Bioequivalence Study Requirements: FDA Guidance and Study Design
Bioequivalence (BE) studies demonstrate that a generic drug product delivers the same rate and extent of drug absorption as the reference listed drug (RLD). Under 21 CFR Part 320, FDA requires that the 90% confidence interval for the geometric mean ratio of AUC (area under the curve) and Cmax (maximum concentration) between test and reference products falls within 80.00-125.00% using log-transformed data. The standard design is a single-dose, randomized, 2-period, 2-sequence crossover study in healthy adult volunteers under fasting conditions (and sometimes fed conditions). Alternatives include BCS-based biowaivers for qualifying immediate-release products, pharmacodynamic BE studies for locally acting drugs, and clinical endpoint BE studies for complex products.
Key Takeaways
Key Takeaways
- Standard BE studies use a single-dose, 2-period, 2-sequence crossover design comparing AUC and Cmax with 90% CI within 80.00-125.00% acceptance range
- BCS-based biowaivers allow in vitro dissolution to substitute for in vivo studies for BCS Class 1 (rapid dissolution) and Class 3 (very rapid dissolution) drugs per 21 CFR 320.22
- Narrow therapeutic index (NTI) drugs require tighter acceptance criteria and additional statistical analyses per FDA guidance
- OGD product-specific guidances (PSGs) define the recommended BE study design for each RLD; deviating without justification is a leading cause of CRLs
- Bioequivalence study requirements are the scientific foundation of the generic drug approval system. Under the ANDA pathway, demonstrating BE to the reference listed drug substitutes for the clinical safety and efficacy trials required in a full NDA. The logic is straightforward: if a generic delivers the same drug exposure as the approved reference product, it will produce the same therapeutic effect.
- Getting BE studies right is critical. Bioequivalence-related deficiencies account for approximately 24% of all ANDA complete response letters, making it the second most common cause of CRLs after chemistry. Deviating from FDA's expectations in study design, bioanalytical methods, or statistical analysis can result in failed studies, wasted resources, and delayed approvals.
- In this guide, you'll learn:
- The regulatory basis for bioequivalence under 21 CFR Part 320
- Standard BE study design for oral solid dosage forms
- In vivo vs. in vitro bioequivalence approaches
- BCS-based biowaiver requirements and eligibility
- Bioanalytical method validation per FDA guidance
- Statistical acceptance criteria (80-125% CI)
- Special requirements for highly variable drugs and narrow therapeutic index drugs
- Fasting vs. fed study requirements
- ---
Regulatory Basis: 21 CFR Part 320
Bioequivalence is defined in 21 CFR 320.1(e) as "the absence of a significant difference in the rate and extent to which the active ingredient or active moiety in pharmaceutical equivalents or pharmaceutical alternatives becomes available at the site of drug action when administered at the same molar dose under similar experimental conditions in an appropriately designed study."
Key Regulatory References
| Regulation/Guidance | Subject |
|---|---|
| 21 CFR Part 320 | Bioavailability and bioequivalence requirements |
| 21 CFR 320.21 | Requirements for submission of in vivo BE data |
| 21 CFR 320.22 | Criteria for waiver of in vivo BE data (biowaivers) |
| 21 CFR 320.23 | Basis for measuring in vivo BE |
| 21 CFR 320.24 | Types of evidence to measure BE |
| FDA Guidance: Bioequivalence Studies with PK Endpoints (2022) | Design and analysis of standard PK BE studies |
| FDA Guidance: Bioanalytical Method Validation (2018) | Validation requirements for bioanalytical methods |
| FDA Guidance: BCS-Based Biowaivers (2021) | Biopharmaceutics Classification System waivers |
| Product-Specific Guidances (PSGs) | Drug-specific BE study recommendations |
Types of Evidence for Bioequivalence (21 CFR 320.24)
FDA recognizes multiple types of evidence to establish BE, listed in order of preference:
| Rank | Evidence Type | Application |
|---|---|---|
| 1 | In vivo pharmacokinetic (PK) study | Most oral solid dosage forms; measures blood/plasma drug levels |
| 2 | In vivo pharmacodynamic (PD) study | Locally acting drugs where blood levels don't reflect effect |
| 3 | Comparative clinical endpoint study | Complex products where PK and PD approaches are not feasible |
| 4 | In vitro study | BCS-based biowaivers, topical solutions, certain injectables |
| 5 | Any other approach adequate to measure BE | Case-by-case determination |
Standard In Vivo PK Bioequivalence Study Design
The standard BE study for oral solid dosage forms follows well-established design principles.
