Bioequivalence and Bioavailability Forum

chidambarajoshi ★ Hyderabad-INDIA, 2008-08-13 12:02 (6117 d 11:30 ago) Posting: # 2173 Views: 16,379	Tmax [Nonparametrics] Post reply
	Dear all, I have a result of delayed release formulation in which Tmax of reference is 6hrs and that of Test is 11.2hrs. Is this results are statistically significant? (sorry I am weak in stat and PK) — Best Regards Joshi

ElMaestro
★★★

Denmark,
2008-08-13 12:25
(6117 d 11:07 ago)

@ chidambarajoshi
Posting: # 2174
Views: 14,586

Tmax

Post reply

Hi Chidambarajoshi,

you need to tell us a little more. Only a stat. analysis* can tell if this is stat. significant. To find out if it is clinically significant you might want to apply common sense and judgment (which regulators seem to hate to apply to their assessments).
Usually Tmax is given some regard in evaluations, unless a protocol for some reason specifes something else.
Having said all that, a 6 hrs. difference does not exactly sound promising.

EM.

*: Many companies actually just use Wilcoxon

chidambarajoshi
★

Hyderabad-INDIA,
2008-08-13 12:49
(6117 d 10:43 ago)

@ ElMaestro
Posting: # 2175
Views: 14,683

Tmax

Post reply

Dear EM,
Here is the detailed question is "Confirm if the differences between Tmax for the reference product, and test are statistically significant. If not, this should be made clear in any subsequent report to the RMS. The critical factor here is to convince the assessor that steady state concentrations are of most clinical relevance in a medicinal product for use in a chronic condition (weeks, months or years), rather than a small difference in time to Tmax (hours)" This is a Venlafaxine study 150mg Delayed realease

—
Best Regards
Joshi

Helmut
★★★

Vienna, Austria,
2008-08-13 15:48
(6117 d 07:44 ago)

@ chidambarajoshi
Posting: # 2178
Views: 15,111

Tmax (EU)

Post reply

Dear Joshi!

❝ "Confirm if the differences between Tmax for the reference product, and test are statistically significant. If not, this should be made clear in any subsequent report to the RMS. The critical factor here is to convince the assessor that steady state concentrations are of most clinical relevance in a medicinal product for use in a chronic condition (weeks, months or years), rather than a small difference in time to Tmax (hours)"

OK, 'RMS' means 'Reference Member State' - so you are talking about a European Market Authorisation...
To quote from the relevant Guideline:

3.6.1 Statistical analysis
If appropiate to the evaluation the analysis technique for t_max should be non-parametric and should be applied to untransformed data.
3.6.2 Acceptance range for pharmacokinetic parameters
Statistical evaluation of t_max only makes sense if there is a clinically relevant claim for rapid release or action or signs related to adverse effects. The non-parametric 90% confidence interval for this measure of relative bioavailability should lie within a clinically determined range.

The statistical method mentioned by ElMaestro:

Hauschke D, Steinijans VW and E Diletti
A distribution-free procedure for the statistical analysis of bioequivalence studies
Int J Clin Pharm Ther Toxicol 28(2), 72-78 (1990)

What software do you have? Paper and pencil is cumbersome for this procedure. I could help you with StatXac (Cytel), EquivTest/PK (Statistical Solutions), and WinNonlin (Pharsight). SAS is not my thing - PROC STATXACT would do the job there.

❝ This is a Venlafaxine study 150mg Delayed release

Just read Section 3.6.2 over. Is a statistical significant difference clinically relevant? How do C_max of the 150mg DR compare to 75mg IR (safety)? Have also a look at EMEA’s MR guideline.

—
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

chidambarajoshi ★ Hyderabad-INDIA, 2008-08-14 08:30 (6116 d 15:02 ago) @ Helmut Posting: # 2182 Views: 14,534	Tmax (EU) Post reply
	Dear HS, Thanks for the reply. Yes it's EU Submission. We have WinNonlin and trying to measure with 90% CI. But what exactly meant by clinically determined rage? Regards, joshi — Best Regards Joshi

Helmut
★★★

Vienna, Austria,
2008-08-14 22:12
(6116 d 01:19 ago)

@ chidambarajoshi
Posting: # 2188
Views: 16,445

Tmax (WinNonlin)

Post reply

Dear Joshi!

