The course of rheumatoid arthritis (RA) is highly variable, ranging from a mild self-limiting to a very aggressive form. To follow and predict the course of the disease in an individual patient, several recognized and proposed prognostic markers, including markers for disease activity, have been considered. However, no individual marker for disease activity has shown satisfactory specificity and sensitivity. Thus an index of disease activity combining several variables is needed. Response criteria based on the Disease Activity Score (DAS) were developed in an open study of 227 patients with RA of recent onset. Response was defined as a combination of a significant change from baseline and the level of disease activity attained. Good response was defined as a significant decrease in DAS (> 1.2) and a low level of disease activity ( < or = 2.4). Non-response was defined as a decrease < or = 0.6, or a decrease > 0.6 and < or = 1.2 with an attained DAS > 3.7. Any other scores were regarded as moderate responses. These response criteria were adopted as the EULAR response criteria and were validated, together with the WHO/ILAR and ACR response criteria, in a 48 week, double-blind trial comparing hydroxychloroquine and sulphasalazine in 60 patients. Response was evaluated against radiographic damage (construct validity) and functional disability (criterion validity); discriminating capacity was also assessed. EULAR response criteria showed significant association with X-ray progression and functional disability, and differentiated between sulphasalazine and hydroxychloroquine. ACR and WHO/ILAR response criteria performed less well, only showing good criterion validity. Several groups are working on the prognosis of early RA and have agreed to collaborate to test DAS and other prognostic markers to better recognize severe, progressive RA, before joint damage takes place.