bioequivalence studies: relevance for veterinary medicine

5

Braz. J. vet. Res. anim. Sci., São Paulo, v. 45, suplemento, p. 5-19, 2008

Bioequivalence studies: relevance for veterinary

medicine

1 - Applied Pharmacology and Toxicology Laboratory, Department ofPathology, School of Veterinary Medicine, University of São Paulo, SãoPaulo - SP, Brazil2 - Molecular Pharmacology Laboratory, Department of PharmaceuticalSciences, School of Health Sciences, University of Brasília, Brasília - DF,Brazil

João PALERMO-NETO1

Dario Abbud RIGHI2

Corresponding author:Prof. João Palermo-Neto, Faculdade deMedicina Veterinária e Zootecnia,Laboratório de Farmacologia Aplicada eToxicologia, Av. Prof. Dr. Orlando Marquesde Paiva, n° 87, 05508-900 CidadeUniversitária, São Paulo, SP, Brasil, Telefone:+55 11 30917957 e/ou Fax: +55 1130917829, [email protected]

Received: 03/06/2008Approved: 25/09/2008

Abstract

Bioequivalence (BE) studies are scientific methods that allowcomparison of different medicinal products containing the same activesubstance, or different batches of the same medicinal products or, ina broad sense, different routes of administration of the same product.Actually, legislation on generic drugs and bioequivalence only exist inBrazil for drugs intended for human purposes. In the field ofVeterinary Medicine, BE is being used in many countries as part ofthe necessary requirements for registration of animal health products,i.e., to provide efficacy and safety animal data and to allow consumerssafety; indeed, they also assure the quality of the food derived fromtreated animals. The present manuscript was designed to review anddiscuss BE; for that, it was divided into three major parts: 1-understanding bioequivalence: importance of BE studies for animaland human health; 2- type of BE studies included; 3- generalconsideration on experimental design involved o BE studies.

Key words:Pharmacokinetic.Pharmacodynamic.Bioavailability.Bioequivalency.Reference drugs.Generic drugs.

Introduction

Bioequivalence (BE) studies arescientific methods designed to comparedifferent health products containing the sameactive constituent or different batches of thesame veterinary medical products, based ontheir formulation, pharmacokinetic andpharmacodynamic properties and residualprofiles in the tissues of treated animals.1,2,3

BE studies are used in a variety of situations,most often when a sponsor proposesmanufacturing a generic version of anapproved off-patent product. A BE studymay also be part of a new animal drugapplication (NADA) or supplementalNADA for approval of an alternativedosage form, new route of administration,or a significant manufacturing change whichmay affect drug bioavailability. In veterinarymedicine, BE studies are necessary andrelevant because they allow the establishment

of the necessary requirements for registrationof generic animal health products that ensureanimal and food safety for human being.

Actually, legislation on generic drugsin Brazil exists only for those to be used inhumans; as a consequence, specific rules forbioequivalence (BE) studies on veterinaryproducts are still lacking. In this regard,legislation on both generic drugs and BE inveterinary medicine are being employed andmight be found in many countries.1,2,3

Consideration on such maters seemsappropriate and necessary; indeed, it maybecome a useful instrument for governmentalauthorities, pharmaceutical industries andveterinary professionals’ managements in thisfield. The present manuscript was themdesigned, to discuss BE; it was divided intothree major parts: 1- understandingbioequivalence: importance of BE studiesfor animal and human health; 2- type of BEstudies included; 3- general consideration on

6


experimental design involved o BE studies.

Understanding BE: importancefor animal and human health

Veterinary pharmaceutical product isa finished dosage form that contains thepharmacologically active substance(s) withor without excipient(s). Changes in activesubstance purity or in excipients in aveterinary medical product as well as in themanufacturing process may have significanteffects on bioavailability.

It is scientifically valid to assume that,if an active substance or therapeutic moietyof a test animal health product reaches thesystemic circulation with the same rate andextent as the active substance or therapeuticmoiety of a reference animal health product,the local availability (concentration in tissue)of the active substance or therapeutic moietywill be similar for the test and referenceproducts. The similarity of availability at thesite(s) of action is the basis of concludingtherapeutic equivalence of the products.However and importantly, it is usuallyaccepted that tissue residue depletion of thegeneric product is not adequately addressedthrough BE studies. For that, sponsors ofdrug products for food-producing animalsare generally requested to include BE andtissue residue studies in their data package.2,3

The in vivo BE of an animal healthproduct is demonstrated if the rate andextent of absorption, as determined bycomparison of measured parametersderived from relevant data (e.g.concentration of the active substance in theblood or pharmacological/clinical effects)do not indicate a biologically relevantdifference in the rate and extent ofabsorption from the reference product. Inother words, two products are consideredto be bioequivalent when they are equallybioavailable; that is, equal in the rate andextent to which the active ingredient(s) ortherapeutic ingredient(s) is (are) absorbed andbecome(s) available at the site(s) of drugaction.1,2,3

As a general rule, the proposedgeneric product should be tested for BE

against the original pioneer product.However, if the original pioneer product isno longer marketed, but remains eligible tobe copied, then the first approved andavailable generic copy of the pioneer shouldbe used as the reference product for BEtesting against the proposed new genericproduct.2 If several approved NADA’s existfor the same drug product, and eachapproved product is labeled differently (i.e.,different species and/or claims), then thegeneric sponsor is asked to clearly identifywhich product label is the intended pioneer2.BE testing should be conduct against thesingle approved product which bears thelabeling that the generic sponsor intends tocopy.

