Back to the Top
Dear All,
Is it mandatory for a generic submission that the dissolution profiles
of the batches used in the bioequivalence studies must meet the F2
similarity criteria in all three media over the pH range of 1 to 6.8?
For an immediate release formulation [with Tmax of about 2.5 - 4.8
hours] what are the regulatory implications if the Bioequivalence test
passes, the profiles are similar in two of those [i.e. 0.1N HCl & pH
4.5 Buffer] but fails in the third [i.e.pH 6.8 buffer]? Shall I have to
re-develop my dissolution method?
I would be grateful to have some comments on the above.
regards,
kaushal
Back to the Top
The following message was posted to: PharmPK
Dear Kaushal,
I am not sure, whether I understand you correctly:
>Is it mandatory for a generic submission that the dissolution profiles
>of the batches used in the bioequivalence studies must meet the F2
>similarity criteria in all three media over the pH range of 1 to 6.8?
>
No. If you are able to show bioequivalence, dissolution testing does not
count at all!
>For an immediate release formulation [with Tmax of about 2.5 - 4.8
>hours] what are the regulatory implications if the Bioequivalence test
>passes, the profiles are similar in two of those [i.e. 0.1N HCl & pH
>4.5 Buffer] but fails in the third [i.e.pH 6.8 buffer]? Shall I have to
>re-develop my dissolution method?
>
Now I am confused. You already have shown bioequivalence in vivo, and
now are going for a biowaiver?
Biowaivers are possible only for BCS class I drugs (high solubility,
high permeability), but if you're BE, so what?
Bioequivalence is a surrogate for therapeutic equivalence, and F2
similarity of dissultion is only a surrogate for bioequivalence.
Regards
Helmut
--
Helmut Schutz
BEBAC
Consultancy Services for Bioequivalence and Bioavailability Studies
Neubaugasse 36/11
A-1070 Vienna/Austria
tel/fax +43 1 2311746
http://BEBAC.at
Bioequivalence/Bioavailability Forum at http://forum.bebac.at
Back to the Top
The following message was posted to: PharmPK
Once you have demonstrated BE, you do not have to worry about the
dissolution profile. However, you do have to work out the appropriate
dissolution condition as a batch release assay during manufacturing.
Just pick the one that is predictive and use that.
Back to the Top
The following message was posted to: PharmPK
Dear Kaushal and Helmut
my guess is that the question comes from this
paragraph of the EU note for guidance
CPMP/EWP/QWP/1401/98, I quote: (page 10)
3.7 In vitro dissolution complementary to a
bioequivalence study
The results of "in vitro" dissolution tests, obtained
with the batches of test and reference
products that were used in the bioequivalence study
should be reported. The results should be
reported as profiles of percent of labelled amount
dissolved versus time.
The specifications for the in vitro dissolution of the
product should be derived from the
dissolution profile of the batch that was found to be
bioequivalent to the reference product and
would be expected to be similar to those of the
reference product (see Appendix II).
For immediate release products, if the dissolution
profile of the test product is dissimilar
compared to that of the reference product and the in
vivo data remain acceptable the
dissolution test method should be re-evaluated and
optimised. In case that no discriminatory
test method can be developed which reflects in vivo
bioequivalence a different dissolution
specification for the test product could be set."
From what I understand is that they would expect that
if the two formulations have shown BE in vivo, the
profiles "would be expected to be similar to those of
the reference product" and if the diss profile of tes
is dissimilar, they recommend you re evaluate and
optimize it...but from the last sentence I understand
that if this is not posible, you can set the diss
spec. based on your dissolution method.
let me know your opinion.
marival
Back to the Top
The following message was posted to: PharmPK
Hello Marival,
We ran across a case recently where matching the in vitro dissolution
profile made it impossible to achieve bioequivalence.
This was because the innovator product had a significant number of
instances wherein the in vivo release of the drug was dramatically
different than the in vitro profile. In a sense, you could say that the
poor quality of the innovator caused a significant shift in Cmax, Tmax,
and AUC for a number of subjects, resulting in a shift of the confidence
intervals. Thus, a generic company trying to demonstrate bioequivalence
would have to bias their release rate to compensate for malfunctions in
the innovator drug product.
This is clearly nonsense. Our regulatory agencies need to take another
look at how bioequivalence is measured. In this case, for example,
throwing out those few subjects who clearly had a malfunction in the
innovator release (e.g., first nonzero data point more than 24-48 hours
after dosing!) would provide a more desirable comparison. The generic
company should be trying to match the innovator product when it performs
correctly, not when it malfunctions.
Walt Woltosz
Chairman & CEO
Simulations Plus, Inc. (AMEX: SLP)
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
Back to the Top
The following message was posted to: PharmPK
Hello Walt,
I strongly support you views.
>Our regulatory agencies need to take another
>look at how bioequivalence is measured. In this case, for example,
>throwing out those few subjects who clearly had a malfunction in the
>innovator release (e.g., first nonzero data point more than 24-48 hours
>after dosing!) would provide a more desirable comparison. The generic
>company should be trying to match the innovator product when it
performs
>correctly, not when it malfunctions.
