Limitations of clinical trials in chronic diseases: is the efficacy of methotrexate (MTX) underestimated in polyarticular psoriatic arthritis on the basis of limitations of clinical trials more than on limitations of MTX, as was seen in rheumatoid arthritis?

Clinical trials are the optimal method to establish efficacy of a drug versus placebo or another drug. Nonetheless, important limitations are seen, particularly in chronic diseases over long periods, although most are ignored. Pragmatic limitations of clinical trials include a relatively short observation period, suboptimal dosage schedules, suboptimal surrogate markers for long-term outcomes, statistically significant results which may not be clinically unimportant and vice versa. Even ideal clinical trials have intrinsic limitations, including the influence of design on results, data reported in groups which ignore individual variation, non-standard observer-dependent interpretation of a balance of efficacy and toxicity, and distortion of a "placebo effect." Limitations are seen in many clinical trials of methotrexate (MTX) in rheumatoid arthritis (RA) and psoriatic arthritis (PsA). The first MTX clinical trial in rheumatology documented excellent efficacy in PsA, but frequent adverse events in 1964, explained by intravenous doses up to 150 kg. MTX was abandoned until the 1980s for RA, while gold salts and penicillamine were termed "remission-inducing," on the basis limitations of clinical trials. In the most recent MTX in PsA (MIPA) trial, all outcomes favoured MTX, but only patient and physician global estimates met the p<0.05 criterion. A conclusion of "no evidence for MTX improving synovitis" appears explained by insufficient statistical power, wide individual variation, no subsets, low doses, and other limitations. MTX appears less efficacious in PsA than RA, but may be underestimated in PsA, similar to historical problems in RA, resulting more from limitations of clinical trials than from limitations of MTX.