Background

• Structural model - Two compartment model for parent and one compartment Model for metabolite with reversible metabolism

• Route of administration - Administration of parent drug and metabolite on two different occasions

• Dosage Regimen - 4.3 Micromol/Kg of parent & 5 Micromol/Kg of metabolite

• Number of Subjects - 1

Learning Outcome

In this model, you will learn how to build a two compartment parent and one compartment metabolite model with reversible metabolism, while parent and metabolite is admininstered on two different occasions

Libraries

call the "necessary" libraries to get start.

using Pumas
using Plots
using CSV
using StatsPlots
using Random

Model

Two compartment model for parent and one compartment model metabolite

pk_45            = @model begin
  @param begin
    tvvcp        ∈ RealDomain(lower=0)
    tvvpp        ∈ RealDomain(lower=0)
    tvqp         ∈ RealDomain(lower=0)
    tvclp        ∈ RealDomain(lower=0)
    tvvcm        ∈ RealDomain(lower=0)
    tvclm        ∈ RealDomain(lower=0)
    tvclpm       ∈ RealDomain(lower=0)
    tvclmp       ∈ RealDomain(lower=0)
    Ω            ∈ PDiagDomain(8)
    σ²_prop      ∈ RealDomain(lower=0)
  end

  @random begin
    η            ~ MvNormal(Ω)
  end

  @pre begin
    Vcp          = tvvcp * exp(η[1])
    Vpp          = tvvpp * exp(η[2])
    Qp           = tvqp * exp(η[3])
    Clp          = tvclp * exp(η[4])
    Vcm          = tvvcm * exp(η[5])
    Clm          = tvclm* exp(η[6])
    Clpm         = tvclpm * exp(η[7])
    Clmp         = tvclmp * exp(η[8])
  end

  @dynamics begin
    Centralp'    = (Qp/Vpp)*Peripheralp - (Qp/Vcp)*Centralp - (Clp/Vcp)*Centralp - (Clpm/Vcp)*Centralp + (Clmp/Vcm)*Centralm
    Peripheralp' = (Qp/Vcp)*Centralp -  (Qp/Vpp)*Peripheralp
    Centralm'    = -(Clm/Vcm)*Centralm - (Clmp/Vcm)*Centralm + (Clpm/Vcp)*Centralp
  end

  @derived begin
    cp           = @. Centralp/Vcp
    met          = @. Centralm/Vcm
    dv_cp        ~ @. Normal(cp, sqrt(cp^2*σ²_prop))
    dv_met       ~ @. Normal(met, sqrt(cp^2*σ²_prop))
  end
end

PumasModel
  Parameters: tvvcp, tvvpp, tvqp, tvclp, tvvcm, tvclm, tvclpm, tvclmp, Ω, σ
²_prop
  Random effects: η
  Covariates: 
  Dynamical variables: Centralp, Peripheralp, Centralm
  Derived: cp, met, dv_cp, dv_met
  Observed: cp, met, dv_cp, dv_met

Parameters

Parameters provided for simulation. tv represents the typical value for parameters.

• tvvcp - Volume of distribution of central compartment of Parent (L/kg)

• tvvpp - Volume of distribution of peripheral compartment of Parent (L/kg)

• tvqp - Intercompartmental clearance of Parent (L/hr/kg)

• tvclp - Clearance of Parent (L/hr/kg)

• tvvcm - Volume of distribution of central compartment of Metabolite (L/kg)

• tvclm - Clearance of Metabolite (L/hr/kg)

• tvclpm - Conversion of Parent to Metabolite (L/hr/kg)

• tvclmp - Conversion of Metabolite to Parent (L/hr/kg)

• Ω - Between Subject Variability,

• σ - Residual error

param = (tvvcp   = 0.563,
         tvvpp   = 0.424,
         tvqp    = 0.115,
         tvclp   = 0.343,
         tvvcm   = 0.932,
         tvclm   = 0.068,
         tvclpm  = 0.015,
         tvclmp  = 0.046,
         Ω       = Diagonal([0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0]),
         σ²_prop = 0.001)

(tvvcp = 0.563, tvvpp = 0.424, tvqp = 0.115, tvclp = 0.343, tvvcm = 0.932, 
tvclm = 0.068, tvclpm = 0.015, tvclmp = 0.046, Ω = [0.0 0.0 … 0.0 0.0; 0.0 
0.0 … 0.0 0.0; … ; 0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0], σ²_prop = 0.001)

Dosage Regimen

• A dose of 4.3 μmol/kg of the Parent is administered as a rapid IV Injection over 15 seconds

• A dose of 5 μmol/kg of the metabolite is administered as a rapid IV Injection over 15 seconds

dr_p  = DosageRegimen(4.3,cmt=1,time=0, duration=0.0041)
sub_p = Subject(id=1,events=dr_p)
dr_m  = DosageRegimen(5,cmt=3,time=0,duration=0.0041)
sub_m = Subject(id=2,events=dr_m)
sub   = [sub_p, sub_m]

Population
  Subjects: 2
  Covariates:

Simulation

We are going to simulate parent and metabolite concentration profile

Random.seed!(123)
sim_sub = simobs(pk_45,sub,param,obstimes=0:0.001:31)
df1     = DataFrame(sim_sub)

Dataframe and Plotting

Convert the simulation to a dataframe and use the necessary for plotting

df1_par = filter(x -> x.id == "1", df1)
filter!(x -> x.time in [0.0042,0.0333,0.1667,0.5,1,2,2.75,5,7,24,31], df1_par)
df1_met = filter(x -> x.id == "2", df1)
filter!(x -> x.time in [0.0042,0.0333,0.1333,0.25,0.75,2,4,7,12,23], df1_met)

@df df1_par plot(:time, :cp,  yaxis=:log, linewidth=2, legend = :topright,
                label="Parent [IV.Par]", title="Concentration vs Time",
                xlabel="Time (hrs)", ylabel="Concentration (uM)",
                xlims=(0,35), xticks=[0,5,10,15,20,25,30,35], yticks=[0.001,0.01,0.1,1,10], ylims=(0.001,10))
@df df1_par plot!(:time, :met, label="Metabolite [IV.Par]",linewidth=2)
@df df1_met plot!(:time, :cp, label="Parent [IV.Met]",linewidth=2)
@df df1_met plot!(:time, :met, label="Metabolite [IV.Met]]",linewidth=2)
@df df1_par scatter!(:time, :dv_cp, marker=[:square], label=false, color=:green)
@df df1_par scatter!(:time, :dv_met, marker=[:square], label=false, color=:green)
@df df1_met scatter!(:time, :dv_cp, marker=[:circle], label=false, color=:red)
@df df1_met scatter!(:time, :dv_met, marker=[:circle], label=false, color=:red)

---