Usage Examples
- The fasting BE study passed: the 90% CI for Cmax and AUC0-t both sat inside 80.00-125.00%.
- We are pursuing a biowaiver on the two lower strengths instead of running three separate BE studies.
- The comparator lot was 6% off the test product on assay content, so we sourced another batch before dosing.
What is Bioequivalence (BE)?
Bioequivalence is the regulatory standard requiring a test product to deliver the same rate and extent of active-ingredient exposure as its comparator, unlike pharmaceutical equivalence, which compares only composition, strength, and dosage form.
Bioequivalence exists because an abbreviated application carries no new efficacy trials. An ANDA must instead show that its product is bioequivalent to the reference listed drug, and on that showing it inherits the originator's safety and efficacy conclusions. Bioequivalence is the evidentiary bridge that makes the inheritance defensible, turning a clinical question about outcomes into a measurable question about exposure.
Bioequivalence covers rate and extent of systemic exposure, measured as Cmax and AUC0-t in a randomised single-dose crossover study normally run in healthy subjects. Bioequivalence does not cover whether the drug works; the reference product already settled that. It also stops at the formulation: ICH M13A addresses immediate-release solid oral dosage forms and explicitly excludes comparator acceptability across regions and bioavailability work supporting labelling for new drugs.
Bioequivalence is applied as a pre-specified statistical test, not a judgement call. Sponsors fix the model and the analysis population in the protocol, dose at least twelve evaluable subjects in a crossover, and compute 90% confidence intervals for the geometric mean ratios. Post hoc, data-driven adjustments are not acceptable, so a study that misses 80.00-125.00% is a failed study, not a negotiation.
Not to be confused with
- Bioavailability
- bioavailability describes how much of one product reaches systemic circulation and how fast. Bioequivalence is the comparative verdict between two products, and it is the comparison, not the measurement, that supports an abbreviated approval.
- Pharmaceutical equivalence
- same active moiety, strength, and dosage form. It is a statement about composition only; two pharmaceutical equivalents can and do fail a BE study, which is precisely why the study is required.
- Therapeutic equivalence
- the Orange Book conclusion that a product can be substituted at the pharmacy. Bioequivalence is one input to that rating, not the rating itself; a passing BE study alone does not confer an AB code.
- Biosimilarity
- the standard for biologics, where sameness cannot be established from PK comparison alone. Biosimilarity rests on a totality-of-evidence package rather than one confidence interval on two PK parameters.
The obligations split between what the CFR demands of the application and what ICH M13A demands of the study.
What you must do
- 1Submit evidence in the ANDA demonstrating the product is bioequivalent to the reference listed drug, or information supporting a waiver of in vivo data21 CFR 320.21(b)
- 2Select the evidence type by the regulatory hierarchy, starting with an in vivo PK measurement and dropping to a dissolution test only where that test is acceptable to FDA21 CFR 320.24(b)
- 3Enrol at least 12 evaluable subjects for a crossover design, or 12 per treatment group for a parallel design, sized by a pre-specified power calculationICH M13A, Section 2.1.3
- 4Show that the 90% confidence interval for the geometric mean ratio of each primary PK parameter lies within 80.00-125.00%ICH M13A, Section 2.2.4
- 5Retain reserve samples of the test article and reference standard for at least 5 years after the application is approved21 CFR 320.38(e)
Common mistakes
Hard-coding a tighter window for narrow therapeutic index drugs
ICH M13A does not cover NTI drugs; data analysis for them, for highly variable drugs, and for complex designs is deferred to ICH M13C. Writing a European 90.00-111.11% criterion into a US protocol imports a limit the applicable product-specific guidance may not use, and the acceptance range is not something to settle after unblinding.
Treating a failed confidence interval as a reanalysis problem
the statistical model must be pre-specified in the protocol, and post hoc, data-driven adjustments are not acceptable for the primary analysis. Dropping subjects once the result is visible turns a failed study into an unusable one and puts the review clock, not just the study, at risk.
Sourcing a comparator batch without checking assay content
the difference in assay content between test and comparator should normally be within 5%. A larger potency gap discovered after dosing is not a statistics problem; it is a repeat of the clinical phase and the cost of the batch.
When This Matters
- The fasting BE study passed: the 90% CI for Cmax and AUC0-t both sat inside 80.00-125.00%.
- We are pursuing a biowaiver on the two lower strengths instead of running three separate BE studies.
- The comparator lot was 6% off the test product on assay content, so we sourced another batch before dosing.
Frequently Asked Questions
The 90% confidence interval for the geometric mean ratio of test to comparator must lie within 80.00-125.00% for each primary PK parameter, per ICH M13A. Primary parameters are Cmax and AUC0-t (or AUC0-72h), plus early exposure data where applicable. Highly variable and narrow therapeutic index drugs are handled separately in ICH M13C.
Related Use Cases
- Pharma Use Cases
Cut NDA and sNDA prep time by 60% with AI-assisted drafting and automated readiness checks
- Biotech Use Cases
Compress IND prep from 8-12 weeks to under 3 weeks with AI-assisted drafting and validation
- Clinical Workflows
Track clinical trial regulations across global markets
- Regulatory Affairs Workflows
Cut regulatory intelligence tracking from 10+ hours/week to automated, real-time alerts
