Compliance Regular India, 20140313 09:43 Posting: # 12612 Views: 7,776 

Dear All, We would like to submit dossier of generic modified release formulation in Europe and we are not cleared about following statement of guideline. The European guidance line no. 593 to 595 states as follows: "A multiple dose study is needed unless a single dose study has been performed with the highest strength which has demonstrated that the mean AUC(0τ) after the first dose covers more than 90% of mean AUC(0∞) for both test and reference, and consequently a low extent of accumulation is expected". In this if you read they are asking to see AUC(0τ) after single dose which confuse us as after single dose administration how this parameter can be evaluated. I think here to cases are cubed which ultimately resulting into the confusion. with addition to this the entire section is not delivering correct understanding to me hence if possible can any one make it simple it to get optimum understanding. regards, Compliance Edit: In 2005 (!) EMEA (European Agency for the Evaluation of Medicinal Products) was renamed to EMA (European Medicines Agency). [Helmut] 
nobody Senior 20140313 09:52 @ Compliance Posting: # 12613 Views: 7,106 

...don't see any problems in calculating AUCtau, could you specify your problem a little more in detail?! — Kindest regards, nobody 
kumarnaidu Regular Mumbai, India, 20141127 11:02 (edited by kumarnaidu on 20141127 12:00) @ fno Posting: # 13928 Views: 6,577 

Dear All, We did BE studies (single dose fast and fed,multiple dose fed) for metformin ER tablets 1000 mg and want to ask Can steady state studies marginally failing at Cτ,ss 80.19 (72.8 to 88.34) with ISCV=23.1) be clinically justified for regulatory acceptance if the single dose fast and fed studies are passing for all the parameters including pAUCs? — Kumar Naidu 
nobody Senior 20141127 11:37 @ kumarnaidu Posting: # 13929 Views: 6,505 

The guideline allows widening for Ctau,ss as for highly variable drugs. Did you go for that in the protocol? Otherwise... — Kindest regards, nobody 
kumarnaidu Regular Mumbai, India, 20141127 12:53 @ nobody Posting: # 13930 Views: 6,536 

» The guideline allows widening for Ctau,ss as for highly variable drugs. Did you go for that in the protocol? Ok thanks for the reply. But I think for applying widening criteria we need to show CV>30% for reference product and the study should be reference replicate (3way or 4way). Can we use Predose concentration and Ctau of steady state to show variability? — Kumar Naidu 
Helmut Hero Vienna, Austria, 20141127 13:47 @ kumarnaidu Posting: # 13932 Views: 6,581 

Hi Kumar, » But I think for applying widening criteria we need to show CV>30% for reference product and the study should be reference replicate (3way or 4way). Correct. I wouldn’t be too optimistic with the CV_{intra} of 23.1% you reported. » Can we use Predose concentration and Ctau of steady state to show variability? This idea was discussed in Bonn last year. Members of EMA’s PKWP seemed to be happy with it. I guess the MRGL was adopted at EMA’s CHMPmeeting last week. Let’s wait for it (and the comments). The analysis is a little bit tricky* because the replicates originate from the same period.
Edit: The MRGL was adopted indeed. See here. Edit: GL, Section 6.8.2.2: Any widening of the acceptance criteria for C_{max} should follow the recommendations on highly variable drug products in the Guideline on the Investigation of Bioequivalence (CPMP/EWP/QWP/1401/98). — Cheers, Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. ☼ Science Quotes 
kumarnaidu Regular Mumbai, India, 20141127 15:12 @ Helmut Posting: # 13937 Views: 6,457 

Thanks Helmut for your valuable comment. — Kumar Naidu 
kumarnaidu Regular Mumbai, India, 20141205 06:03 @ Helmut Posting: # 13979 Views: 6,300 

Hi Helmut thanks for the comment. » Calculation of the intrasubject variability in multiple dose studies can be based on two consecutive administrations of the same product after reaching steady state. That means we have to use the 4 way (Fully replicate) method here but with a different sequence RRTT or TTRR — Kumar Naidu 
kumarnaidu Regular Mumbai, India, 20141210 11:15 @ kumarnaidu Posting: # 14044 Views: 6,032 

Hi All, I am little bit confused with this statement. In all other cases, where accumulation is likely (AUC(0τ) after the first dose covers less than 90% of mean AUC(0∞)) a multiple dose study is required. Generally, steadystate studies should be performed under the conditions concerning concomitant food intake recommended in the SmPC for the originator product. If for any drug concomittant food intake is reccomended and it shows no accumulation (AUC(0τ) after the first dose covers less than 90% of mean AUC(0∞)) then in that case should we go for steady state or not? — Kumar Naidu 
nobody Senior 20141210 17:04 @ kumarnaidu Posting: # 14048 Views: 6,031 

Hi! Read the red sentence again, and again, and again... You will understand it... Otherwise show it to a colleague... — Kindest regards, nobody 
Dr_Dan Senior 20141210 17:21 @ kumarnaidu Posting: # 14049 Views: 6,088 

Dear kumarnaidu » That means we have to use the 4 way (Fully replicate) method here but with a different sequence RRTT or TTRR No, you do not need to use a 4 period replicate design: You will have 2 periods and 2 sequences, you just need to dose until you reach steady state, take then the PK profile for the dosing interval and after an afreshed dosing for the consecutive interval. e.g.: 1st dosing drug1 Kind regards Dr_Dan — Kind regards and have a nice day Dr_Dan 
Helmut Hero Vienna, Austria, 20141210 18:40 @ Dr_Dan Posting: # 14051 Views: 6,034 

