Bioequivalence and Bioavailability Forum

Dear all,

We did a typical 2×2 cross-over study for BE study in 23 subjecct. But the results was not BE. We handled the data by PHX 6.3 and got its inter- and intra-subject CV(>30%) (see below). We considered it is a highly variable drug (HVD) and also realized the sample size was not enough. So, we want to add some more subjects and evaluate the BE again with all the data. I am confusing about the sample size calculation.

Dependent   Units    Parameter        Estimate
Ln(Cmax)    ng/mL    Var(SEQUENCE*ID) 0.07
Ln(Cmax)    ng/mL    Var(Residual)    0.35
Ln(Cmax)    ng/mL    Intersubject CV  0.22
Ln(Cmax)    ng/mL    Intrasubject CV  0.65
Ln(AUClast) hr*ng/mL Var(SEQUENCE*ID) 0.38
Ln(AUClast) hr*ng/mL Var(Residual)    0.17
Ln(AUClast) hr*ng/mL Intersubject CV  0.67
Ln(AUClast) hr*ng/mL Intrasubject CV  0.41

1. Which PK parameter should I use to calculate the sample size? I chose the intra-sub CV of AUClast and used PowerTOST to calculate the sample size (see results below):

sampleN.TOST(CV=0.41)
+++++++++++ Equivalence test - TOST +++++++++++
            Sample size estimation
-----------------------------------------------
Study design:  2x2 crossover
log-transformed data (multiplicative model)

alpha = 0.05, target power = 0.8
BE margins        = 0.8 ... 1.25
Null (true) ratio = 0.95,  CV = 0.41

Sample size (total)
 n     power
70   0.810715

As I understand n=70 means that there are actually only 35 subjects participate the study. So we only need 35-23=12 subjects more to perform a 2-way cross design. Is it correct?

2. If I consider it as a two-stage design, should I use alpha=0.0294 to calculate the sample size (see below)? If so, we need more subjects(40-23=17).
sampleN.TOST(alpha=0.0294,CV=0.41)
+++++++++++ Equivalence test - TOST +++++++++++
            Sample size estimation
-----------------------------------------------
Study design:  2x2 crossover
log-transformed data (multiplicative model)

alpha = 0.0294, target power = 0.8
BE margins        = 0.8 ... 1.25
Null (true) ratio = 0.95,  CV = 0.41

Sample size (total)
 n     power
82   0.802106

3. Is the Null (true) ratio the one calculated from the first 23-subject study? PHX 6.3 results below. If so, I should set "theta0=0.98" in the sampleN.TOST function.
Dependent    Ratio_%Ref_
Ln(Cmax)     102
Ln(AUClast)  98

Thank you in advance for any comments.

---

Edit: Category changed. [Helmut]