There is no ideal slow-acting antirheumatic drug. Therapy of rheumatoid arthritis (RA) is currently being modified, with strong recommendations to abandon the traditional pyramidal approach. The call is for a more aggressive, earlier approach to suppress inflammation. Combination therapy rather than the use of a single agent is advocated by some. Improved methods for assessing disease activity as well as measurement of outcome have been developed. Markers of poor prognosis have helped to define patients for earlier treatment. Comparison of toxicity among such a diverse group of drugs is probably best achieved with a toxicity index measuring the number of episodes expressed in terms of patient-years of exposure. Toxicity remains the commonest reason for discontinuing an agent, while remission beyond 36 months on therapy is uncommon, except with methotrexate. The profile of toxicity is clearly defined for individual agents, but combination therapy may reveal an entirely different set of toxic manifestations. There is an urgent need to develop a set of risk factors to predict toxicity in an individual patient. Juvenile chronic arthritis behaves differently from adult RA. Drug toxicity profiles are similar, but less common. Outcome is more difficult to measure, with the major impact of disease and therapy being on growth retardation.