Bioequivalence and Bioavailability Forum

Pharma_88 ☆ India, 2020-08-27 18:07 (1502 d 19:08 ago) Posting: # 21893 Views: 8,007	Design [Regulatives / Guidelines] Post reply
	Dear All, We are thinking to perform steady state BE study for XXX submission where Test product is ER formulation and Reference product is IR formulation. Further, Its HVD product. So, question is whether is this feasible to conduct replicate BE IR vs ER and if yes then what are the parameters for conclusion? Thanks. — Regards, Pharma_88

Helmut
★★★

Vienna, Austria,
2020-08-27 19:32
(1502 d 17:42 ago)

@ Pharma_88
Posting: # 21894
Views: 7,270

Jurisdiction?

Post reply

Hi Pharma_88,

❝ We are thinking to perform steady state BE study for XXX submission…

Unfortunately my crystal ball is in the laundry and reading tea leaves turned out to be an insufficient substitute. ;-)

Hence, which XXX?

❝ … where Test product is ER formulation and Reference product is IR formulation.

<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>

❝ Further, Its HVD product. So, question is whether is this feasible to conduct replicate BE IR vs ER and if yes…

Is it a highly variable drug¹ or are one – or both – products² highly variable? But, in principle, yes.

❝ … then what are the parameters for conclusion?

Once you tell us what XXX is, we can possibly help.

HVDs exhibit highly variable clearances (CV_wR ≥30% if administered as a solution).
HVDPs may additionally – or solely – show highly variable absorption.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Pharma_88
☆

India,
2020-08-31 12:26
(1499 d 00:48 ago)

@ Helmut
Posting: # 21898
Views: 7,184

Jurisdiction?

Post reply

❝ ❝ We are thinking to perform steady state BE study for XXX submission…

❝

❝ Hence, which XXX?

Its for EMEA.

❝ ❝ … where Test product is ER formulation and Reference product is IR formulation.

❝

❝ <nitpick>

</nitpick>

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

❝ ❝ Further, Its HVD product. So, question is whether is this feasible to conduct replicate BE IR vs ER and if yes…

❝

❝ Is it a highly variable drug or are one – or both – products highly variable? But, in principle, yes.

❝

❝ ❝ … then what are the parameters for conclusion?

❝

❝ Once you tell us what XXX is, we can possibly help.

EMEA

—
Regards,
Pharma_88

Helmut
★★★

Vienna, Austria,
2020-08-31 14:21
(1498 d 22:53 ago)

@ Pharma_88
Posting: # 21901
Views: 7,366

EMA: ABEL for HVD(P)s, PK metrics

Post reply

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>

</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 studies³ 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|RTR⁶) and not the partial replicate design (TRR|RTR|RRT). For the pros and cons see this post.

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.
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).
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).
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.
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.
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).

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Pharma_88
☆

India,
2020-09-08 11:05
(1491 d 02:10 ago)

@ Helmut
Posting: # 21910
Views: 7,049

EMA: ABEL for HVD(P)s, PK metrics

Post reply

Thanks For your detailed response.

I have a few questions and need your suggestion in this regards.

I have kept following parameters in SS study (IR VS ER).

Primary pharmacokinetic parameters: AUC0-τ,ss
Secondary pharmacokinetic parameters:
Day 01 to 04-Morning dose: Cpd
Day 04 (Morning dose): Cτ,ss, Cmax,ss, Cmin,ss, Tmax,ss, accumulation index, %Fluctuation and Cav

Please correct me.

and also want to know for Cpd and Cτ,ss:

Cpd: All pre-dose samples needs to be considered including profiling day sample or only samples before profiling day?

Cτ,ss: How to calculate it? Whether morning dose of Day 04 is also need to be taken in to account to calculate.

Thanks.

—
Regards,
Pharma_88

Helmut
★★★

Vienna, Austria,
2020-09-08 19:25
(1490 d 17:50 ago)

@ Pharma_88
Posting: # 21911
Views: 7,196

EMA: MD PK metrics

Post reply

Hi Pharma_88,

❝ I have kept following parameters in SS study (IR VS ER).

PK metrics, pleeze.

❝ Primary pharmacokinetic parameters: AUC0-τ,ss

OK.