Thus, a generic animal health drugmay be defined as a product that containthe same active ingredient, in the same doserange, in the same pharmaceuticalformulation and that is administered to thesame animal species, for the same route andfor the same therapeutic/prophylactic/feedaddictive purposes. For that, they present thesame efficacy, same security for animal andhuman health and, the same quality of thereference standard, as statically proven by BEand by tissue depletion residue studies.

According to the Australian Pesticides& Veterinary Medicines Authority -APVMA1- a proposed product is consideredto be generic to a reference product if: 1-ingredients are of equivalent pharmaceuticalcompendial standard, and 2- the activeconstituents are the same substances and arewithin ± 5% of that in the reference productor the dose range to the animal is within ±5% of the reference product, and 3- the non-active constituents are the same substancesand are within ± 5% of that in the referenceproduct, except for those constituentsdefined in the Australian Guidelines onMinor Formulation Changes, where they canvary more than ± 5%, and 4- the product isin the same dosage form, and finally 5- theproduct has the same physico-chemicalproperties as the reference product (includingpH, particle size, crystal form, and dissolutionprofile where applicable).

7


Type of BE studies includedBE can be demonstrated in vivo or,

under specific conditions, in vitro. Figure 1depicts the more relevant studies related toBE determination. The best approachwould be that one that uses all methodsavailable to display BE, i.e., that employsboth in vitro and in vivo methods. Howeverand in particular terms, the blood level studyis generically preferred above all others beingconsidered, since it is considered the mostsensitive measure of BE. For that, thesponsor should provide justification forchoosing either a pharmacologic or clinicalend-point study over a blood-level (or otherbiological fluids or tissues) study. It shouldbe noted, however, that tissue residuedepletion studies were also included in thisfigure; this was done because, as alreadystated, it is usually accepted that tissue residuedepletion of a generic product is notadequately addressed through BE studies.

In vivo BE studies must be conductedusing the most accurate, sensitive andreproducible approach. It may bedetermined by one of several direct orindirect methods. Selection of the methoddepends upon the purpose of the study, the

analytical method available, and the natureof the drug product. BE testing should beconducted using the most appropriatemethod available for the specific use of theproduct. A classification of such approachesproposed by the EMEA Committee forVeterinary Medical Products in descendingorder of accuracy, sensitivity andreproducibility2 is listed bellow:

a) In vivo testing in target species inwhich the concentration of the activesubstance or therapeutic moiety or itsrepresentative metabolite(s) in blood, plasma,serum or if justified, other biological fluidor appropriate tissues is measured as afunction of time. Reliance on in vivo BE datarelies upon an assumption that the measuredconcentrations of active substance havemeaning with respect to the objective of thetrial and the intended label claim. It alsonecessitates that adequate drug concentrationsare achieved to allow for the determinationof product concentration vs time profile inthe blood or other biological fluid.

b) In vivo testing in target species in whichan appropriate acute pharmacological/clinicaleffect of the active substance or therapeuticmoiety or its metabolite(s) is measured as a

Figure 1 - Studies implicated in bioequivalence determination. Note that equivalence between a generic and apioneer product should be obtained using in vitro pharmaceutical equivalence and/or in vivo tests suchas the blood level, the pharmacological and clinical end-points studies. Note also that residue depletionstudies are also necessary

8


function of time. This approach can be usedwhen analytical methods are not available.Its use requires demonstration of dose-related response. For veterinary medicalproducts intended for local effect,pharmacological/clinical end-points can bethe most relevant approach for thedemonstration of BE.

According to FDA BE Guidance forIndustry3, when absorption of the drug issufficient to measure drug concentrationdirectly in the blood (or other appropriatebiological fluid or tissues) and systemicabsorption is relevant to the drug action, thena blood (or other biological fluid or tissue)level BE study should be conducted.However, when the measurement of the rateand extent of absorption of the drug inbiological fluids can not be achieved or isunrelated to drug action, a pharmacologicend-point (i.e., drug induced physiologicchange which is related to the approvedindications of use) study may be conducted.Lastly, and in order of preference, if drugconcentration in blood (or fluids or tissues)are not measurable or are inappropriate, andthere are no appropriate pharmacologiceffects that can be monitored, then a clinicalend-point study may be conducted,comparing the test (generic) product to thereference (pioneer) product and a placebo(or negative) control.

Finally, In vitro equivalence studiescould support in vivo BE studies in thefollowing case2:

a) In vivo BE has been demonstratedfor the highest dosage strength and in vitrodissolution data is used to support the BEof the lower dosage strengths for that genericformulation. In these cases, the use of in vitromethods requires that the followingconditions are all met: 1- the dosage strengthsdiffer only by the mass of the activesubstance; 2- the drug is known to beassociated with linear pharmacokinetics; 3-the composition of all formulations arequalitatively identical; 4- the ratios betweenactive substances of the test- and referenceproducts are identical.

b) In vitro comparability might be

adequate to confirm the comparability ofthe reference product and its generic productto be administered orally. This appliesparticularly to immediate release oral dosageforms that are rapidly dissolving and containdrug substances that are both highly solubleand high permeable.

c) There is a very minor formulationchange to an approved product (or a minorpre-approval change to a product that hasundergone extensive clinical trials), and hasbeen determined that the change requiresonly in vitro confirmation of comparabilityto the formulation that underwent theclinical trials.

d) Ensuring batch to batch consistencywithin a product.

