or medicine. Regulation of these professions is properly within a state's purview. The states are free to use FDA's bioeuivalence determinations and therapeutic equivalence evaluations or not use them. Depending on individual product selection laws, some states develop their own formularies of interchangeable drug products. Each state has its own requirements regarding substitution. Thirty-eight states and the District of Columbia have permissive substitution and 12 have mandatory substitution. Forty states require patient consent for substitution. Commissioner Young stated that ". . . we can in no way substitute federal judgment for state responsibility, and in no way substitute, in my opinion, state and federal responsibility for the responsibility of the physician. To do so would be, in my opinion, utter folly." All states prohibit pharmacists from making substitutions of products when "DAW" (Dispense as Written) or some similar instruction is written by the physician. The Task Force concludes that allegations of violations of these prohibitions are appropriately within the jurisdiction of the states. Unauthorized substitution should be brought to the attention of the appropriate state authorities for any necessary action. These issues are not federal issues. RECOMMENDATIONS: The Task Force recommends that the Agency encourage physicians and pharmacists to make patients aware of the possibilities of drug substitution with their patients to avoid potential patient confusion when a different color tablet is dispensed, for example. The Task Force encourages pharmacists and physicians to discuss with their patients the potential nontherapeutic differences of different brands and trade dress of products before the patient receives the product. Special efforts should be made to inform the elderly patient. The Agency should work with organizations like the AARP to disseminate advice to pharmacists and physicians and should encourage the states and drug industry to do so. The Agency could also accomplish this in a program similar to the "Talk About Prescriptions" campaign sponsored by the National Council on Patient Information and Education. 6. What should the requirements be for approving better formulated copies of poorly bioavailable innovator products? CONCLUSION: The Task Force is aware of only a handful of approved products on the market that are considered to be poorly bioavailable. This does not appear to be a major and immediate public health problem because: (1) each of the poorly formulated products was approved on the basis of clinical investigations demonstrating its safety and efficacy; and (2) each of these poorly formulated products has been on the market for several years without documented safety and efficacy problems. FDA is prepared to approve generic products where the innovator is poorly available and the generic product matches the bioavailability of a more fully bioavailable formulation and produces blood levels equivalent to those of the innovator product when given at a lower dose. Such a product would be considered bioinequivalent to the innovator product under the petition provisions [Section 505(j)(2)(C)] of the Act. RECOMMENDATION: The Task Force recognizes that each of the few cases in which the innovator product is poorly bioavailable presents a unique circumstance and the Agency should use the petition [Section 505(j)(2)(C)] or other procedures to encourage the marketing of fully bioavailable products. 7. What would be the significance of one documented generic failure? CONCLUSION: The Task Force concludes that there is no reason to doubt the fundamental principle that drug products delivering comparable blood levels of a therapeutic moiety in bioequivalence tests in normals will generally yield comparable therapeutic results. There are known differences among patients, such as gut transit time or gastric pH that could, combined with differences between products, such as pH dependency of dissolution, theoretically yield differences in performance of products in certain patients. whether this hypothesis actually is manifested clinically in any significant way has not been shown. A distinction must be drawn between a single case of a patient who does not respond to a drug product and evidence that a drug product is not performing. Virtually all products are, from time to time, the subject of isolated reports of therapeutic failures. The Agency looks particularly for patterns of such reports or cases which may indicate a generalized problem with a drug product or a batch of the product. The documentation of a single instance of clinical inequivalence does not, in the Task Force's view, undermine the much wider experience that shows bioequivalence testing to be an excellent predictor of clinical performance. A product failure, on the other hand, would necessitate that the Agency investigate thoroughly and take steps to deal with the particular case and others that might arise from similar circumstances.