Bioequivalence and Bioavailability Forum

Hi Beholder,

❝ An article which might serve as "guideline" for conducting telmisartan BE studies in Russia written by expert organization representative. Open access and abstract in English is available.

THX; interesting. In Table 3 three studies are give twice each. Could reproduce the pooled CV of C_max (43.57%). The upper limit 44.84% is based on a 75% CI. I would use a 80% CL. Given.

library(PowerTOST)
design <- "2x2x2"
n      <- c(85, 36, 40, 59, 60, 50, 40)
CV     <- c(43.27, 32.72, 33.67, 48.35, 49.45, 34.12, 55.34)/100
df     <- n - 2
study <- data.frame(design = "2x2x2", n = n,
                    df = df, CV = CV, CV.CL = NA,
                    stringsAsFactors = FALSE)
for (j in seq_along(CV)) {
  study$CV.CL[j] <- signif(CVCL(CV = study$CV[j], df = study$df[j],
                                side = "upper", alpha = 0.25)[["upper CL"]], 4)
}
print(study); print(CVpooled(study, alpha = 0.25), verbose = TRUE)

#   design  n df     CV  CV.CL
# 1  2x2x2 85 83 0.4327 0.4677
# 2  2x2x2 36 34 0.3272 0.3701
# 3  2x2x2 40 38 0.3367 0.3781
# 4  2x2x2 59 57 0.4835 0.5331
# 5  2x2x2 60 58 0.4945 0.5449
# 6  2x2x2 50 48 0.3412 0.3778
# 7  2x2x2 40 38 0.5534 0.6280
# Pooled CV = 0.4357 with 356 degrees of freedom
# Upper 75% confidence limit of CV = 0.4485

The recommended sample of 20 subjects is for a 2×2×4 design and ABEL based on a GMR of 0.95 and CV 0.45). I would not be that optimistic and suggest a GMR of not better than 0.90 (42 subjects).
Not sure whether it makes sense to work with the pooled CV at all (different sampling schedules and bioanalytical methods).

I like that the experts recommend 2-sequence 4- (TRTR|RTRT) and 3-period (TRT|RTR) full replicate designs and not the lousy partial replicate. Kudos!