It should be noted, however, that theAustralian Guidelines on BE1 states that forcertain dosage forms (intravenous solutions,oral solubilized forms, topical dose forms,non-absorbed antacids and radio opaquemedia and oral tablets), an applicant maywish to address BE by providing in vitro datademonstrating pharmaceutical equivalence ofa ´product to a reference product. In thiscase, such data should include: 1- nature ofdosage form; 2- solubility of activeconstituent(s) in water; 3- relevantpharmaceutical characteristics includingparticle size, crystal form, and dissolutionprofile where applicable; 4- rate limiting stepsin absorption of the active constituent(s) e.g.disintegration, dissolution, gut absorptionwhere applicable or in access to the site ofeffect; and 5- relevant scientific argumentregarding clinical consequences ofinequivalence.

General consideration onexperimental design involved on BEdetermination

As depicted in Figure 1, BE can bedemonstrated in vivo or, under specificconditions, in vitro.

In vivo BE studiesBE may be demonstrated in vivo by

one of several direct or indirect methods. Indescending order of sensitivity: blood levels

9


studies, pharmacological end-point studiesand clinical end-point studies. As statedbefore, the selection of the method is linkedto the purpose of the study, the analyticalmethod available and, among many othervariables, the nature of the product analyzed.

Basic principlesIn vivo BE studies usually involve an

experimental part conducted in the field,where the active substance or therapeuticmoiety are given to the target animal species,an analytical part where the active substance(s)or its representative metabolite(s) areanalyzed in the blood, plasma, serum orother biological fluids or appropriate tissuesand a statistical part. Figure 2 depicts theseparts, the main steps they involve and thequality compliance test protocols that shouldbe employed in each of them.

Veterinary Good Clinical Practice(GCP) should be used in all blood level,pharmacological end-point, clinical end-point and residue depletion studies if theproducts intendment is food-producinganimals use. The global requirementsreferred to as VICH-GCP guidelines4 areusually employed in the field experiments,because they were specially prepared toensure the accuracy, integrity and correctnessof the obtained data. VICH is a tripartite(EU, Japan and USA) programme aimed atharmonizing technical requirements for

veterinary product registration. Its full title isthe International Cooperation onHarmonization of Technical Requirementsfor Registration of Veterinary MedicalProducts.

Some relevant extracts from theVICH- GCP guidelines4 are: 1- theinvestigator responsible for the study shouldhave sufficient knowledge, scientific trainingand experience, as evidenced by a currentcurriculum vitae and other credentials, toconduct clinical studies to investigate theeffectiveness and in-use safety ofinvestigational veterinary products in thetarget species; 2- the investigator shoulddelegate authority and work only toindividuals qualified by training andexperience to perform the assigned duties;3- the field experiment should comply withthe study protocol, being conductedaccording to GCP and applicable regulatoryrequirements; 4- the equipment and facilitiesused to conduct the study should beadequate and well-maintained; 5- thehousing, feeding, and care of all animals atthe study site should be carefully supervised;6- all data should be collected and recorded,including unanticipated observations, inaccordance with the study protocol andapplicable regulatory requirements in anunbiased manner that accurately andcompletely reflects the observations of thestudy; 7- the investigator should ensure that

Figure 2 - The three different steps of bioequivalence studies. Note the requirements on quality compliance appliedto each of them

10


all specimens required to be retained by thestudy protocol and any applicable regulatoryrequirements are identified in a manner thatis complete, accurate, legible and precludesloss of identification from the specimen; 8-the special analysis and/or tests to beperformed including the time of samplingand interval between sampling, storage ofsamples, and the analysis of testing shouldbe carefully described; and, among others,9- all study documentation should be securelystored for the period of time required bythe relevant regulatory authorities, beingprotected from deterioration.

To ensure VICH-GCP, selectedanimals for BE studies must be from thetarget population for which the product isintended. Ordinarily, BE studies should beconducted with health animals representativeof the species, class, gender, breed, age,weight, hormonal, nutritional andphysiological maturity for which the drugwas approved. Where possible, it is advisableto restrict the studies to the gender for whichthe pioneer product is approved, unless thereis known interaction of formulation withgender. An attempt should be made torestrict the weight of the test animals to anarrow range and animals should not receiveany medication prior to testing for a periodof two weeks or more, depending upon thebiological half-life of the ancillary drug.

Feeding may either enhance orinterfere with drug absorption, dependingupon the characteristics of the drug and theformulation. Feeding may also increase theinter- and intrasubject variability in the rateand extent of drug absorption. The rationalefor conducting each bioequivalence studyunder fasting or fed conditions should beprovided in the protocol. In this respect, theprotocol should describe the diet and feedingregime which will be used in the study.However, and importantly, if a pioneerproduct label indicates that the product islimited to administration either in the fed orfasted state, then the BE study should beconducted accordingly.

Concerning the analytical part of theBE studies, it should be conducted and

reported in conformity with the internationalprinciples of GLP and quality compliancemethods. Specifically, it should be conductedin accordance with requirements stated in the32nd, 47th and 52nd meetings of the JointFAO/WHO Expert Committee on FoodAdditives (JECFA)5 or with the EuropeanCommunity legislation reported in Directives87/18/EEC and 88/320/EEC. The FDAconsiderations included on Assay methodsspecified in the BE Guidance to Industrydocument3 and the VICH guideline GL1Validation of analytical procedures (CVMP/VIVH/97/076) might also be useful.

