conclusionsand

recommendations1Current standards and regulatory practices do notinsure bioequivalence for drug products.

2Variations in the bioavailability of drug productshave been recognized as responsible for a fewtherapeutictherapeuticorigin have

3Most of the

failures. It is probable that otherfailures (or toxicity) of a similarescaped recognition.

analytical methodology and experimentalprocedures for the conduct of bioavailabilitystudies in man are available. Additional work maybe required to develop means of applying them tocertain drugs and to special situations of drug use.

4It is neither feasible nor desirable that studiesof bioavailability be conducted for all drugs ordrug products. Certain classes of drugs for whichevidence of bioequivalence is critical should beidentified. Selection of these classes should bebased on clinical importance, ratio of therapeuticto toxic concentration in blood, and certainpharmaceutical characteristics.

5Present compendia standards and guidelines forCurrent Good Manufacturing Practice do not insurequality and uniform bioavailability for drugproducts. Not only may the products of differentmanufacturers vary, but the product of a singlemanufacturer may vary from batch to batch or maychange during storage.

New compendia standards for active ingredients,excipients and finished drug products should bedeveloped and revised on a continuing basis toreflect the best available technology to insurequality and uniform bioavailability. Appropriatestatistical procedures should be specified to makecertain that the purposes of the standards areobjectively satisfied. The guidelines for CurrentGood Manufacturing Practice should be expanded toinclude specific descriptions of all significantaspects of manufacturing processes from the rawmaterials to the final product.

7Additional research aimed at improving the assess-ment and prediction of bioequivalence is needed.This research should include efforts to developin vitro tests or animal models that will be validpredictors of bioavailability in man.

Current law requiring manufacturers to maintainrecords and make information available to the FDAis ambiguous or inadequate and should be clarifiedand strengthened. In particular, manufacturersshould be required to submit all informationrelating the tests they conduct to the bioavailabilitydata they develop in order to help provide informationon the factors that modify the bioavailability ofdrug products. This information should be availableto aid in the establishment of compendia standards.

3

9Exemptions provided in current law for some drugproducts based on their year of introduction inrelation to amendments in the Food, Drug, andCosmetic Act (so-called grandfather clauses) haveimpeded improvement in the quality of theseproducts. Such exemptions should be eliminated.

A single organization capable of setting standardsadequate to insure the quality and uniformbioavailability of drug products should beestablished to replace the present USP and NF asthe official standard-setting organization of theFederal Government.

1 1A system should be organized as rapidly as possibleto generate an official list of interchangeabledrug products. In the development of the list,distinctions should be made between two classes ofdrugs and drug products:

1. Those for which evidence of bioequivalenceis not considered essential and that could beadded to the list as soon as standards ofpharmaceutical equivalence have beenestablished and satisfied.

2. Those for which evidence of bioequivalenceis critical. Such products should belisted only after they have been shown tobe bioequivalent or have satisfied standardsof pharmaceutical equivalence that have beenshown to insure bioequivalence.

chargeto the

panelThe basic charge made to the Drug BioequivalenceStudy Panel* by the Office of TechnologyAssessment (OTA) was to examine the relationshipsbetween the chemical and therapeutic equivalenceof drug products and to assess the capability ofcurrent technology--short of therapeutic trialsin man--to determine whether drug products withthe same physical and chemical composition producecomparable therapeutic effects.

As members of the Panel, we agreed that a reviewof the system of regulations and practicesnow used to insure quality and uniformity--thecompendia standards, Current Good ManufacturingPractice guidelines, and the manufacturers’procedures for quality control and formulation--would have to be an inherent part of our assessment.After examining these factors, we would makerecommendations, if necessary, for any modificationsthat we believed might be needed to improve thepresent system for insuring the therapeuticequivalence of drug products.

It was clear to us from the outset that certainchemically equivalent drug products have producedclinically important and measurable differences intherapeutic effect and that these differences werethe result of differences in bioavailability.Conversely, we recognized that differences inbioavailability (bioinequivalence) among some drugproducts may not be a critical concern with regardto the equivalence of their therapeutic effects.

*The Drug Bioequivalence Study Panel met for thefirst time on April 12, 1974 to clarify itscharge and to formulate a plan for its study.A chronological review of the activities under-taken by the Panel is presented as the Appendix.

6

As our study progressed, it became clear to us thatwe should recommend a system that, through improvedassurance of quality and rational scientificjudgments, would insure the therapeutic equivalenceof drug products.

Our conclusions and recommendations are basedon considerations of current technology and of thetechnology that we believe could be developedwithin the next few years. Our report also includesa consideration of the need and potential for newresearch relevant to the development of thattechnology and to other aspects of the problemsunder consideration.*

*The Panel communicated extensively with varioussources, requesting information and receivingdocumented replies in response to questionsraised. A bibliography of the documentationreceived and reviewed by the Panel appears onpage 61.

7

scopeof the

studyBecause of the limited time available to us, itwas necessary to restrict the scope of our studyto solid drug products (capsules and tablets, butexcluding timed-release and coated products) thatare administered by mouth and produce a systemiceffect. These are the products for which consid-eration of the relationships between chemical andtherapeutic equivalence is most important. Inaddition, tablets and capsules constitute the greatmajority of all drug products used in the UnitedStates.

The therapeutic effect of a solid drug productis assumed to be a function of the concentration ofthe active ingredient in the systemic circulationand is thus related to its bioavailability.Accordingly, we directed our attention to theprinciples and methodology involved in estimatingand comparing the bioavailability of drug products.This methodology involves measurement of theconcentration of the active ingredient in thesystemic circulation, either directly in blood orindirectly through studies of urinary excretion.

We also attempted to identify the generalcategories of drug products for which measurementsof bioavailability may be critical to the assuranceof therapeutic equivalence. The therapeuticequivalence of some categories of drug products canbe insured, however, without a direct assessment ofbioavailability, and we tried to identify thegeneral characteristics of such products.

Certain drug products contain the same activemoieties but do not have the same salt, ester ordosage form. These products, although similar, arenot chemically equivalent but may be therapeuticallyequivalent. For example, tetracycline hydrochloridetablets, capsules and syrup all may be expected toproduce the same therapeutic effect as tetracyclinephosphate capsules, but they are different in dosage

8

or salt forms. Drug products which are notchemically equivalent but have the same thera-peutic effect may be referred to as pharmaceuticalalternatives but not as chemical or pharmaceuticalequivalents. We believe this distinction issignificant and that it requires further elabora-tion; however, we did not examine pharmaceuticalalternatives or the extent to which they do or donot produce equivalent therapeutic effects.

It must be recognized that beyond characteristicsof drug products that influence absorption ofthe active drug from the gastrointestinal tract,a number of other factors have an important rolein determining the therapeutic effect produced inany individual. Among these factors are thefollowing:

1.

2.

3.

4.

Variations from one individual to another inabsorption, metabolic conversion and excretionof the drug. Some individuals have rates ofintestinal absorption and metabolism thatdiffer markedly from the average rates, andthese differences may be exaggerated by thedisease state. Consequently, the concentrationof the drug in the blood of such individualsmay be well above or below the average, evenif the drug is rapidly and completely dis-solved in gastrointestinal fluids.

Appropriateness of tile prescriber’s choiceof drug.

Differences among patients in complying withinstructions for taking the drug product.

The simultaneous use of other drugs that canaffect or alter the action of the-prescribeddrug.

Although such factors may greatly influencethe therapeutic effect obtained from drug products,our major concern was directed at factors affectingthe efficacy of drug products that are within thecontrol of the manufacturer.

discussionof

conclusionsand

recommendations

11

CURRENT STANDARDS AND REGULATORY PRACTICES DO NOT

INSURE BIOEQUIVALENCE FOR DRUG PRODUCTS,

Although there is a spectrum of opinion about thefrequency and importance of differences in thebioavailability of chemically equivalent drugproducts, there can be no dispute about the factthat well-documented and significant differencesin bioavailability have been demonstrated inchemically equivalent products representing anumber of drug categories.

Problems of bioequivalence have received seriousinvestigative attention only during the past fewyears. In this brief period, however, a numberof studies of marketed drug products containingthe same therapeutic ingredient have revealedmarked differences in the rate and extent ofabsorption. A considerable body of literature hasaccumulated in this period that indicates theexistence of demonstrable differences in thebioavailability of products involving roughly ascore of drugs. A partial list of studies demon-strating bioinequivalence of chemically equivalentdrug products might include the following: tetra-cycline (MacDonald et al, 1969; Barr et al, 1972);chloramphenicol (Glazko et al, 1968); digoxin(Wagner, 1973; Lindenbaum et al, 1971) ; phenyl-butazone (VanPetten et al, 1971; Chiou, 1972); andoxytetracycline (Blair et al, 1971; Brice et al,1971).

Not only has bioinequivalence been shown to existin products of different manufacturers but therealso have been substantial variations in thebioavailability of different batches from the samecompany (for example, see Lindenbaum et al, 1971).It is difficult to determine whether differences

..- . . .

12

such as those cited are the exception or the rule,since positive results (in this case a demonstrabledifference) are far more likely to be publishedthan negative findings. Furthermore, in somestudies in which several chemically equivalentproducts have been tested, some have been found tohave the same bioavailability while others havenot.

