drmuneesh
★    

2008-10-18 11:17
(6088 d 05:45 ago)

Posting: # 2553
Views: 9,191
 

 sample size for BA/BE studies [Regulatives / Guidelines]

Post reply
Dear all
I would like to know the globally accepted formula used sample size calculation for Bioequivalence studies. Which are the parameters should be available for the calculation of the sample size? Are there any valid software available to do so? In case we have only the data like Mean +/- S.D. then how to calculate N? What will be the minimum N to be considered for pivotal studies? Kindly provide me detailed answer for my clarification on sample size calculation. Thanks to everybody in advance...

With kind regards

Muneesh
 
shankar
★    

India,
2008-10-18 13:47
(6088 d 03:16 ago)

@ drmuneesh
Posting: # 2554
Views: 8,298
 

 sample size for BA/BE studies

Post reply
Dear Dr.Muneesh,

I think no hard and fast rules for calculating sample size in particular method. Whatever method you are following it should be justifiable and you have to follow standard SOP/s. Now a day's many software available in market eg SAS.

As per FDA recommendation minimum N size to be 12 and the sample size will depend on variability of the study drug/s.

I hope my answer will help you.
—
Shankaranarayanan
 
drmuneesh
★    

2008-10-18 14:18
(6088 d 02:44 ago)

@ shankar
Posting: # 2555
Views: 8,302
 

 sample size for BA/BE studies

Post reply
Dear Shankar
Thanks for the answer but SAS can be used when we have Intra-subject CV and T/R ratio obtained from the previous pilot study. If we have not done the pilot study and we want to use the published literature data where we usually get Mean +/- SD, in that case what is the globally accepted formula to give proper sample size. This was my query.

Thanks

Dr Muneesh

--
Edit: Full quote removed. Please see this post! [Jaime]
 
Ohlbe
★★★

France,
2008-10-18 15:32
(6088 d 01:31 ago)

(edited on 2008-10-18 17:18)
@ drmuneesh
Posting: # 2556
Views: 8,269
 

 sample size for BA/BE studies

Post reply
Dear Dr Munesh,

The Intra-subject CV is needed to calculate the sample size, and there is not much you can do without it. The T/R ratio can be just an assumption.

A minimum sample size of 12 subjects is needed for submissions in the EU.

Regards
Ohlbe

--
Edit: See also one of Helmut's posts. [Jaime]
 
drmuneesh
★    

2008-10-19 12:42
(6087 d 04:20 ago)

@ Ohlbe
Posting: # 2558
Views: 8,194
 

 sample size for BA/BE studies

Post reply
Dear Ohlbe

Thanks very much for the reply. I'will get back to you

Regards
Dr Muneesh

--
Edit: Full quote removed. Please see this post! [Helmut]
 
drmuneesh
★    

2008-11-17 15:10
(6058 d 00:53 ago)

@ Ohlbe
Posting: # 2692
Views: 8,126
 

 sample size for BA/BE studies

Post reply
Dear Ohlbe

In continuation to your reply, we have done a pilot study with 11 subjects and results from the pilot study are as follows:

intra subject CV: 0.166203

%T/R Ratio for Cmax: 115.89

and we want to calculate sample size for pivotal study. Now my query is that should we consider only the Intra-subject CV or both the Intra-subject CV and T/R ratio observed from the pilot study or the expected T/R ratio (0.95-1.05) to calculate the sample size.

Because if we consider both the Intra-subject CV and observed T/R ratio, sample size comes much more than what is coming from considering only the Intra-subject CV.

Please guide us what is the globally accepted method of calculating sample size.

Best Regards

Dr Muneesh
 
ElMaestro
★★★

Denmark,
2008-11-17 15:32
(6058 d 00:31 ago)

@ drmuneesh
Posting: # 2694
Views: 8,006
 

 sample size for BA/BE studies

Post reply
Hehe,
Dear dr. muneesh,
I think your situation is sort of potentially borderline nasty.
If the true T/R (CV) is close to the value you have observed then 'there might be an issue'.
Could I ask, how do your formulations perform comparatively in the dissolution assay (for the batches you have evaluated in the human pilot study) ???