In this respect, it’s undisputable that aproperly validated assay method is pivotalto the acceptability of any pharmacokineticstudy. For that, the submission should containadequate information necessary for thereviewer to determine the validity of theanalytical method used to quantitate the levelof the drug under study in the biologicalmatrix (e.g., blood or tissues). Thus, thefollowing aspects should be addressed inassessing method performance: 1-concentration range and linearity, defined byat least five concentrations; 2- limit ofdetection (LOD), estimated as the responsevalue calculated by adding three times thestandard deviation of the backgroundresponse to the average backgroundresponse; 3- limit of quantification (LOQ),at least ten times the standard deviation ofthe background response to the averagebackground response; 4- specificity,demonstrated by the analysis of sixindependent sources of control matrix; 5-accuracy (recovery), based upon the meanvalue of six replicate injections, at three doselevels (high, mid-range and LOQ); 6-precision, evaluated using at least threeknown concentration of analyte freshlyspiked in control matrix; 7- analyte stability,determined with incurred analyte in thematrix of dosed animals in addition tocontrol matrix spiked with pure analyte; 8-analytical system stability, to assure that thesystem used remains stable over the timecourse of the assay; 9- quality control (QC)samples, to assure that the complete analytical

11


method, sample, preparation, extraction,clean-up, and instrumental analysis wasperformed according to acceptable criteria;10- replicate and repeat analysis, determinedprior to running the study and recorded inthe method protocol.

Finally, Good Statistical Practice (GSP)is necessary to analyze the obtained data andto allow their interpretation.6

Blood level studiesThese studies are the most sensitive

measures of BE. They are performed whenabsorption of the drug is sufficient tomeasure drug concentration directly in theblood and systemic absorption is relevantto the drug action.1,2,3 These studies comparea test (generic) product to a reference(pioneer) product using parameters derivedfrom the concentrations of the drug moietyand/or its metabolites, as a function of time,in whole blood, plasma, serum (or in otherappropriate biological fluids or tissues).Generally, the study should encompass theabsorption, distribution, and depletion(elimination) phases of the drugconcentrations vs time profiles.

The potency of the pioneer andgeneric products should be assayed prior toconducting the BE study to ensurecompendial specifications are met. FDAauthorities recommend that the potency ofthe pioneer and generic lots should differby no more than ± 5% for dosageproducts3. The animals should be dosedaccording to the labeled concentration orstrength of the product, rather than theassayed potency of the individual batch (i.e.,the dose should not be corrected for theassayed potency of the product). The BEdata or derived parameters should not benormalized to account for any potencydifferences between the pioneer andgeneric lots.

A single dose study at the highestapproved dose will generally be adequatefor the demonstration of BE. However, amultiple dose study may be appropriate whenthere are concerns regarding poorlypredictable drug accumulation, (e.g., drug

with nonlinear kinetics) or a drug with anarrow therapeutic window. A multiple dosestudy may also be needed when assaysensitivity is inadequate to permit drugquantification out to three terminalelimination half-lives beyond the time whenmaximum blood concentrations (CMAX)are achieved, or in cases where prolongedor delayed absorption exist.

Figure 3 exemplifies a time vs. plasmaconcentration curve. As depicted, the pivotalparameters for BE estimation on blood levelstudies are: CMAX, the maximalconcentration or peak concentration achievedby the drugs under test in plasma; TMAXthe corresponding time to reach CMAX and,area under the time vs. concentration curveAUC(0-t).

As presented in figure 3, the rate ofabsorption will be estimated by themaximum observed drug concentration(CMAX) and the corresponding time toreach this maximum concentration (TMAX).By another way, the extent of productbioavailability is estimated by the area underthe blood or plasma concentration vs. timecurve (AUC). AUC is most frequentlyestimated using the linear trapezoid rule.Other methods for AUC estimation may beproposed in the study protocol and shouldbe accompanied by appropriate literaturereferences. In the case a multiple dose BEstudy is used, the AUC should be calculatedover one complete dosing interval AUC(0-t).

For a single dose BE study, AUCshould be calculated from time 0 (predose)to the last sampling time associated withquantifiable drug concentration AUC(0- LOQ).The comparison of the test and referenceproduct value for this noninfinitive estimateprovides the closest approximation of themeasure of uncertainty (variance) and therelative bioavailability estimate associated withAUC(0- inf), the full extent of productbioavailability7. However, because of thepossibility of multifunctional absorptionkinetics, AUC can not always be determinedwhen the available drug has been completelyabsorbed. Therefore, it is generally acceptedthat the duration of sampling should be

12


extended until such a time that AUC(0-LOQ)/AUC(0-inf) > = 0.80. For that, the samplingtimes should extend to at least three multipleof the drug’s apparent elimination half-life,beyond the time when CMAX are reached.Further detail on this subject might be foundelsewhere.8,9

Other relevant pharmacokineticmeasures that should also be part of BEstudies (but preclude statistical comparisonbetween the pioneer and the drug under test)are: apparent volume of distribution (Vd),elimination half-life (t1/2) and depuration orclearance (D). Vd is a measure that links thequantity of drug within the organism (thedose administered) and its concentration inthe plasma (C). The elimination half-life (t1/

2) of a drug is the time necessary to reach50% of CMAX. Finally, the capacity of anorganism to eliminate a drug is measured byits clearance that is obtained multiplying theapparent volume of distribution (Vd) by thedrug’s constant of elimination (Ke). Thefollowing formulas might be used tocalculate those values.10

In general, product BE should bebased upon total (free plus protein bound)concentrations of the parent drug (ormetabolite, when applicable). However, ifnonlinear protein binding is known to occurwithin the therapeutic dosing range (asdetermined from literature or pilot data),then both the free and total drugconcentrations for the generic and pioneerproducts should be included. Similarly, if thedrug is known to enter blood erythrocytes,the protocol should address the issue ofpotential nonlinearity in erythrocyte uptakeof the drug administered within the labeledtherapeutic dosing range.3

The total number of sampling timesnecessary to characterize the blood levelprofiles will depend upon the curvature ofthe profiles and the magnitude of variabilityassociated with the bioavailability data(including pharmacokinetic variability, assayerror and interproduct differences inabsorption kinetics). Anyway, the sampling

Figure 3 - Pivotal parameters analyzed in a Blood Level Study of Bioequivalence. CMax is the maximal concentrationor peak concentration in the blood, TMax the corresponding time to reach CMax and AUC(0-t) the areaunder the time vs. concentration curve

13


times should adequately define peakconcentration(s) and the extent ofabsorption. As stated, it is generally assumedthat the sampling times should extend at leastthree terminal elimination half-lives beyondTMAX, as stated above.