Nevertheless, the number of positive findings hasbeen sufficient to establish that the problem ofbioinequivalence in chemically equivalent productsis a real one. Since the studies in which lack ofbioequivalence was demonstrated involved marketedproducts that met current compendia standards,these documented instances constitute unequivocalevidence that neither the present standards fortesting the finished product nor the specificationsfor materials, manufacturing process, and controlsare adequate to insure that ostensibly equivalentdrug products are, in fact, equivalent inbioavailability.

13

2VARIATIONS IN THE BIOAVAILABILITY OF DRUG PRODUCTS

HAVE BEEN RECOGNIZED AS RESPONSIBLE FOR A FEW

THERAPEUTIC FAILURES, IT IS PROBABLE THAT OTHER

THERAPEUTIC FAILURES (OR TOXICITY) OF A SIMILAR

ORIGIN HAVE ESCAPED RECOGNITION,

The fact that drug products differ in bioavailabilityis not, in itself, evidence that the use of suchproducts will produce significant practical problemsin the treatment of patients. However, it is alsoa fact that therapeutic inequivalence has beenobserved among certain chemically equivalent drugproducts. One example of therapeutic failurearising from variations in the bioavailability ofostensibly equivalent products involved theimportant and highly potent cardiotonic drug,digoxin. A number of patients were observed torequire unusually large maintenance doses of digoxindespite the absence of any condition that might haveexplained a high tolerance to the drug. Uponinvestigation, the patients were found to have lowdigoxin concentrations in their blood plasma(Lindenbaum et al, 1971). A crossover study revealedstriking differences in bioavailability among fourdigoxin preparations available in the same hospitalat the time. The peak concentration after a singledose was found to vary among the four drug productsby a factor of as much as seven. It is noteworthythat the margin of safety of this drug issufficiently narrow that serious or even lethaltoxic effects can result if the dose given andabsorbed is as little as twice that needed to achievea therapeutic effect.

14

Another instance of therapeutic failure with adrug product meeting compendia standards wasdemonstrated in the case of a thyroid preparationthat met those standards but was therapeuticallyinactive (Catz et al, 1962).

While these therapeutic failures resulting fromproblems of bioavailability were recognized andwell documented, it is entirely possible thatother therapeutic failures and/or instances oftoxicity that had a similar basis have escapedattention. The variability among individuals--in the absorption, excretion and metabolicconversion of drugs; in the individual physiologicalor toxic response; and even in the regularity withwhich the prescribed dose is actually taken--issuch that the source of an abnormal or inadequatetherapeutic effect in an individual patient is notreadily identifiable. Differences owing to thebioavailability of the administered product arelikely to be recognized only when information iscollected in an organized manner or effects arenoted in a significant number of patients undersufficiently close observation by an unusuallyalert and observant physician. More important,for drugs that, because of relatively narrow marginsof safety, are generally administered in doses thatproduce plasma levels not much higher than theminimum required for efficacy, therapeuticinequivalence must certainly occur whenever thereare substantial differences in bioavailability.

Therefore, although the number of instances ofdemonstrable therapeutic inequivalence is small,the problem is an important one and, in the caseof drugs with narrow margins of safety, assuranceof bioequivalence is vital.

15

3MOST OF THE ANALYTICAL METHODOLOGY AND EXPERIMENTAL

PROCEDURES FOR THE CONDUCT OF BIOAVAILABILITY STUDIES

IN MAN ARE AVAILABLE. ADDITIONAL WORK MAY BE

REQUIRED TO DEVELOP MEANS OF APPLYING THEM TO CERTAIN

DRUGS AND TO SPECIAL SITUATIONS OF DRUG USE,

The conduct of bioavailability studies in manrequires that a drug product be administered to agroup of individuals and that the time-course ofthe concentration of the drug in the blood beevaluated either directly or indirectly.* It isnecessary, therefore, that there be available(1) analytical methods for determining the concen-tration of the active ingredient in body fluids;(2) standardized procedures for administering thedrug product and obtaining appropriate blood and/orurine samples; and (3) adequate methods forstatistical analysis and interpretation of theresults.

ANALYTICAL METHODSIt may be necessary to measure, in a small volume ofbiological fluids, an amount of the intact drug that

* In 1972, the Academy of Pharmaceutical Sciencesof the American Pharmaceutical Associationpublished the Guide for Biopharmaceutical Studiesin Man, which presents a systematic approach tothe conduct of bioavailability studies based onanalytical determination of drug in blood and/orurine.

16

is only one millionth to one billionth of the doseadministered. Fortunately/, such recent advancesas gas-liquid chromatography, high pressure liquidchromatography, fluorescence techniques, massspectrometry, radioimmune assays, and microbiologicalassays have greatly increased our capability ofmeasuring such minute amounts of drugs. The methodsused must have not only adequate sensitivity andaccuracy, but also the selectivity that will make itpossible to quantify the drug in the presence of itsmetabolizes or of endogenous compounds that mayinterfere with the determination of the compound inbiological fluids.

In instances in which no sufficiently sensitivechemical method is available to detect the activeingredient, radioactively labeled molecules may beutilized. It must be verified, however, that themeasured radioactivity is contained in the intactcompound that has been separated from its meta-bolites. One must also be assured that the dosageform containing the radioactive drug to beadministered possesses, insofar as possible,physical and chemical properties identical to thoseof the usual (unlabeled) dosage form.

Some of these methods are cumbersome and rathertime-consuming, but they are capable of providingrelatively accurate measurements at the requiredlevels. Unfortunately, some of the earlypharmacokinetic studies were based on methodssubject to ambiguities. Continuing efforts,therefore, are still required to simplify andimprove the existing methods and to develop newones.

PROCEDURES FOR ADMINISTERING THE DRUG PRODUCTThe most common experimental plan for comparing thebioavailability of two drug products is a simplecrossover study. In this design, each individualin a group of subjects receives both drug products(at different times) so that there is a directcomparison of the absorption of each product in thesame individual. Special care must be taken toallow sufficient time to elapse between the adminis-tration of the first and second drug products sothat there are no carryover effects. In order tominimize the influence of such effects on theoutcome of the study, good experimental designrequires that each drug product be administeredinitially to half of the subjects.

In some instances, more than two drug products areto be compared in a single investigation. If threeproducts are to be studied, all three can beadministered to each individual in a suitablybalanced order. However, investigations involvingmore than three drugs may require alternativeexperimental designs, as it may be impractical togive more than three drug products to the sameindividual. A suitable way of planning suchinvestigations is to utilize statistical experimentalplans called incomplete-block designs. Thestatistical methodology for such experimental plansis readily available.

For some applications of some drug products, thesteady-state concentration of the drug in the blood,attained after repeated administration of the drugproduct at regular intervals, may be a more appropriateindex of bioavailability than the time-concentrationcurve after a single dose.

STATISTICAL METHODOLOGY FOR BIOAVAILABILITY STUDIESThe statistical methods to be used in bioavailabilitystudies should be chosen with careful attentionbeing given to the effect of the variations amongindividuals and among batches of nominally identicalmanufactured drug products. The planning, analysisand interpretation of these experiments are notroutine problems but, rather, require considerablecare, consonant with the purpose for which the dataare to be used.

I N D I V I D U A L V A R I A B I L I T Y A N D B I O E Q U I V A L E N C E .

When drug products are administered to individuals,the investigator inevitably finds differences inone or more of the variables measured. Thesedifferences are due partly to factors related todosage form and partly to biological factors uniqueto each individual, since each person has his owncharacteristics for absorption, metabolism andexcretion of each drug. Through appropriate use ofstatistical procedures, it is possible to identifythe variations that result from differences amongindividuals and thus to isolate those that resultfrom differences in the bioavailability of the drugproducts.

18

BATCH-TO-BATCH VARIABIL ITY AND B IOEQUIVALENCE

Some drug products may exhibit substantial batch-to-batch variation with respect to importantbioavailability characteristics. If the batchesare manufactured under adequately controlledconditions, then one could regard the differentbatches to be theoretically bioequivalent. However,one would expect the average levels of some of thebioavailability characteristics to vary from batchto batch. The measurement of the effects of batch-to-batch variations on bioavailability is importantin assessing the effect of changes in the manu-facturing process which may take place in the future.

The problem arises as to how to judge whether twochemically equivalent drug products are bioequivalent.The solution to this problem depends on the batch-to-batch variation of the bioavailability. Aworking rule for judging the bioequivalence of twodrug products might be that the two are consideredequivalent if the differences between them aresimilar to what one would expect from the batch-to-batch variation of the original product.Unfortunately, there is little information concerningthe effects of batch-to-batch variation onbioavailability. Studies to examine this questionhave not often been carried out, since they requirein vivo tests of bioavailability unless there arein vitro tests whose results have been shown tohave a high correlation with bioavailability.

SAMPLE-S IZE CONSIDERATIONS

One of the most important and difficult problems inplanning bioavailability investigations is theselection of the appropriate sample size. Ifconventional statistical tests of significance areused to analyze the data, then it is possible thatstudies involving small numbers of observations(subjects) may fail to yield differences that arestatistically significant even if the drug productsbeing compared are, in fact, different. Alternatively,if large numbers of observations are used, then onemay find statistically significant differencesbetween drug products, even if the real differencesare small and of no pharmaceutical or therapeuticsignificance.

Therefore, in planning bioavailability investigations,one must determine the difference in mean valuesof the parameters of bioavialability that it ispractically (pharmaceutically or therapeutically)

19

important to detect. The choice of sample sizerequires that the probability of failing to detectimportant differences be small, when such differencesexist.