❝ We have WinNonlin and trying to measure with 90% CI.

In the workbook you need the following columns: subject, sequence, treatment, Tmax.
Tools > Crossover Design... drag subject, sequence, treatment to the named fields and Tmax to Response Calculate
WinNonlin has an unreported bug, giving differences always in alphabetical order... :angry:

(Just a side note for people who like software 'validation' by developers: if I remember it correctly the 'validation pack' costs about 75% of the the initial price of the software or 4× the annual licence - but nonparametrics are not a strength of WinNonlin - to say it politely).
If your treatments are coded 'test' and 'reference' WinNonlin will come up with:
Treatment_Diff_(reference - test) (!!)
in the column Median: the point estimate
in the columns CI_Lower and CI_Upper you find the confidence interval - in the second row the 90% Confidence_Level.
Again, don't forget to change the sign of these three values, if the coding of your test treatment is in lexical order before the code of the reference treatment!
Example:

Test coded 'T' or 'test', reference coded 'R' or 'reference'
→ wrong, because T>R: change signs of PE and CL!
Test coded 'A', reference coded 'B'
→ correct, because A<B...

Don't be surprised that the point estimate is not the one you probably expect; many people think it's just the difference of medians of test and reference, but actually it's the median of all pairwaise differences (for details see the reference given already).
The confidence level given in the 4^th column (Heading Exact) is rarely what it seems to be, namely exact. I had many conversations with Pharsight's support over the years and simply gave up, because I don't use WinNonlin for nonparametrics myself.
E.g., for 12 subjects (balanced 6/6) WinNonlin comes up with 0.9000; correct is 0.9069…

The underlying test statistic (Mann-Whitney U) is discrete (different from parametric ones, which are continuous); therefore it's rather unlikely that α is exactly 0.05 (p=1-2α=0.90). Since the test is conservative, the nonparametric confidence interval is slightly wider if applied to normal distributed data – but tighter, if applied to nonnormal distributed ones.

If you know FORTRAN or somebody with a little experience in programming, you can compute exact values of U for any combination of subjects within sequences (very useful, because the table given by Hauschke et al. covers only up to 24 subjects).*

❝ But what exactly meant by clinically determined rage?

This has nothing to do with statistics; you must ask a specialist in the medical field. For instance a difference of 30 minutes may be clinically relevant for a pain-killer (I want my headaches to go away – now!); for venlafaxine in steady state even hours may be irrelevant (but I'm neither a physician nor a patient – so I can't tell).

Dinneen LC, Blakesley BC. Algorithm AS 62: A Generator for the Sampling Distribution of the Mann- Whitney U Statistic. J Roy Stat Soc C. 1973;22(2):269–73. doi:10.2307/2346934.

—
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

chidambarajoshi ★ Hyderabad-INDIA, 2008-08-25 11:47 (6105 d 11:44 ago) @ Helmut Posting: # 2244 Views: 14,367	Tmax (WinNonlin) Post reply
	Dear HS, Thanks for the reply and we could able to resolve it. Regards, Joshi — Best Regards Joshi

SDavis
★★

UK,
2008-11-18 20:47
(6020 d 01:45 ago)

@ Helmut
Posting: # 2722
Views: 14,320

Tmax (WinNonlin)

Post reply

❝ WinNonlin has an unreported bug, giving differences always in alphabetical order... :angry:

Hi Helmut - Today I've been going through some s/w issues you've noted here in the forum which I've been finding informative, entertaining and at times shaming!

I just want to clarify if anyone finds a bug (or other concern) they should report it to the Developer - then it is no longer unreported :surprised:

In Pharsight's case that is through support (at) pharsight.com or at http://pharsight.com/support/support_home.php in the latter case you will automatically be given a tracking ID so you can see how the case, hopefully progresses.

