Log in
Register|
VK ☆ Russia, 2013-07-17 12:35 (4079 d 15:29 ago) Posting: # 10991 Views: 9,954 |
|
|
Dear HS and colleagues, I am a newcomer, I work in BE bioanalytics lab in BRIC country (right btw Brazil and India). We recently had a BE project (single dose, two-way crossover) with a cytostatic which is prescribed based on body surface area. Thus c.a. 70% of the sample size received Х mg of either R/T in each period, while 30% received 0.75Х mg or 1.25Х mg. We did regular calculations however authorities require different approach. They want us to use only X mg data for statistical evaluation, saying that we cannot use 30% because of the different dose. I am currently trying to think of argumentation persuading authorities to accept calculations based on the whole data set. Unfortunately I cannot refer to common sense and basic statistical considerations that is why I need more like a reference to a GL or an article. I would greatly appreciate if you could help me with few references if any. Sincerely, VK |
|
Dr_Dan ★★ Germany, 2013-07-17 13:09 (4079 d 14:56 ago) @ VK Posting: # 10993 Views: 9,081 |
|
|
Dear VK What do you mean by "We did regular calculations"? In a cross-over study BE is based on the intra-individual comparison of test and reference and therefore it is theoretically possible to administer different doses (dose adjusted PK parameters). However, I guess in your case you administered a combination of different strengths (=different products), right? For example: 70% received 1x100mg 30% received either 1x50mg+1x25mg or 1x100mg+1x25mg Then indeed only 70% of the subjects could be evaluated since a pure comparison is only possible between your 100 mg product and the originator. If you had choosen 70% received 4x25mg 30% received either 3x25mg or 5x25mg Then 100% of the subjects could be evaluated. Kind regards Dan — Kind regards and have a nice day Dr_Dan |
|
VK ☆ Russia, 2013-07-17 18:17 (4079 d 09:47 ago) @ Dr_Dan Posting: # 11000 Views: 9,028 |
|
|
Dear Dr_Dan, Thank you for the prompt reply. Please let me comment below: ❝ What do you mean by "We did regular calculations"? We calculated PK parameters (Cmax, AUClast and Cmax/AUClast), correspodning 90% CIs and ANOVA (as required in Russia) for 100% of patients. ❝ If you had choosen ❝ 70% received 4x25mg ❝ 30% received either 3x25mg or 5x25mg ❝ Then 100% of the subjects could be evaluated. This is our case. To shed more light on the design - 70% receieved 4*500mg, while 30% received either 3*500mg or 5*500 mg. Hence the only strength was 500, while the doses were different. So the regulatory body experts tell us that we should use 70% for the calculation because the received one dose. That is why I looking for some kind of GL/textbook I could refer our experts. Cheers, VK |
|
Helmut ★★★   Vienna, Austria, 2013-07-17 14:18 (4079 d 13:47 ago) @ VK Posting: # 10994 Views: 8,906 |
|
|
Hi Vasily, welcome to the club.  ❝ I am currently trying to think of argumentation persuading authorities to accept calculations based on the whole data set. Unfortunately I cannot refer to common sense and basic statistical considerations that is why I need more like a reference to a GL or an article. Agree with what Dan wrote about different strenghts. Since I’m on my way to the airport, no references (I’m not too optimistic to find anything in GLs). Out of curiosity: What did you specify in the protocol – were there any comments by the authority and/or IEC in approval? — 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 |
|
VK ☆ Russia, 2013-07-17 18:31 (4079 d 09:33 ago) @ Helmut Posting: # 11002 Views: 8,869 |
|
|
Dear Helmut, ❝ welcome to the club. Thanks! ❝ Out of curiosity: What did you specify in the protocol – were there any comments by the authority and/or IEC in approval? Nothing was stated in the protocol wrt to data analysis and there were no comments from the authority in the submission stage. However during the BE study final report review the expert was surprized by the study design: "how come that in BE study subjects receive different amount of drug. You should re-design and re-do the study in such a way thay everybody receive the same amount of drug". We said that it is not ethical since some subjects will get more drug than needed (risk of AEs), some subjects will get less drug (risk of tumor resistance). We also provided a link to the FDA capecitabine GL. Experts said that ok, the design is now acceptable and commented that we have to analyze the subset of the sample with the similar amount of drug taken. I understand that this expert's position does not have strong scientific background, however obviously I cannot tell it to the expert. That is why I need a kind of reference I could present to the expert explaining the situation. Sincerely, Vasily |
