Minocycline in Early Rheumatoid Arthritis


Synopsis: Minocycline produced more improvement than did placebo when used in patients with rheumatoid arthritis beginning less than a year after disease onset. Sixty-five percent of those receiving minocycline improved by 50% or more, compared to only 13% of those receiving placebo.

Source: O’Dell JR, et al. Arthritis Rheum 1997;40:842-848.

Two previously published randomized placebo controlled studies of the efficacy and safety of minocycline in rheumatoid arthritis (RA) involved patients with established disease who had not responded to treatment with other agents.1,2 In those studies, minocycline’s efficacy was statistically significantly better than placebo, though the magnitude of the difference in efficacy was clinically modest.

O’Dell and colleagues report their experience with minocycline vs. placebo in patients with RA, all rheumatoid factor positive, none with serologic evidence of recent factor positive, none with serologic evidence of recent parvovirus infection or Lyme disease, and all with disease of less than one year duration. None of the 23 subjects in each group had received corticosteroid or disease-modifying antirheumatic drug (DMARD) therapy prior to enrolling in the study. The primary outcome assessed was achieving and maintaining a 50% or better composite improvement from baseline. The study population was predominantly female, with mean disease duration of morning stiffness of a little more than two hours. After randomization, subjects received either minocycline 100 mg bid or placebo, in addition to a steady dose of NSAID. Blinded evaluation after three months determined whether subjects met the 50% improvement criterion. If so, they continued the minocycline or placebo. If not, they were withdrawn from the blinded portion of the study. After three months, 65% of the minocycline and 17% of the placebo group achieved the target 50% improvement. At six months, 64% of the minocycline group and only 13% of the placebo group continued to have 50% or better improvement. Of the seven variables followed, the patient global assessment (change of 2.0 vs 0.5 on 0-10 scale; P = 0.006), physician global assessment (change of 1.6 vs 0.3 on 0-10 scale; P = 0.01), and the duration of morning stiffness (change of 70 vs 16 minutes; P = 0.03) all demonstrated greater improvement in the minocycline group than the placebo group. There were trends that favored minocycline in decrease in ESR, decrease in total joint score, decrease in tender joint score and in swollen joint score, but none of the differences in these measures between the two groups were statistically significant.

There were no withdrawals due to adverse effects in the minocycline group, though O’Dell et al do not specifically state how many subjects in each group had dizziness. The only serious adverse event noted was GI bleeding in a patient receiving placebo.

Following the six-month blinded phase, most patients continued to be followed. Results at one year of follow-up were reported. Remission had been achieved in five of those subjects originally in the minocycline group and only in one of the placebo group, a result that was not statistically significant.

Eighty-seven percent of those in the minocycline group met the 50% improvement criterion vs. only 18% of those who started in the placebo group. DMARDs, other than minocycline, were being used in 30% of the minocycline group and in 85% of the 20 patients left of the original placebo group.


The efficacy of minocycline in this study was impressive, and frankly, a surprise to me. The number of patients in each group is relatively small, and it is disappointing that the intergroup differences in the change in the joint scores (tender, swollen, and total) were not significantly different.

The efficacy of minocycline after six months of treatment in this study is less than that reported previously by the same group of investigators for a three-drug regimen (methotrexate, sulfasalazine, and hydroxychloroquine).3 Although the three drug regimen required more monitoring and carried a risk of more serious toxicity than minocycline, it produced 50% or better improvement in about 90% of patients at six months.

In addition, the subjects in the three-drug combination group had much longer duration of disease and more active synovitis. Total joint scores (tender and swollen) at entry of subjects in the authors’ minocycline trial were about two-thirds of the scores of patients entering their previous combination therapy trial. By excluding patients with prior DMARD or corticosteroid therapy from their minocycline trial, the investigators appear to me to have selected patients with milder, as well as with early, RA.

Although RA is not an FDA-approved indication for minocycline, based upon the results of the prior controlled trials1,2 and this study, "off label" use may be an option for patients with RA, especially for those with early, milder disease and those who prefer not to risk the adverse effects of other DMARDs. It is likely to be a popular choice for the primary care and family physicians, who are likely to encounter patients during the first year of disease, because minocycline does not require cumbersome or costly monitoring and rarely causes serious adverse effects. However, the goal of treatment of patients with RA is to achieve remission, or, failing that, a minimum of active synovitis. Patients with more severe disease at onset, or who do not respond to treatment as expected, may require the use of more aggressive drug regimens.


1. Kloppenburg M, et al. Arthritis and Rheum 1994; 37:629-636.

2. Tilley BC, et al. Annals Intern Med 1995;122:81-89.

3. O’Dell JR, et al. New Engl J Med 1996;334:1287-1291.