Log in
Register|
dixit ★ India, 2009-04-07 13:17 (6279 d 01:20 ago) Posting: # 3492 Views: 7,797 |
|
|
Dear all, Can any one please tell me about Finney approach for calculating the confidence Intervals in a dose response Bioequivalence study of acarbose. Please do provide me with a relevant literature if possible. Thanks and Regards,  DIXIT. |
|
earlybird ☆ 2009-04-08 10:10 (6278 d 04:27 ago) (edited on 2009-04-08 10:35) @ dixit Posting: # 3498 Views: 6,617 |
|
|
Dear DIXIT, This was the first introduction a member of the forum (I do not know the name!) gave to me. And that was really helpful. Consider a situation where you give two levels of test (T) and two dose levels of reference (R), and measure the effect of all four treatments. This gives you four data points: Effect of T at dose level 1 Effect of T at dose level 2 Effect of R at dose level 1 Effect of R at dose level 2 Now imagine a line drawn through the points for T and also a line drawn through the points for R. Importantly, to assure that there is dose-response-relationship (in other words: to ensure sensitivity) both lines must have a slope that is significantly different from Zero. In addition, it is necessary to test that two slopes are not significantly different (departure from parallelism). This test should not be significant. If there is departure from parallelism then the two products are not similar. Now for the really interesting stuff: Finney's approach is to compare on the dose-axis, not on the effect axis. This contrasts with many people's idea of comparability, but it makes good sense. It is simply "relative potency" and it avoids the problem of defining an equivalence margin in terms of effect. When there is no departure from parallelism then the horizontal difference between the two lines projected onto the logarithmic dose axis is constant. From this we can calculate the T:R ratio, and furthermore with some statistical black magic(  ) also a confidence interval for the T:R ratio through the ANOVA's root mean square error.And that's in fact that. It is actually almost the same principles that are laid out in the Ph Eur test for biological substances (reference below). In don't think they use the term Finney Bioassay there but the principle is exactly the same. Have Fun Earlybird References: Ph Eur: 5.3. Statistical analysis of results of biological assays and tests Formula for crossover-design could be found: Guidance to establish equivalence or relative potency of safety and efficacy of a second entry short-acting beta-2 agonist metered dose inhaler -- Edit: Full quote removed. Please see this post! [Jaime] |
|
dixit ★ India, 2009-04-10 09:56 (6276 d 04:41 ago) @ earlybird Posting: # 3504 Views: 6,573 |
|
|
Thanks for Your reply, Now coming to my study design it is a Anti diabetic drug, dose response cum bioequivalence study. As per the study design it is a four treatment four period four sequence crossover study. We will administer three different doses of reference (25mg, 50mg and 75mg) and a test drug (50mg) in different periods. And each period has two days. On day1 we will be administering 75g sucrose solution and on day2 we will be administering one of the study formulation along with sucrose solution and collect the blood samples. From these samples we are measuring the glucose content. From these measured glucose contents we are going to calculate three parameters Cmax (Max observed glucose content), AUEC0-2 and AUEC0-4 for day1 and day2 separately and calculate the difference of these parameters (Day1-Day2). And now how can we identify the minimum effective dose (reference) for comparing the test product for bioequivalence. As per the study design we have to apply ANCOVA on the ln-transformed differences of the above calculated parameters (Cmaxdiff, AUCE0-2diff & AUEC0-4diff) by considering the day1 parameters as covariates for determination of minimum effective dose (Reference 25mg/50mg/75mg). Can you please suggest me what is the procedure for determination of minimum effective dose and how can we determine the homogeneity of slopes between formulations. Pls do guide me with relevant literature if possible. Regards, DIXIT. |
|
earlybird ☆ 2009-04-15 12:10 (6271 d 02:27 ago) @ dixit Posting: # 3531 Views: 6,463 |
|
|
Dear DIXIT! As far as I could understand Finney, I miss a second dose of test drug in your design. Otherwise I think it is not possible to calculate dose potency. ANCOVA is possible of course as you can include co-variables in your model. I already mentioned two literatures, one more could be: Pharmaceutical Statistics, Applications and Clinical Applications by Sanford Bolton, Charles Bon. This book describes Finney in general. But I think the already mentioned should be all what you need! In European Pharmacopoeia there are a lot of examples for parallel and cross-over, which can be used and programming can be done step by step and there is also described how to test slopes. Regards, Earlybird -- Edit: Full quote removed. Please see this post! [Ohlbe] |
Ing. Helmut Schütz