Bioequivalence and in vitro-in vivo correlations Martin Čulen Faculty of Pharmacy, University of Veterinary and Pharmaceutical Sciences, Palackeho 1-3, 612 42 Brno, Czech Republic Generic vs. Innovator drugs Bioequivalence IVIVC (In vitro-in vivo correlations) Lecture Outline Generic vs. Innovator drugs Bioequivalence IVIVC Lecture Outline 1/1000 new drug candidates is approved for the market Cost = 300-1000 million USD Time spent = 12-15 years 20-year patent (approx. 10 y during development + 10 y during marketing) Innovator drugs Basic facts By 2016, generics constituted 89% of total prescriptions, while only accounting for 27% of total drug costs* *https://www.optum.com/resources/library/7-fast-facts-generic-drugs.html?o=optum:soc:RX_8.1_2017:tw:rx:lib:17imtkv04cm24 Generic drugs Identical: • Active ingredient • Dosage form type (e.g. immediate release tablet) • Route of administration • Strength • Indication • Quality • Performance Generic drugs vs. originator drugs Differences: • Formulation shape, color… • Packaging • Release mechanisms • Clinical tests – only bioequivalence study required! – safety and efficacy data provided by innovator Generic drugs vs. originator drugs Generics production – copy & paste?? Not exactly… Obstacles implied by intellectual property patents: • Active ingredient - e.g. patented polymorphs (different crystal structures) • Dosage form type (e.g. immediate release tablet) - e.g. patented excipients, formulation content Generic drugs vs. originator drugs innovator patent patent X X X ?? ?? Making the same product, but different way… Generic vs. Innovator drugs Bioequivalence IVIVC Lecture Outline Equal performance of generic vs. innovator drug = bioequivalence = equal bioavailability Two products or formulations containing the same active ingredient are bioequivalent if their rates and extents of absorption (bioavalabilities) are the same (within predefined limits). Bioequivalence may be demonstrated through in vivo or in vitro test methods, comparative clinical trials, or pharmacodynamic studies. Bioequivalence Absolute bioavailability Bioequivalence Bioavailability oral iv iv oral DOSE DOSE AUC AUC F = Relative bioavailability Bioequivalence Bioavailability f ormB f ormA rel AUC AUC F = Identical bioavailability = bioequivalence Investigated parameters: - AUC0-t - CMAX Acceptance criteria: - 80-125%* range for 90% confidence interval of test/reference product Bioequivalence In vivo studies *in specific cases, more narrow (90-111%) or wider (75-133%) intervals are acceptable DOES NOT MEAN THAT: “GENERIC CAN BE ±20% DIFFERENT FROM THE ORIGINAL” 90% confidence interval has to fit the 80-125% interval Bioequivalence In vivo studies What is the 90% confidence interval??? If the BE study was repeated 100 times - 90 times the population value (mean Cmax or AUC) would fall inside this interval, and 10 times outside (biological variability). Bioequivalence In vivo studies In practice (data from 2070 BE studies): • the generic/innovator ratios were 1.00 ± 0.06 for Cmax and 1.00 ± 0.04 for AUC (mean ± SD) • the average difference in Cmax and AUC between generic and innovator products was 4.35% and 3.56%, respectively • in nearly 98% of the BE studies, the generic product AUC differed from that of the innovator product by less than 10% Davit et al., Ann. Pharmacother. 2009, Comparing Generic and Innovator Drugs: A Review of 12 Years of Bioequivalence Data from the United States Food and Drug Administration. Bioequivalence In vivo studies EMA Guideline on the Investigation of Bioequivalence, 2010 STUDY DESIGN : • standard design: randomized, two-period, two-sequence, single dose cross-over design • alternative designs: parallel design (substances with very long halflives) and replicate designs (in case of highly variable drugs or drug products) Bioequivalence In vivo studies Standard 2×2 Crossover design • standard design: randomized, two-period, two-sequence, single dose, crossover design Subjects Randomization Sequence 1 Sequence 2 Period I II Reference Test Test Reference Washout Bioequivalence In vivo studies Replicate design • randomized, four-period, two-sequence, single dose, cross-over design (highly variable drugs or drug products) Subjects Randomization Sequence 1 Sequence 2 Period I II Reference Test Test Reference Washout III Washout IV Test Reference Washout Reference Test Bioequivalence In vivo studies Parallel design (substances with very long half-lives) Subjects Randomization Group 1 Group 2 I Reference Test Bioequivalence In vivo studies EMA Guideline on the Investigation of Bioequivalence, 2010 STUDY SUBJECTS: • ≥ 12 subjects - more subjects = better homogeneity, “more accurate result” • healthy volunteers to reduce variability (patients, e.g. for chemotherapy) • strict inclusion/exclusion criteria, • subjects could belong to either sex, • preferably non-smokers and without a history of alcohol or drug abuse Bioequivalence In vivo studies EMA Guideline on the Investigation of Bioequivalence, 2010 SAMPLING TIMES • frequent sampling around the predicted tmax • Cmax should not be the first point of the concentration-time curve INAPPROPRIATE STUDY DESIGN IS ONE OF THE MOST COMMON CAUSES OF FAILURE personal communication, Helmut Schutz Bioequivalence In vivo studies Bioequivalence In vivo studies Bioequivalence In vivo studies Bioequivalence In vivo studies Generic vs. Innovator drugs Bioequivalence IVIVC Lecture Outline FDA definition: “a predictive mathematical model describing the relationship between an in-vitro property (dissolution) of a dosage form and an in-vivo response (PK curve)” Purpose: to utilize