Bone strain index (BSI) is a tool measuring bone strain, derived from dual x-ray photon absorptiometry. It is able to characterise an aspect of bone quality that, joined to the quantity and quality parameters of bone mineral density (BMD) and trabecular bone score (TBS), permits an accurate definition of fracture risk. As no data are available about BSI precision, our aim was to assess its in vitro reproducibility.A Hologic spine phantom was used to perform BSI scans with three different scan modes: fast array (FA), array (A), and high definition (HD). Different soft tissue thicknesses (1, 3, 6 cm) of fresh pork rind layers as a surrogate of abdominal fat were interposed. For each scan mode, the phantom was consecutively scanned 25 times without repositioning.In all scan modes (FA, A, HD) and at every fat thickness, BSI reproducibility was lower than that of BMD. The highest reproducibility was found using HD-mode with 1 cm of pork rind and the lowest one using HD-mode with 6 cm of pork rind. Increasing fat thickness, BSI reproducibility tended to decrease. BSI least significant change appeared to be about three times that of BMD in all modalities and fat thicknesses. Without pork rind superimposition and with 1-cm fat layer, BSI reproducibility was highest with HD-mode; with 3 or 6 cm fat thickness, it was higher with A-mode.BSI reproducibility was worse than that of BMD, but it is less sensitive to fat thickness increase, similarly to TBS.