OUTCOME IN PATIENTS WITH RHEUMATOID-ARTHRITIS RECEIVING PREDNISONE COMPARED TO MATCHED CONTROLS

Objective. To determine the longterm outcome including disease activity, mortality, and adverse events in patients with rheumatoid arthritis (RA) treated with prednisone. Methods. A case-control study was performed, based on our cohort of 893 mostly Caucasian patients with adult onset RA, followed since 1966. Data collection was based on protocols and included single physician global assessment. Prednisone was started in 122 patients (85 women, 37 men) after 1966. All were matched for age, sex, disease duration, and global assessment to 122 controls from the same cohort who have never received prednisone. Results. Mean disease duration before prednisone was 14.1 years. Mean duration of use was 6.9 years with a mean dose of 8.0 mg/day. Prednisone was eventually stopped in 34% of patients. Life expectancy and causes of death were similar in both groups. No differences in hemoglobin, erythrocyte sedimentation rate, global assessment, Lansbury index, functional class or Health Assessment Questionnaire (HAQ) disability index were seen between the 2 groups before or 5 years after starting prednisone. Ten years after starting prednisone, HAQ scores were similar but Lansbury and global assessment were worse in the prednisone treated group. As expected, adverse events, notably cataracts and fractures, were observed more often in the prednisone group. Conclusion. Case-control matching can only reduce, not eliminate, potential selection bias. Nonetheless, the lack of demonstrable longterm benefit with prednisone use in this and other studies is disconcerting. Caution and further studies are required before the more aggressive use of longterm prednisone therapy in RA is embraced.