Abstract
OBJECTIVE: To determine whether or not N of 1 trials with diclofenac/misoprostol (Arthrotec) are superior and cost-effective compared with standard treatment in osteoarthritis (OA). METHODS: We randomized subjects with OA who were "uncertain that nonsteroidal antiinflammatory drugs (NSAID) were helpful" to 2 different groups. One group received conventional treatment whereby they were told to stop their NSAID and to wait and see what happened. If necessary, treatment with other NSAID and all other usual OA treatment strategies were used. The other group received a series of crossover trials with diclofenac 50 mg and misoprostol 200 micro g twice a day or an identical placebo for 2 weeks each in a random double-blinded manner. Every 4 weeks they chose which treatment they preferred. By 3 months, if there was no clear preference, the N of 1 trials were discontinued. All trial participants were seen monthly for 3 months and at 6 months. All costs (direct and indirect) were collected for both groups. Costs of research-generated visits were not counted in the "conventional treatment" group. RESULTS: Fifty-one subjects were randomized (stratified by most symptomatic OA area): 25 with knee, 7 with hip, and 19 with hand OA. Twenty-four were randomized into the N of 1 group. There were no differences in the baseline and followup variables including age, income, education, past and current NSAID use, global assessment, Health Assessment Questionnaire (HAQ), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), and Medical Outcome Study Short Form-36 (SF-36) scores. In the N of 1 group, 11 patients preferred diclofenac in total (of whom 7 did only one round, 2 did 2 rounds, and 2 did 3 rounds). None of the N of 1 patients preferred placebo and 11 had no preference (2 dropped out at baseline). At 6 months, 15 of 19 in "conventional therapy" and 17 of 21 in N of 1 were taking NSAID. That is, NSAID appeared to be effective in 81% of N of 1 subjects and 79% of conventionally treated patients, even though subjects were initially uncertain that their NSAID were helpful. The total OA-related costs in Canadian dollars per patient (in 1996) for N of 1 treated patients at 6 months were: 551.66 dollars +/- 154.02 dollars (SD) versus 395.62 dollars +/- 226.87 dollars for controls, excluding 2 research visits for controls (p < 0.009). The HAQ pain and disability, WOMAC scales, and physician global assessments improved more in the N of 1 group (at greater cost), but no between-group differences in efficacy were seen, possibly due to small numbers. CONCLUSION: N of 1 trials were time-consuming in these patients and are more expensive, but with slightly better outcomes. In addition, NSAID seem to be effective in a majority of subjects with OA who have been uncertain of their benefit.
In this issue
