Background In the two-year Computer Assisted Management in Early Rheumatoid Arthritis-trial II (CAMERA-II) (1), patients with early rheumatoid arthritis were randomized to a methotrexate (MTX-) based tight control strategy with either an additional 10mg/d prednisone (MTX+pred) or placebo (MTX+plac) from start. Until the first sustained (≥3 months) period of remission, time (in months (SD)) was shorter in the MTX+pred group than in the MTX+plac group (6 (5) vs.11 (5) resp., p<0.001). In the MTX+pred group more patients achieved at least one period of sustained remission, although not significantly different (72% vs. 61%, p=0.09). There is a need for methods to additionally describe both the duration of remission as well as the interruptions in this desired state that is simple to use. The continuity rewarded (“ConRew”) score and patient vector graph were developed with that in mind.(2)
Objectives Posthoc analysis of the the CAMERA-II remission data in two alternative ways as compared to a simple sum score.
Methods All 236 patients of the CAMERA-II trial (MTX+pred n=117, MTX+plac n=119) were studied to calculate their individual scores. Remission was defined as a swollen joint count of 0 (range 0-38 joints), and at least 2 of the following factors: tender joint count ≤3 (range 0-38 joints), VAS score ≤20mm and ESR ≤20mm/h. Remission was evaluated in one-month time periods. The ConRew score was calculated as follows: for every patient it counts (as points) the number of remission periods of a predefined length, adding a point for every period that is directly followed by another period in remission or for being in remission at the end of the observation period. To study the additional effect of continuity rewarding, a sum score was calculated by adding the number of remission periods. The visual pattern of remission periods in both groups were compared through patient vector graphs.
Results The mean (SD) ConRew score differed statistically significantly (22.6 (15.7) vs.17.1 (13.9), p=0.004) between the two strategy arms in favour of the MTX+pred group. The mean (SD) plain sum score, in favour of the MTX+pred group, did not differ significantly (13.3 (8.1) vs. 11.3 (7.1), p=0.051). Effect size of the strategy arms measured through ConRew was larger than the effect size measured through the sum score (Cohen's d=0.374 vs. Cohen's d=0.255). The MTX+pred strategy led to a pattern of more and longer periods of remission in the patient vector graphs.
Conclusions There is a significant difference in sustained remission between the two treatment arms, favouring MTX+pred. Although both the effect size values are in the small to medium range, the fact that the ConRew effect size is almost 1.5 times larger shows a more sensitive way of addressing sustained remission. The ConRew score and the patient vector graph appeared to be good alternatives to the sustained remission variable in CAMERA-II.
Bakker MF et al. Low-dose prednisone inclusion in a Methotrexate-Based, Tight Control Strategy for Early Rheumatoid Arthritis. A Randomized Trial. Ann Intern Med 2012;156:329-39.
Boers M et al. A new graph and scoring system simplified analysis of changing states: disease remissions in a rheumatoid arthritis clinical trial. J Clin Epidemiol 2010;63:633-7.
Disclosure of Interest None declared