Dissolution Profile f2 Similarity Factor: Worked Examples
Contents
- What f₂ measures
- Check suitability before calculating
- Calculation steps
- Worked example 1: identical means
- Worked example 2: different means
- When f₂ may be unsuitable
- Common mistakes
- Related reading
- Frequently asked questions
- Does an f₂ result above 50 prove that two products are bioequivalent?
- Can I calculate f₂ from two published mean curves?
- Can I use the same time points for every product?
- Does this calculation set a routine dissolution release limit?
- References
In short: The f₂ similarity factor compares the mean dissolution results of a test product and a reference product at matched sampling times. It is useful only when the study design and data meet the applicable conditions; an f₂ result by itself does not prove bioequivalence or authorise a batch decision.
What f₂ measures
The f₂ factor is a model-independent summary of the distance between two dissolution profiles. It uses the mean percentage dissolved for each product at each matched time point. Identical mean profiles produce a value of 100; larger differences lower the value. FDA’s August 1997 guidance presents f₂ for comparing profiles of immediate-release solid oral dosage forms, and EMA’s 2010 bioequivalence guideline describes its use in Appendix I.
The formula is:
f₂ = 50 × log₁₀{[1 + (1/n) Σ(Rₜ − Tₜ)²]⁻⁰·⁵ × 100}
| Symbol | Meaning |
|---|---|
| n | Number of matched, non-zero sampling time points included in the calculation. |
| Rₜ | Mean percentage of the reference product dissolved at time t. |
| Tₜ | Mean percentage of the test product dissolved at time t. |
| Σ(Rₜ − Tₜ)² | Sum of squared differences between the two means at the included time points. |
Use the same sampling times and test conditions for both profiles. The calculation is based on the mean at each time, but the individual unit results are still needed to assess variability and confirm whether the method is suitable. FDA’s 1997 guidance and EMA’s 2010 guideline set out similar but not identical conditions, so do not combine them into a universal rule.
Check suitability before calculating
The sources below describe a framework for immediate-release products and specific comparison purposes. Confirm which current guideline, protocol, product-specific instruction, and regulatory context apply before using an f₂ result for a submission or quality decision.
Under FDA’s Dissolution Testing of Immediate Release Solid Oral Dosage Forms guidance (August 1997, section V.A), the profile comparison uses twelve units each for the test and reference products, matched time points and conditions, and generally three to four or more time points. FDA also recommends considering only one measurement after both products exceed 85% dissolved and limits the coefficient of variation for using mean data to 20% at early points and 10% at later points.
EMA’s Guideline on the Investigation of Bioequivalence, CPMP/EWP/QWP/1401/98 Rev. 1 (2010), Appendix I, describes at least three non-zero matched time points, twelve individual values at each time point for each formulation, no more than one mean value above 85% for any formulation, and relative standard deviation or coefficient of variation below 20% at the first point and below 10% from the second point onward. It says an f₂ between 50 and 100 suggests similar profiles and allows justified alternative methods when f₂ is unsuitable.
These conditions are source-specific recommendations, not universal batch-release limits. EMA states that ICH M13A took effect on 25 January 2025 and supersedes applicable parts of the older EMA guideline for non-replicate immediate-release bioequivalence study design and data analysis; topics not addressed by M13A continue under the relevant EMA guidance. This article teaches the calculation and does not determine which approach applies to a particular study.
Calculation steps
- Confirm the comparison. Identify the test and reference products and the purpose of the comparison. Confirm that both profiles were generated under the same method and conditions.
- Check the raw data. Use the individual unit results to calculate each product’s mean and variability at every time point. Investigate transcription errors, missing results, or unusual variation before summarising.
- Choose eligible time points. Follow the applicable protocol and guidance. Keep the same non-zero time points for both profiles and apply that source’s rules for the late points after high dissolution.
- Calculate the difference at each point. Subtract the test mean from the reference mean and square the difference.
- Average the squared differences. Add them and divide by n, the number of included time points.
- Apply the formula and interpret cautiously. A value above 50 is commonly used in the cited guidance as an indication of similar profiles under the stated assumptions. It is not a stand-alone finding of bioequivalence or product quality.
Worked example 1: identical means
The following fictional mean values are only for arithmetic practice. They do not define a test method, acceptance criterion, or product specification.
| Matched point | Reference mean (%) | Test mean (%) | Difference | Squared difference |
|---|---|---|---|---|
| A | 30 | 30 | 0 | 0 |
| B | 55 | 55 | 0 | 0 |
| C | 75 | 75 | 0 | 0 |
| D | 90 | 90 | 0 | 0 |
The squared differences sum to 0. With n = 4, the term inside the brackets is 1. Therefore, f₂ = 50 × log₁₀(100) = 100. This is the maximum value and reflects identical means in this illustration. It says nothing about unit-to-unit variability, method suitability, or the representativeness of the batches.