Core Study Design Elements
| Element | Standard Requirement |
|---|---|
| Design | Randomized, single-dose, 2-period, 2-sequence crossover |
| Population | Healthy adult volunteers, 18-55 years (typically) |
| Sample size | 24-48 subjects (based on power calculation; minimum 12 per sequence) |
| Dosing | Highest marketed strength, single dose |
| Conditions | Fasting (and fed, if required by PSG) |
| Washout period | Minimum 5 half-lives between periods (typically 7-14 days) |
| Blood sampling | Sufficient time points to characterize full PK profile |
| Reference product | RLD purchased in the United States per Orange Book listing |
Study Conduct Requirements
Subject Selection:
- Healthy adult volunteers (no relevant disease states)
- No concomitant medications that could affect drug absorption
- Fasting for at least 10 hours before dosing (for fasting studies)
- Standardized meals provided during study (timing and composition controlled)
- Subjects serve as their own controls (crossover design)
Dosing:
- Single dose of test (generic) and reference (RLD) products
- Administered with 240 mL of water
- Subjects remain upright for a specified period after dosing
- No food for a specified period after dosing (fasting studies)
Blood Sampling:
- Pre-dose sample to confirm no carryover from previous period
- Sufficient samples to capture Cmax, absorption phase, and elimination phase
- Typically 12-18 time points over 72 hours (varies by drug half-life)
- Sampling extends to at least 3 half-lives or until AUC0-t covers at least 80% of AUC0-inf
Analytical Phase:
- Plasma samples analyzed using validated bioanalytical method
- Incurred sample reanalysis (ISR) required for at least 7% of samples
- Method must distinguish between parent drug and metabolites
Always check the product-specific guidance (PSG) before designing a BE study. PSGs may specify non-standard requirements such as: multiple-dose studies, additional PK endpoints, specific subject populations, additional study conditions (fasting + fed), or alternative study designs. Deviating from a PSG without prior FDA agreement is the single most common cause of BE-related complete response letters.
Fasting vs. Fed Bioequivalence Studies
FDA may require BE studies under fasting conditions, fed conditions, or both. The requirement depends on the drug and dosage form.
When Each Condition Is Required
| Condition | When Required | Study Specifics |
|---|---|---|
| Fasting only | Products with no food effect or where PSG specifies fasting only | 10-hour overnight fast; no food for 4 hours post-dose |
| Fed only | Products labeled to be taken with food; products where BE under fasting is not informative | Standardized high-fat, high-calorie meal per FDA guidance |
| Both fasting and fed | Products where food effect is significant or unknown; most modified-release products | Two separate studies or one study with additional periods |
FDA Standard High-Fat Meal for Fed BE Studies
| Component | Specification |
|---|---|
| Total calories | Approximately 800-1000 calories |
| Fat content | Approximately 50% of total calories from fat |
| Meal timing | Administered 30 minutes before dosing |
| Meal completion | Must be consumed within 30 minutes |
| Example meal | 2 eggs fried in butter, 2 strips of bacon, 2 slices of toast with butter, 4 oz hash browns, 8 oz whole milk |
Modified-Release Products
For modified-release (extended-release, delayed-release) products, FDA typically requires:
| Study | Condition | Additional Requirements |
|---|---|---|
| Study 1 | Fasting | Single-dose crossover |
| Study 2 | Fed | Single-dose crossover |
| Study 3 (sometimes) | Multiple-dose | Steady-state crossover (for some products) |
BCS-Based Biowaivers: In Vitro Alternatives
For certain drug products, FDA allows in vitro dissolution data to substitute for in vivo BE studies. This BCS-based biowaiver approach is codified in 21 CFR 320.22 and detailed in FDA's 2021 guidance.