Best regards
EM.
 
drmuneesh
★    

2008-11-18 11:34
(6057 d 04:29 ago)

@ ElMaestro
Posting: # 2710
Views: 7,945
 

 sample size for BA/BE studies

Post reply
Dear EM

Thanks very much for the reply. I'will get back to you with dissolution assay

Regards

Dr Muneesh Garg
 
drmuneesh
★    

2008-11-28 08:50
(6047 d 07:13 ago)

(edited on 2008-11-28 10:07)
@ ElMaestro
Posting: # 2827
Views: 7,742
 

 sample size for BA/BE studies

Post reply
Dear EM

As required, the Comparative dissolution profile of the drug is as follows:
Apparatus: USP Type II (Paddle)

Medium: 0.1% SLS in Water

       0 min. 10 min. 15 min. 20 min. 30 min. 45 min. 60 min. 90 min.
Test   0      85      89      89      89      89      88      88
Ref    0      45      56      62      71      77      81      86

Medium: 0.1% SLS in Acetate buffer PH 4.5

       0 min. 10 min. 15 min. 20 min. 30 min. 45 min. 60 min. 90 min.
Test   0      92      91      98      98      97      96      98
Ref    0      48      63      68      75      83      86      89

Medium: 0.1% SLS in 0.1N HCL

       0 min. 10 min. 20 min. 30 min. 45 min. 60 min. 90 min.
Test   0      94      95      94      93      92      91
Ref    0      65      81      86      87      88      90

Regards
Dr Muneesh Garg

--
Edit: Formatted using BBCode. [Jaime]
 
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2008-11-17 16:47
(6057 d 23:15 ago)

@ drmuneesh
Posting: # 2695
Views: 8,504
 

 sample size for BA/BE studies

Post reply
Dear Muneesh!

❝ […] we have done a pilot study with 11 subjects and results from the pilot study are as follows:

❝ intra subject CV: 0.166203 %T/R Ratio for Cmax: 115.89

❝ Now my query is that should we consider only the Intra-subject CV or both the Intra-subject CV and T/R ratio observed from the pilot study or the expected T/R ratio (0.95-1.05) to calculate the sample size.


If you want to talk about an expected deviation of 5%, it's always better to go with 1-0.05, rather than with 1+0.05. In other words you get the same power for T/R-ratios of 1-0.05=0.95 and (1-0.05)-1=1.053.
Why do you expect a deviation of 5% after a pilot study where you have already observed almost 16%?
Not only the point estimate, but also the CVintra obtained in the pilot study carry some degree of uncertainty.

❝ Because if we consider both the Intra-subject CV and observed T/R ratio, sample size comes much more than what is coming from considering only the Intra-subject CV.


So if we take both numbers as carved in stone - which they aren't - we get n=62, right?
If you believe in a T/R-ratio of 0.95 you'll get n=14, but if you calculate an upper 80% confidence limit of CVintra you'll get 0.21596 and therefore n=24.
How will you convince an Ethics Committe to run a pivotal study based on a sample size estimation based on believes rather than your pilot study's results? With the actual PE from your pilot study you'll get n=104.
So that's a nice demonstration about the steepness of power curves if moving away from unity and the uncertainty of CVintra from small pilot studies.
I would reformulate.

❝ Please guide us what is the globally accepted method of calculating sample size.


Sorry, there's none. See the Guidance page for examples. As an example in the European Draft Guideline in the section 'Number of subjects' detailed specifications in the current NfG (variance, α, expected T/R ratio, power) are replaced by a laconic 'The number of subjects to be included in the study should be based on an appropriate sample size calculation.'
I see!
—
Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
 
drmuneesh
★    

2008-11-18 08:44
(6057 d 07:19 ago)

@ Helmut
Posting: # 2701
Views: 8,039
 

 sample size for BA/BE studies

Post reply
Dear HS

Thanks very much for the reply. What I understand is that if I have done a pilot study and when I have intra-subject CV and %T/R ratio, I should always consider both the intra-subject CV and observed %T/R ratio to calculate the appropriate sample size for the pivotal study. And If I have only intra-subject CV and not the %T/R ratio, in that case only I should assume T/R ratio of 0.95, right?

