Bioequivalence and Bioavailability Forum

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 -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)