Hi All,
I have a question about the dosing setting in NCA.
Drug A was dosed 300 mg on Day 1 and 30 mg on Days 2 to 6. We collected the blood sample on Day 4 (pre-dose and several time points post-dose) for PK analysis. I assume that PK concentration is at steady state at that time points. How to set dose and tau in NCA? Or making the situation more complex, if drug A was dosed 150 mg twice on Day 1 and 30 mg on Days 2 to 6, how to set up the dose and tau?
Thank you very much for your help!
LLLi
LLLi, this regimen is somewhat problematic for an NCA analysis since the assumption is that the same dosing will have been given to attain steady state, whereas your proposed regimens have loading doses. This will affect your confidence in asssement of dose related parameters such as CL and V. That said since after the loading dose your dosing is regular and constant I would propose that you enter that information in your analysis but be sure to discuss your assumptions in your report. Incidentally why did you make your PK profile day on Day 4 when you dosed until Day 6? Same with your second design - the loading dose will have less and less impact on your profile as time passes. Simon. EDIT : added a plot of simulations of your two designs versus a consistent 30mg once per day. I chose 0.05 as my Ke value since that gave a half-life of nearly 14h because you said at day 4 (assume would dose at 72h) you assumed steady state- so this would be after 5 half- lives. 5*13.86h= 69.3h. Simon.
Hi Simon,
Thank you for your reply!
According to your figure, it seems the effect of loading dose is very small on Day 6 for all three different regimens. It should be better to collect the samples on Day 6 to analyze PK as you said. And I can use 30 mg as dose and 24 hr as tau to do the NCA. Please correct me if I am wrong.
Furthermore, I have a question about the report of t1/2. I read some PK papers. Some papers compared the arithmetic mean of t1/2 under different conditions while some compared geometric means of t1/2. I know that we compare the geometric mean of Cmax and AUCs because these PK parameters are log normal distributed. But for t1/2, I am not sure. Any input is appreciated!
Thank you!
LLLi
Hi LLLi,
Yes, simple modelling and simulation is useful to plan your NCA studies and should be used more than it is in my opinion. So yes if your halfife was around 14h (or less) I would be pretty confident to perform my NCA on Day 6 as ‘steady state’ with 30 mg as dose and 24 hr as tau tor all three different regimens. I would still discuss briefly in the report that the dosing started with a loading dose and there might be a small impact on some parameter’s estimation but as the plot showed this should be minimal.
Regarding the reporting of t1/2, of those two, I would prefer geometric over arithmetic. In a previous role the statisticians recommended a harmonic mean as it is more stable regarding outliers, and i do try to report that now. See also;
http://forum.bebac.at/forum_entry.php?id=2920
Simon
Hi Simon，
Thank you for your reply! I never used harmonic mean in my study. 
Talking about the simulation of drug A, what method did you use to do the simulation? NPS? There several simulation methods in WinNonlin, including built-in PK model, NPS and NLME. Is there any difference or preference when we choose the simulation methods?
Thank you!
LLLi
Hi LLLi, it depends really, I would prefer to use a model versus NPS since the latter is so dependent on the sampling of the original profile. by default all of them make assumptions about e.g. linear kinetics, but with a model you can account for that more easily.
Between WNL classic models and Phoenix models it is really a case of personal preference for built-in models, in this case I just used WNL classic because I’ve done this a lot over the last 15 or years so it was familiarity. If I needed a custom model then I would definitely build it as a phoenix model as much easier to simulate these.
we learn about these tricks in the introduction to WinNonlin course which is available onsite, as a public course or online;
http://www.certarauniversity.com/lms/index.php?r=course/details&id=26
Simon.
attached file in v6.4
LLLi_simSS profiles.phxproj (285 KB)
Hi Simon,
Thank you for your reply during the weekend!
In your project, the results show that the VIFs for V and K10 of 1x regimen is the smallest among the three regimens. Can I say that 1x regimen is the best design?
Thank you!
LLLi
I can’t recall if my other settings were the same between the project i.e. same number of samples (or same number of doses if comparing sampling regimens), otherwise you can’t determine which element of your design contributed to the change in VIF. But yes, all other points being equal, for designing a future experiment that would be the best design for estimating V and K10. However if you were interested in other parameters e.g. Cmax then it might be another design.
Simon.
Hi Simon,
I have some question about the setting of dose, time and tau in NCA again 
I got some data to analyze by using NCA. The study design is as follows.
X was given from day 1 to 7. Its dose is 225 mg daily (75 mg X 3, and the dose interval within one day is 6 hr).
On day 1 and 7, the first dose in the morning (75 mg) was given and the PK samples were collected at pre-morning dose, 1, 2, 3, 4, 6, 8, 10 and 12 hr post-morning dose.
The concentration-time curve has two peak concentrations since the second dose 75 mg was given at 6 hr. Any suggestions about the setting of the dose and tau for day 1 and 7 in NCA? Or split the data into 2 sets: the first 6 hr for the first 75 mg, and the remaining 6 hr for the second 75 mg?
