Evaluation of combined effects in dose-response studies by statistical comparison with additive and independent interactions.

An improved method for the evaluation of combined drug effects by means of dose-response curves (DRCs) is described. A drug, A, was tested in the absence and presence of a fixed concentration of another drug, B, mainly in organ-bath experiments with smooth muscle strips from bovine coronary arteries and tracheal muscle. The results of such experiments are expressed in terms of percent of maximum response. Median values, rather than mean values, for each concentration of drug A were determined in order to allow a comparison of observed with expected frequencies above or below median DRCs of additive and independent interactions. This comparison was done with the chi-square goodness-of-fit statistic. Advantage was taken of the curve-fitting program ALLFIT to construct DRCs. However, the method presented does not require a computer program. The results indicate that additive interactions point to actions of drugs at the same site inasmuch as they differ significantly from independent interactions. Overadditive interactions reflect differences between the sites of action. This may either be due to independent actions or to some kind of synergistic "cooperativity", for example, sequential interaction. The effects of the latter significantly exceed the effects expected for independently interacting compounds. This method appears applicable to compounds exerting all kinds of responses that can be described by DRCs.