Log in
Register|
myy ☆ Germany, 2014-03-07 17:19 (4081 d 23:03 ago) Posting: # 12575 Views: 11,507 |
|
|
Sorry for this question but I am totally new to the field of bioequivalence and I have the following case: A single dose, two sequence and two period- crossover bioequivalence study with 32 subjects fails to show bioequivalence because the criteria for AUC 0-t were not met (76.62-102.75%) but Cmax is clearly within the limits (81.71-109.23%). T/R of Cmax was 94.47% and T/R of AUC 0-t was 88.73%. CV% was 42.89% for Cmax and 46.99% for AUC 0-t of the test product. The dissolution profiles have already revealed a tendency of slower dissolution rates of the test formulation although statistically from the f2 values similarity was concluded. Since the test product fails to meet only the lower end of the criteria for AUC 0-t it is not truely bioinequivalent. Could it be possible that bioinequivalence was demonstrated accidentally although the two products are actually bioequivalent? How could this be proven? In the study protocol I have read: “A formal statistical calculation of the study’s power was not carried out (given the scarce literature on intraindividual variability and confidence intervals for ivermectin).” Is it therefore really only possible to guess the sample size? Would it be advisable to perform another bioequivalence study with the same design? Edit: Category changed. [Helmut] |
|
Helmut ★★★   Vienna, Austria, 2014-03-07 17:47 (4081 d 22:36 ago) @ myy Posting: # 12576 Views: 10,341 |
|
|
Hi myy, ❝ The dissolution profiles have already revealed a tendency of slower dissolution rates of the test formulation although statistically from the f2 values similarity was concluded. Unfortunately in vivo sometimes does not match in vitro. ❝ Since the test product fails to meet only the lower end of the criteria for AUC 0-t it is not truely bioinequivalent. Could it be possible that bioinequivalence was demonstrated accidentally although the two products are actually bioequivalent? How could this be proven? Don’t worry, it is too late. SCNR. I’m not a friend of a posteriori power, but even Cmax passed by luck. With your ratios and CVs power is just 32.2% for Cmax and 14.6% for AUC. ❝ In the study protocol I have read: “A formal statistical calculation of the study’s power was not carried out (given the scarce literature on intraindividual variability and confidence intervals for ivermectin).” Is it therefore really only possible to guess the sample size? Well, to perform a study “out of the blue” is against the BE-GL Section 4.1.3 (“appropriate sample size calculation”) and ICH-E9. It is surprising that both the IEC and the regulatory authority (may I ask which one?) approved the protocol. If the CV is unknown, you should have performed a pilot study. ❝ Would it be advisable to perform another bioequivalence study with the same design? No. If I take your ratio and CV for AUC as “carved from stone” (cave: there is no guarantee that you will be able to get exactly these numbers in the new study) you would need 232 (!) subjects to show BE with 80% power. — 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 |
|
myy ☆ Germany, 2014-03-08 22:32 (4080 d 17:50 ago) @ Helmut Posting: # 12581 Views: 10,282 |
|
|
❝ Well, to perform a study “out of the blue” is against the BE-GL Section 4.1.3 (“appropriate sample size calculation”) and ICH-E9. It is surprising that both the IEC and the regulatory authority (may I ask which one?) approved the protocol. The Spanish authority. |
|
ElMaestro ★★★ Denmark, 2014-03-07 18:20 (4081 d 22:03 ago) @ myy Posting: # 12577 Views: 10,194 |
|
|
