Sirs

Fidock et al.1 are to be congratulated on their comprehensive and informative review of antimalarial drugs, drug targets, and models for drug evaluation. They rightly argue that drug combinations are generally preferable to monotherapy and that for maximal efficacy there should be a positive interaction between the two (or more) elements of the combination — that is, synergism. However, methods are currently available for the interpretation of laboratory drug-interaction data that are more sophisticated than that presented by Fidock et al. in Box 2 of their article, and which make statistically sound evaluations of the degree of synergism or antagonism.

Fractional inhibitory concentrations (FIC) do indeed express in a quantitative way the degree of interaction between two drugs. However, when they are presented with no estimates of precision (for example, confidence intervals), dependable conclusions about the degree of synergism or antagonism cannot be drawn. In situations of considerable uncertainty, in which there are few repetitions of the experiment or the degree of variability is high, a single number such as FIC can be potentially misleading. Moreover, the often-used 'cut-offs' of ΣFIC >2 for antagonism and ΣFIC <0.5 for synergism are completely arbitrary. For example, two drugs demonstrating a consistent ΣFIC of 0.6 might be exhibiting synergism that is of a small degree, but both mechanistically and clinically significant. On the other hand, a ΣFIC of 0.4 obtained without sufficient reproducibility might not represent genuine synergism at all. The answer to this problem is rigorous statistical analysis of drug-interaction data, as done by Gavigan et al.2 and others3,4 (and see references therein). In this way, the degree of synergism or antagonism can be estimated and significance tests can be made (at usual levels such as p = 0.05) of the null hypothesis of additivity, thereby permitting firm conclusions regarding synergy or antagonism.

Finally, many authors, unfortunately, claim findings of synergism and antagonism between antimalarial (and other antimicrobial) agents but do not provide statistical justification for their conclusions. It is to be hoped that this will change.