Bioequivalence and Bioavailability Forum

Hi Pharma_88!

❝ Its for EMEA.

The EMEA (European Agency for the Evaluation of Medicinal Products) was renamed to EMA (European Medicines Agency) fifteen years ago.

❝ ❝ <nitpick>

Not BE, but comparative BA. You can only hope for similar extent of absorption but never for rate of absorption in a comparison of ER vs IR. Furthermore, BE means similarity in PK metrics of interest if equimolar doses are administered. Sometimes one has to increase the ER dose to get similar AUCs… Whatever is applicable in your case, BE is the wrong term.

</nitpick>

❝

❝ Can you please elaborate to understand. If you can provide me any supportive literature to go in detail its really helps me.

Very nitpicking: Even if release characteristics of products are identical (say, IR vs IR, MR vs MR) there are no “BE studies”. Bioequivalence is the desired outcome of a comparative BA study. One assessor of Poland regularly did not accept study protocols which mentioned BE in the title. His response was essentially:

“If you already know the outcome, why do you want to perform a study at all?”

But to reiterate: Bioequivalence means similar rate and extent of absorption after equimolar doses. The PK metrics for rate are generally C_max and for extent AUC_0–t. For modified release products AUC_0–∞ is required as well (contrary to the FDA where all three are always required). Depending on the product characteristics additional PK metrics are required. ¹
When comparing MR to IR, the rate of absorption might be similar (i.e., for delayed release products, where only t_max will be different) or different (i.e., for prolonged / controlled / extended / sustained release products).
In the latter case, this is a desired property:

• Less fluctuations in steady state (which was called »the flatter is better« ² in the late 1980s).
  
• Enhanced compliance (o.a.d. instead of b.i.d. or even t.a.d.).

❝ ❝ ❝ … then what are the parameters for conclusion?

See the applicable MR guideline, Section 5. Note that if you want to compare a new MR product to an established IR product, this is not a generic application acc. to Directive 2001/83/EC, Article 10(2)(b) but a hybrid application. For the required PK studies ³ see Section 5.1 and for the therapeutic studies Section 5.2 (only if you are extremely lucky, they might be waived). For the required PK metrics see Section 6.8.2.
In case of high variability, reference-scaling according to the EMA’s ABEL (Average Bioequivalence with Expanding Limits) is acceptable for the following PK metrics:

• Single dose studies
  C_max and C_τ.
  
  
• Multiple dose studies
  C_max,ss and C_τ,ss. ^4,5

In case of multiphasic release products additionally partial AUCs, C_max, and C_max,ss in all phases (where the cut-off times have to be pre-specified).

For details about ABEL see the IR-GL Section 4.1.10 and the Q&A document Section 8. Note that in order to plan for ABEL, you have to give a sound justification that expanding the common BE limits is clinically not relevant (contrary to RSABE for the U.S. FDA and China’s CDE/NMPA; don’t forget this step).
I recommend a full replicate design (4 periods TRTR|RTRT or 3 periods TRT|RTR ⁶) and not the partial replicate design (TRR|RTR|RRT). For the pros and cons see this post.

---

1. 1. Prolonged release products
      Since quite often we have flip-flop PK (k_a < k_el) the late part of the profile represents absorption. For the same reason the truncated AUC_0–72 (as for IR products) is not acceptable.
      Forget the “option” to waive the multiple dose study if AUC_0–τ ≤10% of AUC_0–∞. I didn’t see a single case where it worked.
      
   2. Delayed release products
      Multiple dose study not required.
      
   3. Multiphasic modified release products
      Partial AUCs and C_max in all phases.
      If t_½ is short (say, ≤4 h), waiving the multiple dose study based on AUC_0–τ ≤10% of AUC_0–∞ generally works.
2. Not that simple. Sometimes rapid onset of effect is important as well, which lead to development of multiphasic products with IR- and prolonged release-components (zolpidem, methylphenidate, dexamfetamine).
   
3. A lot. Not only comparative BA but food-effect, dose proportionality,
   Additionally between-subject variability of the MR product should be compared with that of the IR product by a one-sided test (non-superiority).
   
4. If you want to expand the limits only for C_τ,ss (which is more variable than C_max,ss), it is not necessary to sample two profiles. Since you are in (pseudo-) steady state, use the concentrations pre-dose and at τ.
   
5. Instead of testing C_max,ss and C_τ,ss for equivalence, you can also use “bracketing”, i.e., one-sided tests: C_max,ss for safety (non-inferiority) and C_τ,ss for efficacy (non-superiority). For examples see the respective vignette of the R-package PowerTOST.
   
6. Although only the 4 period full replicate and the partial replicate designs are given as examples in Section 8 of the Q&A document, Section 19 made clear that the 3 period full replicate design is acceptable as long as at least twelve eligible subjects are in sequence RTR. However, that’s not relevant in practice (see this post).