Bioequivalence BE
The statistical demonstration that two drug formulations deliver the same total exposure (AUC) and peak concentration (Cmax) within defined limits — the regulatory basis for generic drug approval worldwide.
Definition
Two drug products are bioequivalent if they are pharmaceutical equivalents (same active ingredient, same salt/ester form, same strength, same dosage form, same route) and their rate and extent of absorption do not differ significantly under the same conditions. This is formalised by the FDA ANDA (Abbreviated New Drug Application) process.
The two primary pharmacokinetic metrics used to establish BE are:
Total drug exposure — reflects extent of absorption (F × Dose / CL). The primary BE metric. See AUC glossary entry.
Peak plasma concentration — reflects rate of absorption. Important for drugs where peak is linked to therapeutic or adverse effects.
The regulatory frameworks governing BE are the FDA ANDA guidance, the EMA BE Guideline, and the ICH M9 biopharmaceutics classification system–based biowaivers guideline.
The 80–125% Criterion
BE is concluded when the 90% confidence interval (CI) for the geometric mean ratio (GMR) of AUC and Cmax falls entirely within 80.00–125.00%:
This uses the two one-sided tests (TOST) procedure (Schuirmann 1987): separately testing H₀: GMR ≤ 0.80 and H₀: GMR ≥ 1.25 at α = 0.05 each. Rejecting both demonstrates equivalence. The 90% CI corresponds to two 5% one-sided tests.
The 80–125% limits are not arbitrary — they are the range within which differences in drug exposure are considered unlikely to be clinically significant for most drugs. The asymmetry of the bounds (80% = 1/1.25) is an artefact of the log-normal distribution used; on a logarithmic scale they are symmetric around 1.0. For narrow-therapeutic-index drugs, FDA tightens the criterion to 90–111.1%.
BE Study Design
A standard BE study per FDA guidance for industry:
- 1Crossover design — each subject receives test (generic) and reference (brand) in random order, separated by a washout of ≥5 half-lives. This controls for inter-subject PK variability.
- 2Subjects — 24–36 healthy adult volunteers, fasted for ≥10 h. A fed-state study is required for formulations with known food effects.
- 3Sampling — 12–20 blood samples over ≥3 half-lives post-dose. Must capture ≥80% of AUC0–∞.
- 4Analysis — log-transformed AUC and Cmax; 90% CI by mixed-model ANOVA. GMR and CI reported. Schuirmann TOST.
- 5Exceptions — highly variable drugs (intra-subject CV >30%) may use scaled average bioequivalence (SABE) with widened limits. See FDA highly variable drug guidance.
Types of Bioequivalence
FDA standard for most drugs. Compares population geometric mean ratios. Does not account for intra-subject variability in switching between products. Sufficient for most non-NTI drugs.
For highly variable drugs (intra-subject CV >30%). Reference-scaled approach widens limits relative to variability of the reference product. Approved by FDA and EMA for specific drug classes.
Accounts for both mean differences and between-subject variability. More relevant when the concern is whether the population distribution of exposures is similar. Less commonly required by regulators.
Most stringent: accounts for within-subject variability on switching. Intended to address the concern that an individual patient switching products might experience clinically meaningful changes. Rarely required in practice.
FAQ
What does bioequivalent mean?
Two products are bioequivalent if they are pharmaceutical equivalents (same API, strength, dosage form, route) and their 90% CI for AUC and Cmax ratios falls within 80–125%. This means any difference in exposure is not considered clinically significant for most drugs.
What is the 80–125% rule?
The 90% CI for the geometric mean ratio of generic:brand AUC and Cmax must fall entirely within 80.00–125.00%. This is assessed by the two one-sided tests (TOST) procedure. In practice, approved generics average ~3.5% difference from brand — far less than the allowed range.
How is a bioequivalence study designed?
A randomised two-period crossover in 24–36 healthy volunteers. Each subject takes test and reference with a washout of ≥5 half-lives between periods. Blood sampled over ≥3 half-lives; AUC and Cmax calculated by NCA. 90% CI by log-ANOVA.
What is the difference between bioequivalence and therapeutic equivalence?
Bioequivalence is a pharmacokinetic standard (same AUC/Cmax). Therapeutic equivalence (FDA Orange Book AB rating) additionally requires pharmaceutical equivalence (same API, strength, dosage form) and assumes BE implies equivalent clinical effect. Therapeutic equivalence permits pharmacist substitution in most US states.