METHODS OF ANALYSIS

The experimental results of bioavailability studiescan be analyzed statistically in many different ways.The statistical methods of analysis depend on (1) thestatistical model of the concentration-time curve,(2) the statistical model of the various sources ofvariation (for example, person-to-person variations,or nonindependent measurements) , and (3) theexperimental plan that specified how the measure-ments were to be made.

Perhaps the simplest way to use concentration-timecurves for comparing two drug ‘products is not tocompare the entire curves but to compare charac-teristics of the curves that are deemed importantwith regard to the drug product under study--forexample, area under the curves, peak heights, orrates of absorption. If only a single characteristicis involved, an appropriate method of analysis isthe method of paired comparisons, in which eachindividual generates a paired difference. Thereare also adequate statistical procedures for thecomparison of two or more sets of variables, andthese, can be used when more than two drug productsare studied.

21

IT IS NEITHER FEASIBLE NOR DESIRABLE THAT STUDIES

OF BIOAVAILABILITY BE CONDUCTED FOR ALL DRUGS OR

DRUG PRODUCTS, CERTAIN CLASSES OF DRUGS FOR

WHICH EVIDENCE OF BI0EQUIVALENCE IS CRITICAL

SHOULD BE IDENTIFIED, SELECTION OF THESE CLASSES

SHOULD BE BASED ON CLINICAL IMPORTANCE, RATIO OF

THERAPEUTIC TO TOXIC CONCENTRATION IN BLOOD, AND

CERTAIN PHARMACEUTICAL CHARACTERISTICS,

More than 20,000 prescription drug products arepresently available from drug manufacturers. Foronly a few of these are there adequate datadocumenting their bioavailability in man. Becauseof the large number of drug products for whichstudies of bioavailability might be conducted, theenormous number of human volunteers that ‘would beneeded, and the large number of clinical investi-gators and other sciefitific personnel who would beneeded to do the work, it is clearly not feasibleto carry out studies of bioavailability in man forall drug products.

Furthermore, even were it feasible to do so, itwould not, in our opinion, be ethically justifiable.The administration of drugs to man is never withoutsome hazard, although in some cases the risk isvery small. In addition, subjects always experiencesome inconvenience and usually some discomfort. Itis axiomatic that these hazards and inconveniencesshould not be incurred unless they are outweighedby the prospective benefits of the studies. Whenstudies of bioavailability are necessary to insure

22

the effectiveness and safety of therapeuticallyimportant drugs, such studies are ethicallyjustified and necessary. For any lesser purpose,they will rarely be justifiable. We do believe,however, that bioavailability studies should berequired for products if the active ingredient inthe product has not yet been introduced on themarket.

In asserting that studies of bioavailability willnot be required for all drug products, it becomesimportant to set forth general criteria to guidethe selection of those products whose bioavailabilityshould be documented by testing in man. A necessary,albeit not a sufficient, condition is that the drugbe one that serves a clinically important purpose,especially if it is used in the treatment or theprevention of severe or life-threatening conditions.However, the clinical importance of a drug alonewould not be an adequate reason for conducting anin vivo study of its bioavailability.

Beyond questions of feasibility and ethicaljustifiability, there are other good reasons torefrain from conducting studies of bioavailabilityof a number of drugs, including many that aretherapeutically important. Many drugs are given infairly standard dosage with little regard to thebody size of the patient or to the titration ofdosage to exactly the desired therapeutic effect.Such practice reflects the fact that for many drugsthere is a wide margin between the concentration ofthe drug in the body fluids needed to produce thedesired therapeutic effect and the concentration atwhich undesirable toxic effects begin to appear.Thus, the standard dose is usually one that willproduce in the vast majority of patients a con-centration in the blood well above the levels neededfor the therapeutic effect without reachingunacceptable levels of toxicity. Clearly, undersuch, circumstances a wide range in bioavailabilitycould be tolerated without hazard of therapeuticfailure.

For example, no one would question the clinicalimportance of penicillin, nor the seriousness ofmany of the conditions for which it is used. Butthe margin between its effective concentration andits toxic level is so great that the prescribeddosage can be aimed at achieving a concentrationin the blood far above the minimum effective level,thus insuring a therapeutically effective level invirtually all users. Moderate differences in theconcentration achieved in the blood owing to

23

differences in the bioavailability of chemicallyequivalent penicillin products would be easilytolerated.

Conversely, drugs that have a relatively narrowrange between the concentration needed for thedesired therapeutic effect and the concentrationassociated with significant toxicity would becandidates for testing of bioavailability, sincerelatively modest changes in the concentrationachieved in body fluids might well be associatedwith large changes in the frequency of therapeuticfailure or significant toxicity. Examples of drugsthat might fall into this category include a numberof cardioactive drugs (digitalis glycosides,quinidine), anticonvulsant agents (diphenylhydantoin) ,some corticosteroids, and certain antibiotics(chloramphenicol and cephalosporins) .

The pharmaceutical properties of drugs and drugproducts affecting volubility and dissolutioncharacteristics in gastrointestinal fluidsconstitute additional factors to be considered inthe selection of chemically equivalent drug productsfor which evidence of bioavailability may be required.The drugs themselves may differ in particle size orcrystal form, and these differences can affectvolubility. In addition, the method of manufactureof other ingredients, such as excipients, diluentsor fillers may alter dissolution characteristics.For many drugs and drug products, however, highvolubility of the drug and rapid dissolution ofthe product make it relatively unlikely that majordifferences in absorption from the gastrointestinaltract will occur. However, there are certainconditions in which even these drugs may showdifferences in bioavailability.

Sustained-release and enteric-coated productsconstitute a separate problem that we have notconsidered in any detail, but it is clear thatthere is a particular need for tests of thebioavailability of many such preparations.

In summary, measurements of bioavailability may becritical to assuring therapeutic equivalence ofdrugs or drug products that:

● Are used for treatment or prevention ofserious illness;

● Have steep dose-response curves or unfavorabletherapeutic indices; or

24

● Contain active ingredients that are relativelyinsoluble or are converted to insoluble formsin gastrointestinal fluids.

In view of (1) the demand on limited resources(particularly the time and effort of trainedinvestigators and other scientific personnel), (2)the hazards, however minimal, of exposure ofvolunteers to additional drugs and the associatedinvestigational procedures, and (3) the unnecessaryaddition to the cost of drug products that suchstudies would entail, we consider it desirablethat studies of bioavailability in man be carriedout only in the case of drugs with characteristicssuch as those described above. We believe thatappropriate panels of experts would have littledifficulty in distinguishing between those drugsfor which bioavailability studies in man should be ●

required and those for which they should not. (Seediscussion on page 59.)

25

5PRESENT COMPENDIA STANDARDS AND GUIDELINES FOR

CURRENT GOOD MANUFACTURING PRACTICE DO NOT INSURE

QUALITY AND UNIFORM BIOAVAILABIL ITY FOR DRUG

PRODUCTS, N O T ONLY MAY THE PRODUCTS OF D IFFERENT

MANUFACTURERS VARY, BUT THE PRODUCT OF A S I N G L E

MANUFACTURER MAY VARY FROM BATCH TO BATCH OR MAY

CHANGE DURING STORAGE,

As we have pointed out, the fact that severalmarketed drug products that are chemically equivalenthave been found to differ significantly in bioavail-ability is a clear indication that presentcompendia standards and guidelines for CurrentGood Manufacturing Practice (CGMP) do not insurebioequivalence. A consideration of these standardsand guidelines may clarify the reasons for theirinadequacy.

COMPENDIA STANDARDS

H I STORY

The official drug compendia originated independentlyof the Federal Government and have been maintainedthrough a convention and committee system composedof interested health professionals. The UnitedStates Pharmacopeial Convention, Inc., (USP) datesback to 1820, and the National Formulary Board(NF), subdivision of the American PharmaceuticalAssociation, was established in 1888. The compendiathat have been published by these organizations

26

since their founding were developed to help insurethat drug products of an acceptable level of qualitywould be available to physicians. They set forthstandards for drugs used in dispensing and compoundingprescriptions. Later, they provided pharmaceuticalmanufacturers with descriptions of recognizedstandards for products conforming to certain criteriaof strength, purity and quality.

Pharmacists in early America were responsible foridentifying and establishing the quality of the rawmaterials they used in compounding prescriptions.In the 1880s, the pharmaceutical industry becamethe primary supplier of drug products and acceptedresponsibility for standardizing the products.However, the compendia tests for strength andpurity still retained their basic simplicity,requiring minimal instrumentation and analyticalskill, so that a pharmacist could still perform theassays.

Reliance on simple approaches to specifications ,ofmaterials and the control of quality persists tothis day. These shortcomings in present-dayassessment of drugs particularly pertain to thetests for identity, purity and potency that formthe legal basis upon which compliance or lack ofcompliance is established (Federal Food, Drug,and Cosmetic Act) . In some cases, physical testsand assay procedures of much greater sensitivitythan” those in the compendia are known-; many ofthese tests are fully automated.. .

