Bioequivalence and Bioavailability Forum • CV from previous study

rasheed
☆

2010-04-08 10:53
(5916 d 12:27 ago)

Posting: # 5034
Views: 7,746

CV from previous study [Power / Sample Size]

It is recommended by the FDA that we can use coefficient of variance (CV) from previous study to calculate the smaple size for the new study, but I cant understand that which PK parameter (AUC(0-t), AUC(0-inf), Cmax) is used from previous study for the sample size of the new study? because I encountered large difference between CV's for AUC and Cmax from previous study.

Edit: Please see the Policy. [Helmut]

bharat
☆

India,
2010-04-08 15:57
(5916 d 07:23 ago)

@ rasheed
Posting: # 5040
Views: 6,742

CV from previous study

Post reply

Dear Rasheed
You should no use the CV. It is Intra Subject CV(ISCV) :-)

Kindly use the ISCV for the parameter which is larger.

Ex:                 ISCV

    Cmax            17%

    AUC(0-t)        14%

    AUC(0-inf)      11%

Then You shoud consider ISCV of Cmax for your Sample size calculations.

Regards
Bharat

Edit: Full quote removed. Please delete anything from the text of the original poster which is not necessary in understanding your answer; see also this post! [Helmut]

d_labes ★★★ Berlin, Germany, 2010-04-08 16:23 (5916 d 06:58 ago) @ bharat Posting: # 5042 Views: 6,470	CV nick name Post reply
	Dear Bharat, ❝ You should no use the CV. It is Intra Subject CV(ISCV) How would you abbreviate inter-subject CV? BTW: Of course you are right to use intra-subject CV if we talk cross-over. Parallel group study is another story. — Regards, Detlew

Helmut
★★★

Vienna, Austria,
2010-04-08 16:55
(5916 d 06:25 ago)

@ d_labes
Posting: # 5044
Views: 6,558

CV nick name(s)

Post reply

Dear D Labes!

❝ How would you abbreviate inter-subject CV? :-P

What I have seen in the literature:

intra (within) subject CV: CV_intra, CV_w, CV_e
inter (between) subject CV: CV_inter, CV_b, CV_s
total (pooled) CV: CV_total, CV_t, CV_p

—
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

Helmut
★★★

Vienna, Austria,
2010-04-08 17:34
(5916 d 05:46 ago)

@ rasheed
Posting: # 5045
Views: 6,521

Sample size based on largest CV for AR

Post reply

Dear Rashed,

if the FDA is concerned, the acceptance range for all metrics is 80%-125%.*
As Bharat already pointed out, base your sample size estimation on the metric with the largest CV_intra.
CV_intras are quite often ranked C_max > AUC_inf > AUC_t.

In other countries different rules may be applicable. If a widened acceptance range (AR) of C_max is allowed (e.g., 75%-133%) plan your study accordingly.
Example (T/R 95%, 80% power), CV_intra AUC 26%, C_max 33%:
AUC: n=30 (AR 80%-125%)
C_max: n=46 (AR 80%-125%), n=26 (AR 75%-133%)

For the FDA you would go with n=46 (based on C_max), but if the widened AR is acceptable you would go with n=30 (based on AUC).

* Unless in product-specific guidances scaling is allowed.

P.S.: Please see the Policy for future posts.

—
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

yjlee168
★★★

Kaohsiung, Taiwan,
2010-04-08 23:44
(5915 d 23:37 ago)

@ Helmut
Posting: # 5055
Views: 6,535

any non BE resulting from AUCs?

Post reply

Dear Helmut,

❝ As Bharat already pointed out, base your sample size

❝ estimation on the metric with the largest CV_intra.

❝ CV_intras are quite often ranked C_max > AUC_inf > AUC_t.

Completely agree with you. In regulatory aspect, we need to look at C_max, AUC_t & AUC_inf. However, in my experiences, all non BE cases mostly result from C_max. Have you seen any non BE because of AUC_t or AUC_inf in your experiences? Thanks.

—
All the best,
-- Yung-jin Lee
bear v2.9.6:- created by Hsin-ya Lee & Yung-jin Lee
Kaohsiung, Taiwan https://www.pkpd168.com/bear
Download link (updated) -> here

Helmut
★★★

Vienna, Austria,
2010-04-09 01:48
(5915 d 21:32 ago)

@ yjlee168
Posting: # 5057
Views: 6,529

non BE resulting from AUCs: rarely

Post reply

Dear bears!

❝ In regulatory aspect, we need to look at C_max, AUC_t & AUC_inf.

Well – in the US, yes. Or did Taiwan’s agency copy & paste here again? Would be even more crazy, since AUC₇₂ is acceptable for them.
AUC_inf (in BE – not in BA!) is stupid, because absorption is over, the T/R-ratio is generally very similar to AUC_t (only variability is higher). Both metrics are highly correlated with difficult multiplicity issues, etc., etc., ad libitum, ad nauseam. See also:

Phillips K. Power for Testing Multiple Instances of the Two One-Sided Tests Procedure.
Int J Biostat. 2009; 5(1): Art. 15. doi:10.2202/1557-4679.1169.

❝ […] non BE cases mostly result from C_max. Have you seen any non BE because of AUC_t or AUC_inf in your experiences?

Yes.
Yes. But these studies mainly failed in C_max as well. Or do you mean studies passing C_max and failing in AUC? I don't keep a database of CIs of my studies, but in the EU the situation was different - especially more than 10 years back: sponsors claimed an acceptance range of 70%-143% for C_max – so it was more likely for a given sample size – even with the higher variability of C_max – a study to fail on AUC.

See also slides of the "Two Lázlós" (Meeting on Highly Variable Drugs of the Advisory Committee for Pharmaceutical Sciences, FDA 2004), where they presented data by Diane Potvin / MDS Pharma on 1300 BE studies. 96 failed on C_max and 61 on AUC.

—
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