Dear Dan, I think that your example is a fully replicate design (i.e., RRTTTTRR) indeed. Just substitute drug1 with R and drug2 with T and you’ll get the first of the two sequences from above. Of course you have to modify the code given in EMA’s Q&A (for RTRTTRTR) accordingly. But that’s an easy task.This concept would work for all PK metrics. Coming back to the subject line (and which was discussed in Bonn): Scaling only C_{min} without the need of sampling four full profiles. treatment dose On profile days you have two values of C_{min}. One prior to dosing and another one at τ. Now the stats become interesting. You would need to introduce an additional code (e.g., “replicate”) to distinguish these two values because they are sampled from the same period(s). I’m working on it. EMA’s preference of an allfixed model is nasty. My current setup is: Sequence+Subject(Sequence)+Treatment+Replicate+Treatment(Replicate)+Period PS @Anders: Sorry for the nesting. I know, I know. — Cheers, Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. ☼ Science Quotes 
d_labes Hero Berlin, Germany, 20141211 13:09 @ Helmut Posting: # 14053 Views: 5,931 

Dear Helmut, » ... I’m working on it. EMA’s preference of an allfixed model is nasty. My current setup is: Sequence+Subject(Sequence)+Treatment+Replicate+Treatment(Replicate)+Period That's weird. Two times Treatment in the model . — Regards, Detlew 
Helmut Hero Vienna, Austria, 20141211 14:53 @ d_labes Posting: # 14056 Views: 5,959 

Dear Detlew, » » ... I’m working on it. EMA’s preference of an allfixed model is nasty. My current setup is: Sequence+Subject(Sequence)+Treatment+Replicate+Treatment(Replicate)+Period » » That's weird. Two times Treatment in the model . Dammit! I’m not sure how to reproduce Liu’s results anyhow. Work in progress. — Cheers, Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. ☼ Science Quotes 
VStus Regular Poland, 20170203 13:04 (edited by VStus on 20170203 15:50) @ Helmut Posting: # 17005 Views: 3,497 

Dear Helmut, Have you solved this puzzle? I must confess that Liu's article is too hard for me. I understand that the idea of scaling is based on assessment of withinsubject variability after the administration of only Reference formulation. But article is targeting repeated crossover studies (RRTTTTRR) for assessment of population/individual bioequivalence. Due to this, a more complicated statistical model was proposed and justified. But we are not interested in population bioequivalence now. Would it be correct to analyze only a reference subset of data to estimate the variability of reference PK parameter only? Let's assume I'm interested only in possibility of widening of acceptance limits for Ctau_ss, do I need test formulation data at all? We will not have different formulations in resulting subset, just factors of period + sequence + subject + repetition. I have reduced Liu's dataset to have: subject  20 levels (120) sequence  2 levels (RT and TR) period  2 levels (1 or 2) repetition  2 levels (1 or 2) AUC for Reference formulation. Which gives me:
anova(lm(log(AUC) ~ prd + seq + rep + subj:seq, data=data, na.action=na.exclude)) Residual MS is higher than reported by Lui for Reference formulation ('MS(R) = 0.03094'), thus it seems that I'm overestimating the variability of the Reference in this simplified model. Thank you very much in advance! Regards, VStus Sorry for confusion: this is not a replicate design 
zizou Junior Plzeň, Czech Republic, 20170203 13:49 (edited by zizou on 20170203 18:47) @ VStus Posting: # 17006 Views: 3,478 

Dear VStus, it looks like you reduced it too much.
When you reduce period for 2 levels  I think the change makes period to be equal to replication. Period should still have 4 levels (1 or 2 or 3 or 4). If I am not mistaken. Edit1: My mistake I don't have a data/reference and I thought the data was from 4 periods. After repeated reading I noticed that there were two periods only with created "repetition" term. I am sorry, I was too fast with replaying and I lacked the quality. Edit2: If I get it right (I hope so) there is something like this scheme: period: 1 2 When we are interested only in R (per1+seq1 and per2+seq2 are not used at all). So in evaluation of only R data, Period 1 is equal to Sequence 2 (in the term of categorizing data) and Period 2 is equal to Sequence 1, so it's also known from subject No, when the data was obtaind (period 1 or 2). So period doesn't have any new information (the information is provided in other terms). You may try: anova(lm(log(AUC) ~ seq + rep + subj:seq, data=data, na.action=na.exclude)) (I'm curious if the results differ against the results with period in.) Best regards, zizou 
VStus Regular Poland, 20170203 15:37 @ zizou Posting: # 17008 Views: 3,454 

Dear zizou, Thank you very much for the feedback! I've tried 4 periods instead of 2 and got even higher variation, which seems to be an overestimation:
summary(data) Removing replication from the model does not change residual MS :) Regards, VStus 
VStus Regular Poland, 20170206 08:55 @ zizou Posting: # 17017 Views: 3,305 

Dear zizou, No change in residual variance if I remove 'period' from the model.
anovadata < lm(log(AUC) ~ prd + seq + rep + subj:seq, data=data, na.action=na.exclude) Maybe MIXED procedure should be used? Regards, VStus 