Best Regards

Dr Muneesh Garg
 
ElMaestro
★★★

Denmark,
2008-11-18 09:49
(6057 d 06:13 ago)

@ drmuneesh
Posting: # 2702
Views: 8,019
 

 sample size for BA/BE studies

Post reply
Dear Dr Muneesh,

Perhaps I missed something, so please allow a technicality for my own clarification: With a pilot study, how could you have an estimated CV but not an estimated T/R - what type of pilot study would provide it and why would such a pilot study be prioritised over others that address T/R more directly?

Thanks in advance and best regards,
EM.
 
drmuneesh
★    

2008-11-18 10:31
(6057 d 05:32 ago)

@ ElMaestro
Posting: # 2704
Views: 7,939
 

 sample size for BA/BE studies

Post reply
Dear EM

I mean to say that if I have done a pilot study, I will definitely have the intra-subject CV and %T/R ratio to calculate the appropriate sample size for the pivotal study. But if from the previous published literature, I get only intra-subject CV and not the %T/R ratio, in that case only should I assume T/R ratio of 0.95.

Best Regards

Dr Muneesh Garg
 
ElMaestro
★★★

Denmark,
2008-11-18 10:59
(6057 d 05:03 ago)

@ drmuneesh
Posting: # 2706
Views: 7,919
 

 sample size for BA/BE studies

Post reply
Dear dr Garg,

I would assume 0.95 for T/R if I had no better idea of the real ratio. But if you actually do have a better idea, then you certainly might want to use that.

EM.
 
drmuneesh
★    

2008-11-18 11:27
(6057 d 04:36 ago)

@ ElMaestro
Posting: # 2709
Views: 7,866
 

 sample size for BA/BE studies

Post reply
Dear EM

Thanks very much for the reply

Best Regards

Dr Muneesh Garg
 
drmuneesh
★    

2008-11-19 11:14
(6056 d 04:49 ago)

@ drmuneesh
Posting: # 2730
Views: 7,951
 

 sample size for BA/BE studies

Post reply
Dear EM

Suppose I have done a two-way cross-over pilot study and I have calculated the sample size for the pivotal study considering intra-subject CV and observed %T/R ratio using SAS. The sample size calculated is X. As SAS gives sample size for two-way cross-over study, right? If the intra-subject CV is > 30% and the drug is known to be highly variable and therefore, I want to use the replicate study design, in that case can I do the study with half the number of subjects i.e dividing the calculated sample size by 2 (X/2).

I have done a pilot study with parallel design and I have calculated the sample size for the pivotal study considering CV and observed %T/R ratio using SAS. The sample size calculated is X. As SAS gives sample size for two-way cross-over study, right? In that case can I do the study with double the number of subjects i.e multiplying the calculated sample size by 2 (2X).

Please suggest.

Best Regards

Dr Muneesh Garg
 
ElMaestro
★★★

Denmark,
2008-11-19 11:48
(6056 d 04:15 ago)

@ drmuneesh
Posting: # 2732
Views: 8,152
 

 sample size for BA/BE studies

Post reply
Dear Dr. Muneesh,

this goes beyond my knowledge unfortunately.
I have no experience with SAS, and I have never powered a study myself. You should rely on people who are more competent than I am.
But you mention you did a parallel pilot study; just to make sure I understand it, did you apply replicated dosing in the parallel arms to get the CV?

EM.
 
drmuneesh
★    

2008-11-20 09:30
(6055 d 06:33 ago)

@ ElMaestro
Posting: # 2744
Views: 7,883
 

 sample size for BA/BE studies

Post reply
Dear EM,
I donot understand what do you mean by replicated dosing in the parallel arms to get the CV.

