Log in
Register|
balakotu ★ India, 2012-01-24 10:44 (5252 d 01:08 ago) Posting: # 7989 Views: 9,859 |
|
|
Dear all, Please Clearify any one. How to calculate Sample Size by Using SAS Program for Partial and full replicate study designs for Europe and USA Regulatory. Europe 90% CI Wider limits USA 95% Upper Confidence Bound. Canada? What procedure can accept Canada Regulatory for partial and fully replicate study designs. Regards Kotu. Edit: Category changed. [Helmut] |
|
d_labes ★★★ Berlin, Germany, 2012-01-24 11:49 (5252 d 00:02 ago) @ balakotu Posting: # 7990 Views: 8,435 |
|
|
Dear balakotu, ❝ How to calculate Sample Size by Using SAS Program for Partial and full replicate study designs for Europe and USA Regulatory. I assume that you would go for scaled ABE with these designs. AFAIK there is no out of the box solution if you insist in using ![[image]](img/uploaded/image2.gif) .To cite the both Laszlo's [1]: "Overall, the statistical properties of the methods proposed by EMA and FDA are rather complex as a result of the additional conditions and requirements (mixed procedure, GMR constraint, and (for EMA) a cap on the limits). Furthermore, the tests required by both EMA and FDA are dependent on each other which makes the theoretical treatment very complicated. Therefore, the required sample sizes were obtained by simulations." (emphasis by me) So the only way to calculate the sample size is to implement the simulation of replicate crossover design data in SAS, implement the statistical evaluation methods of scaled ABE according to the regulatory body and vary the sample size until your simulations with that sample size give you your desired power. I suspect that implementation in SAS is a hard job and I suspect further that the run time is tremendous. There must be a reason that the Laszlo's had implemented theirs in Matlab or earlier in Fortran. Thus I don't see a great chance that anybody will come out with SAS code to accomplish this task  .At the moment you have to stick with the sample size tables given in the paper [1] of the two Laszlo's. If you won't go with scABE but with conventional ABE I recommend you to have a look at R-package PowerTOST. It should not so hard to implement the methods used in that package using the undocumented SAS function OwenQ(t,delta,0,R,df).[1] Laszlo Tothfalusi and Laszlo Endrenyi "Sample Sizes for Designing Bioequivalence Studies for Highly Variable Drugs" J Pharm Pharmaceut Sci (www.cspsCanada.org) 15(1) 73 - 84, 2011 — Regards, Detlew |
|
balakotu ★ India, 2012-01-24 12:17 (5251 d 23:35 ago) @ d_labes Posting: # 7992 Views: 8,436 |
|
|
Dear d_labes, Thanking you for your response Already i have seen that document. suppose i got cv 32.5 and t/r ratio is 97%. how to decided the sample size for further study. Europe follows 90% CI Wider limits, USA 95% Upper Confidence Bound. What approach Canada Regulatory follows? Regards Kotu |
|
d_labes ★★★ Berlin, Germany, 2012-01-24 13:13 (5251 d 22:39 ago) @ balakotu Posting: # 7993 Views: 8,490 |
|
|
Dear Kotu, ❝ ... suppose i got cv 32.5 and t/r ratio is 97%. how to decided the sample size for further study. Since the sample size tables do not contain your values I would go with a conservative approach and use CV=35% and GMR=0.95. That would give N= 3-period 4-periodif your desired power is 80% according to the tables in the paper. If you are not satisfied with this approach interpolate in the tables. ❝ What approach Canada Regulatory follows? Did you had a look into the Guidance documents the Forum's Admin has collected gratifyingly for You and all of us  ?AFAIK the current in effect guidance doesn't contain scaled ABE. But there is no Acceptance range for the CI of Cmax in the guidance. Only the point estimator (GMR of T/R) has to be in the range 80% - 125%. The Draft Guidance Document "Comparative Bioavailability Standards: Formulations used for Systemic Effects" states: "7. Drugs with highly variable pharmacokinetics: For the purpose of bioequivalence testing, there is no compelling need for a distinct category of "highly variable" drugs, given that there is sufficient permitted flexibility in study design to address exceptional cases. Notwithstanding the potential need for relatively large numbers of subjects in some bioequivalence studies, the current requirements do not present an unreasonable barrier to product approval. Furthermore, the ethical concern surrounding the exposure of a relatively large number of healthy subjects to study drugs does not outweigh the potential risk of exposing the patient population to a bio-inequivalent drug." (emphasis by me) Thus it seems scaled ABE will not be an issue for the Canadians also in the future. You have to plan your study based on AUC as usual for conventional ABE, given your pilot estimate of the GMR for Cmax is within 80-125%, regardless of whichever design you use. — Regards, Detlew |
|
Helmut ★★★   Vienna, Austria, 2012-01-24 14:26 (5251 d 21:25 ago) @ balakotu Posting: # 7994 Views: 8,317 |
|
|
Dear Kotu! ❝ suppose i got cv 32.5 and t/r ratio is 97%. how to decided the sample size for further study. Some remarks following D. Labes’ post. HVDs/HVDPs show also fluctuating PEs across studies. Quoting the two Lászlós: Larger absolute differences between the two logarithmic means can be noted in the various BE studies when the within-subject variation is higher. Therefore, it is recommended that a 10% deviation between the means, i.e. a true GMR = 1.10, be considered when the sample size tables in the Appendix are used. Since power curves are not symmetrical in raw scale I would recommend the more conservative 0.90 instead of 1.10 – which would give for CV 35% and 80% power: 3-period 4-periodSince you expect a CV of 32.5% it might well be the case that in the actual study the reference does not classify as a HVD/HVDP – which would change the regulatory requirement to unscaled ABE! CV% 3-period 4-period— Dif-tor heh smusma 🖖🏼 Довге життя Україна! ![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png) Helmut Schütz ![[image]](https://static.bebac.at/img/CC by.png) The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
Ing. Helmut Schütz