Usually, maximum sampling timeefficiency is achieved by conducting a pilotinvestigation. In this case, the pilot studyshould identify the general shapes of the testand reference curves, the magnitude of thedifference in product profiles, and the noiseassociated with each blood sampling time(e.g. variability attributable to assay error andthe variability between subjects, for parallelstudy designs, or within subjects, forcrossover designs). This information shouldbe applied to the determination of anoptimum blood sampling schedule. Furtherinformation on this subject might be foundelsewhere.11

Pilot studies are also recommendedas a means of estimating the appropriatenumber of subjects to be used, i.e., thesample size for pivotal BE study.Furthermore, this number depends onseveral factors including variance of theresponse, differences in the two formulationsand level of the rejection of the hypothesis.Estimated sample size will vary dependingupon whether the data are analyzed on a logor linear scale. Useful references on samplesize estimates might be found elsewhere.6,12

A two-period cross-over design iscommonly used in blood level studies. The useof cross-over designs eliminates a majorsource of study variability: between subjectdifferences in the rates of drug absorption,drug clearance, and the volume of drugdistribution. Crossover design has advantagesin terms of power and number of animalsneeded. In a typical two period cross-overdesign, subjects are randomly assigned to eithersequence A or sequence B with the restrictionthat equal numbers of subjects are initiallyassigned to each sequence. The design shouldbe as follows:

However, a one-period parallel designmay be preferable in the followingconditions: 1- the drug induces physiologicalchanges in the animal (e.g. liver microssomalenzyme induction); 2- the drug has a verylong terminal elimination half-life; 3- theduration of the washout time for the two-crossover study is too long; and, 4- the drugfollow a delayed or prolonged absorption.Anyway, when using a cross–over study, thewashout period should be sufficiently longto allow the second period of the study tobe applicable in the statistical analysis.

Concerning the statistical analysis, theuse of 90% confidence intervals is usuallytaken as the best available method forevaluating BE study.1,2,3,10,13 Thus, twoproducts are considered to be equivalentwhen the 90% confidence intervals foundfor the reasons AUCGeneric/AUCPioneer andCMAXGeneric/CMXPioneer are within 80% and120% or when generic and pioneer data arenot different from each other at P<0.05.Further data on statistical analysis of bloodlevel BE studies might be foundelsewhere.6,13,14

Pharmacological end-point studiesWhere the direct measurement of the

rate and extent of absorption of the newanimal drug in biological fluids isinappropriate or impractical, the evaluationof a pharmacologic end-point related to thelabeled indications for use is normallyacceptable. These studies may be conductedanalyzing the physiological changes inducedby the test drug and the reference or pioneerproduct.2,3

Typically the design of a pharmacologicend-point study should follow the sanegeneral considerations as the blood levelstudies. However, specifics such as thenumber of subjects or sampling times willdepend on the pharmacologic end-pointmonitored. The parameters to me measuredwill also depend upon the pharmacologicend-points and may differ from those usedin blood level studies. As with blood levelstudies, when pharmacologic end-pointstudies are used to demonstrate BE, a tissue

14


residue study will also be required in food-producing animals.

For parameter which can be measuredover time, a time vs effect profile is generated,and equivalence is statistically determined witha method of analysis that is essentially thesame as for the blood level BE study. Thus,the use of 90% confidence intervals is usuallytaken as the best available method forevaluating the obtained data. Forpharmacologic effects for which effects vstime curves can not be generated, thenalternative procedures for statistical analysisshould be discussed and justified in theexperimental study protocol.

Clinical end-point studiesIf measurement of the drug or its

metabolites in blood, biological fluids ortissues is inappropriate or impractical, andthere are no appropriate pharmacologicalend-points to monitor (e.g., most productiondrugs and some coccidiostats andantihelmintics), then well-controlled clinicalend-point studies are acceptable for thedemonstration of BE.3 In this case, a parallelgroup design with three groups should beused. The groups should be a placebo (ornegative control), a positive control(reference/pioneer product) and the test(generic) product. The purpose of theplacebo (or negative control) is to confirmthe sensitivity or validity of the study.

Some clinical end-point studies,however, may not include a placebo (ornegative control) for ethical and/or practicalconsiderations. If a placebo is omitted, thenthe response(s) to the test and referenceproducts should each provide a statisticallysignificant improvement over baseline.

Clinical end-point studies should beconducted using the target animal species,with consideration for the sex, class, bodyweight, age, health status, and feeding andhusbandry conditions, as described on thepioneer product labeling. The dosage(s)approved for the pioneer product shouldbe used in this study. Finally, the length oftime that the study is conducted should alsobe consistent with the duration of use on

the pioneer product labeling. As with otherstudies, when clinical end-point studies areused to demonstrate BE, a tissue residuedepletion study will also be required in food-producing animals.

In general, the response(s) to bemeasured in a clinical end-point study shouldbe based upon the labeling claims of thepioneer product. In this respect, it may notbe necessary to collect data on someoverlapping claims (e.g., for a feed additivewhich is added as the same amount per tonof feed for both growth rate and feedefficiency, data from only one of the tworesponses need to be collected).

