Abstract
OBJECTIVE: The use of disease modifying antirheumatic drugs (DMARD) for rheumatoid arthritis (RA) is predicated on the expected value of the treatment course. Most clinical data are generalized from randomized controlled trials (RCT), which may result in estimates that are discordant with clinical experience and cannot address the effects of sequence of drugs. We computed estimates of relative DMARD effectiveness from a large observational database using area under the curve (AUC) data. METHODS: We examined data collected over a 20 year period on 1160 patients who were followed at the Wichita Arthritis Center. We utilized Health Assessment Questionnaire (HAQ) disability index data to quantify the effect of methotrexate (MTX), hydroxychloroquine (HCQ), and injectable gold (gold) on subsequent patient outcome. Using an AUC analysis, we compared length of treatment course, total disability averted, annual disability averted, and percentage of possible disability averted across drugs, and examined differences between first courses of therapy in DMARD naive patients and subsequent courses of the same and different DMARD in patients. RESULTS: Patients treated with MTX, HCQ, and gold improved at a rate of -0.33, -0.18 and -0.38 annualized HAQ area units, respectively. Since duration taking drug was greatest for MTX, then HCQ, then gold, the cumulative improvement was greatest with MTX (-1.07) versus gold (-0.74) versus HCQ (-0.47) in disability unit years. All 3 drugs were better cumulatively with earlier disease (MTX-1.74 for < 1 yr vs -0.95 for > 1 yr; HCQ -0.68 vs -0.43; gold -1.71 vs -0.49). A second trial of the same drug was far less effective than the first course. On a percentage of possible improvement basis, these drugs were nearly equal since HCQ is given to less severely affected patients. CONCLUSION: MTX is the most effective DMARD of these 3 because of the length the therapeutic segment. In terms of disability averted, none of the agents decrease disability by more than 25% of the theoretically possible improvement. We documented that effectiveness of RA treatment is a function of drug sequence, duration of disease, whether it is a first or second course, and severity of disease. None of these clinically relevant observations have emerged from clinical trials. These methodologic approaches provide important quantitative comparative data and will be useful in further assessment of the relative effectiveness of present and future DMARD.
