Bioequivalence and Bioavailability Forum

Hi Louis,

like ElMaestro I’m a little bit confused. Without the additional information he requested we would have to fall back to our crystal ball…
Some remarks/questions from my side:

❝ Two pilot studies (2x2) passed for AUC and Cmax […]

Since both passed, why didn’t you stop and submit them as pivotal evidence of BE? Until very recently the FDA was happy with that (given the usual , like appropriate design and execution, at least 12 sub­jects). Inter­esting that the FDA removed the guidance from its website (last archived 2015-02-26) and left us alone with a draft where such a statement is missing.

Why did you opt for a full replicate design in the pivotal(s)? You wrote that the CV was

❝ ~ 25% in past studies

(and you passed in two 2×2 crossovers). Were you aiming at RSABE as a kind of “safety net”? You have to sub­mit all studies’ data to the FDA. Which impression might a reviewer get? If I play the devil’s advocate:

“They had moderate variability in two pilot studies. Both showed BE. Now they want to convince us (me!) that the drug / drug product is highly variable. Maybe they opted for a cheaper (bad?) CRO in order to cut costs. I know the equation: Bad CRO = lousy study conduct = higher vari­abi­lity. In order to be on the ‘safe side’ they want us (ha, me‼) to accept RSABE. I don’t buy it.”

(The grumpy reviewer grabs the Rejected-stamp from his desk and seals the application)

❝ […] a pivotal study (4x2) failed for both AUC and Cmax (AUC and Cmax ~130%, U95%CI ~1.40).

Oops. Terrible. Which were the PEs (CIs, samples sizes) in the two pilot studies? Did you use different batch(es) of reference and/or test?

❝ CV ~ 35% for both AUC and Cmax in failed study

OK, let’s see: 90% CI ~1.21–1.40. With ~35% CV that would translate into a sample size of 62 subjects in a 4-period full replicate. You could apply RSABE (s_w >0.294), but failed on the PE (above 1.25). Is my guess about the sample size correct? If yes, why was it that high?

❝ Another large scale study passed.

Fine.

❝ Would like to invalidate the failed pivotal study which had significant period effects (AUC and Cmax); […]

Irrelevant in crossovers (unless sequences are extremely [sic] unbalanced).

❝ […] a significant product*period*sequence effect; […]

Here I’m with ElMaestro. Where does this term come from – and what was the intended purpose? The FDA is pretty clear about preferred models for replicate studies.

❝ […] significant product differences were seen in periods 1 & 3, not in 2 & 4; within sequence product differences were significant within period 1 but not in sequence 2;

I would say your model is a bit over-specified. On another note, significant treatment effects are irrelevant in BE. Generally they pop up if the sample size is high for the given variability (the magic “overpowered” studies). Again: Which was your sample size?

❝ CV ~ 35% for both AUC and Cmax in failed study; ~ 25% in past studies.

Shit happens. If you tell us the sample sizes (), we could calculate the respective CIs of the CVs. I bet that they overlap.

❝ I would like to use the second large study as pivotal. How can I do this - do I need to invalidate the pivotol study.

   The combination of some data and an aching desire
for an answer does not ensure that a reasonable answer
can be extracted from a given body of data.     John W. Tukey

By “invalidating” a study you might shoot yourself in the foot – unless you find clear reasons (clinics, bio­analytics). Statistics alone is not sufficient.

The FDA in the past was fine if a failed study was repeated due to low power. Here we have a tricky case. The CV was not the problem (would have be taken care by RSABE), but the PE. Difficult to predict whether the agency will accept your justification along these lines: “Two pilots showed BE. In the pivotal the PE was 1.30. Might have been pure chance. We repeated it.” Duno.
Across the pond (read: EMA) you could run a meta-analysis of the two pivotals. At least one of them must have demonstrated BE. You can try.

“If for a particular formulation at a particular strength multiple studies have been performed some of which demonstrate bioequivalence and some of which do not, the body of evidence must be con­sidered as a whole. […] The existence of a study which demonstrates bioequivalence does not mean that those which do not can be ignored. The applicant should thoroughly discuss the results and justify the claim that bioequivalence has been demonstrated. Alternatively, when relevant, a combined analysis of all studies can be provided in addition to the individual study analyses. It is not acceptable to pool together studies which fail to demonstrate bioequivalence in the absence of a study that does.”