The aim of this study was to study the short-term functional and anatomical prognosis of rheumatoid arthritis (RA) in a series of Spanish patients and to identify different subsets of patients as well as possible baseline factors associated with specific outcomes. All patients seen in our division who met the ACR criteria for RA and with disease duration between 2 and 7 years were eligible for the study. Available patients were further evaluated at the clinic for disease activity using biological tests and joint indices as joint counts and Thompson's index, functional capacity using the ACR functional classification (ACR-FC) and the modified Health Assessment Questionnaire (M-HAQ) and radiologic damage by the Sharp's radiologic scoring method. Cluster analysis was used to identify different clinical subsets of patients. One hundred and sixty-three patients were eligible for the study, 13 could not be located or refused to participate and 12 had died. Mean (+/-SD) age at disease onset and mean disease duration were, respectively, 56(+/-14) years and (55+/-20) months. Median (interquartile range) of M-HAQ was 0.4 (0.1-1.1) and 41% of patients were in III or IV ACR-FC. The majority of patients (93%) showed radiologic lesions and 65% had erosions. Cluster analysis identified three subsets: cluster I (70% of patients) was characterised by a good prognosis, cluster II (13%) by a high level of disease activity, and cluster III (17%) by a greater anatomic damage and longer disease duration. No baseline predictive markers were found for these different outcomes. We concluded that RA portends an overall poor short-term prognosis in a relative large percentage of our patients with significant anatomic and functional sequelae. Aggressive management is specially indicated in this subgroup of patients, although definitive prognostic markers for its early identification are still lacking.