If in this case you meant undocumented then yes I guess that is an oversight. Unfixed/known issues should be noted in e.g. Release notes. I'll go and take a look at this one now. 4040.

thanks, Simon. http://www.pharsight.com/training

Edit: Support's e-mail address modified to prevent spamming. [Helmut]

Helmut
★★★

Vienna, Austria,
2008-11-18 21:15
(6020 d 01:16 ago)

@ SDavis
Posting: # 2723
Views: 14,307

Tmax (WinNonlin)

Post reply

Hi Simon!

❝ ❝ WinNonlin has an unreported bug, giving differences always in alphabetical order... :angry:

❝ In Pharsight's case that is through [...]

❝ http://pharsight.com/support/support_home.php [...]

Yeah, that's a funny website telling me that I never posted before 2007, and some IDs give in the field 'Solutions' the comment 'None Found', while I received one by e-mail… (see this post or that one).

❝ If in this case you meant undocumented then yes I guess that is an oversight.

Oh, yes undocumented is correct! I reported it a couple of times up to WNL v4 (first time according to my files for v3.3 on 08 July 2002) and then gave up. I even provided Pharsight's support with a reference of FORTRAN-code to calculate the CDF. I didn't really need it myself (I use StatXact or EquivTest/PK for nonparametrics)… The output – besides the lexical order – is flawed in another sense:
In the 'Crossover Design Workbook' – 'Confidence Intervals' tab the column 'Confidence_Level' gives the aimed one, but numbers given in the Column 'Exact' are wrong (WinNonlin 5.2.1 always comes up with something like 'Confidence_Level 0.90, 'Exact' 0.9000). Below is a table of exact confidence levels for a couple of balanced sequences n₁=n₂ (N=n₁+n₂):

N alpha p{U_n1,n2} N alpha p{U_n1,n2} 6 0.0500 0.9000 36 0.0485 0.9029 8 0.0286 0.9429 38 0.0482 0.9036 10 0.0476 0.9048 40 0.0482 0.9036 12 0.0465 0.9069 42 0.0486 0.9028 14 0.0487 0.9027 44 0.0491 0.9018 16 0.0415 0.9170 46 0.0499 0.9002 18 0.0470 0.9061 48 0.0486 0.9028 20 0.0446 0.9108 50 0.0497 0.9006 22 0.0440 0.9121 52 0.0490 0.9020 24 0.0444 0.9113 54 0.0485 0.9030 26 0.0454 0.9092 56 0.0499 0.9002 28 0.0469 0.9061 58 0.0498 0.9004 30 0.0488 0.9025 60 0.0498 0.9004 32 0.0469 0.9062 62 0.0499 0.9002 34 0.0493 0.9013 64 0.0487 0.9026

Due the discrete nature of the U-distribution α is ≤0.05, or in other words, nonparametric confidence intervals are generally slightly wider (>0.90) than ones based on normal theory (if the underlying 'true' distribution is normal).

—
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

Helmut
★★★

Vienna, Austria,
2009-06-12 20:55
(5814 d 02:36 ago)

@ SDavis
Posting: # 3852
Views: 14,434

Tmax (Phoenix WinNonlin 6.0/6.1)

Post reply

Hi Simon,

the bug is still there (Phoenix 6.0.0.1648); logged an incident (#00065507) again...

Edit 1: Incident was closed on June 24^th, 2009 with the note:

[…] this issue has been logged in our tracking system for review and to rectify if necessary in a future Phoenix WinNonlin release (QC 7904).

So what about previous versions of WNL?

Edit 2: The issue was partly resolved in Phoenix 6.1 (build 6.1.0.173, October 2009). Now in the Crossover Setup there is a combo-box to select the 'Reference Treatment'. No need any more to change signs or recode treatments to lexical order. The confidence level in the column 'Exact' is still wrong – maybe in the next decade… ;-)

The Release Notes give it as a known issue: Crossover design: "Exact" heading for the p-value (QC 7904): In the Crossover Design tool, the confidence levels reported as "Exact" for the crossover differences (e.g., Treatment_Diff(T-R)) are the target levels rather than exact levels.

—
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

Tmax [Nonparametrics]

Tmax

Tmax

Tmax (EU)

Tmax (EU)

Tmax (WinNonlin)

Tmax (WinNonlin)

Tmax (WinNonlin)

Tmax (WinNonlin)

Tmax (Phoenix WinNonlin 6.0/6.1)