Background: The risk of osteoporosis in patients with rheumatoid arthritis (RA) is frequently overlooked, and investigating a simple indicator in routine care may be beneficial to motivate osteoporosis examination. The aim of this retrospective, case-controlled study was to identify the correlation between serum albumin concentrations and the prevalence of osteoporosis in postmenopausal patients with RA.
Methods: This study enrolled 197 patients who underwent dual-energy X-ray absorptiometry of lumbar spine (LS) and proximal femur without osteoporosis treatment [mean age, 67.5 years; disease duration, 12.8 years; Disease Activity Score assessing 28 joints with C-reactive protein, 2.0; prednisolone dose, 4.9 mg/day (usage, 42.6%); and LS T-score, -1.9]. Patients were classified into 2 groups: osteoporosis, defined as ≥ 1 part bone mineral density T-score ≤ -2.5 or history of fragility fracture of the vertebra or proximal femur (121 patients), and non-osteoporosis (76 patients). Groups were then matched by propensity score using clinical backgrounds affecting bone metabolism.
Results: In non-matched model, serum albumin concentration was significantly associated with osteoporosis-related factors such as aging, inflammation, physical disability, and glucocorticoid dose. Multivariate logistic regression revealed that serum albumin concentration was independently and significantly associated with osteoporosis risk (odds ratio = 0.22, 95% confidence interval = 0.08, 0.61, p = 0.0033). After propensity score matching, 57 patients for each group showed that in addition to the LS and femoral neck T-scores (p < 0.001), serum albumin concentrations (p = 0.01) remained lower in the osteoporosis group compared to non-osteoporosis group. Receiver operating characteristic curve analysis in non-matched model revealed that when cut-off value of serum albumin concentration for indicating osteoporosis was set at 4.2 g/dl, the area under the curve was 0.69, sensitivity 0.74, and specificity 0.58.
Conclusions: Low serum albumin concentration was significantly and independently associated with the prevalence of osteoporosis, which may be considered as one of the osteoporosis-related factors in postmenopausal patients with RA.
Copyright © 2021 The Japanese Orthopaedic Association. Published by Elsevier B.V. All rights reserved.