How to define different treatment route in mrgsolve

[JoanaXu]

Hi Everyone,
I am trying to use mrgsove(v1.5.2) to conduct a poppk simulation. My popPK model describes PK conc after both Iv infusion and PO administration. Besides, the clearance differs with different gender. any I tried to integrate the change in PK in $MAIN with ifelse wording but failed with either clearance or F_CENT don't change as desired, do you have any solution for this case? here is my "useless" code:

code <- '$THETA @Annotated
2.88 : 1. Female_Clearance (L/h)
2.72 : 2. Male_Clearance (L/h)
4 : 3. Central volume (L)
0.2 : 4. Absorption rate constant (1/h)
0.6 : 5, ALAG1, Absorption lag (h)
0.5 : 6. Bioavalibility PO

$PARAM @Annotated
MWT : 32.4 : Median Bodyweight(kg)

$CMT
GUT
CENT

$GLOBAL
#define CP (CENT/V)

$MAIN
//initial dose regimen setting
double ROUTEN = 0; //Route of administration (0=IV infusion, 1=PO)
double GENDER = 0; //Gender(0=Female, 1=Male)
double WT = 30; // Body weight(kg)
double amt = 480000; //Dose amount (ug)

double TVCL_F = THETA1;
double TVCL_M = THETA2;
double TVV2 = THETA3;
double TVKA = THETA4;
double TVALAG1 = THETA5;
double TVF1 = THETA6;

double CL_F = TVCL_F * pow(WT/MWT, 0.75);
double CL_M = TVCL_M * pow(WT/MWT, 0.75);
double V2 = TVV2 * pow(WT/MWT, 1);
double ALAG1 = TVALAG1;
double KA = TVKA;
double F1 = TVF1;

if(GENDER==0) {
double CL=CL_F;
} else {
CL= CL_M;}

if(ROUTEN==0) {
F_CENT=1;
} else {
ALAG_GUT=ALAG1;
F_CENT=F1;}

$ODE
if(ROUTEN>0){ // PO
dxdt_GUT = -KA * GUT;
dxdt_CENT = KA * GUT - (CL/V2)*CENT;
}else{// //iv INFUSION
dxdt_GUT = 0;
dxdt_CENT = -(CL/V2)*CENT;
}

$CAPTURE@annotated
CL: clearance used
CL_F: clearance_female
CL_M:clearance_male
CP=CENT/V2: Central concentration
'
mod <- mcode("model_test", code)

[kylebaron]

Hi @JoanaXu -

Thanks for reaching out and for sharing the code. I think the model is really close. I've put some suggestions in revised model code below:

1. Make GENDER and WT parameters for the model; this will let you update them either from your R code or from the data (see this page in the user guide for more info).
2. Doses (amt) should be set in the data set (see code below)
3. For this model, you don't need to tell the model itself the ROUTE (iv or po); just set up the model for either and then dose into GUT or CENT (compartment 1 or 2) as bolus (RATE = 0) or infusion (RATE > 0); all of this is handled in the input data set
4. Specifically, there is a single set of differential equations for doses of any route

Check out the code below and let me know what you think!

Kyle

library(mrgsolve)
#> 
#> Attaching package: 'mrgsolve'
#> The following object is masked from 'package:stats':
#> 
#>     filter
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

code <- '
$THETA @annotated
2.88 : 1. Female_Clearance (L/h)
2.72 : 2. Male_Clearance (L/h)
4    : 3. Central volume (L)
0.2  : 4. Absorption rate constant (1/h)
0.6  : 5, ALAG1, Absorption lag (h)
0.5  : 6. Bioavalibility PO

$PARAM @annotated
MWT : 32.4 : Median Bodyweight(kg)

$INPUT @annotated
GENDER :  0 : Gender
WT     : 30 : Body weight (kg)

$CMT
GUT
CENT

$GLOBAL
#define CP (CENT/V)

$MAIN
//initial dose regimen setting
//double ROUTEN = 0; //Route of administration (0=IV infusion, 1=PO)
//double GENDER = 0; //Gender(0=Female, 1=Male)
//double WT = 30; // Body weight(kg)
//double amt = 480000; //Dose amount (ug)

double TVCL_F = THETA1;
double TVCL_M = THETA2;
double TVV2 = THETA3;
double TVKA = THETA4;
double TVALAG1 = THETA5;
double TVF1 = THETA6;

double CL_F = TVCL_F * pow(WT/MWT, 0.75);
double CL_M = TVCL_M * pow(WT/MWT, 0.75);
double V2 = TVV2 * pow(WT/MWT, 1);
double ALAG1 = TVALAG1;
double KA = TVKA;
double F1 = TVF1;

if(GENDER==0) {
  double CL = CL_F;
} else {
  CL = CL_M;
}

// Alternatively, you can use this syntax:
// double CL = GENDER==0 ? CL_F : CL_M;

F_CENT = F1; 
ALAG_GUT = ALAG1;

$ODE
dxdt_GUT = -KA * GUT;
dxdt_CENT = KA * GUT - (CL/V2)*CENT;

$CAPTURE @annotated
CL: clearance used
CL_F: clearance_female
CL_M:clearance_male
CP=CENT/V2: Central concentration
'
mod <- mcode("model_test", code)
#> Building model_test ...
#> done.

data <- expand.ev(amt = 480000, rate = c(0,480000/2), GENDER = c(0,1))

data <- mutate(
  data, 
  cmt = ifelse(rate > 0, 2, 1), 
  route = factor(rate > 0, labels = c("po", "iv")), 
  sex = factor(GENDER, labels = c("female", "male"))
)

data
#>   ID time    amt   rate cmt evid GENDER route    sex
#> 1  1    0 480000      0   1    1      0    po female
#> 2  2    0 480000 240000   2    1      0    iv female
#> 3  3    0 480000      0   1    1      1    po   male
#> 4  4    0 480000 240000   2    1      1    iv   male
  
out <- mrgsim(mod, data, delta = 0.1, recover = "route,sex")

plot(out,  CP ~ time |route*sex, scales = "same")

plot(out,  CL ~ time |route*sex, scales = "same")

^{Created on 2025-07-14 with reprex v2.1.1}

[JoanaXu]

Hi @kylebaron , Thank you so much for your quick reply, the code works well and it was exactly what I want! The working way here much like NONMEM used. Besides, what is the difference between $PARAM @input and $INPUT?

[kylebaron]

Hi @JoanaXu -

$PARAM @input and $INPUT are equivalent; they both let you define some parameters with the ability to check an input data set to make sure those parameters are present.

Glad the code is working well. Please let me know if I can help with anything else.

Kyle

[JoanaXu]

Thank you Kyle, that's very helpful!

[kylebaron]

You're welcome; best wishes for your simulation.

[JoanaXu]

Hi Kyle,
I found the code may be not useful for F1? Take the female + PO combination for example, I tried to use the pk2 example with parameter updated and dataset include the F1, the out are not same, here is my code, and the two output plots:

mod <- modlib("pk1")%>%
update( param = list(CL=2.88,V=4,KA=0.2))
data <- ev(amt = 480000*0.5) #F=0.5 for PO

mrgsim(mod, data, delta = 0.1)%>%
plot( CP ~ time )

[kylebaron]

Hi @JoanaXu -

Can you change F_CENT to F_GUT and try it again?

if(ROUTEN==0) {
F_CENT=1;
} else {
ALAG_GUT=ALAG1;
F_CENT=F1;}

Kyle

[JoanaXu]

Hi @kylebaron ,
Thanks, yes, the F_GUT works! Thank you so much~