|
Helmut ★★★ Vienna, Austria, 2013-07-18 01:17 (4079 d 02:47 ago) @ VK Posting: # 11011 Views: 8,832 |
|
|
Hi Vasily! ❝ We said that it is not ethical since some subjects will get more drug than needed (risk of AEs), some subjects will get less drug (risk of tumor resistance). Exactly. What else? ❝ We also provided a link to the FDA capecitabine GL. Nice. ❝ Experts said that ok, the design is now acceptable and commented that we have to analyze the subset of the sample with the similar amount of drug taken. Shit. ❝ I understand that this expert's position does not have strong scientific background, however obviously I cannot tell it to the expert. You have to say it in a polite way.  The expert should not loose his/her face. Ask ElMaestro (our expert for clever wordings) for help. Since you used only one strength I don’t see a problem. The capecitabine guidance is a nice reference.— 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 |
|
luvblooms ★★ India, 2013-07-18 08:42 (4078 d 19:23 ago) @ VK Posting: # 11013 Views: 8,998 |
|
|
Dear Vk and Others Good morning!! Just a thought!! why cant we use all the volunteers data and add the different dose based on volunteers BSA while calculating the the PK parameters (While doing NCA in Model → Dosing Regimen)?  We have done same in Capacetabiene for US and ANVISA submission and never got a query. I may be wrong but this idea is worth pursuing! Check how comfortable your data looks like with this method!! — ~A happy Soul~ |
|
VK ☆ Russia, 2013-07-18 17:28 (4078 d 10:37 ago) @ luvblooms Posting: # 11020 Views: 8,843 |
|
|
Dear Luvblooms, Thank you for the excellent advice: ❝ why cant we use all the volunteers data and add the different dose based on volunteers BSA while calculating the the PK parameters (While doing NCA in Model → Dosing Regimen)? I am now updating the final report with comments wrt the amount of drug. Additionally I am trying to include administered dose into ANOVA part. Sincerely, Vasily |
|
kumarnaidu ★ Mumbai, India, 2013-11-21 15:00 (3952 d 12:05 ago) @ VK Posting: # 11935 Views: 8,241 |
|
|
Hi all, I got one protocol for review. In that they have given some different design which I want to share with you. They are having two cohorts. In cohort 1 15 patients will receive T (5 subjects in each dose category) and 15 will receive R T1, T2 and T3 are different doses. In Cohort 2 - Subject received T in cohort 1 will receive R and Subject received R will receive T (T1 or T2 or T3 based on some set criteria ) Cohort 1 (n = 15) T1 (n1 = 5) T2 (n2 = 5) T3 (n3 = 5) R (n = 15) Cohort 2 (n = 15) R(n = 15) T1 or T2 or T3 (n = 15) Bioequivalence will be evaluated in cohort 2. And cohort one will be used for dose identification. According to me we need to use parallel design here for analysis in cohort 2 but what about period and sequence effect.  — Kumar Naidu |
|
Dr_Dan ★★ Germany, 2013-11-21 16:23 (3952 d 10:41 ago) @ kumarnaidu Posting: # 11937 Views: 8,140 |
|
|
Dear kumarnaidu You are talking about a two period cross-over study with 4 treatments, right? I am not a statistician but I guess it would be reasonable to divide R also into 3 groups with 5 subjects each so that you have an intra-individual comparison T1 vs R1, T2 vs R2 and T3 vs R3. If possible I would prefer a 4 way cross over design, three times test one reference with 5 subjects each. Bioequivalence: Two medicinal products are bioequivalent if they are pharmaceutically equivalent and if their bioavailabilities after administration in the same molar dose are similar to such degree that their effects, with respect to both efficacy and safety, will be essentially the same. I guess in your case we are talking about comparative bioavailabilities and the test products are no pure generic formulations, right? Kind regards Dan — Kind regards and have a nice day Dr_Dan |
|
kumarnaidu ★ Mumbai, India, 2013-11-22 06:34 (3951 d 20:31 ago) @ Dr_Dan Posting: # 11940 Views: 8,135 |
|
|
Dear Dr_Dan, Thanks for your prompt reply. ❝ I guess in your case we are talking about comparative bioavailabilities and the test products are no pure generic formulations, right? — Kumar Naidu |
Ing. Helmut Schütz