Date: Thu, 30 Sep 1999 10:18:38 -0400
From: Sriram Krishnaswami <>
Subject: Forecasting Individual PK and PD

Dear NMusers

I would like to ask a general question about population PK/PD and individual forecasting. I would greatly appreciate your comments and/or references .

Thanks in advance.

Is it possible to accurately (reasonably) forecast an individual's PK and PD?

I have this question because a lot of drug effects are quantified as %suppression or %induction from a certain baseline e.g circadian rhythm

In other words, using prior population PK information and a couple of plasma samples from a particular patient, we are able to predict his/her pharmacokinetic profile with good accuracy.

Assuming that we also have prior population PD information (both baseline and drug induced), I wonder if it is possible to forecast an individual's drug effect without ever knowing how his baseline PD profile looked like (i.e in the absence of drug) simply because, at the time when the patient comes to

the clinic, he/she is already on drug.

So even if we obtain PD information from him / her, we would still be limited in making predictions only based on prior population values without using his/ her observed values because we have no way to find out where his/her baseline (in the absence of drug) would have been.

For instances where drug effect is quantified as % AUC difference between baseline and the suppressed/induced profile (e.g cumulative endogenous cortisol suppression after administration of exogenous corticosteroids) we might be way off in our predictions.


Sriram Krishnaswami
Graduate student
Dept of Pharmaceutics
University of Florida