in vitro dissolution profiles as a surrogate for in vivo bioequivalence Application: - supporting biowaivers (approval without in vivo BE) - Scale-Up and Post-Approval Changes (SUPAC) and line extensions (e.g., different dosage strengths) - support of dissolution methods IVIVC In vitro-in vivo correlations Scale-Up and Post-Approval Changes (SUPAC) 1) Components or composition 2) Manufacturing site 3) Scale-up (increasing production) 4) Manufacturing (process or equipment) × Effect on quality and performance: LEVEL1 – UNLIKELY any detectable effect LEVEL2 – COULD HAVE significant effect LEVEL3 – LIKELY to have significant effect Bioequivalence During development Example 2: Changing direct compression to wet granulation: Type of change: ? LEVEL: ? BE required: ? Example 1: Changing a coloring agent in IR tablet Type of change: ? LEVEL: ? BE required: ? Bioequivalence SUPAC Components or composition 1 NO Manufacturing process 3 YES 14 examples from FDA database: - Change in dissolution method and specifications - Level 3 site manufacturing change - Waiver for lower strengths - Waiver for higher strengths - To support dissolution method - Batch-to-batch variation in the particle size, coating weight, process changes, test product composition do not impact the BE - Change in dissolution specifications - Change in dissolution specifications - Challenge the results of a failed BE study - Batch-to-batch variation in pellet coating does not impact the BE - Change in dissolution specifications - Exploratory to guide the development of pivotal formulation IVIVC submission examples Kaur et al., The AAPS Journal. 2015, Applications of In Vitro–In Vivo Correlations in Generic Drug Development: Case Studies. Description/procedure: 1) Obtaining dissolution (in vitro) and PK (in vivo) data for “input” formulations 2) Building a mathematical IVIVC model using the “input” formulations 3) Testing the predicition power of the established model IVIVC Procedure LEVEL A - highest level of correlation. - point to point relationship between in vitro dissolution rate and in vivo input rate LEVEL B - mean absorption time is plotted against mean dissolution time for ≥ 3 formulations LEVEL C - single point correlation for ≥ 3 formulations - % drug dissolved in X min vs. AUC or Cmax or Tmax IVIVC Division Dissolution data • Idealy 3 formulations with different release rates • any in vitro dissolution method can be utilized (the preferred dissolution apparatus is USP apparatus I or II) • the same for all formulations tested • an aqueous medium either water or buffered solutions not exceeding pH 6.8 is recommended IVIVC 1) Obtaining dissolution (in vitro) and PK (in vivo) data for “input” formulations PK data • > 6 subjects • Crossover design preferred • 3 formulations + inclusion of a reference treatment is advised: - IV solution - Oral solution - Immediate release product IVIVC 1) Obtaining dissolution (in vitro) and PK (in vivo) data for “input” formulations a) Making the PK and dissolution data “comparable” IVIVC 2) Building a mathematical IVIVC model using the “input” formulations ? b) Correlating the mathematically processed PK and dissolution curves a) Making the PK and dissolution data “comparable” IVIVC 2) Building a mathematical IVIVC model using the “input” formulations %absorbed Absorption + elimination Absorption %dissolved Conc.Conc. Elimination oral formulation i.v. bolus deconvolution convolution a) Making the PK and dissolution data “comparable” Deconvolution - calculating the fraction absorbed from PK curve Wagner-Nelson method One compartmental method Loo-Riegelman method Multi-compartmental method Numerical deconvolution Model independent method Commercial software (e.g. Gastroplus) Convolution - calculating the PK curve from fraction absorbed/dissolved Weibull function IVIVC 2) Building a mathematical IVIVC model using the “input” formulations b) Correlation of PK and dissolution data IVIVC 2) Building a mathematical IVIVC model using the “input” formulations %absorbed %dissolved linear non-linear In vitroIn vivo IVIVC 3) Testing the prediction power of the established model Prediction of Cmax and AUC from dissolution data using the established model. Comparing the predicted vs. real PK data For Cmax: For AUC: IVIVC 3) Testing the prediction power of the established model Prediction of Cmax and AUC from dissolution data using the established model. Comparing the predicted vs. real PK data Acceptance criteria: According to FDA guidance • ≤ 15% for absolute prediction error (%P.E.) of each formulation. • ≤ 10% for mean absolute prediction error (%P.E.) IVIVC 3) Testing the prediction power of the established model Internal predictability - 2-3 different formulations used for model building - identical mathematical processing External predictability - 1 formulation not used for model building - identical mathematical processing IVIVC Additional considerations – BCS classification Class Solubility Permeability Absorption rate control step IVIVC I High High Gastric emptying time Correlation (if dissolution is slower than GET) II Low High Dissolution Correlation III High Low Permeability Little or no correlation IV Low Low Case by case Little or no correlation IVIVC Additional considerations – possible issues Dissolution - inaccurate in vitro dissolution data - complex in vivo dissolution processes (precipitation, API binding, poorly identified release mechanisms/kinetics) Pharmacokinetics - absorption rate limitations - non-linear elimination or elimination kinetics - enterohepatic recycling or second peak - inter-individual variability Thank you for attention!