The FDA, at its National Center for Drug Analysisin St. Louis, has found it necessary to establishits own set of tests for monitoring the quality anduniformity of drug products. Many of these testsare also automated and reflect more advancedtechnology than do the compendia tests. Neverthe-less if a given drug product fails to pass theseadvanced automated tests, the FDA is still requiredto repeat the analyses using official, andfrequently less accurate, manual methods in orderto establish a lack of legal compliance. In anycase, it is difficult to defend the current selectionof methods when it is apparent that they are, insome cases, inaccurate, insensitive and nondiscrim-inating, as can be seen by a comparison of severalof the present compendia monographs with the datapublished in more recent sources.. .-

*One such current source is Analytical Profilesof Drugs, edited by Klaus Florey and (continued)

27

C O M P E N D I A M O N O G R A P H S

The selection of active ingredients and finished drugproducts for inclusion in the United StatesPharmacopoeia (USP) and the National Formulary (NF)is based primary on their therapeutic importanceas judged by the revision committees; the basisfor selection of excipients is not clearly specified.A large number of active ingredients and a number ofexcipients are not included in the compendia.Certain excipients are included, but specific testsand standards applicable to their special use ascomponents of tablets and capsules are not listed.For excipients not described in the compendia, thedrug manufacturer may set his own specific require-ments or accept the specifications of the firmthat supplies these ingredients to him. Sincethese so-called inert substances can influence thestability and/or release of the active compoundfrom the dosage form, the lack of standards forexcipients may well lead to variation in importantproperties of the final drug product.

The dosage form monographs contained in the officialcompendia deal with end products. They do notdescribe inactive components or processes ofmanufacture. Thus, they do not assure exclusion ofthe possibility that drugs might interact withexcipients, some of which can cause decomposition,and do not exclude the formation of complexes notdetectable by present official tests. There arealso no specifications for the granulation orprecompression mixture as a preliminary means ofinsuring uniformity of content and dissolutionproperties. Nor do the monographs include adequatetests to specify characteristics of the activeingredients needed to insure quality and standardizedlevels of bioavailability of the finished dosageform.

Beyond the tests for weight and uniformity of content,the most important tests included in the monographsfor tablets or capsules are those for disintegrationand dissolution. The disintegration tests currentlycalled for in the compendia are relatively crude andrequire that an observer record the time necessaryfor the disintegration of six tablets. Since theaverage time for disintegration is about 30 minutes,

(Continued)published by Academic Press. TWO volumes have” beenpublished to date, the first in 1972 and the secondin 1973.

‘lf7-6~0 () - 74 - 3

28

a period long enough for the manufacture of hundredsof thousands of tablets, disintegration tests areclearly unsuitable for use in the monitoring andcontrol of the manufacturing process.

The dissolution tests, which have been specifiedin the compendia since 1970, are carried out inapparatus that may introduce extraneous sources ofvariation in the dissolution process. The testspecifications for individual drug products havebeen established somewhat arbitrarily withoutsufficient consideration of the physical chemicalproperties of the active ingredient and of bio-logical factors that may be critical in the dis-solution process.

While uniformity in testing equipment is desirable,a single apparatus will not cover all test conditionsfor meaningful studies of dissolution. It should benoted that in a New Drug Application (NDA), the FDAallows the pharmaceutical industry to use allavailable analytical techniques and to establishunique anlytical and quality control specificationsfor a new drug product. Before approval of theNDA, the Washington laboratories and the districtlaboratories of the FDA establish their own capa-bilities to conduct these analytical tests, in casethey should be needed for their regulatory activities.It is unfortunate that the official compendia arerestricted to measurement systems with severelimitations in design.

STATISTICAL IMPLICATIONS

Irrespective of the technical soundness of the tests,the adequacy of which we have questioned above, thecompendia standards for finished drug products havebeen established without any serious consideration ofthe adequacy of the statistical procedures for thedetection of defective finished products. In fact,the sampling procedures are generally inadequate.No consideration is given to (1) the method ofdrawing the sample, (2) the rationale for specifyingthe number of units of the product to be testedand (3) the statistical criteria for permitting abatch or lot of a finished product to pass. Thus,current compendia standards offer insufficientprotection against the possible marketing ofsubstantial batches of defective products.

A. THE OFFICIAL TABLET DISINTEGRATION TEST

The USP and NF tablet disintegration tests requirethat six tablets be tested initially. If all tabletsdisintegrate under the conditions described, then it

29

is assumed that the lot or batch of tablets fromwhich the sample of six was drawn is satisfactory.Alternatively, if one or two of the initial sixtablets fail the test, then the test is repeatedon 12 additional tablets. The criterion ofacceptability is then that at least 16 of the totalof 18 tablets disintegrate completely within thetime specified. Table A summarizes the probabilitythat the sample will meet the specified requirementsas a function of the percentage of defective tabletsin the batch from which the sample is drawn.

Table A. Probability of Acceptance as Functionof Percentage of Defective Tablets(Disintegration Test)

Percent Defective Probability of Acceptance

1% . 9955% .96

10% . 7 920$% . 3930% . 1 540% . 0 0 4

Thus, for example, batches of tablets of which 20%are defective with regard to the disintegrationtests would pass the specified disintegration test39% of the time. This test procedure offers ahigh order of assurance of detecting defectivelots only if the lots contain at least 40%defective tablets.

B . D I S S O L U T I O N T E S T

The USP and NF dissolution tests state that sixtablets initially are to be tested for dissolutionproperties. If all six meet the specifications,it is assumed that the batch from which the sampleis drawn is acceptable. If one or two units failto meet requirements, then the test is repeated onsix additional units. If the testing reaches thispoint, at least 10 of the 12 units tested must meetthe requirements of the dissolution test in orderto comply with compendia standards. Table Bsummarizes the probability of acceptance in thistest as a function of percentage of defectivetablets or dosage units in the batch from whichthe sample is drawn.

30

Table B. Probability of Acceptance as a Function of Percentage ofDefective Tablets or Dosage Units (Dissolution Tests)

Percent Defective

1%5%

10%20%30%40%50%60%

Probability of Acceptance

.9998

.98

.90

. 5 8

. 2 8

. 1 0

. 03

. 006

Note that the criteria for acceptance in this testresult in even poorer protection than in thedisintegration test. A batch having 20% defectivetablets or dosage units will be accepted 58% of thetime. Only when the percentage of defective unitsin a batch is 60% or greater do these requirementsoffer high assurance of detecting defective batches.

c . CONTENT UNIFORMITY

The requirements for satisfying compendia standardsfor uniformity of content are somewhat morecomplicated than the dissolution and disintegrationtests. The standards specify the following -

procedures:

● Ten capsules or tablets from batch areinitially assayed. The requirements aresatisfied if nine or 10 of these units fallwithin the limits of 85%” to 115% of thecontent specified and all results fallwithin 75% to 125%.

● If two or three of the initial 10 unitsyield results that fall outside the rangeof 85% to 115% of the specified contentand none fall outside the range of 75%to 125%, an additional 20 units are tested.The requirements are then met if all 30results fall within the limits of 75%to 125% and if not more than three resultsfall outside the limits of 85% to 115%.Table C summarizes representative valuesof the percent defective versus theprobability of accepting the batch fromwhich the samples are drawn.

31

Table C. Probability of Acceptance Versus Percentage of Defective(content Uniformity Test)

Probability of

Acceptance

. 9 7

.92

.88

.82

.72

.56

.25

.17

.13

.01

.009

.006

This sampling procedure for content uniformitycannot, with any assurance, be depended upon fordetecting defective batches. For example, notethat a batch of capsules or tablets having 25%defective units would meet compendia standardsas frequently as 25% of the time.

D. S A M P L I N G P R O C E D U R E S

The compendia standards are vague with regard tohow the samples of tablets or dosage forms are tobe chosen for testing. If the standards are tohave any utility for judging the quality of abatch, then it is essential that careful attentionbe given to the drawing of a sample. In order tobe able to draw inferences about a batch beingtested, it is absolutely essential that the sampleof dosage units from that batch be randomlyselected. All dosage units in the batch must havean equal chance of being selected for the sample.Yet, the compendia standards makea random sample of dosage forms isare guidelines given as to how thebe drawn.

C O M P E N D I A R E V I S I O N P R O C E S S

An examination of the USP revision

no mention thatnecessary, norsample should

process “revealsthe inherent weaknesses in the development ofpresent compendia standards. As a nonprofit,independent organization, the USP draws its member-ship from a broad base of organizations and

3 2

institutions concerned with druqs. Its Committeeon Revision, composed of 20 medical and 40 pharmaceutical scientists, is selected once eachfive years by the 300 USP Convention members. Eachmember of the Committee on Revision draws upon theexpertise of other scientists, thus expanding thenumber of scientists contributing to the effort toabout 200. The contributions of these persons, insome cases, may be very superficial, since both theUSP and the NF lack the funding and the scientificstaff needed to permit serious scientific pro-spective searches to be conducted for potential orsuspected problems.

Except for out-of-pocket expenses, all work doneby committee members is voluntary. The USP (andNF) are dependent on funds derived primarily fromthe compulsory purchase of their published com-pendia by pharmacies in most states and from thesale of reference standards of official compendiasubstances. Major committee revisions are too in-frequent to stay abreast of advancing technology.One factor that delays compendia revision and thedevelopment of test methods is the shortage of per-sonnel to accomplish these tasks; in 1973 therewas a full-time-equivalent staff of only 4.2persons involved in the development of compendiastandards. (In 1967, the corresponding figurewas 1.7.) Thus, the USP and the NF both lack theinternal and external resources required to developrational and modern standards. Their continuedreliance upon such limited resources is anachronisticand inadequate for the scientific activities necessaryto insure pharmaceutical equivalence of drug products.