Biopharmaceutics Classification System
| BCS Class | Solubility | Permeability | Biowaiver Eligible |
|---|---|---|---|
| Class 1 | High | High | Yes (with rapid dissolution) |
| Class 2 | Low | High | Generally no |
| Class 3 | High | Low | Yes (with very rapid dissolution, per 2021 guidance) |
| Class 4 | Low | Low | No |
BCS Class 1 Biowaiver Requirements
| Criterion | Requirement |
|---|---|
| Solubility | Highest single dose strength dissolves in 250 mL or less at pH 1.0, 4.5, and 6.8 (37 +/- 1C) |
| Permeability | Fraction absorbed >= 85% (based on mass balance or absolute BA studies) |
| Dissolution | Rapid dissolution: >= 85% dissolved in 30 minutes in all three media (pH 1.0, 4.5, 6.8) using USP Apparatus I (100 rpm) or II (50 rpm) in 900 mL |
| Excipients | Excipients do not affect rate or extent of absorption |
| Product type | Immediate-release solid oral dosage form |
| Exclusions | Not an NTI drug; not a prodrug with absorption dependent on GI enzymes |
BCS Class 3 Biowaiver Requirements (2021 Guidance)
| Criterion | Requirement |
|---|---|
| Solubility | Same as Class 1 |
| Permeability | Fraction absorbed < 85% but demonstrates high solubility |
| Dissolution | Very rapid dissolution: >= 85% dissolved in 15 minutes in all three media |
| Excipients | Same excipients in similar amounts as the RLD (qualitatively and quantitatively similar) |
| Product type | Immediate-release solid oral dosage form |
| Exclusions | Not an NTI drug; excipients that might affect absorption must be same as RLD |
In Vitro Dissolution Comparison
For BCS-based biowaivers, comparative dissolution testing must demonstrate similarity:
| Parameter | Requirement |
|---|---|
| Apparatus | USP Apparatus I (basket, 100 rpm) or II (paddle, 50 rpm) |
| Media | pH 1.0 (0.1N HCl), pH 4.5 (acetate buffer), pH 6.8 (phosphate buffer) |
| Volume | 900 mL |
| Temperature | 37 +/- 0.5C |
| Sample size | 12 units each of test and reference |
| Similarity criterion | f2 similarity factor >= 50 (or >= 85% dissolved in 15 or 30 minutes for both products) |
According to FDA data, BCS-based biowaivers account for approximately 10-15% of ANDA bioequivalence submissions, with increasing adoption following the 2021 expansion to BCS Class 3 products.
Bioanalytical Method Validation
Every in vivo BE study must use a bioanalytical method validated per FDA's 2018 Guidance for Industry: Bioanalytical Method Validation.
Required Validation Parameters
| Parameter | Acceptance Criteria |
|---|---|
| Selectivity | No significant interference at the retention time of the analyte or internal standard |
| Accuracy | Within +/-15% of nominal concentration (within +/-20% at LLOQ) |
| Precision | CV <= 15% (CV <= 20% at LLOQ) |
| Linearity | Calibration curve with correlation coefficient (r) >= 0.99 |
| LLOQ (Lower Limit of Quantification) | Minimum 5x signal-to-noise ratio; CV <= 20%; accuracy within +/-20% |
| Recovery | Consistent and reproducible (not required to be 100%) |
| Matrix effect | Assessed in at least 6 lots of blank matrix |
| Stability | Short-term, long-term, freeze-thaw, stock solution, post-preparative |
| Dilution integrity | Demonstrated for dilutions of high-concentration samples |
| Carryover | Response in blank sample after highest calibrator <= 20% of LLOQ |
Incurred Sample Reanalysis (ISR)
| Requirement | Specification |
|---|---|
| Percentage of samples | At least 7% of total study samples (selected from near Cmax and elimination phase) |
| Acceptance criterion | At least 67% of ISR results must be within 20% of the original result |
| Purpose | Confirms reproducibility of the bioanalytical method using actual study samples |
Common Bioanalytical Method Deficiencies in ANDAs
| Deficiency | Frequency | Prevention |
|---|---|---|
| Inadequate stability data | High | Complete all stability evaluations per FDA guidance |
| ISR failure | Moderate | Investigate and resolve before study report finalization |
| Insufficient selectivity testing | Moderate | Test selectivity in at least 6 individual lots of blank matrix |
| Inappropriate internal standard | Low | Use stable isotope-labeled IS whenever available |
| Missing dilution integrity | Moderate | Validate dilution for samples exceeding calibration range |
Statistical Acceptance Criteria
Standard BE Criteria
The primary statistical criterion for bioequivalence is based on the two one-sided tests (TOST) procedure:
The 90% confidence interval of the geometric mean ratio (test/reference) for log-transformed AUC and Cmax must fall entirely within 80.00-125.00%.
