Meta- and mega-analysis of randomised controlled trials indicate reduction in tender joint counts and decreased use of non-steroidal anti-inflammatory drugs with fish-oil supplementation in long-standing rheumatoid arthritis (RA). Since non-steroidal anti-inflammatory drugs confer cardiovascular risk and there is increased cardiovascular mortality in RA, an additional benefit of fish oil in RA may be reduced cardiovascular risk via direct mechanisms and decreased non-steroidal anti-inflammatory drug use. Potential mechanisms for anti-inflammatory effects of fish oil include inhibition of inflammatory mediators (eicosanoids and cytokines), and provision of substrates for synthesis of lipid suppressors of inflammation (resolvins). Future studies need progress in clinical trial design and need to shift from long-standing disease to examination of recent-onset RA. We are addressing these issues in a current randomised controlled trial of fish oil in recent-onset RA, where the aim is to intervene before joint damage has occurred. Unlike previous studies, the trial occurs on a background of drug regimens determined by an algorithm that is responsive to disease activity and drug intolerance. This allows drug use to be an outcome measure whereas in previous trial designs, clinical need to alter drug use was a 'problem'. Despite evidence for efficacy and plausible biological mechanisms, the limited clinical use of fish oil indicates there are barriers to its use. These probably include the pharmaceutical dominance of RA therapies and the perception that fish oil has relatively modest effects. However, when collateral benefits of fish oil are included within efficacy, the argument for its adjunctive use in RA is strong.