- On 31 Jan 2001 at 16:11:13, "Tata, Prasad N" (Prasad.Tata.at.MKG.COM) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Colleagues:

The recent BA/BE guidance issued by FDA requires the estimation of early

exposure for all the immediate release formulation. Guidance recommends

"the use of partial AUC as an early exposure measure. The partial area

should be truncated at the population median of Tmax values for the

reference formulation."

My dilemma is that if this reference population median of Tmax values for

the reference formulation happens to fall in between two actual blood draws

how can we calculate the partial AUC up to this reference Tmax. Are there

any special formulae such as we calculate AUCinf from AUCt or do we have to

extrapolate/interpolate the partial AUC from the earlier time point?

I appreciate any sensible approach.

Prasad NV Tata, Ph.D.

Manager-Pharmacokinetics

Mallinckrodt, Inc.

P.O. Box: 5840

675 McDonnell Blvd.

St. Louis, MO 63043

Tel: (314) 654-5325

Fax: (314) 654-9325

e-mail: prasad.tata.aaa.mkg.com - On 31 Jan 2001 at 22:34:47, Walt Woltosz (walt.-a-.simulations-plus.com) sent the message

Back to the Top

The following message was posted to: PharmPK

At 04:11 PM 1/31/01 -0600, you wrote:

>My dilemma is that if this reference population median of Tmax values for

>the reference formulation happens to fall in between two actual blood draws

>how can we calculate the partial AUC up to this reference Tmax. Are there

>any special formulae such as we calculate AUCinf from AUCt or do we have to

>extrapolate/interpolate the partial AUC from the earlier time point?

One way is to build a predictive model that incorporates all of the important

phenomena affecting AUC and Tmax. Our GastroPlus(tm) software is used for this

purpose for orally dosed drugs, accounting for such interactive effects as

permeability/absorption, dissolution, and solubility. Each of these can have

strong pH dependencies for ionized compounds, and such effects are included in

the model.

If you have plasma concentration-time data for one or more subjects, fitting an

appropriate model is usually straightforward. We provide both the software and

consulting services for such modeling for those who do not wish to do it

themselves.

In a recent consulting contract, we took plasma concentration-time data from a

series of subjects for a variety of dosage forms, fit the absorption and

pharmacokinetic model, and then showed that several new controlled release

dosage forms that were ready to go into human trials would not perform as

required - saving the cost and time of the trials. We then optimized a

controlled release design to meet the target plasma concentration-time profile.

The total effort required about two weeks.

Walt Woltosz

Chairman & CEO

Simulations Plus, Inc. (SIMU)

1220 W. Avenue J

Lancaster, CA 93534-2902

U.S.A.

http://www.simulations-plus.com

Phone: (661) 723-7723

FAX: (661) 723-5524

E-mail: walt.aaa.simulations-plus.com - On 31 Jan 2001 at 22:55:13, Laszlo Endrenyi (l.endrenyi.-at-.utoronto.ca) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Dr. Tata,

1. The new guidance requires the estimation of early exposure only when it is

relevant, and not for all immediate-release formulations. 2. The median of the

Tmax values is, by definition, one of the times at which blood samples were

taken. 3. The partial AUC is reported as estimated and is not extrapolated.

Regards.

Laszlo Endrenyi - On 1 Feb 2001 at 12:00:24, Arturo Soto (arturo.soto.-a-.uam.es) sent the message

Back to the Top

Dear Dr Tata:

If there is an odd number samples, the median of Tmax is the central

value. If the number is even, Tmax falls between two times. If the

estimation of early exposure is relevant in your study, one could do

this (use of PK models in bioequivalence evaluation is not

recommended):

1. Calculate Tmax truncated to first Tmax value (AUCTmax1)

2. Calculate Tmax truncated to second Tmax value (AUCTmax2)

3. AUC 0-tmax=AUCtmax2-((AUCtmax2-AUCtmax1)/2)

Regards.

Arturo Soto - On 1 Feb 2001 at 12:01:18, "Jeff Wald" (jwald.-at-.pharsight.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Calculation of partial AUC's is supported in WinNonlin and it does perform

an interpolation to bridge between observed points when a requested interval

does not correspond to a sample time.

The median value for Tmax must fall at an observed value. If all the actual

sample times are very close to the nominal times....end of story. However,

if the actual-nominal differences are not negligible the median value may

not be at an observation point for all subjects in the study. In this case,

I would at minimum calculate the interpolated (i.e, to the median

Tmax)partial AUC values for comparison to the analyses based on nominal

values.

