CH 20 · notes
How it works
FDA's August 1997 guidance on dissolution testing of immediate-release solid oral dosage forms compares two mean profiles with a difference factor, f1, and a similarity factor, f2. The guidance cites Moore and Flanner, Pharmaceutical Technology 1996. f1 = 100 × Σ|R − T| / ΣR. f2 = 50 × log10(100 / √(1 + Σ(R − T)² / n)). R is the reference mean percent dissolved, T is the test mean at the same time, and n is the number of time points in the comparison. Log is base 10. This page reports f1 and f2 half-up to 2 decimal places and compares the unrounded values with the limits.
FDA 1997 says f1 values up to 15 (0–15) and f2 values of 50–100 generally ensure sameness of the two curves. The same guidance says the method is most suitable when three to four or more time points are available, that only one measurement should be considered after 85% dissolution of both products, and that the measurements should be made under the same conditions at the same times. To use mean data, the percent coefficient of variation should be no more than 20% at earlier time points (the guidance gives 15 minutes as the example) and no more than 10% at other time points. Profiles are determined on 12 units of each product. This page does not calculate %CV.
EMA's guideline on the investigation of bioequivalence (CPMP/EWP/QWP/1401/98 Rev. 1/Corr **) uses the same f2 equation in Appendix I. An f2 between 50 and 100 suggests the profiles are similar. The evaluation uses at least three time points (zero excluded), the same times for both products, and twelve individual values at each time. Not more than one mean value of > 85% dissolved is used for any formulation. A mean of exactly 85% is not above 85%, so it does not end the EMA series. The FDA cutoff instead keeps points through the first time both means are at least 85%, and drops later points.
For immediate-release products, both documents allow profiles that are more than 85% dissolved within 15 minutes to be treated as similar without an f2 calculation. When the entered means meet that description, this page says so and still reports f1 and f2. Content uniformity is a different calculation: see the USP 905 AV calculator. Percent relative standard deviation of replicates is on the RSD calculator. Last checked 2026-10-31.
Not a substitute for regulatory judgment. Confirm the dissolution protocol and the current FDA or EMA text.
Worked example
Load example uses the six time points in Diaz et al., AAPS J. 2016;18(1):15–22, Table IV (adapted from a WHO workshop): reference 20, 40, 60, 75, 86, 90 and test 7, 33, 39, 67, 77, 87 at 10, 15, 20, 30, 45, and 60 min. That paper reports f2 = 47 for these six points, which is the nearest integer of the unrounded value. This page shows f1 16.44 and f2 46.62. Both means are at least 85% only at 60 min, so the default FDA cutoff keeps all six points. The same paper reports f2 = 57 when the protocol uses only 15, 30, 45, and 60 min.
Sources
- FDA Dissolution Testing of IR Solid Oral Dosage Forms (Aug 1997) — f1 and f2; f1 0–15 and f2 50–100; one point after both products reach 85%; %CV limits for using means
- EMA Guideline on the Investigation of Bioequivalence (CPMP/EWP/QWP/1401/98 Rev. 1) — Appendix I f2; at least three time points with zero excluded; not more than one mean >85% for any formulation; f2 between 50 and 100
FAQ
What are the dissolution difference factor f1 and similarity factor f2?
FDA's August 1997 immediate-release dissolution guidance defines f1 as the percent difference between two mean curves: f1 = 100 × Σ|R − T| / ΣR. f2 is a base-10 log transformation of the sum of squared differences: f2 = 50 × log10(100 / √(1 + Σ(R − T)² / n)). R is the reference mean percent dissolved, T is the test mean at the same time, and n is the number of time points in the comparison. The guidance cites Moore and Flanner, Pharmaceutical Technology 1996.
When are two dissolution profiles similar?
FDA 1997 says f1 values up to 15 (0–15) and f2 values of 50–100 generally ensure sameness of the two curves. EMA's bioequivalence guideline says an f2 between 50 and 100 suggests the profiles are similar. This page calls the profiles similar only when the unrounded f1 is in 0–15, the unrounded f2 is in 50–100, and at least three time points are compared. Reported values are half-up to 2 decimal places.
Which time points are included?
Time zero is excluded by default, because EMA requires at least three time points with zero excluded. The FDA 1997 cutoff keeps earlier points and only one point after both means are at least 85% dissolved. The EMA cutoff keeps points through the first time either mean is above 85%; a mean of exactly 85% does not end that series. You can also compare every complete time point. Each row needs a time, a reference percent, and a test percent. Unequal or partial rows are not paired.
Does this replace a regulatory decision?
No. Not a substitute for regulatory judgment. Confirm the dissolution protocol and the current FDA or EMA text. FDA 1997 recommends 12 units of each product and says mean data are appropriate when the percent coefficient of variation is not more than 20% at earlier time points and not more than 10% at other time points. This page does not calculate %CV. Values were last checked 2026-10-31.