Hi all,
I came across a paper describing a method for estimating the time to steady state analysis in SAS (see the attachment). One of the methods, NOSTASOT (stepwise test of linear trend), can be executed using the Linear Mixed Effects model in WinNonlin (WNL). In WNL, there are “Contrasts” and “Estimates” tabs; however, I am unsure how to use these tabs to perform the analysis in the same way SAS does.
Does anyone have experience assessing the time to steady state using WNL?
Thanks,
Peggy
Estimating Time to Steady State Analysis in SAS.pdf (437 KB)
Hi Peggy,
I’m pretty sure you could code this approach in PHX. However, may I ask why qou are interested in it?
All [sic] statistical approaches to assess achievement of steady-state are problematic.Maybe this article helps to understand why.
Hi Helmut,
In the FDA labeling content and format guideline (https://www.fda.gov/media/74346/download) p. 8 " The introduction should also include clinically useful information on the expected exposure of the drug (e.g., the maximum plasma concentration (Cmax), the area under the plasma drug concentration over time curve (AUC)) at the approved recommended dosage, time to steady state, accumulation ratio following multiple dosing, and changes in PK over time." Also, p. 10 “The half-life value reported should usually be the half-life based on the time to reach steady state (i.e., the effective half-life).”
For a new drug clinical pharmacology study, I sometimes see the big pharma assessing this using the linear regression (they may use the linear mixed effect model, but I am not quite sure) to assess the time to steady-state and present this in the product labeling information.
Besides, when I confirm the steady-state is achieved, then the steady-state PK parameters (e.g., CLss) can be reported.
Does it sound reasonable to test on a group of subjects with sufficient number (n)? I’m not sure what the n should be, but I often see them perform this on a group of subjects (n>= 12) receiving the same dosing regimen.
Thanks,
Peggy