3 replies to this topic

[sjpan]

Dear all,

 

Our drug was administered as a single IV bolus followed by constant rate IV infusion. The IV infusion started 30 min after the bolus injection and lasted 7-14 days (non-stop). PK samples were taken during the infusion and after the end of infusion.

The drug contains endogenous compound, thus, the observed concentration is a combination of our drug and the endogenous compounds. The concentration of the endogenous compound also varies over time (based on the data from the control group). 

 

We tried to model the data as 1-compartment IV bolus + IV infusion (with lag time). The model text is shown below:

 

test(){

            deriv(A1 = - (A1 * Ke))

            urinecpt(A0 = (A1 * Ke))

            C = A1 / V

            dosepoint(A1, idosevar = A1Dose, infdosevar = A1InfDose, infratevar = A1InfRate)

            dosepoint2(A1, tlag = (0.5), rate = (41666667))

            error(CEps = 0.1)

            observe(CObs = C * (1 + CEps))

            stparm(V = tvV)

            stparm(Tlag = tvTlag)

            stparm(Ke = tvKe)

            fixef(tvV = c(, 50000000, ))

            fixef(tvTlag = c(, 0.5, ))

            fixef(tvKe = c(, 0.000962777, ))

}

 

the initial estimates (tvK, tvV)  came from the non-compartmental analysis of the plasma concentration time profile. 

 

However, we got the following error:

 

"Model execution failed.

 Initial parameter values result in -LL = NaN

 Restart from new values

 Error: Model Exception: Fortran Exception"

(the full error message is attached).

 

Can someone please explain what these error messages mean and how to troubleshoot?

 

Thank you very much.

Attached Files

•  Warnings and Errors.txt   1.78KB   175 downloads

[Simon Davis]

Hi SJPan, there is no need to have two dosepoints for this situation, simply another dosing record. I am also not sure if you are entering the rate correctly, is that the total amount infused over 168 hours?

Also when you say you have an endogenous compound you can model that e.g. circadian rhythm and we talk about that in our https://www.certarau...oenix-winnonlin

and also in the examples
https://onlinehelp.c...excircadian.htm

I would suggest contacting us for some training or if it is specifc to one project; our office hours;
https://www.certarau...id-office-hours.

as to what is specifically casuing your errors we would need to see the data and would suggest you send your project to support@certara.com if you still have issues after trying my suggestion using a single dose point and checking "infusions possible" ina built in model to start with.

you can see below the initial bolus to 0.5 h then the infusion out to 168 h.
Simon.

Attached Thumbnails

• 

Edited by Simon Davis, 10 June 2022 - 06:25 AM.

[sjpan]

Hi Simon,

 

Thank you for the feedback. The infusion rate is 1000 mcg per day (24 hr) and we entered the total amount infused over the entire treatment period. 

We re-ran the individual model fitting using another dosing record (as you suggested) followed by simulation (the project file is attached).

 

Is there a way to get the Css and AUCtau from the simulation data?

 

Thank you very much for your help.

 

SJ

 

 

 

Attached Files

•  IV bolus followed by IV infusion_20220614.phxproj   1.28MB   183 downloads

[Simon Davis]

Do you have an NLME license as that may help your fitting ?

 

I am assuming you are considering your dosing interval to be 1 day, however you may not be at steady state, I am getting a Ke half life of 10 days, whilst you would normally say SS is achieved after 5-7 half lives.

 

I think you could also look at re-paremtising you model.  Additive error seems better with this small range in concs and also clearance seems to improve loglikelihood and CV%

 

think about what you are really trying ot assess, a 'typical conc' or  peak on last day of your dosing?

 

try this code below, ast line (ranef) can be comment out if you have no NLMElicense.

 

then request a table with AUCinf and AUCtD7

 

 Simon.

test(){
    deriv(A1 = - Cl * C)
    dosepoint(A1, idosevar = A1Dose, infdosevar = A1InfDose, infratevar = A1InfRate)
    C = A1 / V
    error(CEps = 15)
    observe(CObs = C + CEps)
    stparm(V = tvV * exp(nV))
    stparm(Cl = tvCl * exp(nCl))
    fixef(tvV = c(, 37393.001078584, ))
    fixef(tvCl = c(, 5000, ))
    secondary(Ke = tvCl/tvV)
    secondary(Tinf = A1InfDose/A1InfRate)
    secondary(AUC = A1InfDose/tvV/Ke)
    secondary(Coef = A1InfDose/(Tinf*tvV*Ke))
    secondary(Cmax = Coef*(1-exp(-Ke*Tinf)))
    secondary(AUMCIV = A1InfDose/tvV/(Ke*Ke))
    secondary(AUMC = AUMCIV+(Tinf*AUC)/2)
    secondary(MRT = AUMC/AUC-Tinf/2)
    secondary(VSS = (A1InfDose*(AUMC/AUC-Tinf/2))/AUC)
    secondary(Ke_hl = log(2)/Ke)
        deriv(AUCtD7 = (t>7)*(t<8)*C)
    deriv(AUCinf=C)                       #AUC Calculation
    ranef(diag(nV, nCl) = c(1, 1))
}

Attached Thumbnails

• 
•