Worked example 2: different means
Assume the fictional data below came from a comparison whose individual results met the selected method’s variability conditions. That assumption is only for practising the arithmetic; the table alone cannot demonstrate it.
| Matched point | Reference mean (%) | Test mean (%) | Difference | Squared difference |
|---|---|---|---|---|
| A | 30 | 27 | 3 | 9 |
| B | 55 | 50 | 5 | 25 |
| C | 75 | 69 | 6 | 36 |
| D | 90 | 86 | 4 | 16 |
The squared differences total 86. Divide by four time points: 86 ÷ 4 = 21.5. Add 1, take the reciprocal square root, multiply by 100, then take log₁₀ and multiply by 50:
f₂ = 50 × log₁₀[100 ÷ √(1 + 21.5)] ≈ 66.2.
The result is above 50, which is consistent with the cited similarity framework if the comparison also meets its design and variability assumptions. It is not proof that every unit is equivalent or that any regulatory requirement has been met.
When f₂ may be unsuitable
Do not rely on the simple factor when the source’s time-point or variability conditions are not met. High variation can make the summary unstable; EMA Rev. 1 permits a statistically valid and justified alternative when f₂ is unsuitable. A suitable analysis may require a statistician and a predefined method. Do not choose a different method only after seeing which result gives the preferred conclusion.
The factor also compresses the whole profile into one number. It cannot explain why curves differ, detect every local pattern, or replace examination of individual results. Plot the profiles and inspect both the means and the spread of unit-level data. Keep the protocol, method, and approved product criteria as the controlling documents for the actual work.
Common mistakes
- Comparing different time points. The formula assumes matched time points; use the same selected times for both profiles.
- Using only the mean and ignoring variation. The mean is an input, but the individual unit data determine whether the approach is suitable.
- Including too many late points. Apply the relevant source’s rule for time points after high dissolution rather than adding every late observation automatically.
- Calling f₂ above 50 proof of bioequivalence. The result is a profile-comparison statistic under specified assumptions, not a clinical or regulatory conclusion by itself.
- Mixing guidance frameworks. FDA’s 1997 and EMA’s Rev. 1 recommendations are similar in places but differ in wording and context. State which protocol or guidance was used.
- Rounding intermediate values. Carry the calculation through before rounding the displayed result; retain the underlying profile data and calculation record.
Related reading
For equipment operation, see the dissolution test apparatus SOP. The process validation guide discusses process consistency, while the failure investigation SOP gives a framework for documenting an unexpected result. For analytical data handling, see the good chromatography practices SOP.
Frequently asked questions
Does an f₂ result above 50 prove that two products are bioequivalent?
No. FDA and EMA describe f₂ as a method for comparing dissolution profiles in specified contexts. Bioequivalence is a broader conclusion and depends on the applicable study design, regulatory requirements, and complete evidence.
Can I calculate f₂ from two published mean curves?
You can practise the arithmetic if both curves report the same time points, but means alone do not show unit-level variability. Check the original study data and the selected framework before interpreting a result.
Can I use the same time points for every product?
No universal schedule applies to every product. Use a justified, predefined protocol and the current product-specific or authority guidance for the intended comparison.
Does this calculation set a routine dissolution release limit?
No. A profile-comparison statistic does not replace an approved release specification or method. Follow the product’s current specification and applicable procedure.
References
- U.S. Food and Drug Administration, Dissolution Testing of Immediate Release Solid Oral Dosage Forms, Guidance for Industry, August 1997, section V.A, “Model Independent Approach Using a Similarity Factor.” Official FDA guidance page · Official FDA PDF.
- European Medicines Agency, Guideline on the Investigation of Bioequivalence, CPMP/EWP/QWP/1401/98 Rev. 1, adopted 20 January 2010, effective 1 August 2010, Appendix I. EMA lists it as the current effective version and notes that applicable parts have been superseded by ICH M13A. Official EMA PDF · EMA version and history page.
- European Medicines Agency, Considerations regarding the implementation of ICH M13A on bioequivalence for immediate-release solid oral dosage forms, EMA/531548/2024, adopted 17 February 2025; ICH M13A effective in the EU from 25 January 2025. Official EMA PDF.
Related articles
How to Read an Indian Pharmacopoeia Monograph
Learn to read an Indian Pharmacopoeia monograph: check the edition and General Notices, follow the test method, and record findings without guessing limits.
Visual Inspection of Parenterals: Defects and Records
Learn how to classify parenteral visual-inspection defects, qualify operators, record batch results and trend rejects using current compendial and GMP guidance.
Pharmaceutical Analysis Practical Viva for Students
Prepare for a pharmaceutical analysis practical viva with a clear answer framework, instrument examples, common questions and safe revision tips for students.
Pharma Calculations: Dilution, Concentration and % Strength
Practise pharmaceutical calculations for dilution, concentration and percentage strength with 18 worked laboratory problems, unit conversions and clear answers.