A new method for vertebral height measurements, morphometric X-ray absorptiometry (MXA) based on dual-energy X-ray absorptiometry, has been proposed. This technique overcomes some limitations of morphometric radiography (MRX): the effective radiation dose is low, some sources of geometric distortion are eliminated, such as dependence on patient position, magnification gradient, and the effect of scoliosis is minimized. The purpose of this study was to compare morphometric parameters obtained by both methods (MXA and MRX), and to evaluate the agreement between morphometric evaluations and qualitative reading for vertebral fracture diagnosis. The evaluation was performed with an Hologic QDR 2000 device in 67 women without vertebral fractures and 31 women with vertebral fractures (according to a qualitative assessment). The reproducibility of the image analysis was <4% and comparable to MRX. The estimated bias between the two methods was on average 10 mm, and was a function of the vertebral height, according to the Bland and Altman method. The agreement between MXA and other methods for vertebral fracture diagnosis was poor for the thoracic level above T7, due to a lack of resolution and rib interposition. Agreement was also low for T10, due to the motion of the diaphragm muscle. MXA cannot currently be used for the diagnosis of thoracic vertebral fracture in clinical practice. Technological improvements are necessary to make this promising method useful as a screening tool to evaluate the presence of thoracic vertebral fractures.