Seeking Advice: Implausible Vmax and Km from Michaelis-Menten Model for Oral Drug in Phoenix NLME

General
1

Dear all,

I am currently developing a population pharmacokinetic (PopPK) model using Phoenix NLME, and I am relatively new to the platform. My goal is to describe the pharmacokinetics of an orally administered drug using a Michaelis-Menten elimination model.

The structural model is a one-compartment model with first-order absorption and non-linear elimination. However, I have encountered an issue where the estimated values of Vmax and Km are extremely large (in the range of millions), which do not appear to be physiologically plausible or consistent with the observed data.

Here is the current model code I am using:
test(){
deriv(Aa = - Ka * Aa)
deriv(A1 = Ka * Aa - Vmax * C / (Km + C))
dosepoint(Aa)
C = A1 / V
error(CEps = 0.1)
observe(CObs = C * (1 + CEps))

stparm(Ka = tvKa)
stparm(V = tvV * exp(nV))
stparm(Km = tvKm)
stparm(Vmax = tvVmax * exp(nVmax))

fixef(tvKa = c(, 4.49, ))
fixef(tvV = c(, 300, 1000))
fixef(tvKm = c(0.1, 6.91, ))
fixef(tvVmax = c(0.1, 254, ))

ranef(diag(nV, nVmax) = c(0.1, 0.1))

}

To address the issue, I have taken the following steps:

  • Verified that all units are consistent across the dataset and the model (e.g., dose in mg, concentration in ng/mL, volume in L).
  • Tried both FOCE ELS and Laplacian estimation methods.
  • Ran the model with and without interindividual variability (IIV) on each parameters
  • Adjusted initial estimates and boundaries for Km and Vmax to reflect plausible values.

Unfortunately, despite these efforts, the estimated values of Vmax and Km remain implausibly high, and model fit does not seem to improve meaningfully.

I wonder if anyone has encountered a similar issue when implementing Michaelis-Menten elimination in an oral PK model. I would greatly appreciate any suggestions, shared experiences, or guidance on how to improve model performance or parameter estimation in this context. As someone new to Phoenix NLME, any practical advice would be immensely helpful.

Thank you very much for your time and support.

Best regards,
Virunya Komenkul
Ph.D Student

2

Virunya, probably we need to see the data to be sure. that said I would probably start by standardising to common units e.g. ug/L and ug for dose perhaps.