It should be noted, however, that the Food, Drug,and Cosmetic Act contains a section enacted byCongress in 1944 entitled The Revision of the USP;Development of Analysis and Mechanical and PhysicalTests in which the Secretary of Health, Education,and Welfare is “authorized hereafter to cooperatewith the associations and scientific societies in therevision of the USP and in the development of methodsof analysis and mechanical and physical tests necessaryto carry out the work of the FDA.” It appears thatvery little has been done to implement this sectionof the Act during the last 30 years. We have beenadvised that the FDA allocated only $75,000 to theimplementation of this section of the Food, Drug,and Cosmetic Act in 1974.

3 3

CURRENT GOOD MANUFACTURING PRACTICEThe regulatory measures of the FDA are based on theFood , Drug, and Cosmetic Act, as amended up toAugust 1972, and the FDA’s supplementary regulationsand guidelines established under the authority ofthe Act. In recent years, the FDA has establishedguidelines under the title Current Good ManufacturingPractice. Unfortunately, these guidelines are basedprimarily on the official compendia standards andare limited to rather general-statements that may besubject to wide differences in interpretation.They correctly emphasize environmental factors,including control of air supply and organization ofmanufacturing plants, to minimize cross-contaminationof products or errors in packaging or labeling;they also focus on the integrity of quality controland production records. These guidelines havegreatly improved the handling of materials,cleanliness, and consistency in production, but theydo little to minimize lot-to-lot variation in theoutput of manufacturers generally or bioinequivalenceamong the chemically equivalent” products of differentmanufacturers.

A manufacturer of a duplicate product can requestand receive approval to market his product on thebasis of an Abbreviated New Drug Application (ANDA)specifying only the official compendia standardsfor controlling the quality of the raw materials andfinal drug product, including both the active . .ingredients and the excipients used in theproduct. Further, the raw materials can be includedby the manufacturer on the basis of an identitytest and a certification from the supplier of theraw materials that they passed the officialcompendia tests. Excipients not covered in anofficial monograph may be accepted from the supplierwithout any particular specifications at all. Themanufacturer using these raw materials is expectedonly to assure himself that the raw materialmanufacturer maintains an acceptable level of qualityassurance.

Current Good Manufacturing Practice guidelines donot mention preformulation tests or intermediatetesting and refer only vaguely to in-processingtests. Although a manufacturer of duplicate productsis required only to meet the compendia specifications,with all their limitations, the original manufacturerof the product is required to continue meeting allspecifications of production and quality controldescribed in the approved NDA. Undoubtedly, someduplicate products undergo tests and are manufactured

34

according to specifications that go beyond thosein the compendia standards. There are, then, twosets of standards: those in the approved NDA andthose established by the official compendia. It ishardly surprising that pharmaceutical equivalenceis not guaranteed by the present procedures.

A million or more individual batches of tablets andcapsules are manufactured each year in approximately800 pharmaceutical manufacturing plants in the U.S.Some of these batches include several million units.Inspection of plants is required only once every twoyears by the FDA, although it usually is done atleast once a year. The National Center for DrugAnalysis in St. Louis has the capacity to test onlya few thousand batches of drug products per year.The large number of different drug products andthe sizes and number of batches produced make itdifficult to establish assurance of quality usingthe present system of inspection.

In many parts of the country having very few.. pharmaceutical manufacturing plants, FDA agentstrained in food inspection also serve “as druginspectors. These inspectors are often trained fortheir ancillary assignments in short courses inpharmacy schools; most of them are not trainedsufficiently to determine whether adequate testingof the final product has been performed. The verynature of the present Current Good ManufacturingPractice-guidelines forces inspectors to concentrateon plant design and maintenance practices and torely on inspection of records to insure manufacturers’compliance with compendia standards.

35

6NEW COMPENDIA STANDARDS FOR ACTIVE I N G R E D I E N T S ,

EXCIPIENTS AND FINISHED DRUG PRODUCTS

DEVELOPED AND REVISED ON A CONTINUING

REFLECT THE BEST AVAILABLE TECHNOLOGY

QUALITY AND UNIFORM BIOAVAILABILITY,

SHOULD BE

BASIS TO

TO INSURE

APPROPRIATE

STATISTICAL PROCEDURES SHOULD BE SPECIFIED TO

MAKE CERTAIN THAT THE PURPOSES OF THE STANDARDS

ARE OBJECTIVELY SATISFIED, THE GUIDELINES FOR

CURRENT GOOD MANUFACTURING PRACTICE SHOULD BE

EXPANDED TO INCLUDE SPECIFIC DESCRIPTIONS OF ALL

SIGNIFICANT ASPECTS OF MANUFACTURING PROCESSES

FROM THE RAW MATERIALS TO THE FINAL PRODUCT,

The philosophy that is applied in regulating thepreparation of sterile solutions for parenteraluse can serve as a model in establishing therequirements for insuring the quality of drugproducts. In the case of parenteral solutions,it has been determined that a zero level ofmicrobial contamination should be the standard ofquality with respect to sterility. No amount oftesting applied only to the final product couldinsure this total freedom from microbialcontamination without destroying and testing everypackage. Instead, a series of tests and procedureshas been specified for every material used and atevery stage of the process of manufacture in orderto minimize the possibility of microbial contamination

36

at each step and thus in the final package.

Similarly, rather than aiming to insure theuniformity of oral drug products by tests appliedonly to the final entity, uniformity of qualityand bioavailability should be promoted by makingas specific as possible the requirements for thecharacteristics of the materials to be used, theprocesses by which they are to be assembled, andthe tests to be applied to representative samplesof the final product. Testing at all stages ofthe manufacturing process should reflect the bestavailable technology and should be based onstatistical procedures which will insure that thepurposes of the tests are objectively satisfied.

COMPENDIA STANDARDS

T E S T S T O B E A P P L I E D T O R A W M A T E R I A L S

Many of the present monographs of the USP and NFfor raw materials in drug products contain outmodedor suboptimal procedures. Revision committees haveadopted the policy of specifying only a singleanalytical method for determining the purity of anactive ingredient, even though the Food, Drug, andCosmetic Act clearly permits more than one suchtest. As we have noted earlier, the policy of thecompendia bodies has severely limited the improve-ment of analytical procedures. Most modernanalytical methods are sufficiently accurate andspecific that results obtained with one can becompared with results obtained by other methods.Although for some active ingredients, a singleanalytical procedure may be desirable, for others,any one of a number of methods might be acceptable.

Beyond its chemical identity and purity, there arecharacteristics of the active ingredient that couldbe specified in order to increase the reproducibilityof the properties of the final drug product. Thesemight include such specifications as the distributionof particle size, a maximally acceptable particlesize, requirements for a certain crystal form,compressibility and a requirement for a rate ofdissolution. Certainly not all such propertiesneed be specified for all drug products, but theymight be critical to the quality of some.

Raw material test requirements for excipients shouldbe expanded in the compendia monographs. The USP

37

and NF do contain tests for some excipients. Thesetests however, tend to be nonspecific, and testsfor many other excipients are not included at all.The same degree of test specificity describedabove for the clinically active material isappropriate for those excipient materials thatinfluence bioavailability of the drug product.

S P E C I F I C A T I O N S F O R I N T E R M E D I A T E S T A G E SO F T H E M A N U F A C T U R I N G P R O C E S S

The properties of the final product will be moreeasily controlled if requirements are imposed atintermediate stages of manufacturing. One suchstage is that of the precompression mix--,the stageafter the active ingredients and various excipientshave been mixed but before the mixture iscompressed into tablets. The qualities of theprecompression mix could be defined by tests suchas those for bulk density, loss of weight on drying,particle-size distribution, compressibility, andrate of dissolution. Again, it is not intendedthat all of these tests be required for all rawmaterials or at all stages of the process. However,certain specifications for the intermediate stagesmay be very important in establishing batch-to-batchreproducibility, especially if any changes are madein the manufacturing process.

Data from such tests might make it possible todetermine whether changes in processes or startingmaterials have caused changes in the properties ofthe final product sufficient to necessitate theperformance of new studies of bioavailability.We have been informed by the FDA that it has nopolicy at present that outlines specific require-ments for bioavailability testing when changes aremade in size of batches, processing equipment orraw materials used in the manufacture of drugproducts. Instead, these decisions are left to thevarious divisions within the Bureau of Drugs to bemade on a more or less arbitrary basis. Theproposed tests for products at intermediate stagesmight provide the FDA an objective basis formaking these decisions. Testing at intermediatestages would be useful when applied to certainvery insoluble active ingredients since it isdifficult to apply meaningful dissolution teststo the final product of such drugs because of theinordinate volumes of fluid needed to dissolveall of the active ingredient.

38

T E S T S APPLIED TO THE F INAL PRODUCT

Dissolution tests should be included in specifi-cations for the final product. The tests specified,however, should relate to the physical and chemicalproperties of the drug and to possible interactionswith the gastrointestinal fluids. This may requirechanges in the solvent to simulate the events thatoccur as the capsule or tablet is exposed first togastric and then to intestinal fluids. Theapparatus currently in use is not readily adaptableto such changes of solvent. The disintegration anddissolution test procedures, as well as tests ofdeaggregation, are in need of serious reevaluation.Automated dissolution apparatus is in use in anumber of laboratories and is adaptable to themeasurement of deaggregation; this apparatus shouldalso be evaluated as a possible alternative todisintegration testing. When appropriate bioavail-ability data have been developed for a specificdrug product, it is essential that the dissolutiontests for that drug product be modified so that theresults bear a high statistical correlation withbioavailability. In the absence of bioavailabilitydata, dissolution tests should be adopted for theirvalue in quality control.

