Bioequivalence and Bioavailability Forum • Need help downloading a publication

selwa ts
☆

Algeria,
2021-09-13 17:25
(1321 d 02:28 ago)

Posting: # 22568
Views: 3,526

Need help downloading a publication [Power / Sample Size]

Hello dear community, I need to download a PDF article which I did not have access to, and at the same time I would like to have references if possible in the same subject: calculation of the number of subjects in bioequivalence and evaluation of sample size.

Référence of the article : Estimated coefficient of variation values for sample size planning bioequivalence studies DOI:10.5414/CPP39037.

I count on your precious help dear members

Helmut
★★★

Vienna, Austria,
2021-09-13 17:52
(1321 d 02:01 ago)

@ selwa ts
Posting: # 22569
Views: 2,837

Old references…

Post reply

Hi Selwa,

❝ Hello dear community, I need to download a PDF article which I did not have access to, and at the same time I would like to have references if possible in the same subject: calculation of the number of subjects in bioequivalence and evaluation of sample size.

A more exhaustive list (also quoted in the article you mentioned) is given by Steinijans et al.*

❝ Référence of the article : Estimated coefficient of variation values for sample size planning bioequivalence studies DOI:10.5414/CPP39037.

I have it. If you want I can send you a scan. However, old references are problematic. You have no clue about the clinical performance and the bioanalytical methods used. Therefore, you should be extremely cautious relying on the reported values.
If you want to design a study, I suggest to find recent publications about the drug you are interested in. The FDA’s ANDA files and European Public Assessment Reports are helpful. An advice: The larger a study is, the more reliable the reported variability. If the CV is not given, you can calculate it from the confidence interval, the design, and the sample size (see there).

Steinijans VW, Sauter R, Hauschke D, Diletti E, Schall R, Luus HG, Elze E, Blume H, Hoffmann C, Franke G, Siegmund W. Reference tables for the intrasubject coefficient of variation in bioequivalence studies. Int J Clin Pharm Therap. 1995; 33(8): 427-430. PMID 8556220.

—
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

selwa ts ☆ Algeria, 2021-09-14 14:51 (1320 d 05:02 ago) @ Helmut Posting: # 22570 Views: 2,850	Old references… Post reply
	Hello Sir, Thank you for your feedback, i will be very grateful if you forward me the scan of the article and others if you have any.

Helmut
★★★

Vienna, Austria,
2021-09-14 19:46
(1320 d 00:07 ago)

@ selwa ts
Posting: # 22571
Views: 2,789

Old references…

Post reply

Hi Selwa,

❝ […] i will be very grateful if you forward me the scan of the article and others if you have any.

Check your inbox. ;-)

But again: Don’t rely on the results. An [image]

-script at the end. Here the 95% confidence interval of the two studies of 500 mg IR metformin reported by Yuen et al.:

study n CV.AUC CI.CV.AUC CV.Cmax CI.CV.Cmax 1 24 15.1% 11.7–21.5% 16.3% 12.6–23.2% 2 12 9.5% 6.6–16.8% 12.1% 8.4–21.4%

Be prepared for the worst.
What shall we think about the three studies of 300 mg IR ranitidine in twelve subjects? The CV of AUC_0–∞ in the third was twice the one of the second, whereas the CVs of C_max were similar. On the other hand, the CV of C_max in the first was twice the one of the third. That’s not helpful.

If you don’t find references for your drug with more consistent results, I recommend a pilot study or – possibly better – a two-stage design.

library(PowerTOST) studies <- 2L n <- c(24L, 12L) CV.AUC <- c(15.1, 9.5) CV.Cmax <- c(16.3, 12.1) res <- data.frame(study = 1:studies, n = n, CV.AUC = CV.AUC, CI.CV.AUC = NA_real_, CV.Cmax = CV.Cmax, CI.CV.Cmax = NA_real_) for (j in 1:nrow(res)) { tmp <- 100 * CVCL(CV = res$CV.AUC[j] / 100, df = res$n[j] - 2, side = "2-sided") res$CI.CV.AUC[j] <- paste0(sprintf("%.1f", tmp[1]), "\u2013", sprintf("%.1f%%", tmp[2])) tmp <- 100 * CVCL(CV = res$CV.Cmax[j] / 100, df = res$n[j] - 2, side = "2-sided") res$CI.CV.Cmax[j] <- paste0(sprintf("%.1f", tmp[1]), "\u2013", sprintf("%.1f%%", tmp[2])) } res$CV.AUC <- sprintf("%4.1f%%", res$CV.AUC) res$CV.Cmax <- sprintf("%4.1f%%", res$CV.Cmax) print(res, row.names = FALSE)

—
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