Much has been written about the potential of pharmacogenetic testing to inform therapy based on an individual's genetic makeup, and to decide the most effective choice of available drugs, or to avoid dangerous side effects. Currently, there is little hard data for either in the field of cardiovascular disease. The usual approach has been opportunistic use of drug trials in unrelated patients, and to look for differences in response or outcome by "candidate gene" genotype, for example genes coding for drug metabolising enzymes (activators and metabolisers), and enzymes and receptors involved in lipid metabolism, adrenergic response, etc. As with all association studies, initially promising results have often failed the test of replication in larger studies, and the relationship between the CETP Taq-I variant and response to statins has now been disproved. The strongest data to date is the report [Chasman, D.I., Posada, D., Subrahmanyan, L., Cook, N.R., Stanton Jr., V.P., Ridker, P.M., 2004. Pharmacogenetic study of statin therapy and cholesterol reduction. J. Am. Med. Assoc. 291, 2821-2827] of a poorer cholesterol-lowering response to Pravastatin in the 7% of patients carrying a certain haplotype of the HMG CoA reductase gene (14% fall versus 19%), but if this is overcome simply by a higher dose, it is of little clinical relevance. Currently, the best example of avoiding side effects is determining genotype at the CYP2C9 locus with respect of warfarin treatment, since carriers for functional variants (>20% of the population) require lower doses for optimal anticoagulation, and homozygotes, although rare, may well experience serious bleeding if given a usual dose. The full potential of this field will only be realised with much further work.