Statistical analysis should be used inclinical end-point studies, to compare the testproduct and the reference product.However, a traditional hypothesis test shouldbe performed before, comparing both thetest and the reference products separately tothe placebo (or negative control). Thehypothesis test is conducted to ensure thatthe study has adequate sensitivity to detectdifferences when they actually occur. If noimprovement (a = 0.05) is seen in parameter(i.e., the mean of the test and the mean ofthe reference products are each notsignificantly better than the mean of theplacebo (or negative control), the studyshould be considered inadequate toevaluate BE.

Assuming that the test and referenceproducts have been shown to be superiorto the placebo (or negative control), thedetermination of BE is based upon theconfidence interval of the differencebetween the two products. The use of 90%confidence intervals is usually taken as thebest available method for evaluating clinicalend-point data.

If the results are ordered categoricaldata (e.g., excellent, good, fair or poor), anon-parametric hypothesis test of nodifference between test product and placebo(negative control) and between the referenceproduct and the placebo (or negativecontrol) should be performed. As above, ifthese tests result in significant differencesbetween the test product and the negative

15


control and the reference product andnegative control, then a non-parametricconfidence interval on the differencebetween the test and the reference productsis calculated. Another acceptable approachfor categorical data is to calculate theconfidence interval on the odds ratiobetween the test and reference products aftershowing that the test and the referenceproducts are significantly better than thecontrol. Further detail on statistical analysisof clinical end-point data might be foundelsewhere.13

Tissue residue depletion studiesThe panel on Human Food Safety at

the 1993 Veterinary Drug BE Workshopaddressed tissue residue depletion studies forgeneric animal drugs13. As stated many timesbefore in the present document, the centerhas concluded that in addition to BE study,a tissue residue depletion study should beconducted for approval of a generic animaldrug product in food producing species.Indeed, two drug products may have thesame plasma disposition profile at theconcentrations used to asses product BE, butmay have different tissue disposition kineticswhen followed out to the withdrawal timefor the pioneer product. Therefore, to showthe withdrawal period at which residues ofthe generic product will be consistent withthe tolerance for the pioneer product, a tissueresidue depletion study is necessary.1,2,3

However, and importantly, for purposes ofcalculating a withdrawal period for a genericanimal drug, only the generic product wouldbe tested (i.e., not the pioneer product), andonly the marker residue in the target tissuewould be analyzed.3

In this respect, it should not beforgotten, that the results of a BE study ortissue depletion study in one animal speciescan not generally be extrapolated to anotherspecies. Possible species differences in drugpartitioning or binding in tissues couldmagnify a small difference in the rate orextent of drug absorbed into a largedifference in marker residue concentrationin the target tissue. Furthermore, differences

in drug metabolism and excretion are alsoknown to exist among different food-producing animals. Therefore, for a pioneerproduct labeled for more than one food-producing species, a BE study and a tissueresidue depletion study will generally berequested for each major food-producingspecies on the label.1,2,3

Traditional withdrawal studies, asthose described in FAO/WHO Codexalimentarius guidelines15 are considered thebest design for collecting data for thecalculation of withdrawal periods for drugsused in food-producing animals. In thisrespect, it was not until the 42nd JECFAmeeting that specific requirement for residuedata were presented.5 The committeerequested detailed reports, includingindividual animal data, on the followings: 1)the chemical identity and properties of thedrug; 2) the use and (recommended) doses;3) pharmacokinetic metabolic andpharmacodynamic studies in experimentaland food producing animals, and in humans,where available; 4) residue depletion studieswith radiolabelled drug in target animals fromzero withdrawal time to periods extendingbeyond the recommended withdrawal time;5) information on total residues, includingfree and bound residues, and major residuecomponents to permit selection of a markerresidue and a target tissue; 6) residue depletionstudies with unlabeled drug for the analysisof marker residue in target animals and ineggs and milk; 7) a review of routineanalytical methods that may be used byregulatory authorities for the detection ofresidues in target tissues; and 8) a descriptionof the analytical procedure that was chosenby the sponsor for the detection anddetermination of residues in target tissues(with sensitivity equal to or less than MRL,ideally ≤0.5 MRL).

Usually, in these withdrawal studies,animals are divided into different groupsbeing slaughtered at carefully preselectedtime points following the last (or single)administration of the test product anddifferent matrix or target tissue samples arecollected for residue analysis. A statistical

16


tolerance limit approach is used to determinewhen, with 95% confidence limits, 99% oftreated animals would have tissue residuesbellow the codified limits.

Figure 4 shows an example of thisstudy; it depicts the depletion of a markerresidue in a given tissue of a food-producinganimal. It is relevant to note that in this curve,the Maximum Limit of Residue (MLR) is apoint on the curve describing the upper one-sided 95% confidence limit while the medianis the corresponding point in the regressionline. In other words, the withdrawal time iscorrectly defined by the upper one sided95% confidence limit.

According to international legislation1,2,3,the generic animal drug will be assigned thewithdrawal time supported by the residuedepletion data, or the withdrawal timecurrently assigned to the pioneer product,whichever is the longer. If the genericsponsor wants to request a shorterwithdrawal period for one specific product,he should provide all data necessary tosupport his request, including those that showno other food safety concerns for the drugare evident.

Finally, the analytical methodologyused to determine the withdrawal period forthe generic product does not have to be more

rigorous than the approved methodologyused to determine the existing withdrawalperiod for the pioneer product. However,it has to be fully validated and performedaccording to GLP and GSP standards, asstated above. Thus, if an analytical methodother than the approved method is used, thegeneric sponsor should provide datacomparing the alternate method to theapproved method.