| Parameter | Statistical Test | Acceptance Range |
|---|---|---|
| AUC0-t | 90% CI of geometric mean ratio | 80.00-125.00% |
| AUC0-inf | 90% CI of geometric mean ratio | 80.00-125.00% |
| Cmax | 90% CI of geometric mean ratio | 80.00-125.00% |
| Tmax | Not a primary endpoint | Reported but not tested |
Statistical Analysis Model
| Component | Specification |
|---|---|
| Data transformation | Natural log transformation of AUC and Cmax |
| Statistical model | ANOVA (Analysis of Variance) for crossover design |
| ANOVA factors | Sequence, subject nested within sequence, period, treatment (formulation) |
| Confidence interval | 90% CI calculated from the ANOVA residual variance |
| Software | SAS, R, or equivalent validated statistical software |
| Point estimate | Geometric mean ratio (test/reference) reported alongside CI |
Sample Size Determination
Sample size for BE studies is calculated based on:
| Input | Typical Values |
|---|---|
| Within-subject coefficient of variation (CV) | 15-40% (drug-dependent) |
| Power | 80-90% |
| Alpha | 0.05 (one-sided, corresponding to 90% CI) |
| Expected geometric mean ratio | 0.95-1.05 (assumed close to 1.0) |
| Resulting sample size | 24-48 subjects (higher CV requires more subjects) |
Use published data or pilot study results to estimate the within-subject CV before finalizing sample size. Underpowered studies that fail to demonstrate BE due to insufficient sample size are a preventable cause of wasted resources. For drugs with known high variability (CV > 30%), consider the reference-scaled average bioequivalence approach for highly variable drugs (see below).
Highly Variable Drugs: Reference-Scaled Average Bioequivalence
Highly variable drugs (HVDs) are drug products with a within-subject coefficient of variation >= 30% for Cmax, AUC, or both. Standard BE criteria may be difficult to meet for these drugs, not because of true clinical differences, but because of inherent pharmacokinetic variability.
Reference-Scaled Average Bioequivalence (RSABE)
FDA's guidance on BE studies for highly variable drugs allows a reference-scaled approach:
| Element | Standard BE | RSABE for HVDs |
|---|---|---|
| Applicable when | Within-subject CV < 30% | Within-subject CV >= 30% |
| Design | 2-period crossover | Replicate design (3-period or 4-period) |
| Acceptance range for Cmax | Fixed 80.00-125.00% | Scaled based on reference variability |
| Scaling formula | Not applicable | exp(+/- 0.8928 * sWR) where sWR is the within-subject SD of the reference |
| Maximum scaling | Not applicable | Widened to a maximum of 69.84-143.19% |
| Point estimate constraint | None beyond CI | Geometric mean ratio must be within 80.00-125.00% |
| AUC criterion | 80.00-125.00% | 80.00-125.00% (not scaled for AUC) |
Replicate Study Designs for HVDs
| Design | Structure | Advantages |
|---|---|---|
| 3-period partial replicate | TRR or TRR/RTR/RRT | Fewer periods than full replicate; reference replicated |
| 4-period full replicate | TRTR/RTRT | Provides both test and reference variability; higher power |
| 2-sequence, 3-period | TRR/RTT | Balanced design with reference replication |
Narrow Therapeutic Index Drugs
NTI drugs have a narrow range between therapeutic and toxic concentrations. FDA has proposed tighter bioequivalence criteria for NTI drugs.
FDA Approach to NTI Drug BE
| Element | Standard BE | NTI Drug BE |
|---|---|---|
| Acceptance range | 80.00-125.00% | Tightened (proposed: 90.00-111.11%) |
| Study design | Standard crossover | Replicate design required |
| Scaling | Not applicable | Reference-scaled, but with tighter scaling limit |
| Additional requirement | None | Comparison of within-subject variability (test vs. reference) |
Examples of NTI Drugs
| Drug | Therapeutic Area | NTI Concern |
|---|---|---|
| Warfarin | Anticoagulation | Bleeding risk with supratherapeutic levels |
| Levothyroxine | Thyroid replacement | Hyper/hypothyroidism with dose changes |
| Phenytoin | Antiepileptic | Toxicity with supratherapeutic levels |
| Cyclosporine | Immunosuppression | Rejection or toxicity with dose changes |
| Tacrolimus | Immunosuppression | Nephrotoxicity with supratherapeutic levels |
| Digoxin | Cardiac glycoside | Arrhythmias with supratherapeutic levels |
| Lithium | Psychiatric | Toxicity with supratherapeutic levels |
FDA has published product-specific guidances with tighter BE criteria for several NTI drugs. As of early 2026, these tighter criteria apply to a defined list of products, with additional NTI guidances under development.
