PREDICTION OF BLOOD LEVELS AFTER MULTIPLE DOSES FROM SINGLE-DOSE BLOOD LEVEL DATA - DATA GENERATED WITH 2-COMPARTMENT OPEN MODEL ANALYZED ACCORDING TO ONE-COMPARTMENT OPEN MODEL

Forty sets of single‐dose blood levels were simulated by varying the parameters in the equation appropriate to the two‐compartment open model with first‐order absorption (Model II). Each set was fit by the method of least squares with an iterative nonlinear program and an IBM 360/30 digital computer to the equation appropriate to the one‐compartment open model with first‐order absorption (Model I). The estimated parameters of Model I were then used to make predictions of maximum, average, and minimum blood levels to be expected after multiple doses of the drug given at uniform intervals of 6 hr. The predicted values were then compared with the actual values derived for Model II. In general, the equation appropriate to Model I fitted the data generated by means of Model II quite well. When V2 of Model II was eight times V1, the fitting of the data generated by Model II with Model I was poor, and the prediction of multiple‐dose blood levels was poor. When 8 ⩽ V1/V2 ⩽1 for Model II, the predictions 01 multiple‐dose blood levels made with the Model I analysis were quite accurate. Literature data suggest that the volume ratio has been in the latter range when the two‐compartment open model has been elaborated from actual blood level data collected after intravenous administration. Hence, in the practical situation, one may expect the mathematical error introduced by use of Model I in making predictions of multiple‐dose blood levels to be relatively small compared with other possible sources of error in such predictions. Copyright © 1969 Wiley‐Liss, Inc., A Wiley Company