Using opportunistic screening with abdominal CT to identify osteoporosis and osteopenia in patients with diabetes

Opportunistic osteoporosis screening involves measuring the attenuation of L1 vertebrae on abdominal computed tomography (CT), which correlates with DXA T-score. We found that this approach is useful for detecting low bone mass in patients with diabetes and propose L1 attenuation <= 135 Hounsfield units (HU) as a threshold for which DXA should be strongly considered. Introduction Attenuation of the L1 vertebrae on computer tomography (CT) images done for other reasons ("Opportunistic Osteoporosis Screening") has been found to correlate well with DXA-derived T-score. However, the method and the thresholds have never been tested specifically in those with diabetes mellitus (DM), in whom the fracture risk is greater than explained by BMD. Methods In a retrospective study of subjects with DM who had both abdominal CT and DXA within 6 months of each other, we compared L1 attenuation and DXA T-score to define the sensitivity and specificity of thresholds previously established in the general population. Results There were 313 subjects among whom 18 (5.8%) had prior major osteoporotic fracture (MOF). Subjects with MOF had lower T-scores (- 2.3 +/- 1.4 vs. - 0.9 +/- 1.4,p < 0.001) and L1 attenuation (104 HU +/- 46 vs. 149 HU +/- 47,p < 0.001) than non-fracture subjects. L1 attenuation <= 160 HU was 91% sensitive for osteoporosis, while <= 110 HU was 80% specific. For a higher T-score of <= - 1.5, L1 attenuation <= 135 HU showed balanced sensitivity and specificity (65% and 69%, respectively). Conclusion Opportunistic osteoporosis screening with abdominal CT is useful in determining the need for DXA screening in subjects with diabetes. We propose L1 attenuation <= 135 HU as a reasonable threshold for detecting the T-score of <= - 1.5, which is likely associated with increased fragility in DM.