This review contrasts the difficulty of differentiating early rheumatoid arthritis (RA) from the many other causes of inflammatory polyarthritis and emphasizes the need for early treatment of RA with drugs that are present considered inappropriate for a mild, self-limiting disease. Pointers to help establish a definitive diagnosis of RA include the use of the American College of Rheumatology (ACR; formerly the American Rheumatism Association) diagnostic criteria, the presence of rheumatoid factor (RF), a raised erythrocyte sedimentation rate (ESR), raised serum levels of C-reactive protein (CRP) and low concentrations of serum sulphydryl (SH) groups. More sophisticated tests that are helpful and indicative of erosive-type disease are genetic markers such as the human leukocyte antigen HLA DR4 (especially the third allelic variable of DR beta 1) as well as an impaired sulphoxidation status. Similarly, the presence of the shared epitope alleles 0401 and 0404 quantitatively increases disease susceptibility and severity as their penetration increases. Experience with treatment indicates that control of the inflammatory process of RA reduces the progression of radiological damage. It is concluded that antimalarials are effective in mild disease, as is auranofin, whilst sulphasalazine is more effective than hydroxychloroquine in patients with disease diagnosed as definitive or classical RA. Intramuscular gold also lessens the development of erosions. The role of corticosteroids has not been defined. A number of combinations of drugs have been used, although they are probably not indicated as the first choice treatment for very early disease. Attitudes towards early treatment are changing; these are briefly reviewed and the authors' opinions on the management of early disease are outlined. Finally some new ideas, both theoretical and practical, on future development are summarized.