>
But: how do you actualy "throw out" those few subjects?
Since to my knowledge post-hoc exclusion of outliers is discouraged in
all guidelines worldwide (unless you have a clinical justification like
vomitting), one option would be the use of nonparametric methods. They
generally give for the whole data set point estimates and confidence
intervals almost identical to the ones we would see after exclusion of
outlying subjects in ANOVA.
Nonparametrical methods compensate for deviations from distributional
assumptions rather than reject - somehow arbitrarily - data from the
study.
Canadian regulators at least consider nonparametrics, if the are planned
a-priori
(http://www.hc-sc.gc.ca/hpfb-dgpsa/tpd-dpt/sac_bb_2004-06-03_rop_e.pdf).
In my experience (500 BE studies) for european regulators it was quite
common to accept studies of the following design:
\0x95 calculation of intra- subject residuals from ANOVA
\0x95 check for assumptions needed for a parametric model (normal
distribution, outliers)
\0x95 if assumptions were rejected, confirmatory analysis by a
nonparametric
method
One caveat: You may run into multiplicity issues, i.e., if you perform a
formal test in the decision procedure (let's say Shapiro-Wilk at alpha
0.05), it may be wise to calculate a 95% confidence interval (instead of
90%) to keep the overall error type I (patient's risk) at 5%.
However, recently a european generic company followed this procedure
(they know from literature and rather large previous studies (n=50) that
the innovator product showed some kind of product failure in 5-10% of
cases), and got a deficiency letter, because the Note for Guidance
(http://www.emea.eu.int/pdfs/human/ewp/140198en.pdf) states
"Pharmacokinetic parameters derived from measures of concentration, e.g.
AUC, Cmax should be analysed using ANOVA."
At a meeting at the agency they made the statement "We do not like
nonparametrics". Period.
What next?
Helmut
--
Helmut Schutz
BEBAC
Consultancy Services for Bioequivalence and Bioavailability Studies
Neubaugasse 36/11
A-1070 Vienna/Austria
tel/fax +43 1 2311746
http://BEBAC.at Bioequivalence/Bioavailability Forum at
http://forum.bebac.at
http://www.goldmark.org/netrants/no-word/attach.html
Back to the Top
The following message was posted to: PharmPK
Dear Marival,
yes, I think you are exactly right!
>From what I understand is that they would expect that
>if the two formulations have shown BE in vivo, the
>profiles "would be expected to be similar to those of
>the reference product" and if the diss profile of tes
>is dissimilar, they recommend you re evaluate and
>optimize it...but from the last sentence I understand
>that if this is not posible, you can set the diss
>spec. based on your dissolution method.
Ok, if we don't have a discriminatory dissolution test, we are expected
to develop one. If that fails, we can set your own specification, yes,
but based on what?
On a non-discriminatory test!
So our new spec's are nice, but they have no "meaning", /i.e./, no
predictive power at all.
So what?
I think it's hypocrisy (if negatively seen) or at least some kind of
"self fullfilling prophecy" (if neutrally seen).
If we don't have a valid iv/iv correlation, similar in-vitro curves
simply
don't mean anything (because formulations /may or may not/ be
bioequivalent
in-vivo).
Just imagine:
(1) *in-vitro curves are similar:
* we are happy and start the BE study, getting one
of two possible outcomes
(a) formulations are bioequivalent:
we are even more happy and present our results to the regulators
(b) formulations are not bioequivalent:
we curse at the people from the galenic development department
we don't present our result to the regulators
we reformulate (this is obvious, because you were not BE)
and try to get a better dissolution method, until
(I) we end up with situation (1), or
(II) we give up and try to life with situation (2)
(2) *in-vitro curves are not similar:**
* we try to change you dissolution method until we end up with either
(a) situation (1), or
(b) we give up and give it a try in-vivo, since either
(I) there exists no in-vitro/in-vivo correlation, and we
are ending up playing around with dissolution testing,
when it already would have been possible to show
bioequivalence...
(II) there is a "true" correlation, but we are too stupid
to find the right conditions ;-)
The crazy thing is, that we all - or better, many of us - silently
assume
some kind of in-vivo/in-vitro correlation for /all drugs across all
formulations/. Nobody would dare to start an in-vivo study with clearly
dissimilar dissolution profiles.
I think regulators are aware of this, since I never heard of any
study which was able to show bioequivalence, resulting in a deficiency
letter concerning the dissolution tests.
Best regards,
Helmut
--
Helmut Schutz
BEBAC
Consultancy Services for Bioequivalence and Bioavailability Studies
Neubaugasse 36/11
A-1070 Vienna/Austria
tel/fax +43 1 2311746
http://BEBAC.at Bioequivalence/Bioavailability Forum at
http://forum.bebac.at
http://www.goldmark.org/netrants/no-word/attach.html
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)