- On 1 Dec 2004 at 08:49:14, Nisha K are (nishavenkatesh.-a-.yahoo.co.in) sent the message

Back to the Top

The following message was posted to: PharmPK

Hello Forum,

Could anyone tell me why and how the limit of 80% to 125% is fixed for

Bioequivalence studies.

Nisha. K.R. - On 1 Dec 2004 at 11:11:50, "Peter Gingras" (pgingras.at.apotex.com) sent the message

Back to the Top

The following message was posted to: PharmPK

I'm sure you will hear many different explanations for how these

standards were developed. I've been involved in various aspects of

Comparative Bioavailability testing since 1978, and have worked with

some pretty bright experts in this field over the years (hi Mike,

Yu-Chung, Laszlo, Radu). Here's one explanation.

Manufacturing standards for individual dose concentrations (Content

Uniformity and/or Uniformity of Dosage Unit Limits per USP/BP)

typically allow for limits between 75 to 125% of label claim.

The bioequivalence limit was established, at least in part, based on

the CU Limits, and the sense that a standard that is more stringent

than the limits on the permissible concentration of the drug in an

individual dose would be unnecessary, and perhaps illogical.

Peter Gingras

Apotex Inc - On 1 Dec 2004 at 08:58:11, "Vuong Trieu" (vtrieu.-at-.AmericanBioScience.com) sent the message

Back to the Top

The following message was posted to: PharmPK

This was derived from the FDA guidance for BE and in their judgment a

20% change will not affect the clinical outcome. - On 1 Dec 2004 at 13:57:49, "Peter Gingras" (pgingras.-at-.apotex.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Vuong:

Actually, 80 - 120% (125%, 130%) limits were in use long before there

were any written regulatory guidelines, American, Canadian or other.

Those limits where incorporated into the guidance documents after many

years of data collection.

An interesting side note, well into the 1980's, many generic drugs were

approved (and continue to be used interchangeably with brand drugs)

based on comparative dissolution rate data alone. It is interesting

that the FDA has recently contemplated this approach again. There is

certainly plenty of historical precedent for this approach, for drugs

with certain properties.

Peter Gingras - On 3 Dec 2004 at 03:34:00, Nisha K are (nishavenkatesh.aaa.yahoo.co.in) sent the message

Back to the Top

Peter, Vuong,

Thank you.

But I am still puzzled. It seems pretty arbitrary for something that

companies spend a lot on to establish.

Also, 80 is -20%, but 125 is... ?

Can someone clarify?

Nisha. K. R.

[But 80/100 = 100/125 i.e. 0.8 - db] - On 2 Dec 2004 at 22:30:57, Dimiter Terziivanov (terziiv.at.yahoo.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Nisha,

When the lower limit, teta1=0.8, the upper limit

teta2, after log transformation is ln(teta2) -ln(teta1),i.e. teta2 =

teta1^(-1). Hence, for teta1 0.8, teta2 results in 1.25.

Greetings,

D.Terziivanov

Dimiter Terziivanov, MD,PhD,DSc, Professor

Head, Clinic for Clinical Pharmacology and

Pharmacokinetics, Univ Hosp "St. I.Rilsky",

15 Acad. I. Geshov st, 1431 Sofia, Bulgaria

Tel:(+ 359 2)8510639;(+ 359 2)5812 828.

Fax:(+ 359 2)8519309. e-mal: terziiv.aaa.yahoo.com - On 3 Dec 2004 at 08:57:10, "Porzio, Stefano" (Stefano.Porzio.-at-.ZambonGroup.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Nisha K.R.,

the reason is that you must consider a NORMAL distribution to compare

data and it's well known that to obtain AUC and Cmax NORMAL

distributions you have to apply a log-transformation!! This implies

that you confidence limits 0.8-1.2 are log-transformed as well and

become 0.8-1.25.

I report exactly what it's written in Bioavailability and

Bioequivalence section of "Applied Biopharmaceutics and

Pharmacokinetics" L.Shargel and A.B.C. Yu :

"The lower 90% confidence interval for the ratio of means cannot be

less than 0.8, and the upper 90% confidence interval for the ratio of

the means cannot be greater than 1.20. When log-transformed data is

used, the 90% confidence interval is set at 80 to 125%, since the log

of 0.8 (except for sign) is 1.25 or 125%...................for AUC and

Cmax values of the test drug product should not be less than 0.8 (80%)

nor greater than 1.25 (125%) of the reference product based on

log-transformed data."

