BACKGROUND Sagittal pelvic balance is a recognized factor influencing targeted acetabular-component anteversion during total hip arthroplasty implantation. However, no studies in the literature have systematically reported pelvic parameters data in the standing, sitting and supine positions. OBJECTIVE Variations in acetabular cup orientation can be traced to eventual pelvic balance changes in one of these three usual positions. METHODS In these three positions (supine, standing and sitting), pelvic anatomical parameters and reference planes were radiologically defined from a group of 67 patients (average age: 70.2+/-3.2 years). The complete X-rays individual sets were digitized and measurements were obtained by a single operator using a Spineview software (previously, strictly validated for these kind of measurements). Positioning according to the Lewinnek pelvic coordinate system, which is considered as a possible source of errors when vertically standing or horizontally lying, was also investigated. RESULTS The average pelvic incidence of 59.6 degrees did not vary in the sitting, supine or standing positions, with no statistically significant difference between sexes. The Legaye equation--pelvic incidence is equals to pelvic version plus sacral slope--was verified. Pelvic version increased by an average 22 degrees from the sitting to the supine or standing positions. Sacral slope varied in a reverse order. Pelvic-femoral angle (PFA) decreased by 20 degrees from the standing to the supine position. The Lewinnek plane was located 4 degrees posterior to the vertical plane. Whatever the position adopted, pelvi-Lewinnek angle appeared constant, averaging 12 degrees. CONCLUSIONS The average pelvic incidence in this series was high, most probably associated with advancing patient age and/or pathology. The concept of functional anteversion appeared critical when taking into account pelvic version variations (according to the position, sitting, supine or standing) positions. The Lewinnek plane, commonly accepted as the reference plane for hip navigation, was individualised to each patient and should not be mistaken with the vertical plane; positioning of the femur in relation to the Lewinnek plane was also specific to each patient. Cumulative approximation on these two parameters at surgery resulted in a combined imprecision of 26 degrees when standing and 36 degrees when lying down. We have thus defined crucial parameters to be integrated in computer-assisted hip surgery softwares: positional variations of the pelvic version (functional anteversion), positioning of the Lewinnek plane, and PFA value (both specifically patient's dependant). If integration of these parameters into new sofwares versions appears possible, this would represent a reliable compromise between maximum prosthetic stability, maximum joint amplitudes and elimination of possible prosthetic conflict.