In my query posted on 19/11/08, in one two-way cross-over pilot study, subjects were dosed two times in the study. And in another pilot study with parallel design for long half-life drug, subjects were dosed only once in the study.

Dr. Muneesh
 
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2008-11-20 12:18
(6055 d 03:44 ago)

@ drmuneesh
Posting: # 2747
Views: 7,994
 

 CVintra # CVinter # CVtotal

Post reply
Dear Muneesh,

❝ I donot understand what do you mean by replicated dosing in the parallel arms to get the CV.


If you want to plan the sample size for a cross-over study you need the intra-subject variability (CVintra). If you have dosed both treatments only once per subject (nonreplicate) in a parallel design, you only have the total variability (CVtotal). See this post (you may have found it searching the forum as DLabes suggested already).

❝ And in another pilot study with parallel design for long half-life drug, subjects were dosed only once in the study.


OK, you are talking about the same drug hopefully? You have one cross-over pilot and one parallel and now want to plan a pivotal cross-over? From the cross-over study you may get a relationship of CVintra and CVtotal, which you may apply to the parallel getting another estimate of CVintra. You may pool these two CVintra-values - but this is some kind of vicious circle… ;-)
—
Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
 
d_labes
★★★

Berlin, Germany,
2008-11-20 14:57
(6055 d 01:06 ago)

@ Helmut
Posting: # 2750
Views: 8,174
 

 Circulus vitiosus

Post reply
Dear Helmut,

❝ [...] - but this is some kind of vicious circle ... ;-)


Die Henne ist der Umweg eines Eies,
ein anderes Ei zu erzeugen (L. STERNE)

(The hen is the way round of an egg to lay another egg :-D )

BTW, there is a Vicious cycle software firm. Its not SAS nor "Dark side".
—
Regards,

Detlew
 
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2008-11-20 19:11
(6054 d 20:51 ago)

@ d_labes
Posting: # 2753
Views: 7,883
 

 Sterne!

Post reply
Dear DLabes,

Yeah, the honorable Lawrence Sterne!

He gave a nice image about the storyline of the first 5 volumes of his
   The Life and Opinions of Tristram Shandy, Gentleman (1760):
[image]
Reminds me sometimes on the way we are doing our studies...
—
Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
 
drmuneesh
★    

2008-11-28 11:37
(6047 d 04:26 ago)

@ Helmut
Posting: # 2830
Views: 7,880
 

 CVintra # CVinter # CVtotal

Post reply
Dear HS,

No, I am not talking about the same drug. I have done two studies with two different drugs. I have done one cross-over pilot with drug A which is a known highly variable drug and now I want to plan a pivotal two-way cross-over study. I have done another parallel Pilot with drug B with very long half and now want to plan a pivotal parallel study. I will elaborate my query as under-
Case-1: I have done a two-way cross-over pilot study with drug A and I have calculated the sample size for the pivotal study using SAS considering intra-subject CV of 0.5363 and observed %T/R ratio of 0.9148. The sample size calculated is 175 for two-way cross-over pivotal study as programmed. As the intra-subject CV is > 30% and the drug is known to be highly variable and therefore, I want to use the replicate (four-way cross-over) study design, in that case can I do the pivotal four-way cross-over study with at least half the number of subjects i.e. 88?

Case-2: I have done another pilot study with parallel design for a very long half-life drug B and I have calculated the sample size for the pivotal study using SAS considering total CV of 0.16 and observed %T/R ratio of 112.08. The sample size calculated is 27 for two-way cross-over study as programmed. In that case can I do the pivotal parallel study with at least double the number of subjects i.e. 54?

Please suggest.

Regards
Dr Muneesh Garg
 
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2008-11-29 11:14
(6046 d 04:49 ago)

@ drmuneesh
Posting: # 2833
Views: 7,770
 

 CVintra # CVinter # CVtotal

Post reply
Dear Muneesh,

❝ Case-1: [...] intra-subject CV of 0.5363 and observed %T/R ratio of 0.9148. The sample size calculated is 175 for two-way cross-over pivotal study as programmed.


