FRAX Score Can Be Used to Avoid Superfluous DXA Scans in Detecting Osteoporosis in Celiac Disease: Accuracy of the FRAX Score in Celiac Patients.

The Fracture Risk Assessment (FRAX) tool has been developed to estimate patients' 10-yr probability of fracture, thus establishing which patients should undergo dual-energy X-ray Absorptiometry (DXA) scan. This study aimed to evaluate if the FRAX tool can replace or optimize the use of DXA scan in celiac disease (CD). We prospectively enrolled all CD patients aged over 40 yr diagnosed at our third-level unit. At time of CD diagnosis, all patients underwent FRAX score calculation for risk of major osteoporotic and hip fractures and DXA scan (used as gold standard) to assess the accuracy of the FRAX score. The FRAX score calculation was based on the following 10 variables: age (>40 yr), sex (M/F), body mass index, history of previous fracture (yes/no), parent fractured hip (yes/no), current smoking (yes/no), use of steroids (yes/no), rheumatoid arthritis (yes/no), secondary osteoporosis (yes/no), and alcohol ≥3 units/d (yes/no). DXA assessment was performed within 1 week from FRAX calculation. The FRAX score was dichotomized as normal or pathologic in accordance with the National Osteoporosis Guideline Group. A total of 160 CD patients were enrolled (M/F = 20/140; mean age 48.7 yr). A pathologic FRAX score was evident in 14 out of 160 patients (8.7%), whereas osteoporosis based on DXA scan was found in 10 patients (6%) (κ = 0.6); 3 patients with osteoporosis (1.9%) showed a 10-yr risk of major fracture >10% according to the National Osteoporosis Guideline Group criteria. With regard to diagnostic accuracy, the FRAX score showed sensitivity of 0%, specificity of 91%, positive predictive value of 0%, and negative predictive value of 94%. The prevalence of osteoporosis in adult CD appears to be quite low and only a small proportion of patients would require a DXA investigation. The FRAX score could be an effective tool to avoid useless DXA scans in CD patients in view of its high negative predictive value.