Pos1078 effects of long-term low dose glucocorticoid treatment for rheumatoid arthritis on body weight and blood pressure: a pooled analysis of individual patient data from five randomised trials

Background High-dose glucocorticoids (GCs) can cause weight gain and hypertension. It is unclear whether GCs at ≤7.5mg/d prednisone equivalent (“low dose”), administered for rheumatoid arthritis (RA), do as well. Prior studies could not definitively answer this research question: Observational studies are confounded by indication, and individual randomized trials (RCTs) are usually underpowered for small safety signals. Objectives To assess the effects of long-term treatment with low dose GCs in RA on body weight and blood pressure by pooling individual patient data from RCTs. Methods Data from five RCTs with two-year interventions were pooled. [1-5] All trials originated in Europe and allowed concomitant treatment with DMARDs. Intervention groups received GCs at a dose of ≤7.5mg/d prednisone equivalent, control groups received placebo or nothing. Co-primary outcomes were the difference in change from baseline in a) body weight (kg) and b) mean arterial blood pressure (MAP; mmHg). A secondary outcome was difference in the change of number of administered antihypertensive drugs. Several sensitivity and subgroup analyses were conducted. All analyses were based on analyses of covariance, trial ID was included as a random effects factor to account for clustering of patients within trials. Multiple imputation was used to account for missing data under the intention-to-treat approach. No imputations were performed for trials with no data collected or available for the given outcome. A detailed prespecified protocol (dx.doi.org/10.17504/protocols.io.x54v9y4d1g3e/v1) with a hierarchical statistical testing procedure was followed. Results 1,112 participants were included (mean ± SD age 61 ± 15 years; 68% female). A mean DAS28 of 4.87 ± 1.16 indicated moderate disease activity at baseline; disability was moderate to severe with a median (interquartile range) health assessment questionnaire score of 1.38 (0.80; 2.25). Median disease duration was 1 year (0.42; 7). Baseline values for weight and MAP were 73kg ± 14 and 98mmHg ± 12; median number of antihypertensive drugs was 0 (0; 1). After two years, patients on GCs gained mean 1.1kg (95%CI 0.5 to 1.8; p < 0.001; Table 1 ) more weight than patients in the control groups. Both groups increased MAP by 2-3mmHg, without difference between the groups (–0.4; 95% CI –3.0 to 2.2 mm Hg; p = 0.19; Table 1 ). The change in number of antihypertensive drugs was low and similar in both groups ( Table 1 ). Results were consistent across sensitivity and subgroup analyses focusing on overweight or hypertensive patients, and when comparing GCs at 5mg/d with 7.5mg/d (data not shown). Conclusion We present robust evidence that low dose GCs, taken over two years for the treatment of RA, lead to about one additional kg of weight gain but do not cause changes in blood pressure. References [1] Boers et al. Ann Rheum Dis 2022 [2] Kirwan et al. NEJM 1995 [3] Choy et al. Ann Rheum Dis 2005 [4] Wassenberg et al. Arthritis Rheumatol 2005 [5] Svensson Arthritis Rheumatol 2005 Table 1. Changes in weight and mean arterial blood pressure in GC and control groups over two years. n GC group n Control group Contrast P value Weight, kg 547 1.9 (0.8) 565 0.7 (0.8) 1.1 (0.5 to 1.8) <0.001 Mean arterial blood pressure, mmHg 429 2.3 (1.3) 433 2.7 (1.4) –0.4 (–3.0 to 2.2) 0.19 Number of antihypertensive drugs* 276 0 (0;0) 282 0 (0;0) 0 (0 to 0) ** Values are least squares mean changes (SE), and differences in least squares mean changes (95% confidence intervals), unless otherwise stated. * Non-normal distribution: the group estimates are median changes (inner quartiles), and effect estimate is the median of differences between the two groups with corresponding 95% CIs, with no imputation for missing data and ignoring the grouping within studies. Analysis on the studies separately, and separate analyses on each imputed dataset showed similar results. ** In line with the study protocol, no significance test was performed for this secondary outcome. Acknowledgements We thank Sascha Marc Seidl (Merck, Inc.) for his help regarding data retrieval and interpretation. Disclosure of Interests Andriko Palmowski: None declared, Sabrina Mai Nielsen: None declared, Zhivana Boyadzhieva: None declared, Linda Hartman: None declared, Judith Oldenkott: None declared, Björn Svensson: None declared, Ingiäld Hafström: None declared, Siegfried Wassenberg: None declared, Ernest Choy Grant/research support from: EC has received research grants and honoraria from Abbvie, Bio-Cancer, Biocon, Biogen, Chugai Pharma, Eli Lilly, Fresenius Kai, Galapagos, Gilead, Janssen, Pfizer, Sanofi, and UCB unrelated to this manuscript., John Kirwan: None declared, Robin Christensen: None declared, Maarten Boers Consultant of: MB has received consultancy fees from Novartis unrelated to this manuscript., Frank Buttgereit Consultant of: FB has received consultancy fees, honoraria and travel expenses from Abbvie, Pfizer, Gruenenthal, and Horizon Therapeutics, all unrelated to this manuscript.