It should be noted that the present regulations forthe certification of antibiotics do not includedissolution tests. The published data indicatingthat some preparations of antibiotics have beensubject to problems of bioavailability clearlydemonstrate the need for standards of dissolution.

STATIST ICAL PROCEDURES

We have indicated above some of the statisticalinadequacies of present sampling and testingprocedures. It should not be difficult to effecta great improvement in these. Statisticalprocedures have reached a high degree of develop-ment and can be used in such a way as to yield ahigh level of assurance that defective batches willbe detected. The detection of batches containinga high proportion of defective dosage forms can besubstantially improved by (1) choosing sample sizesin accordance with the average quality that isdeemed necessary, (2) adopting sequential samplingand inspection procedures, and (3) recognizing thatthe acceptance or rejection of a batch should bedetermined not on the basis of the individual testsalone, but on the totality of the tests on allimportant characteristics.

39

In order to utilize the appropriate statisticalprocedures for judging the quality of batches, itis necessary to recognize that no sampling plan canguarantee that there will be zero defective dosageforms in every acceptable batch. Statisticalsampling plans can yield an estimate of the averagelevel of quality of accepted batches and thepercentage of defective dosage forms that may bepresent in accepted batches. Sampling plans should .not be the same for all drugs but should be designedto take account of the cost of testing and the harmthat might ensue when defective dosage forms areused for therapeutic purposes. Clearly, thestatistical sampling plans for a drug such asdigoxin should be different from tests on many ofthe antibiotics. Drugs that have narrow therapeuticmargins and/or are capable of producing serious sideeffects should be more tightly controlled than drugswithout these characteristics.

GUIDELINES FOR CURRENT GOOD MANUFACTURING PRACTICE

The present guidelines are too general and non-specific to insure the uniformity of drug products.For example, Section 133.8 of the Code of FederalRegulations states:

To insure uniformity and integrity ofproducts, there should be an adequatein-processing control, such as checkingthe weights and the disintegrationtimes, adequacy of mixing, the homo-geneity of the suspensions and theclarity of solutions. In-processingsampling shall be done at intervalswith suitable equipment. Representativesamples of the final dosage forms shallbe tested to determine their conformitywith the specifications of the productbefore distribution.

A considerable period of time will probably berequired to complete the total revision of theindividual compendia monographs and to expandthem to include as many of the marketed drugproducts as possible. However, the FDA has theauthority to modify the CGMP guidelines and shoulduse it as rapidly as possible to establish morecomprehensive standards for drugs and drug products.This could be accomplished by establishing acomprehensive list of appropriate tests for use

—

4 0

at three stages--raw material, intermediate, andfinished product. The individual manufacturers could

“-then be ’required to develop specific standards for.their individual preparations, and a reasonable periodof time should be allowed for the implementationof these standards. - “

Since different manufacturers use different equip-ment and processes for the manufacture of tabletsand capsules and include different excipients inthem, certain aspects of these specifications willhave to be individualized. It will be necessaryto develop specific deaggregation and/or dissolutiontests for the products at intermediate steps of themanufacturing, process as well as certain other uniquetest procedures in order ‘to insure that the qualityand uniform bioavailability of the product can bemaintained.

Finally, consideration should be given to theproblem of determining how batches of fabricatedtablets that do not meet specifications should bereprocessed. The availability of sufficientlydetailed specifications for intermediate stagesshould allow evaluation of whether the reworkedmaterial possesses sufficiently similar specifi-cations to qualify for continued processing.

We recommend that the CGMP guidelines be amended torequire that every manufactured drug product haveits own quality assurance plan, with attention ineach plan given to each major step of the manu-facturing process. Suitable testing proceduresshould be specified at each key stage of manu-facturing to insure identification of any stagethat is out of control. The design of such qualityassurance plans should be regarded as an importantcomponent of the preformulation investigation ofeach product.

Quality assurance plans should be drawn up by themanufacturer and need not follow a standardizedformat. The plans should take into account specialfeatures associated with the plant, its personnel,and the drug product. The plans need not be staticand should be modified by the manufacturer wheneverappropriate. Thus, the CGMP would shift in emphasisfrom a concern for a clean and safe working environ-ment to the control of all major design factors inthe manufacturing process.

We also recommend more frequent inspection ofmanufacturing plants by the FDA to assess the degreeof compliance with new compendia standards and

41

improved CGMP guidelines. As we recommend belowon pages 46-48ff, the FDA should also be given theauthority to request the submission of a company'srecords that demonstrate compliance with standards.To conduct the kind of inspection that we recommend,on a frequent basis, substantial increases will berequired in both the number and capability ofinspectors in the FDA.

The recommendations that we have made in thissection all have the goal of making far morespecific the requirements for the manufacture ofeach drug product, including coverage of all rawmaterials to be used, processes by which they areto be assembled, and tests to be applied at everystage to make certain that the requirements arebeing met.

The resulting set of new standards should not beconsidered to be a fixed ideal. Rather, it shouldbe revised on a continuing basis as new informationand advances in the state of the art are developed.On the other hand, we could caution against thecapricous introduction of new requirements; changesshould be made only when the benefit to be derivedwarrants the disruption of manufacture that theymight impose. We believe that the kinds of chanqesin procedure that we have recommended will greatlyincrease the quality and uniform bioavailability ofdrug products and warrant the new designation ofpharmaceutical equivalents.

43

7ADDITIONAL RESEARCH AIMED AT IMPROVING THE

ASSESSMENT AND PREDICTION OF BIOEQUIVALENCE IS

NEEDED, THIS RESEARCH SHOULD INCLUDE EFFORTS TO

DEVELOP IN VITRO TESTS OR ANIMAL MODELS THAT

WILL BE VALID PREDICTORS OF BIOAVAILABILITY IN MAN,

An effective technological base is required for theprovision of reasonable assurance that patients willexperience adequate and predictable blood levels oforally administered drugs. Although much of thenecessary methodology is already developed,additional research is needed to strengthen thistechnology in several areas.

REFINEMENT AND DEVELOPMENT OF ANALYTICAL METHODOLOGYThe ultimate proof of bioequivalence is achieved bydemonstrating equivalent concentrations of drugsin plasma over an appropriate period of time afteradministration. Methods are now available formeasuring the concentration of most drugs in bloodand/or urine. Further, presently availableanalytical techniques could be applied to theanalysis of some drugs for which methods are notnow available. For some drugs, however, newanalytical methods will have to be developed andpresent capabilities refined to permit reliablemeasurement of their concentration in plasma.For example, certain drugs used in cancer chemo-therapy, some steroid drugs and reserpine cannotcurrently be measured adequately in studies ofbioavailability. Whereas these analytical problemsare potentially solvable through application of

44

existing technology, their solution is not likelyto be easy. It will require substantial initialinvestment in the synthesis of drugs labeledwith stable, and radioactive isotopes and developmentof appropriate antibodies for radioimmunoassay.

SPECIAL POPULATION GROUPSAt present, there is inadequate informationregarding the prediction of bioequivalence in illpatients from bioavailability studies in normaladults. The uncertainty hinges on the importanceof differences in absorption rates. It is possiblethat two drugs which are totally absorbedbut have different absorption rates in the normaladult might differ significantly in the degree oftheir absorption in sick patients. Similarly, drugsthat are absorped in the normal adult may beincompletely absorbed in infants since infants haveimmature digestive enzyme systems and differ fromadults in other aspects of gastrointestinalfunction. Additional research is required todetermine whether factors such as differences inabsorption rate do, in fact, constitute a problemin the application of data derived from studies innormal adult subjects. Sensitive legal and ethicalissues make research on drug absorption in younghumans difficult, but we believe that, wherepossible, attempts should be made to obtain usefulbioavailability information for populations otherthan healthy adults.

IN VITRO CORRELATIONSIt is not practical (or desirable) that all batchesof drugs be tested in man to assure adequate andpredictable bioavailability. Accordingly, in vitromethods that are predictive of bioavailability mustbe developed to confirm the bioequivalence of(1) different batches of the same drug product madeby the same manufacturer, (2) drug products made bydifferent manufacturers and (3) drug productsin storage for a relatively long time.

Several investigations have indicated that intests (for example, dissolution tests) may bewith bioavailability. However, such data arelacking for most drug products. In addition,is needed on the development of animal modelscan be used for predicting bioavailability in

held

vitrocorrelated

researchthatman.

4 5

The development of such predictive systems wouldgreatly simplify the establishment and monitoringof standards insuring adequate and uniform levelsof bioavailability. -

RELATIONSHIP BETWEEN PLASMA LEVELS AND THERAPEUTIC EFFECTSDecisions about which drugs do indeed require directproof of bioequivalence in man will be based in parton knowledge regarding the relationship of plasmalevels of the drug to its therapeutic effect. Thistype of pharmacologic information in man is notwell developed for many drugs. Research on therelationships between plasma levels and efficacyand toxicity will facilitate rational regulatorydecisions.

RESOURCES FOR ADDITIONAL RESEARCHThe foregoing discussion has described several kindsof research needed to strengthen the assurance thatproducts will provide adequate bioavailability. Theparticular agency appropriate for sponsoring thisresearch will depend upon the kind of researchthat is initiated. For example, research directedtoward the improvement of analytical methodologyfor in vivo studies of drugs, or toward investigationsof the relationships between plasma levels ofdrugs and’ their therapeutic effect, is alreadyincluded in the program objectives of several ofthe National Institutes of Health, particularlythe National Institute of General Medical Sciences.These efforts should be encouraged and supportedwith clear recognition of their importance.