Thank you very much and have a nice weekend!
LLLi
Sorry LLLi, but it’s hard to know what ot advise when you’ve already run a study with inequal dosing intervals. it sounds like your 3 doses happened at 0, 6, 12 and then again at 24 (or 0h for the next day).
AS I’ve said before NCA formulas are for assessment of single dose events at either naive first dose or at steady state. ASsunming you’ve dosed for 5-7 half-lives you’re probably at steady state but I would still be careful to assume complete accuracy about clearance and other dose dependent parameters since you they don’t truly follow the formula’s assumptions.
Then yes you could assess, on Day 7, two intervals 0-6 and 6+ as you described.
However on day 1 I think you will struggle to make all the assessments you are asking with confidence. in future i would define day in the protocol as a single dose day out to e..g 48 and then commence multiple dosing (ideally at equal intervals).
Simon
Hi Simon,
Thank you for your prompt reply! I was very frustrated when I looked through the data
For the data of day 7 (assuming steady state at that time), If I split the data into two sets (0-6 hr and 6-12 hr), we can get two sets of PK parameters (AUCtau, Cavg, Cmax,ss ect) by NCA? The dose is 75 mg and the tau is 6 hr??
For the data of day 1, if I only use the 0-6 hr data to do NCA (dose = 75 mg, time = 0) and if the time range for lambadz is larger than 2 half-lives, can I still say that the parameters are reliable?
Thank you!
LLLi
I personally would tend to agree with you that for Day 7 (assuming steady state at that time), you can split the data into two sets (0-6 hr and 6-12 hr), and get two sets of PK parameters (AUCtau, Cavg, Cmax,ss ect) by NCA with dose or 75 mg and the tau is 6 hr. You can also compare to the twoo to see if your assumption of steady state is supported.
And for the data of day 1, if you only use the 0-6 hr data to do NCA (dose = 75 mg, time = 0) and the time range for lambadz is greater than 2 half-lives, then probably the parameters are reliable, but remember if you have more than one compartment; what you see at low doses as elimination may in fact be the distribution half life.
Simon
Dear LLLi and Simon,
I personally would tend to agree with you that for Day 7 (assuming steady state at that time), you can split the data into two sets (0-6 hr and 6-12 hr), and get two sets of PK parameters (AUCtau, Cavg, Cmax,ss ect) by NCA with dose or 75 mg and the tau is 6 hr. You can also compare to the twoo to see if your assumption of steady state is supported.
I suppose the apparent half life is relatively small (otherwise we do not have to dose subjects 3 times per day, right?)
in this case the Ctrough at 0hr cannot be alike Ctrough at 6hr even without C-dependent parameters. I would not name this condition steady state.
If Ctrough is relatively small, than parameters could be similar. But in this case what is the purpose of 3-unequal-times-per-day dosing?
Mittyright
Thank you Simon and Mittyright!
Mittyright: We have the data of Ctrough on Day 1, 2, and 4 before the morning dose. I will check Ctrough to see if steady state reached on Day 7.
Simon: Would you please provide more information about “if you have more than one compartment;what you see at low doses as elimination may in fact be the distribution half life” above. Why did you emphasize low dose?
Thank you!
LLLi
Just because it is often not apparent in an SAD study until you get to higher dose levels as the elimination phase is lost below the LLoQ
Simon
Hi Simon,
I have some question about the dosing setting in NCA again .
I have PK data from pre-dose to 72 hr and the dosing time is 200 mg at time=0 and 200 mg at time=12 hr (BID). For this situation, if i want to perform NCA using all data, I can only set dose = 400 mg and time = 0? Or any other method for setting dosing I didn’t find?
Thank you!
LLLi
yes there is an option to enter TAU; the dosing interval, when assessing the profile following a dose at steady state.
So in your NCA, for estimates of Clearance and Volume to be valid you can split up your data to assess two profiles.
Day 1, 0h dose 200mg (leave tau blank)
and on Day 4 you can assess data from your last dose using a tau of 12h and dose of 200mg if you think steady state is likely.
But, as I’ve said before in this thread, you cannot use NCA to assess data from a profile with more than one dose without being aware of the potential inaccuracies for any dose related parameters. You could put all the data in one profile in NCA and get some results but without a valid dose input, what conclusons can you draw from that?
What are really trying to assess/report with this study? It’s much better to have considered the study analysis when designing the study rather than being left with less useful data after the event.
Simon.
Hi Simon,
Sorry for the confusion. The subjects took 200 mg twice (12 hours apart). The blood samples were collected from pre-1st dose to 72 hrs.
We want to determine the dose regimen for next cohort via our current PK data. For this purpose, we may only need Cmax and AUCs, which is not related to dose? But for the other dose-related parameter, like clearance and Vz, the results are not “real”.
Please correct me if I am wrong.
Thank you!
LLLi
