RESUMO
OBJECTIVE: Duration of therapy has been suggested to represent a measure of effectiveness. Life table analyses of therapy with methotrexate (MTX) in rheumatoid arthritis (RA) have indicated a longer duration than with other drugs. However, individual patients continue taking MTX for different periods of time. We assessed the influence of patient variables at treatment onset upon subsequent duration of MTX therapy. METHODS: Patients with RA (n = 437) from 8 North American databank centers beginning MTX therapy after January 1, 1988, were followed prospectively. Age at onset of MTX treatment, sex, years of education, age at onset of disease, years with disease, number of comorbid conditions, number of disease modifying antirheumatic drugs (DMARD) and nonsteroidal antiinflammatory drugs (NSAID) taken just prior to MTX. disability level, pain, and global assessment prior to starting MTX were used in univariate Kaplan-Meier analyses to predict number of months taking MTX alone. An index that divided the patients into risk strata for predicting duration of therapy was constructed to be clinically useful. RESULTS: The median number of months continuing MTX without addition of other DMARD was 41 months and the median for the total course taking MTX was 52 months. The retention rate was lowest for patients with the most negative initial health state. High level of initial pain, long duration of disease, and not using a DMARD just prior to MTX were associated with low retention rate and can be used to predict expected durations of MTX treatment ranging from 17 to 52 months. For practical guidance in clinical decisions an index was computed based on the predictor variables: level of initial pain, duration of disease, and number of DMARD; this index identifies subgroups with very different durations taking MTX alone. Disease duration at baseline was strongly related to time taking MTX alone and could therefore also be used as a simplified rule in clinical work. CONCLUSION: Expected duration of MTX treatment is influenced by clinical variables, and these may suggest those patients likely to have more or less satisfactory experiences with MTX. The time taking drug alone (therapeutic segment) may be a more logical and sensitive indicator of effectiveness than the total course on the medication.