Epidemiological studies published since 1996 have established that hypertriglyceridemia can predict risk of cardiovascular disease in a manner statistically independent of HDL cholesterol. Nevertheless, the relationship of concentrations of plasma triglycerides to risk of cardiovascular disease remains less than straightforward, partly because triglycerides are carried in lipoproteins of different atherogenicity, partly because hypertriglyceridemia is associated with non-lipid atherogenic and thrombogenic processes. For example, the association of highest risk of cardiovascular disease to moderate rather than to severe hypertriglyceridemia can be understood in terms of the distribution of triglycerides between different classes of plasma lipoproteins. It is counter-intuitive to most clinicians, however, and hence it can result in the misdirection of clinical efforts including drug therapy. Fibrates lower plasma triglycerides, and raise HDL, efficiently and with few immediate side-effects. Central to their mode of action is activation of certain nuclear receptors in cells. There is no necessary connection, however, between that fascinating biochemistry and clinical benefit as defined by reductions in rates of death by coronary artery disease. A review of trials of cholesterol-lowering by diet and drugs, published between 1966 and 1996, included 12 trials of therapy with fibrates or placebo in more than 21000 patients. Overall, these trials indicated no benefit in terms of reduction in risk of coronary deaths. The period since 1996 has seen the publication of four additional trials of treatment of 6144 patients with fibrates or placebo. Two of them were major trials. The VA-HIT was very encouraging, because treatment with gemfibrozil produced a signficant reduction in the combined incidence of fatal and non-fatal coronary events. There was no significant reduction in coronary deaths, however. The results of BIP were frankly disappointing, because they demonstrated no significant effect of treatment with bezafibrate on either the primary end-point of the trial or on rates of coronary death. Clinical indications for the use of fibrates can obviously not be based on biochemical insights, however intriguing in their own right, but they have also not been satisfactorily defined by the randomized clinical trials published to date. Hope remains, however, that some clarification will result from ongoing trials of fibrate treatment of patients with type II diabetes.