## BE parallel design [General Statistics]

Dear PKPDPKPD!

» Should AUC and Cmax be also calculated from the log transformed data?

See my first post. You should only transform the calculated PK

Let's misuse Helmut's data (download here)

Although his data are from a cross-over study, we will use only period 1.

83.657% (63.514% - 110.189%)

I checked the 'manual' calculation in WinNonlin and EquivTest:

WinNonlin: 83.6572% (63.5100% - 110.1958%)

EquivTest: 83.66% (63.51% - 110.18%)

Slight differences seen in results are not uncommon...

» Should AUC and Cmax be also calculated from the log transformed data?

See my first post. You should only transform the calculated PK

**parameters**(not the concentrations).- Calculate AUC from your analytical results by any method you like (trapezoidal rule preferred, but not limited to).

- Cmax is simply the highest measured concentration.

- For the comparison you log-transform AUC and Cmax.

Let's misuse Helmut's data (download here)

Although his data are from a cross-over study, we will use only period 1.

- Change the header of the first column from 'Seq' to 'Trt'

- Delete the column 'Rand'

- Delete the column 'P2'

- log-transform 'P1'

- calculate
*separately*for each treatment:- arithmetic mean: (1: 3.56227, 2: 3.38383)

- exp(arithmetic mean): (1: 35.24321, 2: 29.48349),

- note: these are the geometric means of untransformed data!

- standard deviations: (1: 0.35950, 2: 0.42377)

- variances = SD²: (1: 0.12924, 2: 0.17958)

- n
_{1,2}(group sizes): (1: 12, 2: 12)

- Q
_{1,2}= variance × (n_{1,2}-1): (1: 1.42165, 2: 1.97539)

- arithmetic mean: (1: 3.56227, 2: 3.38383)
- calculate
`R = sqrt[(n`

: 0.16042_{1}+n_{2})/(n_{1}×n_{2})×(Q_{1}+Q_{2})/(n_{1}+n_{2}-2)]

- look up the critical value of the
*t*-distribution for alpha=0.05 with n_{1}+n_{2}-2 degrees of freedom:*t*1.71714

- calculate
*t*× R: 0.27547

- calculate Delta (difference of means Trt 2 - Trt 1): -0.17844

- antilog Delta (= point estimate): 0.83657

- calculate lower/upper 90% confidence limit = Delta ±
*t*× R: lo: -0.45391, hi: 0.09702

- antilog lo and hi: exp(lo): 0.63514, exp(hi): 1.10189

83.657% (63.514% - 110.189%)

I checked the 'manual' calculation in WinNonlin and EquivTest:

WinNonlin: 83.6572% (63.5100% - 110.1958%)

EquivTest: 83.66% (63.51% - 110.18%)

Slight differences seen in results are not uncommon...

—

Regards, Jaime

Regards, Jaime

### Complete thread:

- BE parallel design pkpdpkpd 2007-04-03 17:45 [General Statistics]
- BE parallel design Jaime_R 2007-04-03 18:58
- BE parallel design pkpdpkpd 2007-04-03 19:07
- BE parallel design Jaime_R 2007-04-03 19:41
- BE parallel design pkpdpkpd 2007-04-03 20:14
- BE parallel design Jaime_R 2007-04-03 20:29
- BE parallel design pkpdpkpd 2007-04-03 21:48
- BE parallel designJaime_R 2007-04-04 15:14
- BE parallel design Helmut 2007-04-04 15:31
- BE parallel design pkpdpkpd 2007-04-04 17:32
- BE parallel design Jaime_R 2007-04-04 20:33

- BE parallel design Sathya 2008-09-01 07:21
- BE parallel design Sathya 2008-09-04 12:37

- BE parallel design Helmut 2007-04-04 15:31

- BE parallel designJaime_R 2007-04-04 15:14

- BE parallel design pkpdpkpd 2007-04-03 21:48

- BE parallel design Jaime_R 2007-04-03 20:29

- BE parallel design pkpdpkpd 2007-04-03 20:14

- BE parallel design Jaime_R 2007-04-03 19:41

- BE parallel design pkpdpkpd 2007-04-03 19:07

- BE parallel design Jaime_R 2007-04-03 18:58