A similarity factor (F2) is commonly described in the drug dissolution testing area to reflect the similarity of two dissolution profiles by a number i.e., if the number is between 50 and 100, then the two profiles are considered similar. The question that arises is in what respect are they similar; do numerically similar profiles show similar dissolution characteristics in the human GI tract- a commonly understood implication. No, they do not, making this implication faulty. This is because the numbers are usually based on results obtained using apparatuses, and experimental conditions, which have never been qualified and validated for dissolution testing purposes. Therefore, the similarity or dissimilarity of in vitro-in vivo profiles based on an F2 value has no meaning or relevance. Hence, presenting it as a useful parameter for bio-relevance is certainly a faulty fact.
It is often suggested, and in fact strongly promoted, that drug dissolution testing is a quality control tool or an aid during the development of a product. This in itself is also a faulty fact. Drug dissolution testing by itself, without its in vivo link, has no meaning because both applications (QC and aid in product development) are derived from in vivo relevance, such as mandatory use of bio-relevant experimental conditions (e.g. 37 ºC, aqueous buffers etc.).
The similarity factor (F2) does not have any added value because a number between 50 and 100 reflects an average difference of dissolution 10% or less. Therefore, by definition, a quick way to establish the similarity of the profiles is to calculate the average differences at different dissolution sampling times. If the value is less than 10% then, the curves meet the similarity (or F2) criteria. Derivation of the F2 value can often add biases and/or errors, e.g., one is restricted to use only one data point beyond 85% drug release. One is required to have two or more dissolution points to be able to calculate F2. However, a 10% difference would be easier to use and applicable irrespective to product type (fast release vs slow release) or the number of points/results available.
The range of 50 to 100 is not in line with current pharmacopeial requirements, even for QC purposes where a Q-based tolerance of 80% certainly allows a difference of 15 to 20% for products to have similar dissolution characteristics. However, F2 approach allows only differences of less than 10%.
Mathematically, the formula or calculations for the F2 does not appear to be more than a fancy skill-testing question, such as “(2 × 4) + (10 × 3)” (Answer: 38) (link) or (8 x 6 – 5 + 9=52, link). The point being, it is a sort of arbitrary arithmetic exercise without any scientific relevance or value. However, it certainly adds a burden on to the resources and interpretation.
In short, a similarity factor (F2) may be considered as not a very useful parameter which can lead to erroneous interpretation. The approach based on different criteria (e.g., 10%) offers perhaps a simpler, logical, and more robust approach for assessing the similarity or dissimilarity of the dissolution curves or results.