Background Patients with Spondyloarthritis (SpA) have an increased risk of osteoporosis, which can be observed at the early stages of the disease, related to a systemic bone effect of inflammation. Diagnosis of SpA can be challenging because several diseases can mimick SpA symptoms and patients may be overdiagnosed especially in the presence of HLA B27.
Objectives The aim of our study was to evaluate the psychometric properties (specificity, sensitivity, positive (PPV) and negative predictive values (NPV) and positive likelihood ratio (LR+)) of low BMD in patients consulting in our department for SpA suspicion.
Methods From January 2008 and December 2012, patients visiting a tertiary Department of Rheumatology for symptoms suggestive of SpA were included. Exclusion criteria were biological therapies and antiosteoporotic treatment intake. Demographic data, disease duration, activity disease (BASDAI, CRP), presence of HLAB27 were assessed. BMD measurements at lumbar spine and total hip were assessed using DXA. Patients were classified into two groups: confirmed SpA and unconfirmed or uncertain SpA according to the diagnosis of the rheumatologist, which was considered the “gold standard”. BMD measurements results and percentage of patients with a low BMD (T score≤-2) at lumbar spine and total hip were compared between both groups. Comparisons of BMD measurements between patients with or without SpA, and performance of BMD (specificity, sensitivity, PPV, NPV and LR+) were performed.
Results During 4 years, 267 patients were included in the study. Compared to patients with uncertain SpA (n=193), patients with confirmed SpA (n=74) were more frequently male (63.2% vs 40.5% p=0.001), with positive HLA B27 (78.5% vs 50.7%, p<0.0001), higher disease duration (11.2±9.8 vs 8.3±7.9 years, p=0.035) and higher CRP (12.2±16.9mg/l vs 4.3±10.7mg/l, p<0.0001). There were no differences in former and current therapies (corticosteroids, NSAIDs and DMARDs), and in risk factors of osteoporosis. Mean lumbar spine BMD of patients with SpA was statistically lower than patients with uncertain SpA (0.98±0.16 g/cm2 vs 1.04±0.15 g/cm2, p=0.023) without any differences for total hip BMD. The specificity, sensitivity, PPV and NPV were respectively for a T score≤-2 at lumbar spine: 90%, 26%, 87% and 32% and LR+ of 2.60. Performances values for a T score≤-2 at total hip were 96%, 14%, 89% and 30% and LR+ of 3.12. In the subgroup of patients (n=128) without radiographic sacroiliitis nor syndesmophytes, lumbar spine BMD was significantly lower for patients with SpA (n=61) 0.98 (±0.18) g/cm2 than patients with uncertain SpA (n=67) 1.04 (±0.16) g/cm2 (p=0.023). In this subgroup the specificity, sensitivity, PPV and NPV were respectively for T score≤-2 at lumbar spine: 90%, 28%, 73%, and 58%, and 95%, 12%, 70% and 54% at total hip. The LR+ were 2.90 for a T score≤-2 at lumbar spine and 2.54 at total hip.
Conclusions In patients having symptoms suggestive of SpA, a low BMD (Tscore≤-2) is an additional tool for the diagnosis of SpA.
Disclosure of Interest None declared