❝ Secondary pharmacokinetic parameters:

❝ Day 01 to 04-Morning dose: Cpd

OK. No comparison – just report them.

❝ Day 04 (Morning dose): Cτ,ss, Cmax,ss, Cmin,ss, Tmax,ss, accumulation index, %Fluctuation and Cav

❝

❝ Please correct me.

If you don’t have a lag-time, C_min,ss (the ‘true’ minimum concentration) should be very similar to both C_pd and C_τ,ss. C_min,ss is required only for originators. Since a hybrid application is closer to a generic, you could drop it.

AFAIK, the accumulation index \(\frac{1}{1-\exp (-\lambda_\textrm{z}\cdot \tau)}\) is not required by any agency.
You would need a reliable estimate of \(\small{\lambda_\textrm{z}}\), which you may not obtain for the ER formulation in all subjects. As long as you don’t have flip-flop PK (see below), \(\small{\lambda_\textrm{z}}\) is a property of the drug and therefore, not relevant for comparing products.
C_av is not informative because it is simply AUC_0–τ/τ. If you want you can report it though without a statistical comparison. The outcome will be exactly the same as for AUC_0–τ. C_av is only needed for the calculation of fluctuation.

❝ Cpd: All pre-dose samples needs to be considered including profiling day sample or only samples before profiling day?

You have to collect at least three pre-dose samples in order to demonstrate achievement of steady-state. In your case mornings of days 2–4. See also this presentation (slide 14).

❝ Cτ,ss: How to calculate it? Whether morning dose of Day 04 is also need to be taken in to account to calculate.

Nope. It’s the concentration at the end of the (last) dosing interval or in your case at 96 hours. Too lazy to search but the EMA states somewhere that this sample should be collected with a time deviation of ≤10 minutes. In my studies I always used the estimated concentration at τ.

Let’s explore three examples. One-compartment model, no lag-time, D 100, V 5, t_½,el 14 h, t_½,abs 1 h (IR). Pseudo-steady-state reached after 5 × 14 = 70 h, results for the profile day 4.

Cpd Ctau AUCtau %PTF Cmax ─────────────────────────────────── IR 9.2 9.4 400.0 87 23.9 ER 11.5 11.7 399.0 48 19.8 ─────────────────────────────────── T/R 125% 125% 100% 56% 83%

C_τ,ss > C_pd. In practice this difference likely will not be evident due to variability. If you prefer a belt plus suspenders, dose for another day. As expected the AUCs are practically identical and the ER produces less fluctuation.
Even for a moderately slower rate of absorption it is not possible to waive the MD study because the extrapolated AUC is 45% of AUC_0–∞.
Cpd Ctau AUCtau %PTF Cmax ─────────────────────────────────── IR 9.2 9.4 400.0 87 23.9 ER 13.2 13.6 395.9 28 18.2 ─────────────────────────────────── T/R 144% 145% 99% 32% 76%

Similar but a further reduction in fluctuation.
Cpd Ctau AUCtau %PTF Cmax ─────────────────────────────────── IR 9.2 9.4 400.0 87 23.9 ER 13.6 14.5 377.8 13 16.6 ─────────────────────────────────── T/R 144% 145% 94% 15% 69%

Now we crossed the border of flip-flop PK (k_abs ≤ k_el) which leads to trouble. Absorption of the ER is the slowest process and hence, steady-state is not reached.

In any case to hope to demonstrate equivalence of C_max and C_τ is futile. Either concentrate on PTF (one-sided test for non-superiority) and/or consider a bracketing approach (see this post).

Since you mentioned before that you have to deal with a HVD(P), you can expand the limits for some PK metrics (see this post) if the study will be performed in a replicate design. You don’t need a treatment-free washout, i.e., the built-up of a subsequent steady states can overlap with the washout from the previous one. Example for a full replicate design, where T = ER and R = IR. Profile days in red.

Sequence/Day 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 T T T T T R R R R R T T T T T 2 R R R R R T T T T T R R R R R

No fun.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Design [Regulatives / Guidelines]

Jurisdiction?

Jurisdiction?

EMA: ABEL for HVD(P)s, PK metrics

EMA: ABEL for HVD(P)s, PK metrics

EMA: MD PK metrics