Best regards

Dr. Stefano Porzio

Pharmacokinetic and Tox. Dept.

Inpharzam Ricerche SA - ZAMBON-GROUP

Taverne - Switzerland - On 3 Dec 2004 at 11:02:59, fabrice_nollevaux.-a-.sgs.com sent the message

Back to the Top

The following message was posted to: PharmPK

Nisha,

The 20% maximum difference between a test (T) and a reference (R)

formulation can be expressed in 4 different ways

(do not forget we are considering a multiplicative model, e.g. T being

20%

higher than are is not equivalent to are being 20% lower than T):

1) Test 20% lower than Reference <=> T/R = 80%

(=0.8/1.0)

2) Test 20% higher than Reference <=> T/R = 120%

(=1.2/1.0)

3) Reference is 20% lower than Test <=> T/R = 125%

(=1.0/0.8)

4) Reference is 20% higher than Rest <=> T/R = 83.3%

(=1.0/1.2)

As you see, T/R within 80-125% covers all cases consistent with a

maximum

20% difference between both formulations.

Hope this helps,

Fabrice

Fabrice Nollevaux,

Senior Pharmacocineticist

SGS Life Science Services, www.sgs.com/life_sciences

SGS Biopharma - Wavre, Belgium, www.sgsbiopharma.com - On 3 Dec 2004 at 08:23:02, "Shilpi Khan" (SKhan.at.cedracorp.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Nisha,

80% to 125% comes from the statistical analysis of natural log

transformed of Pharmacokinetic parameters. The 90% CI of the estimate

of the difference of treatment should between -0.223 (ln(0.8)) and

0.223 (ln(1.25)). The difference in the PK parameters due to treatment

is of transformed data. To show the ratio on original scale we have to

exponentiate the point estimate and the CI. Hence exp(-0.223)=0.8 and

exp(0.223)=1.25. I think this is the reason why we have 80 to 125%

rather than 80 to 120%.

Hope it clarifies.

Thanks,

Shilpi

Shilpi Khan, M.S.

Staff Scientist/Biostatistician

CEDRA Corporation

Austin, TX 78754 - On 3 Dec 2004 at 09:51:10, RPop.aaa.pharmamedica.com sent the message

Back to the Top

Hi Nisha,

Once upon the time, the clinicians and pharmacologists were asked what

kind of fluctuation in the blood/plasma/serum concentration levels of a

drug can be tolerated. Meaning, what kind of changes in the

concentration

levels should not trigger significant/observable changes in the

clinical/pharmacological end point. The agreement was that a 20% change

can be considered tolerable.

Based on this, the 80-120% (0.8-1.2 in ratio terms) range was

established

as a regulatory standard. It corresponds to a symmetrical interval of

+-20% around 100%.

Later, based on observation of the pharmacokinetic data, it was decided

that the PK parameters (AUC, Cmax, Cmin) follow a normal distribution

only

after log-transformation. The assumption of normal distribution is

important because all the formulas used in the statistical analysis were

derived based on this distribution.

Consequently, it was recommended that statistical analysis should be

applied to log-transformed PK parameters. On the logarithmic scale the

80-120% range does not correspond to a symmetrical interval: equal

"distances" on both sides of the central tendency.

The central tendency of 100% (1.0) from the raw scale correspond to the

value 0.0 on the logarithmic scale.

The lower limit of 80% (0.8) corresponds to approximately -0.223:

ln(0.8)=-0.223143551...

The location of the upper limit at equal distance, but on the opposite

direction from 0.0 on the logarithmic scale is +0.223 that corresponds

on

the raw data scale to 1.25 (125%): exp(0.223...)=1.25.

I hope this help.

radu - On 3 Dec 2004 at 09:00:52, "Vuong Trieu" (vtrieu.-a-.AmericanBioScience.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Nishra:

It is pretty arbitrary; however, most or all ANDAs approved has only 10%

variation. So in practice, you do try to be as close to the innovator

as possible.

Want to post a follow-up message on this topic? If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "Bioequivalence Limit" as the subject

PharmPK Discussion List Archive Index page

Copyright 1995-2010 David W. A. Bourne (david@boomer.org)