Assessment of Vertebral Microarchitecture in Overt and Mild Cushing's Syndrome Using Trabecular Bone Score

Helene Vinolas; Virginie Grouthier; Nadia Mehsen-Cetre; Amandine Boisson; Renaud Winzenrieth; Thierry Schaeverbeke; Charles Mesguich; Laurence Bordenave; Antoine Tabarin


Clin Endocrinol. 2018;89(2):148-154. 

In This Article

Abstract and Introduction


Objective: Osteoporotic fractures associated with Cushing's syndrome (CS) may occur despite normal bone mineral density (BMD). Few studies have described alterations in vertebral microarchitecture in glucocorticoid–treated patients and during CS. Trabecular bone score (TBS) estimates trabecular microarchitecture from dual–energy X–ray absorptiometry acquisitions. Our aim was to compare vertebral BMD and TBS in patients with overt CS and mild autonomous cortisol secretion (MACE), and following cure of overt CS.

Setting: University Hospital.

Design: Monocentric retrospective cross–sectional and longitudinal studies of consecutive patients.

Patients: A total of 110 patients were studied: 53 patients had CS (35, 11 and 7 patients with Cushing's disease, bilateral macronodular adrenal hyperplasia and ectopic ACTH secretion respectively); 39 patients had MACE (10 patients with a late post–operative recurrence of Cushing's disease and 29 patients with adrenal incidentalomas); 18 patients with non–secreting adrenal incidentalomas. 14 patients with overt CS were followed for up to 2 years after cure.

Results: Vertebral osteoporosis at BMD and degraded microarchitecture at TBS were found in 24% and 43% of patients with CS, respectively (P < .03). As compared to patients with nonsecreting incidentalomas, patients with MACE had significantly decreased TBS (P < .04) but not BMD. Overt fragility fractures tended to be associated with low TBS (P = .07) but not with low BMD. TBS, but not BMD values, decreased with the intensity of hypercortisolism independently of its aetiology (P < .01). Following remission of CS, TBS improved more markedly and rapidly than BMD (10% vs 3%, respectively; P < .02).

Conclusion: Trabecular bone score may be a promising, noninvasive, widely available and inexpensive complementary tool for the routine assessment of the impact of CS and MACE on bone in clinical practice.


Exposure to chronic glucocorticoid excess has adverse effects on bone resulting in an increased risk of fracture.[1] As endogenous Cushing's syndrome (CS) is a rare disease, most of our knowledge of this condition has been indirectly obtained by studying patients treated with exogenous glucocorticoids.[2] However, exogenous CS is associated with underlying primary conditions (inflammatory, haematological and autoimmune disorders) that may directly affect bone metabolism.

Several studies conducted in both exogenous and endogenous CS have shown that fractures frequently occur despite normal or only slightly decreased bone mineral density (BMD) suggesting a decreased bone strength due to qualitative deterioration of bone tissue.[1,3,4] The hypothesis that glucocorticoid excess has a greater impact on bone microarchitectural texture than on areal BMD has been confirmed by histomorphometric studies showing that glucocorticoid treatment induces a thinning of the trabeculae and a loss of trabecular bone volume.[5] Assessment of bone microarchitecture would, therefore, be of interest accurately evaluating bone alteration in CS. To date, the gold standard for studying bone microarchitecture is histomorphometric analysis of bone biopsy samples, a tool that is not compatible with routine clinical management. Alternatives are computed tomography systems (high resolution peripheral quantitative computed tomography (HRpQCT) and central quantitative computed tomography (QCT)) or micromagnetic resonance imaging (MRI).[6] In accordance with histomorphometric studies, qualitative assessment of the radius and tibia by HRpQCT in postmenopausal women using glucocorticoids revealed pronounced abnormalities in microarchitecture and lower stiffness compared with controls contrasting with no difference in BMD.[7] However, these techniques are not readily available in a clinical setting and are impractical for routine clinical management.

The lumbar spine trabecular bone score (TBS) is a grey–level texture measurement based on the use of experimental variograms. It is computed from the two–dimensional (2D) projection images acquired during a dual–energy X–ray absorptiometry (DXA) lumbar spine scan.[8] Low TBS values are associated with thinner trabeculae, low trabecular number, low connectivity density, that is, with poor microarchitectural bone quality. TBS has been shown to correlate with measures of HRpQCT and QCT in pre– and postmenopausal women.[9] Several prospective studies involving large numbers of patients with primary and secondary osteoporosis have shown that TBS is an independent predictor of fragility fractures (review in[10]). In addition, a recent meta–analysis of 14 prospective studies confirmed that TBS predicted fracture risk independently of FRAX probability and that their combined application enhance the probability prediction of fractures.[11]

It has been strongly suggested in a consensus report in 2015 by the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis that TBS might have a role in the assessment of fracture risk in conditions of secondary osteoporosis wherein BMD readings are often misleading.[10] Indeed, subsequent studies conducted in patients receiving glucocorticoids have shown that TBS was more frequently altered than BMD and assessed more accurately the qualitative properties of the bone.[12–14] Very few clinical studies using TBS in endogenous CS are available: one showed that TBS is more frequently altered than BMD in overt CS[15] and one showed similar results in patients with mild autonomous cortisol excess (MACE) associated with adrenal incidentalomas.[16] However, these studies deserve confirmation and the evolution of TBS after remission of hypercortisolism has not been studied.

The purposes of our study were to compare lumbar spine BMD and TBS in a large cohort of cases of overt CS and MACE of various aetiologies and to assess the evolution of BMD and TBS in subset of patients after remission of overt CS.