A mathematical model, involving three consecutive first order reactions, has been applied to already existing experimental material in order to explain the hourly variation in serum levels obtained after administration of varying repeated doses of different preparations of p-aminosalicylic acid to human volunteers. The model closely agrees with the experimental data for all the orally administered preparations concerned, and the differences in blood levels for different preparations can be explained in terms of the values attributed to the parameters involved in the model.The same model adequately describes the elimination of aminosalicylic acid from the blood after a single intravenous infusion provided that the initial blood level attained does not exceed about 11 mg/100 ml., but fails at higher values. In the latter instance a second order model gives a good fit to the experimental curves. The increase in serum levels found experimentally for the same subjects during the first weeks of prolonged treatment with phenyl-p-aminosalicylate corresponds to a change in the values of the parameters for the elimination from the blood only.First order reactions have previously been proposed to explain the elimination of various drugs from the blood stream, and a combination of two consecutive first order processes has been used to account for the absorption into and subsequent elimination from the blood of orally administered preparations. Wiegand & Taylor (1960) have developed formulae for the combination of three consecutive first order reactions in order to explain the effects of sustained-release tablets.Various methods for estimating the rate constants in the first order reactions involved have been described in the literature. Dominguez & Pomerene (1945), using a model not specifying the absorption process, determined the rate constant for elimination from that part of the experimental curve where absorption has become negligible, and calculated the absorption rate from the fitted curves for blood level and excretion rate as found by urine determinations. Nelson (1960), using a similar model calculated the absorption rate from urine values only, while Wiegand & Taylor (1960) determined the rate constant for the sustained-release process from in vitro experiments.
METHODSIn the present study none of these methods were applicable because the already existing experimental material did not include urinary determinations or in vitro experiments for the sustained-release process; moreover the scheme involving three doses per day prevented the elimination process from becoming dominant for a sufficiently long time to obtain a reasonably good estimate of the elimination rate constant. Therefore, it was decided to obtain direct estimates of the three rate constants and the apparent distribution volume by non-