Background The international guidelines recommend to use the average bone mineral density (BMD) over L1 to L4 in the management of osteoporosis and prediction of fracture. Exclusion of certain vertebrae is recommended according to specific rules (ISCD position). The spine Trabecular Bone score (TBS), a surrogate of bone micro-architecture, has been newly introduced into international guidelines and the FRAX® tool for clinical use in conjunction with BMD and clinical risk factors. Objectives The aim of this study is to test several TBS vertebrae combinations in regard to major osteoporotic fracture prediction. Methods The osteoLaus cohort (Lausanne, Switzerland) included 1500 woman 50 to 80 years old. All women had a detailed questionnaire related to clinical risk factors and treatment known to influence bone metabolism, BMD measurement (hip, spine and whole body), vertebral fracture assessment and TBS. The primary outcome was the prevalence of major osteoporotic fracture (MOF) according to TBS per-vertebral combination. Appropriate statistics and necessary adjustment for confounding factors were performed. L1-L4 TBS was used as reference value. Results Out of 19466 women included in the study (mean age 64.5±7.6 years, BMI 25.7±4.4), 12.7% suffered from MOF. The odd ratios per standard deviation decreased (OR) were 1.53 (1.29 – 1.80) and 1.80 (1.50 – 2.15) for the spine and total femur BMD respectively. Adjusted (age & glucocorticoids status) OR and area under the curve of different combination of vertebrae can be found in the following table for TBS. Conclusions It seems that excluding L4 tends to improve the fracture risk prediction. TBS is very sensitive to vertebra positioning (e.g. projection). To compensate for the natural lordosis of the spine, one has to lift the legs of the patient. L4 is often still angled which could explained such results. Further prospective studies are needed to confirm these results. Disclosure of Interest None declared