Skip navigation

Treatment termination rates in rheumatoid arthritis


Clinical bottom line

Patients with rheumatoid arthritis stay significantly longer on methotrexate other disease modifying drugs (DMARDs). Parenteral gold withdrawals are mainly due to toxicity, while withdrawals with sulphasalazine and hydroxychloroquinone are due to lack of efficacy.


A Maetzel et al. meta-analysis of treatment termination rates among rheumatoid arthritis patients receiving disease-modifying antirheumatic drugs. Rheumatology 2000 39: 975-981.


The review used extensive searching of MEDLINE and a Cochrane database for studies (RCTs or observational studies) that looked at termination rates for four common treatments for rheumatoid arthritis - methotrexate, parenteral gold, sulphasalazine and hydroxychloroquinone. Studies had to evaluate patients from treatment inception, and patients had to have documented diagnosis of rheumatoid arthritis. The outcomes were withdrawal for any reason, plus lack of efficacy or toxicity. The analysis was a survival analysis, necessarily done in a rather complex way because of the way in which information is reported (by particular times, like at six months, rather than individual dropout points for individual patients).


There were 110 studies included, with information up to 72 months on three of the treatments. Hydroxychloroquinone data were only up to 24 months.

The median survival time for methotrexate was longer than that for parenteral gold or sulphasalazine (Table 1). Sulphasalazine withdrawal for lack of efficacy was higher than that for parenteral gold or methotrexate. Sulphasalazine and parenteral gold withdrawals because of toxicity were higher than that for methotrexate.

Table 1: Continuing treatment on DMARDs


Still on treatment at 60 months (%)

Median survival time (months)




Parenteral gold







This is a really interesting, if complicated paper. It is complicated because it seeks to provide an answer to a question infrequently addressed in clinical trials, that is how long patients remain on treatment. It is interesting because treatment withdrawal, in rheumatoid arthritis and other medical treatment, is a major determinant of effectiveness in clinical practice over efficacy in clinical trials.

Adverse event analysis is not a subject that is well worked out methodologically, and this paper gives some real clue as to how it may be addressed in systematic reviews. It is a must-read for those wanting to follow this excellent North American example.