In vitro pharmaceutical equivalenceA product is considered to have

pharmaceutical equivalence to a referenceproduct if the formulation ingredients arethe same and package in a size container madefrom identical material.2,3 Specifically, 1-ingredients in both products should be ofequivalence pharmaceutical compendialstandard; 2- the active constituents shouldbe the same substance and should be within± 5% of that in the reference product orthe dose range to the animal should be within± 5% of the reference product; 3- the non-active constituents should be the samesubstances and should be within ± 5% ofthat in the reference product; and 4- theproduct should have the samephysicochemical properties as the referenceproduct.

Figure 4 - Depletion of a marker residue in a given tissue of a food producing animal. Note that in curves like thisone, the Maximum Limit of Residue (MRL) is a point on the curve describing the upper one-sided 95%confidence limit while the median is the corresponding point in the regression line. Note also that thewithdrawal time is correctly defined by the upper one-sided 95% confidence limit

17


According to the EMEA guidelinesfor the conduct of BE studies for VeterinaryMedical products2 in vitro BE studies couldsupport in vivo BE in the following cases: 1-in vivo BE has been demonstrated for thehighest dosage strength and in vitrodissolution data is used to support the BEof the lower dosage strengths for thatgeneric formulation; 2- in vitro comparabilitymight be adequate to confirm thecomparability of the reference product andits generic product to be administered orally;this applies particularly to immediate releaseoral dosage forms that are rapidly dissolvingand contain drug substances that are highlysoluble and highly permeable; 3- there is avery minor formulation change to anapproved product, and it has beendetermined that the change requires only invitro confirmation of comparability to theformulation that underwent the original trials;and, 4- ensuring batch to batch consistencywithin a product.

According to EMEA guidelines2,data for in vitro pharmaceutical equivalenceshould include: 1- the nature of the dosageform, the solubility of the activeconstituent(s) in water; 2- the pharmaceuticalcharacteristics of the product, includingparticle size, crystal forms, and dissolutionprofile where applicable; 3- information onrate limiting steps in absorption of the activeconstituent(s) e.g. disintegration, dissolution,gut absorption where applicable or in accessto the site of effect; and, 4- relevant scientificargument regarding clinical consequences ofinequivalence.

It should be noted, however, that thein vitro pharmaceutical equivalence test mustbe a validated predictor of the in vivodissolution of the product, i.e., the in vitrotest conditions has to have been previouslyrelated to in vivo conditions. An in vitro testcannot be used when the mean dissolutiontime is higher than the mean absorption time.Moreover, the longer is the dissolution time,the more difficult will be the extrapolationbetween the in vitro and in vivo conditions. Itis therefore not recommended to performin vitro test when the dissolution time is too

long. On the contrary, when the process ofdissolution is not the rate limiting step withrespect to the rate and extent of absorption,the equivalence between the dissolutionprofiles of the generic and the referenceproducts would not be required.

The experimental units collected forin vitro test are tablets, defined quantities of apaste, or a powder in a specified packing.These units should be collected using asampling plan, based on a randomizationprocedure. This sampling procedure shouldbe the same for the pioneer formulation andfor the generic formulation, and should berepresentative of the entire productpopulation. Finally, the number of batchesfrom which experimental units are sampledfor the in vitro test should be related to theexpected variability between batches.

In vitro pharmaceutical equivalenceshould take into account the main sourcesof variation, which are likely to influence thefinal result: product batch, time ofconservation and apparatus used in the test.Precautions to avoid bias must be taken, suchas an equal repartition of units of eachformulation in each analytical run.. Whenrelevant, replicates of measures should bemade, in order to take into account thevariation inherent to the analytical method.

In the case of a study designcomparable to the in vivo BE study design isrelevant, the sample size should bedetermined to provide sufficient power inthe demonstration of equivalence. Theresidual error (coefficient of variation) usedin the calculation of the sample size shouldbe obtained from pilot studies, or estimatedfrom the variance reproducibility of theanalytical method. These points must bedocumented in the experimental protocol.

In vitro pharmaceutical data to bestatistically analyzed must be selected a prioriand must be justified with regard to thecorrelation with the pharmacokinetic. Itcould be sufficient to discuss the relationbetween the dissolution time and theabsorption rate for the products compared2.In vitro BE could be then demonstrated bycomparison of dissolution profiles after

18


fitting to a mathematical model or bycomparison of parameters like 50%dissolution time and 90% dissolution timeand total amount dissolved and AUC.Statistical analysis could be comparable tothe analysis used in the case of an in vivo BEstudy. However, GSP should be employedand the predetermined equivalence intervalshould be carefully justified. It should be keptin mind that exemptions of in vivo studiesare only possible when results of in vitrostudies could lead to the deduction of similarpharmacokinetic behavior between the twoproducts compared.

Conclusion

It is indisputable. High efficiency inthe field of Veterinary Medicine, particularlyin animal production is achieved when newtechnologies are used, quality assurancemethods are always an objective and, animalhealth products, such as veterinary medicinesand feed additives, are employed. The useof animal health products in animalproduction however, should be donefollowing the standards on Good Clinical

Practice (GCP) on Veterinary Drugs Use.Thus, only products approved for food-producing animals use and their genericformulation (that have statistically provenBE) should be employed. However, therequirements for in vivo BE study may bewaived for certain products. According toFDA Guidance to Industry (FDA, 2002)categories of products which may be eligiblefor waivers include, but are not limited to,the followings: 1- parenteral solutionsintended for injection by intravenous,subcutaneous, or intramuscular routes ofadministration; 2- topically applied solutionsintended for local therapeutic effects.3- other topically applied dosage formsintended for therapeutic effects for non-food animals only; 4- inhalant volatileanesthetic solutions. For those waivers,in v i tr o pharmaceutical equivalenceand residue depletion studies would beenough to guarantee their registration.Finally, it is felt relevant to point out againthat in spite of the relevance and thecommon use in other countries, BE studiesare still not requested for veterinaryproducts registration in Brazil.