OK, confirmed your calculations (actually 176...) in two programs.

❝ I want to use the replicate (four-way cross-over) study design,


Why?

❝ in that case can I do the pivotal four-way cross-over study with at least half the number of subjects i.e. 88?


Yes, but you should:
  • accept that both the T/R and CV from the pilot study are not natural constants like the speed of light. Allow for a safety margin!
  • If you go with a replicate design, your safety margin should be even higher (more drop-outs are likely due to the doubled number of periods).

❝ Case-2: […] parallel design […] total CV of 0.16 and observed %T/R ratio of 112.08. The sample size calculated is 27 for two-way cross-over study as programmed. In that case can I do the pivotal parallel study with at least double the number of subjects i.e. 54?


What? Your sample size is 27 per group in parallel study (no X-over!), or 54 in total. Remember: in a parallel design, any differences in the two groups will be treated as a treatment effect - be cautious!
—
Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
 
drmuneesh
★    

2008-11-30 08:18
(6045 d 07:44 ago)

@ Helmut
Posting: # 2838
Views: 7,628
 

 CVintra # CVinter # CVtotal

Post reply
Dear HS

Thanks very much

Regards
Dr Muneesh Garg
 
ElMaestro
★★★

Denmark,
2008-11-29 16:42
(6045 d 23:20 ago)

@ drmuneesh
Posting: # 2835
Views: 7,650
 

 is something fishy in case 2?

Post reply
Dea Dr Garg,

I am not sure I follow you completely regarding case 2:

❝ Case-2: I have done another pilot study with parallel design for a very long half-life drug B and I have calculated the sample size for the pivotal study using SAS considering total CV of 0.16 and observed %T/R ratio of 112.08. The sample size calculated is 27 for two-way cross-over study as programmed. In that case can I do the pivotal parallel study with at least double the number of subjects i.e. 54?


Not considering the last sentence there, I get the impression that you first did a parallel pilot which gave you a CV and a T/R estimate, which you then used to power a crossover trial?
This does not sound right. As indicated in previous question and further elaborated on by DLabes, the CV you get in a parallel study is not the intra-CV you need for powering the crossover trial.
As a rule of thumb a parallel pilot study will give you info you can use to plan another parallel study; a crossover pilot will give you what you need to power another crossover trial (or optionally even a parallel one).

Best regards
EM.
 
drmuneesh
★    

2008-11-30 08:17
(6045 d 07:45 ago)

@ ElMaestro
Posting: # 2837
Views: 7,635
 

 is something fishy in case 2?

Post reply
Dear EM
Thanks very much

Regards
Dr Muneesh Garg
 
d_labes
★★★

Berlin, Germany,
2008-11-19 12:33
(6056 d 03:29 ago)

@ drmuneesh
Posting: # 2733
Views: 8,030
 

 SAS and sample size

Post reply
Dear Muneesh Garg

❝ Suppose I have done a two-way cross-over pilot study […] As SAS gives sample size for two-way cross-over study, right? […]


SAS is a multi-purpose statistical packet. It will give you the sample size for many study designs, not only for 2x2 x-over, if properly programmed.

For the rest of your questions have a look into the forums category "Sample size".
—
Regards,

Detlew
 
drmuneesh
★    

2008-11-20 09:31
(6055 d 06:31 ago)

@ d_labes
Posting: # 2745
Views: 7,889
 

 SAS and sample size

Post reply
Dear d_labes

Thanks very much

Dr Muneesh Garg
 
UA Flag
Activity
 Admin contact
23,424 posts in 4,927 threads, 1,677 registered users;
31 visitors (0 registered, 31 guests [including 8 identified bots]).
Forum time: 17:03 CEST (Europe/Vienna)

Philosophy, like medicine, has plenty of drugs, few good remedies,
and hardly any specific cures.    Sebastien-Roch Nicolas de Chamfort

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5