Research concerned with the improvement of methodologyfor in vitro tests related to the establishment ofnew compendial standards might be sponsored bythe proposed new standard-setting organization,discussed on pages 53 ff., or the FDA. Theseorganizations might also become increasinglyinvolved in sponsoring research concerned with theapplication of existing technology to the formula-tion, production and quality assurance for new drugproducts.

In order for research to be sponsored by either thenew organization or the FDA, substantial new resourceswill be required because support for such researchis almost completely lacking at the present time.

46

The FDA has been only minimally involved in researchactivities of these kinds and, of course, the newstandard-setting organization has yet to bedeveloped. Irrespective of the agency sponsoringthe research, the base of scientific staff will haveto be expanded in order to design proposals andmonitor research activities adequately.

47

8CURRENT LAW REQUIRING MANUFACTURERS TO MAINTAIN

RECORDS AND MAKE INFORMATION AVAILABLE TO THE

FDA IS AMBIGUOUS OR INADEQUATE AND SHOULD BE

CLARIFIED AND STRENGTHENED, IN PARTICULAR,

MANUFACTURERS SHOULD BE REQUIRED TO SUBMIT ALL

INFORMATION RELATING THE TESTS THEY CONDUCT TO

THE BIOAVAILABILITY DATA THEY DEVELOP IN ORDER

TO HELP PROVIDE INFORMATION ON THE FACTORS THAT

MODIFY THE BIOAVAILABILITY OF DRUG PRODUCTS,

THIS INFORMATION SHOULD BE AVAILABLE TO AID IN

THE ESTABLISHMENT OF COMPENDIA STANDARDS,

Under the Food, Drug, and Cosmetic Act, authorityis provided to the FDA to obtain proof of safetyand efficacy (and, in some instances, bioavailability)of certain drug products. Section 505 of the Actauthorizes the extensive collection of data on newdrug products with respect to both the safety andefficacy of the products as labeled. There iscontinued surveillance by the FDA of new drugproducts after they enter the marketplace.Manufacturers with New Drug Applications arerequired to establish records and make reportsconcerning clinical experience and other informationneeded for regulation by the FDA.

However, current law is ambiguous with respect toproviding the FDA with general authority to requiredrug manufacturers to maintain records and to make

48

reports demonstrating their compliance withcompendia standards for drug products. Further-more, it is not clear that the law provides authorityto require manufacturers to submit informationconcerning quality control procedures and manu-facturing processes.

It has been indicated in testimony presented beforethis Panel and elsewhere that manufacturers oftenperform tests of product quality in addition tothose set forth in compendia standards. Althoughthe FDA is informed of these tests, the frequencywith which manufacturers set standards that aremore stringent than official standards is unknownto us at this time, but it appears to vary amongdrug products and among manufacturers. Informationis also lacking about the specific kinds of testsconducted beyond those required in the officialstandards.

In addition, there is currently no generallegislative authority to require manufacturers tosubmit data, generated in their research anddevelopment activities, concerning formulation,bioavailability, or new procedures correlatingbioavailability with in vitro tests. In itsinvestigations of a new drug and the developmentof dosage forms, a pharmaceutical company mayconduct studies of bioavailability on a seriesof formulations and carry out related in vitrotests. The FDA does not have access to the resultsobtained with these research formulations.

Certain kinds of information currently made availableto the FDA by manufacturers are required undercurrent law to be held confidential and cannot beused in establishing either new compendia standardsor enforcement requirements. We believe thisprovision is too restrictive. The fact that suchinformation may not be made public does not meanthat relevant portions of it should not be used byappropriate experts to establish new standards orrequirements or to revise existing ones. Forexample, if a company develops information that canbe used to improve procedures or methods for testingdrug products, relevant portions of this informationshould be made available to appropriate experts toaid them in establishing new or revised compendiastandards or enforcement requirements. Thisobjective can be achieved even though the confi-dentiality of the specific proprietary informationis maintained.

49

We make the following recommendations:

The ambiguity in current law regardingrequirements for keeping records andreporting results of compendia testsshould be removed; such requirementsshould be specifically mandated.Results of batch-to-batch testing ortesting of single batches over timeshould be made available to the FDAas requested.

The FDA should be given the authorityto request (on a routine basis if itelects to do so) information describingthe methodology, specifications andresults of tests conducted on commercialproducts.

A standardized and simple format thatfacilitates clear understanding shouldbe designed for the presentation ofthe aforementioned data to the FDA.

Manufacturers should, upon the requestof the FDA, be required to submit anyinformation in their files thatrelates manufacturing processes andin vivo and in vitro tests to thebioavailability of any drug product.This requirement should apply totechnical information obtained duringresearch related to the developmentof new drug products, even when suchinformation bears only an indirectrelationship to the final dosage form.

When the FDA finds that particularin vitro or in vivo tests are useful—and deserve wider application, it .should be empowered to share theinformation with appropriate expertsand to incorporate those tests intothe compendia standards to be appliedto all manufacturers of that drugproduct.

51

EXEMPTIONS PROVIDED IN CURRENT LAW FOR SOME DRUG

PRODUCTS BASED ON THEIR YEAR OF INTRODUCTION IN

RELATION TO AMENDMENTS IN THE FOOD, DRUG, AND

C O S M E T I C A C T S O - C A L L E D G R A N D F A T H E R C L A U S E S )

HAVE IMPEDED IMPROVEMENT IN THE QUALITY OF THESE

PRODUCTS, SUCH EXEMPTIONS SHOULD BE ELIMINATED,

The Food, Drug, and Cosmetic Act provides authorityto establish standards and to regulate drugs anddrug products under the new drug provisions (Section505) and under the drug adulteration and misbrandingprovisions (Section 501 and 502). However, drugproducts that were marketed prior to 1938 are notcovered by the “new drug” provisions of the Act,but are subject to so-called grandfather provisions.Thus, unless the product or the labeling is changed,these products are exempt from regulations toinsure their quality. Similarly, for a drug productmarketed after 1938, but prior to 1962, regulatoryauthority is limited to a qualitative descriptionof the product’s composition in which a concentrationrange of the adjuvants, instead of an exact formu-lation, may be submitted.

We believe that these limitations have interferedwith the process of establishing standards to insurethe quality of drugs and drug products marketedbefore 1962, especially those marketed before 1938.Such exemptions provide an incentive for drugmanufacturers to maintain the status quo for thosedrug products rather than to make improvementsbased on more recent research or advances intechnology, since any change in formulation orlabeling may open the door to reclassification of

36.690 0 - 7 4 - 4

5 2

the product as a “new drug, ” subject to morerigorous regulatory standards.

Drug products available before 1962, and especiallythose available before 1938, were formulated underconditions of relatively limited technologycompared with that of today. As a result, theolder products, as indicated on pages 11-12, aremore likely to be subject to bioavailabilityproblems than products that became available after1962 when the New Drug Application and AbbreviatedNew Drug Application requirements were implemented.The most important example of such a problem withan older drug product arose with the therapeuticallyimportant and highly potent drug, digoxin. Aswe discussed on page 13, the bioavailability of abatch of this drug was found to be greatly reduced,resulting in a potentially dangerous situation forpatients using the drug. However, the only legalauthority which the FDA was able to use in actingto correct the situation was a provision regardinglabeling and formulation. Such ambiguous legalauthority may be inadequate to prevent delaysbrought about by court actions at a time whenimmediate action is required to protect theAmerican public.

We believe there should be a clear legal mandateto establish and apply regulatory standardsuniformly to all drug products, irrespective ofwhen they were introduced on the market. Westrongly recommend that the legal authority beprovided to establish standards for all drugs anddrug products, introduced both before and after1938, based on the best available technology inorder to insure that the drug products madeavailable to the American public are of the highestquality possible.

53

10A S INGLE ORGANIZATION CAPABLE OF SETT ING STANDARDS

ADEQUATE TO INSURE

B I O A V A I L A B I L I T Y O F

ESTABLISHED TO REPL

THE QUALITY AND UNIFORM

DRUG PRODUCTS SHOULD BE

ACE THE PRESENT USP AND NF

AS THE OFF IC IAL STANDARD-SETT ING ORGANIZATION

OF THE FEDERAL GOVERNMENT,

To carry out the task of establishing drugstandards adequate to insure the quality anduniform bioavailability of drug products, westrongly recommend that a single organization beestablished to supersede the USP and NF as theofficial standard-setting organization of theFederal Government. We do not believe thatcontinuation of the current organizations with anadmonishment to “do better” would be adequate toinsure the level of quality of drug products thatthe public deserves. We believe there are toomany weaknesses in the structure of the presentorganizations for them to be able to do the jobadequately.

The basic function of the new standard-settingorganization should be to establish and revise drugand drug product standards continuously on the basisof the best available technology in order to insurethat drug products meeting these standards wouldhave the highest quality and uniformity.

The establishment of improved standards for drugsand drug products will require an expanded researchprogram as discussed on page 43 ff. This researchshould be conducted through both grants andcontracts with outside organizations as well as by

54

the scientific staff of the new standard-settingbody. To conduct the research, the new organizationmust have an adequate staff of scientists activelyengaged in research. This staff must be capableof evaluating research conducted by outside groupsand developing proposals for additional research.