Estudos de bioequivalência: importância para Medicina Veterinária

Resumo

Os estudos de Bioequivalência (BE) são utilizados para acomparação de diferentes produtos farmacêuticos que contêm omesmo princípio ativo, de diferentes lotes de um mesmo produtoou, ainda e de uma maneira ampla, de diferentes vias deadministração de um mesmo medicamento. No Brasil dos diasde hoje, encontramos legislações sobre medicamentos genéricose bioequivalência apenas na área de Medicina Humana. No campoda Medicina Veterinária, os testes de BE têm sido considerados,em muitos países, como requerimentos necessários para o registrode produtos destinados aos animais visto que eles asseguram, aomesmo tempo, a eficácia do produto, a saúde dos animais tratadose a qualidade dos alimentos provenientes desses animais. Opresente trabalho faz uma revisão crítica sobre BE. Para tanto, oassunto foi dividido em três grandes partes: 1- Entendendo abioequivalência: importância de estudos de BE para a saúde animale humana; 2- tipos de estudos de BE; 3- considerações geraissobre delineamentos experimentais que envolvam estudos debioequivalência.

Palavras-chave:Farmacocinética.Farmacodinâmica.Biodisponibilidade.Bioequivalência.Medicamento de referência.Medicamento genérico.

19


References

1 AUSTRALIAN PESTICIDES AND VETERINARYMEDICINES AUTHORITY. Bioequivalence guidelinesfor veterinary chemical products. [s.l.]: APVMA, [s.d.].Disponível em: <http://www.apvma.gov.au/guidelines>.

2 EUROPEAN AGENCY FOR THE EVALUATION OFMEDICINAL PRODUCTS. Guidelines for the conduct ofbioequivalence studies for veterinary medicinal products.London: European Medicines Agency, 2001. 11 p.

3 FOOD AND DRUG ADMINISTRATION. Centerfor Veterinary Medicine. Bioequivalence guidance.Rockville, 2002. 28 p.

4 HUTCHINSON, D.; GARDEN, C. Golden GCP rulesfor veterinarian studies. Guildford, Surrey: CanaryPublications, 2001. 26 p.

5 JOINT FAO/WHO EXPERT COMMITTEE ON FOODADDITIVES. Procedures for recommending maximumresidue limits: residue of veterinary drugs in food.Rome: FAO/WHO, 2000. 51p.

6 WESTLAKE, W. J. Bioavailability and bioequivalenceof pharmaceutical formulations. In: PEACE, K. E. (Ed.).Biopharmaceutical statistics for drug development.New York: Marcel Dekker Inc., 1988. p. 329–352.

7 MARTINEZ, M. N.; JACKSON, A. J. Suitability ofvarious noninfinity area under the plasma-concentrationtime curve (AUC) estimates for use in bioequivalencedeterminations - relationship to AUC from zero to timeinfinity. Pharmaceutical Research, v. 8, n. 4, p. 512-517, 1991.

8 LOVERING, E. G.; MCGILVERAY, I. J.; MCMILLAN,I.; TOSTOWARYK, W. Comparative bioavailabilitiesfrom truncated blood level curves. Journal of

Pharmaceutical Sciences, v. 64, n. 9, p. 1521-1524,1975.

9 RONFELD, R. A.; BENET, L. Z. Interpretation of plasmaconcentration-time curves after oral dosing. Journalof Pharmaceutical Sciences, v. 66, n. 2, p. 178-180,1977.

10 AGÊNCIA NACIONAL DE VIGILÂNCIASANITÁRIA. Gerência-Geral de Inspeção e Controlede Medicamentos e Produtos. Manual de boas práticasem biodisponibilidade – bioequivalência. Brasília:ANVISA, 2002. 2 v.

11 DARGENIO, D. Z. Optimal sampling times forpharmacokinetic experiments. Journal ofPharmacokinetics and Biopharmaceutics, v. 9, n. 6, p.739-756, 1981.

12 HAUSCHKE, D.; STEINIJANS, V. W.; DILETTI, E.;BURKE, M. Sample-size determination forbioequivalence assessment using a multiplicative model.Journal of Pharmacokinetics and Biopharmaceutics, v.20, n. 5, p. 557-561, 1992.

13 STEINIJANS, V. W.; HAUCK, W. W.; DILETTI, E.;HAUSCHKE, D.; ANDERSON, S. Effect of changingthe bioequivalence range from (0.80, 1.20) to (0.80,1.25) on the power and sample-size. InternationalJournal of Clinical Pharmacology and Therapeutics, v.30, n. 12, p. 571-575, 1992.

14 MARTINEZ, M. N.; RIVIERE, J. E. Review of the1993 veterinary drug bioequivalence workshop - preface.Journal of Veterinary Pharmacology and Therapeutics,v. 17, n. 2, p. 85-119, 1994.

15 SCHUIRMANN, D. J. A comparison of the 2one-sided tests procedure and the power approachfor assessing the equivalence of average bioavailability.Journal of Pharmacokinetics and Biopharmaceutic,v.15, n.6, p. 657-680, 1987.

bioequivalence studies: relevance for veterinary medicine

Documents