Objective: To determine prognostic factors in rheumatoid arthritis (RA).
Methods: One hundred thirty-two women with definite RA were followed yearly from an early phase of the disease (symptoms < 5 years) for a mean duration of 6 years. The prognostic value of the first available clinical and laboratory variables and assessments of functional ability was related to several outcome measures (physician's opinion of disease severity, disease activity, radiological abnormalities, functional ability and number of prescribed 2nd-line drugs) by single predictor analysis and by logistic regression.
Results: The variables most predictive for one or more of the outcome measures were number of swollen joints, Ritchie score, health assessment questionnaire score, radiographical abnormalities, positive IgM rheumatoid factor (RF), positive IgG-RF, HLA-DR4, and an elevated percentage serum agalactosyl IgG. The accuracy of predicting outcome was calculated from several combinations of these variables, and varied between 70 and 80%. The accuracy based on a combination of the commonly available variables (number of swollen joints, IgM-RF and the erosion score), closely approximated the maximal accuracy that could be achieved.
Conclusion: The outcome of RA can be predicted by a combination of variables that are commonly available in the clinical setting.