Background Glucocorticoid-induced osteoporosis is the commonest cause of secondary osteoporosis and related fracture. The current gold-standard for the diagnosis of osteoporosis is dual x-ray absorptiometry (DXA) that measures areal bone mineral density (aBMD). However, aBMD has a low predictive value for vertebral fracture. High–resolution peripheral quantitative computed tomography (HR-pQCT) allows assessment of volumetric BMD (vBMD) and bone macro/microstructure. Whether HR-pQCT can better predict vertebral fractures in patients on long-term glucocorticoids (GC) compared to DXA remained uncertain.

Objectives To determine if peripheral bone parameters evaluated by HR-pQCT can discriminate patients on long-term GC with and without vertebral fracture independent of aBMD.

Methods This is an ongoing, cross-sectional study of 1100 rheumatic disease patients on long-term GC to access prevalent vertebral fracture radiographically (Genant semi-quantitative assessment). 110 patients with fracture (fracture group) and 110 patients without fracture (control group) underwent assessment with 1) aBMD using DXA; and 2) vBMD/structural assessment of the non-dominant distal radius and tibia using HR-pQCT. We hereby report the interim results of the first 78 and 77 patients in the fracture and control groups respectively.

Results Both groups were matched in gender [female: 63 (81%) vs 59 (77%); p=0.528], disease type [SLE/RA/others: 28 (36%)/27 (35%)/23 (29%) vs 36 (47%)/16 (21%)/25 (32%); p=0.143], and cumulative GC dose (17.3±13.6 vs 20.2±18.1g, p=0.447). Patients in the fracture group were older (62±14 vs 57±11 years, p=0.009), more postmenopausal [50 (79%) vs 37 (63%), p=0.048], of lower body mass index (BMI) (22.7±3.5 vs 24.2±3.4 kg/m², p=0.010), and more likely to be on vitamin D [53 (68%) vs 35 (45%), p=0.005] and bisphosphonates [19 (24%) vs 6 (8%), p=0.005]. Only 26 (35%) of the vertebral fracture patients were classified as osteoporotic (femoral neck T-score≤2.5). Fracture patients had a lower vBMD and impaired bone architecture compared with controls (Table). In multivariate logistic regression analysis, after adjusting for age, gender, BMI, vitamin D and bisphosphonate use, and aBMD, lower cortical thickness in the distal tibia (odds ratio: 1.38 for per 0.1mm decrease, 95%CI 1.13–1.67; p=0.001) and current use of vitamin D (3.78, 1.47–9.71; p=0.006) were independently associated with vertebral fracture.

Conclusions Our results suggest that skeletal abnormalities in patients with vertebral fracture are much more pronounced than those indicated by DXA. In particular, reduction in cortical thickness may be an important mechanism of vertebral fracture in rheumatic patients on long-term GC.

Disclosure of Interest None declared