8 replies to this topic

[Angus McLean]

I have been working with an extended release formulation and I have performed NCA on the data set. After doing my best to extract to infinity extrapolated AUC information I have decided not to use the information and go with to AUC last as the total exposure metric. It is not possible to get reliable elimination rate constants (and half lives) from the profiles, since late absorption is interfering with the elimination phase.

 

 

So my question is how do I tell the computer not to calculate a slope and an elimination rate constant and AUC 0-INF for these profiles. I only need AUC last for these profiles.

 

There is an immediate release treatment in the study so I can calculate the slopes e.t.c for the immediate release profiles in the study.

 

 Angus

[Simon Davis]

Angus,

Note that calculating AUCinf won't affect the calculation AUClast so you could just ignore those values.

There are a couple of ways to do this but what I normally do is simply set LzLower and Lz Upper to 99 and 100 respectrively (or some other pair of points outside the studied time range)

ALternatively I would send the Final Parameters Pivoted results to Data Wizard and there write a transformation to find and filter those AUCinfs I considered not to have been assessed over a sufficient interval e.g. where AUC%inf was greater than 25%

Simon.

Attached Thumbnails

• 

[Helmut Schütz]

Hi Simon!

There are a couple of ways to do this but what I normally do is simply set LzLower and Lz Upper to 99 and 100 respectrively (or some other pair of points outside the studied time range)

Hhm, why don't you simply check the box [X] Disable Curve Stripping and disable both (noninformative) plots?

@Angus: No regulatory assessor would accept using the estimated lambda-z of one formulation for another. If you are dealing with controlled release you might fall into the trap of flip-flop PK. What you ‘see’ in the later part of the profile is actually absorption. I would be wary to compare CR with IR based on AUClast. Contrary to the FDA the EMA requires only AUClast for IR formulations, but according to the recent draft they will require AUCinf for controlled (EMA’s term: prolonged) release formulations. Makes sense.

[Angus McLean]

Thanks Helmut: MR formulations has been an activity of mine for many years. here is the position I take. When comparing with an IR formulation there are obviously situations where you see that late absorption process are interfering significantly with achieving terminal elimination phase. As such you have to judge whether you have a proper terminal elimination phase with the MR formulation. Often you do not so you should not calculate the terminal elimination phase and the dependent variables e..g AUC to inf. Instead I go with AUC to t and make sure (upfront) that the clinical protocol has late enough times to describe the exposure. I do that consistently with the MR formulations in the study, but for the IR formulation I do estimate elimination rate and AUC to inf.

 

So how do you feel about that?

 

Angus

[Helmut Schütz]

Hi Angus!

So how do you feel about that?

Mixed feelings.

In BE of IR vs. IR we can be pretty sure that the slowest phase is elimination. Absorption is essentially completed after 2–4×t_max. That’s why for EMA AUC_t is the only metric for extent of absorption. Alternatively, you can truncate the AUC at 72 hours (contrary to the FDA independent from half-life). The 72 hours are based on GI-studies showing this cut-off as the maximum ever observed gastric transit time. Never understood why the FDA additionally to AUC_t requires AUC_∞. Both metrics are highly correlated – only AUC_∞ is more variable. IMHO, it is hypocritical to require both. Based on the Intersection-Union tests I cannot imagine a study passing AUC_∞ and failing on AUC_t. So why all that fuzz?

MR is a different story (leaving pulsatile / multiphasic / delayed release aside). It might be possible that you cross the Rubicon of k_a=k_el (flip-flop). Remember that in BE we compare extent/rate of absorption. IMHO, once the slowest phase represents absorption you cannot truncate the AUC anymore (regardless whether at t_last or at 72 hours) – otherwise (if absorptions rates differ) your T/R-ratios will be biased. Therefore – as you said – we should sample long enough to reliably describe absorption. Generally estimation of λ_z will not be difficult then. Even if we fail in getting an estimate, long sampling should protect us against bias. Just my two cents.

BTW, in their recent draft on BE of MR formulations EMA added AUC_∞. They are considering whether this will be the only metric for extended (EMA’s term: prolonged) release, AUC₇₂ is not acceptable. Maybe a (much) later truncation time point will be. Let’s see…

[Angus McLean]

Thanks 99-100 works very well................quite often youn cannot get a reliable,estimate of AUC to inf from the data yopu ha ve.

 

 

Angus

[Angus McLean]

Thanks Helmut: some formulations I have worked on are still absorbing after ~20 hours so it is indeed necessary to extend the the samping time to go out further and get a longer AUC to t.

 

When comparing the half lives of the MR fromulations in a study (as you are aware) we see that the true elimination phase has very often not been achieved. One sees ridiculous half lives quoted for drugs such as methyl phenidate. Really the most reliable metric is AUC to t. The trouble is it is more expensive to implement such a study so peole are not inclined to do that.

 

Here in the US I see that people do quote AUC inf values routinely for MR formulations that are obviously not correct. Thye are trying to have a uniform way of treating the data.

 

Angus

[Helmut Schütz]

Hi Angus,

When comparing the half lives of the MR fromulations in a study we see that the true elimination phase has very often not been achieved. One sees ridiculous half lives quoted for drugs such as methyl phenidate.

When it comes to CR/ER/SR what we “see” in the later part of the profile is not elimination but absorption if k_a < k_el.

Really the most reliable metric is AUC to t.

…only if you measure long enough. ;-)

Here in the US I see that people do quote AUC inf values routinely for MR formulations that are obviously not correct.

Well, in the US you have to show BE both for AUC_t and AUC_∞. What do you mean by “not correct”? Staying with MPH: Elimination half life is ~two hours, but the apparent (!) half life for OROS might be as long as five hours since the formulation is still pumping out the API until it ends up in the toilet bowl.

[Angus McLean]

Helmut: I agree that the samping time must be long enough. If you regard the terminal elimination phase as the phase where only elimination is going on then there will be many situations where you cannot quote a ke. The late absoprtion processs interfere with this determination. The predicament becomes "when do you measure and quote an AUC to inf and when not to" . I do not have an answer to that question except to use scientific judgement for each profile.

 

Most people in the US will tend to quote AUC to t and AUC to inf for all subjects if they can. I myself would like to go with a longer sampling time (e.g. 48 hours for Methyl phenidate) and use AUC to t.exclusively ss the metric of choice.

 

 

Angus