Alternative BE Study Approaches
Not all drug products can use standard PK-based BE studies. FDA accepts alternative approaches for specific product categories.
Pharmacodynamic (PD) BE Studies
| Application | When Used | Design |
|---|---|---|
| Locally acting drugs | Blood levels do not reflect local drug action | PD endpoint reflecting drug effect at site of action |
| Examples | Inhaled corticosteroids, nasal sprays, certain topicals | Dose-response curve comparison between test and reference |
Clinical Endpoint BE Studies
| Application | When Used | Design |
|---|---|---|
| Complex products | Neither PK nor PD approaches are feasible | Controlled clinical study comparing efficacy outcomes |
| Examples | Topical dermatologic products, certain ophthalmic products, complex injectables | Randomized, double-blind, parallel or crossover, active-controlled |
In Vitro BE (Beyond BCS Biowaivers)
| Application | When Used | Design |
|---|---|---|
| Solutions | Oral solutions, parenteral solutions with same formulation | Comparative dissolution or physicochemical testing |
| Topical solutions | Same active, same vehicle, same concentration | Formulation comparison |
| Certain gaseous products | Inhaled anesthetic gases | Physicochemical equivalence |
Product-Specific Guidances: The Definitive BE Reference
OGD publishes product-specific guidances (PSGs) that specify the recommended BE study design for individual reference listed drugs. PSGs are the single most important reference for BE study planning.
What PSGs Contain
| PSG Element | Information Provided |
|---|---|
| Study type | In vivo PK, in vitro (biowaiver), PD, clinical endpoint |
| Study design | Crossover, parallel, replicate; number of periods |
| Study conditions | Fasting, fed, or both |
| Dosing | Which strength(s) to study; single or multiple dose |
| Endpoints | AUC, Cmax, and any additional PK parameters |
| Statistical criteria | Standard 80-125% or alternative (e.g., RSABE for HVDs) |
| Population | Healthy volunteers or patients |
| Additional requirements | Dissolution profiles, stability, specific analytical considerations |
Where to Find PSGs
PSGs are published on FDA.gov under "Product-Specific Guidances for Generic Drug Development." As of early 2026, OGD has published over 2,000 PSGs covering the majority of approved oral solid dosage forms and an increasing number of complex products.
If no PSG exists for your target product, request a pre-ANDA product development meeting with OGD to discuss the appropriate BE approach. Designing a study without FDA alignment on methodology is a significant risk, particularly for complex products, locally acting drugs, or products with unusual PK characteristics.
Common BE Study Failures and Prevention
| Failure Mode | Cause | Prevention |
|---|---|---|
| 90% CI outside 80-125% | High variability, insufficient power, formulation issue | Adequate sample size calculation; pilot study; formulation optimization |
| Study not per PSG | Outdated or incorrect PSG referenced | Always use current PSG from FDA.gov |
| Bioanalytical failure | Method validation deficiencies, ISR failure | Follow 2018 FDA bioanalytical guidance rigorously |
| Protocol deviations | Subject compliance, sampling time deviations | Tight clinical site oversight, trained staff |
| Wrong reference product | Used non-RLD or foreign reference product | Verify RLD designation in Orange Book before purchasing |
| Insufficient sampling duration | AUC0-t < 80% of AUC0-inf | Extend sampling to at least 3 half-lives |
| Food effect not assessed | PSG requires fed study but only fasting conducted | Check PSG for fasting/fed requirements |
References
The 80-125% acceptance range is applied to the 90% confidence interval of the geometric mean ratio (test/reference) for log-transformed AUC and Cmax. It does not mean the individual subject ratios must fall within 80-125%. Rather, the statistical procedure (TOST) determines whether the population mean ratio, with 90% confidence, falls entirely within this range. In practice, the average point estimate (geometric mean ratio) is typically very close to 100% (95-105%), but the CI must also fall within bounds.