Bioequivalence and Bioavailability Forum

Hi Thalita,

Since you posted in two different threads, I reply your questions with this thread all together.

❝ You recommend me this article! I request the Excel worksheet but the authors did not answer to my request. Did you have this worksheet ?

Once I read your post, I tried to find the excel file with no luck. It's been a long time ago. Sorry about that. I think it should be not that difficult to re-create that excel file. As far as I can remember, the excel file was just used to validate Wagner-Nelson method used in ivivic for R. You should be able to find it from some textbooks, such as Shargel and Yu's Applied Biopharmaceutics & Pharmacokinetics. Even if you can access that excel file, it may not be helpful for you since you do not have iv data in your case. Wagner-Nelson approach requires iv/solution/suspension data to calculate 'Kel' first, and then use 'Kel' to back-stripping the drug conc. vs time curve for oral dosage forms to get 'Ka' for each subject. If you really need Wagner-Nelson method, that should be very simple. You can dig in ivivc for R and find the file 'WagNel.R' from source tarball. If you had different sampling times for in-vitro and in-vivo, you can find another nice R package called Rivivc. Rivivc uses numerical convolution/deconvolution approach to establish ivivc. So it does not require same sampling times for in-vitro and in-vivo data. Please let me know if you are not familiar with R, although I do not think you need ivivc at all. I am very happy to help.

❝ It is not the classical "FDA-IVIVC" because it is an immediate release drug, and I just have one formulation, with different dissolutions conditions.

See? Here is the question. One IR formulation and dissolution profiles with different pH media? You consider to establish ivivc for your case? I doubt it.

❝ ... I just tried to calculate Fabs based on Wagner Nelson equation, but the values ​​of Fabs were higher than 1. The Kel is too little. On my work group we have a study that used same in vivo data with GastroPlus and establish the correlation based on one-compartment model, because we do not have intravenous data.

Again, if you do not have iv data for each subject, you cannot predict 'Ka' or 'Fabs' correctly. The value of 'Kel' should be estimated from the same subject. Otherwise, the result may not be reliable.

❝ ... I tried the ivivc for R but Mr. Yung-jin Lee told me that may be not works.

That is correct.

❝ There is another thing, the time scale problem. My in vivo date has 192h and my in vitro date has just 2h30. I read other articles that used Levy's plot to scale time, could any one give me an MS Excel example for this?

Did you imply that different sampling times (points) be used for in-vitro and in-vivo data in your study? If yes, look for Rivivc. However, even Rivivc requires iv/solution/suspension data. There should be no excel example for this, as far as I know.

Hi Thalita,

If you do not have iv data for your subjects and you still want to get Fabs (fraction cumulative absorption in %) and then do Levy plots, I think probably you can do the following:

1. fit your data with one-compartment PK model, to get Ka, Kel & Vd;
   
2. use obtained 'Kel' to calculate Fabs (R codes as follows);
   
3. plot Fabs vs. FRD (from your dissolution profiles, varied pH) and get Levy plots.

In this case, you do not need iv data and no more IVIVC or Wagner-Nelson. However, I am not sure if this is what you want.

❝ ... I just tried to calculate Fabs based on Wagner Nelson equation, but the values ​​of Fabs were higher than 1. The Kel is too little. On my work group we have a study that used same in vivo data with GastroPlus and establish the correlation based on one-compartment model, because we do not have intravenous data.

R codes
for (j in 1:length(W.split)){  ### 'W.split' contains in-vivo data for all subj., j = # of subj.
  auc<- 0;Ft<- 0;Fabs<- 0
  for(i in 2:length(W.split[[j]][["time"]])){
# calculate AUC for each time point
  auc[i]<- (W.split[[j]][["time"]][i]-W.split[[j]][["time"]][i-1])*(W.split[[j]][["conc.obs"]][i]+W.split[[j]][["conc.obs"]][i-1])* 0.5
  auc[i]<- auc[i]+auc[i-1] 
# calculate F(t): fraction of absorption at time t
  Ft[i]<-W.split[[j]][["conc.obs"]][i]+kel*auc[i]}
# calculate AUC (0~INF)
  auc.infinity<-W.split[[j]][["conc.obs"]][length(W.split[[j]][["conc.obs"]])]/kel
  aucINF<-auc[length(W.split[[j]][["conc.obs"]])]+auc.infinity
# calculate Fabs(t): cumulative absorption fractions at time t
  Fabs<- 0
  for(i in 2:length(W.split[[j]][["time"]])) Fabs[i]<-(Ft[i]/(kel*aucINF))*100