For interested parties: WinNonlin 3.2 (to be released this spring) performs

individual, and population BE analyses. The wizard scans the data to figure

out which analysis is appropriate and switches models depending on the

nature of the data; i.e., 2x2 crossover gets one model, replicated crossover

gets the full linear mixed effects model requested in the draft 1999 FDA

guidance. Users additionally have the option of bypassing the wizard and

specifying their own linear mixed effects model.

Regards, Jeff - On 1 Feb 2001 at 12:02:04, "Ronald Kavanagh 301-827-4254 FAX 301-827-4264" (KAVANAGHR.-a-.cder.fda.gov) sent the message

Back to the Top

I have forwarded and discussed the question with my management at FDA.

There is currently a proposed response to the issue from the BA/BE

guidance committee. This proposal will be discussed by FDA management

next week.

When there's internal consensus and approval, the response will be made

available.

My management has suggested that I post the response to this newsgroup

at that time.

Ron Kavanagh

Ronald E. Kavanagh. B.S. Pharm., Pharm.D., Ph.D.

Food & Drug Administration

Office of Clinical Pharmacology and Biopharmaceutics

5600 Fishers Lane HFD-870

Rockville, MD 20857

Phone: 301-827-4254 FAX: 301-827-4264

Office: Parklawn (PKLN) 17B-30 e-mail: kavanaghr.-at-.cder.fda.gov - On 1 Feb 2001 at 16:14:03, "Weller, Steve" (sw46385.-at-.GlaxoWellcome.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Although there may not be a need for performing this partial AUC analysis

routinely, this seems like a good question to me. Indeed, if analysis of

Tmax is performed using protocol-specified sampling times and the number of

subjects is an even number (usual for a BE study), then the median will fall

midway between two Tmax values. And when PK analysis is performed based on

actual sampling times (which I think is much more appropriate), then the

chance that the Tmax value to be used for truncation in partial AUC

determinations occurs at an actual sampling time is perhaps even more

remote. One could determine the partial AUCs by interpolating

concentrations at the median Tmax, or perhaps using a nearest neighbor

approach, but this seems to be something that needs to be clarified with

FDA. - On 2 Feb 2001 at 13:04:43, bangarurk.at.drreddys.com sent the message

Back to the Top

The following message was posted to: PharmPK

Before answering the question, I would like to point out the FDA's change

of philosophy from direct or indirect measures of absorption rate to

measures of systemic exposure ( which is mentioned in the guidance

document). In this context Cmax or AUC can continue as measures of product

quality BA and BE but more in terms of capacity to assess exposure than

their capacity to reflect rate and extent of absorption.

The partial areas are mentioned in the Guidance document under IIIA9(a)

subheading of early exposure measures. As it clearly mentioned in the

document it is aimed at finding any deviations in the early response( may

be beneficial or otherwise). The T max to be considered as the population

median of the Tmax, if it is ANDA you need to get the T max value from

innovators SBOA documents, where the Tmax values are cited. It is the

median value not Mean Value, i.e the Tmax value reported highest number of

times. actually it is an additional parameter they are recommending to be

reported in addition to the existing parameters. There is no question of

intrapolation or extrapolation, only thing is that at least two points are

measured before the Tmax( highly significant for drugs whose tmax is less

than 30 minutes).The method of calculation is similar to AUC and upto time

point where T max occurs more frequently ( take the values from reference

product values incase you dont have the SBOA information on this). - On 5 Feb 2001 at 13:39:38, "Edward Newman" (enewman.-a-.coh.org) sent the message

Back to the Top

The following message was posted to: PharmPK

One statement below (previous message - db] is in error. The MEDIAN

value is not the value

"reported the highest number of times." That value is the MODE. The

MEDIAN is the value which 50% of the samples are greater than or equal to

and 50% of the samples are less than or equal to. I don't which the FDA

guidance document requests. If it is the median, interpolation could be

required if there are and even number of samples and the two central values

are not the same. - On 18 Feb 2001 at 19:56:16, William Heybroek (william.heybroek.-at-.kinetics.demon.co.uk) sent the message

Back to the Top

The following message was posted to: PharmPK

You may be confusing median with mean. The median is the value that

divides a population in two. It is only ever a non-integer when the

number of data in the population is even. In this case you will have to

interpolate from the mid-point of two AUCt values.

Hope this helps.

********

William Heybroek PhD

PK Pharma,

PO Box 45, Marple, Stockport,

Manchester SK6 5FX, UK

Tel: +44 (0)161 427 3241

Fax: +44 (0)161 449 0155

*********

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 Guidance Question" as the subject

PharmPK Discussion List Archive Index page

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