Although the new standard-setting organizationshould have a sizable scientific staff, we do notbelieve that it would be efficient to try toencompass within this staff all the expertisenecessary for setting all drug standards. Instead,a series of expert panels should be assembled toestablish standards for drugs and drug products inthe various classes. Individuals who serve onthese panels should have expertise in pharmaceuticaland analytical chemistry, statistics and otherdisciplines that are related to analytical method-ology, materials to be used in the manufacture ofdrug products, techniques of testing, etc. Asingle panel should not be expected to have theknowledge required to deal with all drugs and drugproducts; nor should a single group be expected todo the extensive amount of work required toestablish standards for all drugs.

In addition to making decisions regarding theestablishment of standards, the expert panelsshould also review the results of relevant researchconducted by outside groups as well as by staffscientists and should recommend additional researchto be conducted where necessary. As discussed onpage 47, the panels should also have access toappropriate information from industry for use inestablishing standards.

Members of these panels should be objective expertswho are not in a position to derive economic orpolitical benefit from the manufacture of drugproducts. They should be appropriately paid fortheir efforts; no longer should drug standards beset by those few individuals who can afford tovolunteer significant amounts of their time toperform this service for the public.

Although scientists from the pharmaceuticalindustry should not be allowed to exert a controllinginfluence on the standard-setting organization,scientific input from industry must be encouraged.One possible form of participation of industrialscientists might be as members of technicaladvisory committees reporting to the panels ofexperts. Individuals from industry who participateon these technical advisory committees should be

55

selected for their technical capabilities ratherthan for the managerial positions they may hold.

Cooperation and coordination of the activities ofthe new standard-setting body with those of severalother groups is essential to accomplish theobjectives set forth in this report. The pharma-ceutical industry will continue to conduct a majorshare of the research related to the developmentof new testing methods and technology. The industrywill also be responsible for applying the newstandards in their manufacturing processes. Inaddition, health professionals, scientists, andothers responsible for prescribing and dispensingdrugs or performing the research needed to meet therequirements for improved drug standards and fordata on bioavailability of drugs will also beaffected by implementation of our recommendations.The functions of the new standard-settingorganization must be closely coordinated with thoseof the FDA in order to insure a meaningful andproductive interchange of information between thosewho establish drug standards and those who monitorfor compliance with those standards. It should bethe responsibility of the new standard-settingOrganization to insure coordination of itsactivities with those of the regulatory and enforce-ment groups.

We do not recommend a specific organizationallocation for the new standard-setting body, althoughwe strongly recommend that it be organized andlocated in such manner as to allow reasonablefreedom from inappropriate industrial or politicalinfluence. Drug standards should be concerned withassurance of the highest quality drug products forthe American people rather than with economic orpolitical considerations. We do not believe thatthe standard-setting organization and the compliance-monitoring organization should be the same; butthere should be a healthy, cooperative relationshipbetween these groups.

Several options are readily apparent for thelocation of the new organization: within the FDAbut separate from the compliance-monitoring group,within HEW but separate from FDA, or outside ofthe Federal Government as an independent organization.The possibility that the USP and the NF could mergeand make sufficient changes in their structures andfunctions to fulfill the criteria for an effectivestandard-setting organization is not precluded, butthe changes necessary would be extensive.

56

Wherever the new organization is located adminis-tratively, it will require federal fundingcommensurate with its responsibilities for settingthe standards that would be employed as a basisfor federal regulation. If the organization isplaced within the structure of HEW, direct federalfunding will be straightforward. If it is esta-blished as an organization independent of thefederal structure, it could still receive directfederal funding, or it could be funded through agrant or contract from HEW. In any case, we stronglyrecommend that sufficient federal funds be suppliedto insure that the objectives are achieved.

57

1 1

A S Y S T E M S H O U L D BE ORGANIZED AS RAPIDLY AS POSSIBLE

TO G E N E R A T E AN OFF IC IAL L IST OF INTERCHANGEABLE

D R U G P R O D U C T S, I N T H E D E V E L O P M E N T O F T H E L I S T,

D IST INCTIONS SHOULD BE MADE BETWEEN TWO CLASSES

DRUGS AND DRUG PRODUCTS:

OF

1 . T H O S E F O R W H I C H E V I D E N C E O F B I O E Q U I V A L E N C E

IS NOT CONSIDERED ESSENTIAL AND THAT COULD BE

ADDED TO THE L IST AS SOON AS STANDARDS OF

PHARMACEUTICAL EQUIVALENCE HAVE BEEN

E S T A B L I S H E D A N D S A T I S F I E D ,

2, THOSE FOR WHICH EVIDENCE OF BIOEQUIVALENCE

I S C R I T I C A L , SUCH PRODUCTS SHOULD BE LISTED

ONLY AFTER THEY HAVE BEEN SHOWN TO BE

BIOEQUIVALENT OR HAVE SATISF IED STANDARDS OF

PHARMACEUTICAL EQUIVALENCE THAT HAVE BEEN

SHOWN TO INSURE BIOEQUIVALENCEI

The question that led to the genesis of this studyand report addressed the economics of drug purchase;and, although we have not given consideration toother possibly significant economic issues, ourcharge does relate to whether drug products are orcan be made sufficiently interchangeable so thatprice can be a major factor in their selection.

5 8

It is clear from the conclusions we have alreadvstated that we do not believe that all chemicalequivalents are, at present, interchangeable. Wedo believe, however, that the goal of interchange-ability is achievable within most, if not all,classes or oral druq products and that a systemshould be organized as rapidly as possible toestablish the conditions that will permit a l i s t i n gof interchangeable products.

Establishment of such a listing should begin with aseries of judgments regarding classes of drugs andindividual drug entities that will yield a divisionof drugs into two groups: those for which evidenceof bioequivalence is considered essential and thosefor which such evidence will not be required. Thegeneral criteria for such decisions have been setforth on pages 21 ff.

Upon classification of a drug as one requiringevidence of bioequivalence, products of that drugshould be included on the list of interchangeable drugproducts only after data supporting their bioequivalencehave been provided and approved. Furthermore, beyondthe question of bioequivalence among interchangeabledrug products in this category, methods to insurepharmaceutical equivalence between batches of thesame drug product deserve special attention.

It is apparent that many drug products aretherapeutically equivalent even though they mayvary somewhat in bioavailability. Included in thiscategory, for example, are many groups of drugs thatare customarily given in doses that insureconcentrations in the blood well in excess of theminimum effective concentration. These drugs havea wide margin of safety between effective and toxicconcentrations. Moderate variation within this widemargin can be considered to have little or notherapeutic significance.

It is our opinion, however, that even within thesegroups of drug products, for which evidence ofbioequivalence can be waived, there is room forconsiderable improvement of standards for thecontrol of manufacturing processes and the testingof products. Such-improvements as we haverecommended will have the effect of minimizingmany important sources of variability amongchemically equivalent products. When improvedstandards of pharmaceutical equivalence have beenestablished, drug products meeting the standards andfalling into categories for which evidence ofequivalent bioavailability is not essential can be

59

considered to be interchangeable and listed as suchfor the guidance of purchasers, prescribingphysicians and dispensing pharmacists.

The official list of interchangeable drug productsneed not be permanently limited to those for whichstudies of bioavailability are now available orconsidered unnecessary. As compendia standardsare developed specifying in vitro tests whose

‘results correlate with bioavailability, it shouldbe possible, in most instances, to use the in vitrotests as a basis for providing ongoing evidence ofbioavailability. Products meeting these conditionscan and should then be added to the list of inter-changeable drug products. The ultimate goal shouldbe that virtually all marketed drug products shouldfulfill the criteria necessary for inclusion in thislist.

It should be recognized, however, that there mayremain a few drugs-- those that have low therapeuticindices and, possibly, unfavorable physicalproperties, but that are critically important in thetreatment of serious illnesses --whose productsshould be considered noninterchangeable even thoughthey meet these conditions, at least until experiencehas shown that improved in vitro testing or testingin animal model systems can be relied upon for theprecise prediction of their bioavailability.

Many decisions requiring well-informed judgmentswill be necessary in establishing a list of inter-changeable drug products and in the continuingoperation of the system. The essential judgmentsshould be based upon the highest level of expertisein many specific areas of medicine, clinicalpharmacology and associated scientific disciplines.The wide range of the necessary fields of specializa-tion and the extensive experience and depth ofknowledge that will be required are not to befound within the staff of any single organization.Further, we do not believe that it would be efficientfor the FDA to try to encompass all the necessaryexpertise within its own ranks. Rather, it will beessential to call upon the most knowledgeableindividuals in each field, wherever they may befound, to provide the advice needed to arrive atappropriate decisions.

To accomplish this goal, a series of groups advisoryto the FDA should be established, each to deal withspecific areas in which decisions must be made. Itis not to be expected that a single advisory groupwould be able to deal effectively with the full range

60

of responsibilities involved. For example, thosewho are best equipped to deal with drugs affectingthe cardiovascular system are not likely to be thesame individuals best able to offer the most usefuladvice about anti-infective agents. It is our beliefthat an appropriate range of advisory groupswill not only make certain that decisions are reachedon the basis of the best informed judgments but thatthe acceptance of these decisions by all interestedparties will also be immeasurably enhanced.

The major expertise required to establish a list ofinterchangeable drug products will be provided byadvisory groups rather than staff personnel. However,capable scientific staff will be essential tosupport the work of the advisory groups. Currentstaffing and funding levels are not adequate forthe FDA to meet the significant new responsibilitiesproposed in this recommendation. Consequently,additional financial and staffing support will berequired to develop and maintain the list ofinterchangeable drug products and to coordinate theseefforts with the agencies involved in settingstandards and supporting research. Such resourcesshould be made available as rapidly as possible.