Hi myy, ❝ A single dose, two sequence and two period- crossover bioequivalence study with 32 subjects fails to show bioequivalence because the criteria for AUC 0-t were not met (76.62-102.75%) but Cmax is clearly within the limits (81.71-109.23%). T/R of Cmax was 94.47% and T/R of AUC 0-t was 88.73%. CV% was 42.89% for Cmax and 46.99% for AUC 0-t of the test product. The relevant CV is not for the T or R product but for the GMR. I assume that's what you meant? Anyways, a CV for Cmax lower than CV for AUCt is quite unusual and I always question it. Please take a look at the production of data rather than just at the data. ❝ Since the test product fails to meet only the lower end of the criteria for AUC 0-t it is not truely bioinequivalent. Could it be possible that bioinequivalence was demonstrated accidentally although the two products are Careful with the terminology. Bioinequivalence was not shown. Your product could be BE, it has just not been demonstrated. Bioinequivalence is only if the CI is entirely outside the acceptance range. ❝ Would it be advisable to perform another bioequivalence study with the same design? Advisable? Perhaps. Justifiable? Yes, probably. But you might wish to take a new look at the sample size with your max likelihood PEs and observed CVs..... after review of data production. — Pass or fail! ElMaestro |
|
nobody nothing 2014-03-07 19:02 (4081 d 21:21 ago) @ ElMaestro Posting: # 12578 Views: 10,208 |
|
|
Ivermectin is kind of special compound, I guess (regarding chemical structure etc.), look here http://www.accessdata.fda.gov/drugsatfda_docs/nda/96/050742ap.pdf According to p.191 and 198 CV% for AUC might well be higher than for Cmax, at least in some cases... Regards — Kindest regards, nobody |
|
myy ☆ Germany, 2014-03-08 22:40 (4080 d 17:43 ago) @ ElMaestro Posting: # 12582 Views: 10,025 |
|
|
❝ Since the test product fails to meet only the lower end of the criteria for AUC 0-t it is not truely bioinequivalent. Could it be possible that bioinequivalence was demonstrated accidentally although the two products are ❝ actually bioequivalent? How could this be proven? ❝ ❝ Careful with the terminology. Bioinequivalence was not shown. Your product could be BE, it has just not been demonstrated. Bioinequivalence is only if the CI is entirely outside the acceptance range. Could this also be argued in front of regulatory authorities? |
|
Dr_Dan ★★ Germany, 2014-03-10 18:36 (4078 d 21:47 ago) @ myy Posting: # 12584 Views: 10,041 |
|
|
❝ Could this also be argued in front of regulatory authorities? Kind regards Dr_Dan — Kind regards and have a nice day Dr_Dan |
|
Dr_Dan ★★ Germany, 2014-03-07 20:21 (4081 d 20:01 ago) @ myy Posting: # 12579 Views: 10,155 |
|
|
Dear myy In order to possibly give you some advice more information would be needed. I guess the study was conducted in humans, right? And the AUC was truncated due to the long terminal half life of the drug? Comparative dissolution showed f2 values >50 at all three pH levels (1.2, 4.5 and 6.8)? The high ANOVA CV was not caused by outlying subjects? Before you start reformulating your product you need to be sure that the study outcome is valid. Kind regards Dr_Dan — Kind regards and have a nice day Dr_Dan |
|
myy ☆ Germany, 2014-03-08 23:04 (4080 d 17:18 ago) @ Dr_Dan Posting: # 12583 Views: 10,107 |
|
|
❝ In order to possibly give you some advice more information would be needed. I guess the study was conducted in humans, right? Yes. ❝ And the AUC was truncated due to the long terminal half life of the drug? AUC 0-i was measured as well. And the values were about the same as for AUC 0-t. (76.76-102.52) ❝ Comparative dissolution showed f2 values >50 at all three pH levels (1.2, 4.5 and 6.8)? Because of difficulties with degradation of ivermectin at acidic pH the drug release rate was not measured at pH 1.2. But also at pH 7.0. ❝ The high ANOVA CV was not caused by outlying subjects? I couldn't find any data about outlier detection. Maybe this was not carried out? ❝ Before you start reformulating your product you need to be sure that the study outcome is valid. I have to assess the study from a regulatory point of view and give a strategic advice. But it seems to me that the study should be reviewed by a competent statistician first... Thank you very much for all your comments! |
|
Dr_Dan ★★ Germany, 2014-03-10 18:44 (4078 d 21:39 ago) @ myy Posting: # 12585 Views: 9,994 |
|
|
❝ Because of difficulties with degradation of ivermectin at acidic pH the drug release rate was not measured at pH 1.2. But also at pH 7.0. ❝ Kind regards Dr_Dan — Kind regards and have a nice day Dr_Dan |
Ing. Helmut Schütz