IntroductionSagittal spinal morphology varies from one individual to the another and is specific to each person. Classifications proposed in the literature are based on the measurements of curvatures, angulations or other parameters, chiefly for a normal population [1-3, 7, 11, 13, 15-17]. They are purely descriptive rather than in any way analytic.Modifications of sagittal spinal curvatures have been connected with changes in pelvic orientation [7,11,12,16,17]. Nevertheless, the role of pelvic anatomy in this positioning has never been investigated. A pelvic anatomic parameter named pelvic incidence is proposed in this study as the key factor for this sagittal balance regulation. Spinal balance is conceived as the result of an optimal lordotic positioning of the vertebrae above a correctly oriented pelvis. "Optimal lordotic positioning" and "correctly oriented pelvis" values are defined in this study, performed on a normal and a scoliotic population. Materials and methods MaterialsThe data of two population groups were analysed. The first group comprised a normal population of 49 adults free of vertebral disease. There were 21 women and 28 men, with a mean age of 24 years (SD 5.8 years; range 19-50 years). Mean height was 173 cm (SD 8 cm): 177 cm (SD 7 cm) for the men and 161 cm (SD 8 cm) for the women). Mean weight was 65.8 kg (SD 10 kg): 71 kg (SD 8.6 kg) for the men and 59.2 kg (SD 7.7 kg) for the women. These 49 adults constituted the control group for assessing the relationship between spine and pelvis for a normal population. They consisted of volunteer medical and physical therapy students.The second group comprised 66 adult scoliotic women, with a mean age of 33 years (SD 19.8 years; range 10-84 years), and a mean height of 161 cm (SD 8 cm) and a mean weight of 55 kg (SD 8 kg). The frontal curves were lumbar in 24 cases and thoracolumbar in 42 cases. MethodsThe anatomical parameters (vertebral curvatures and pelvic parameters) were measured on orthogonal plain radiographs in standAbstract This paper proposes an anatomical parameter, the pelvic incidence, as the key factor for managing the spinal balance. Pelvic and spinal sagittal parameters were investigated for normal and scoliotic adult subjects. The relation between pelvic orientation, and spinal sagittal balance was examined by statistical analysis. A close relationship was observed, for both normal and scoliotic subjects, between the anatomical parameter of pelvic incidence and the sacral slope, which strongly determines lumbar lordosis. Taking into account the Cobb angle and the apical vertebral rotation confers a three-dimensional aspect to this chain of relations between pelvis and spine. A predictive equation of lordosis is postulated. The pelvic incidence appears to be the main axis of the sagittal balance of the spine. It controls spinal curves in accordance with the adaptability of the other parameters.

Pelvis and spinal curves were studied with an angular parameter typical of pelvis morphology: pelvic incidence. A significant chain of correlations between positional pelvic and spinal parameters and incidence is known. This study investigated standards of incidence and a predictive equation of lordosis from selective pelvic and spinal individual parameters. One hundred and forty nine (78 men and 71 women) healthy adults, aged 19-50 years, with no spinal disorders, were included and had a full-spine lateral X-ray in a standardised upright position. Computerised technology was used for the measurement of angular parameters. Mean-deviation section of each parameter and Pearson correlation test were calculated. A multivariate selection algorithm was running with the lordosis (predicted variable) and the other spinal and pelvic parameters (predictor variables), to determine the best sets of predictors to include in the model. A low incidence (<44°) decreased sacral-slope and the lordosis is flattened. A high incidence (>62°) increased sacralslope and the lordosis is more pronounced. Lordosis predictive equation is based on incidence, kyphosis, sacral-slope and ±T9 tilt. The confidence limits and the residuals (the difference between measured and predicted lordosis) assessed the predicted lordosis accuracy of the model: respectively, ±1.65 and 2.41°w ith the 4-item model; ±1.73 and 3.62°with the 3-item model. The ability of the functional spine-pelvis unit to search for a sagittal balance depended both on the incidence and on the variation section of the other positional parameters. Incidence gave an adaptation potential at two levels of positional compensation: overlying state (kyphosis, T9 tilt), underlying state (sacral slope, pelvic tilt). The biomechanical and clinical conditions of the standing posture (as in scoliosis, low back pain, spondylisthesis, spine surgery, obesity and postural impairments) can be studied by comparing the measured lordosis with the predicted lordosis.

The anatomic pelvic parameter "incidence" -the angle between the line perpendicular to the middle of the sacral plate and the line joining the middle of the sacral plate to the center of the bicoxofemoral axis -has been shown to be strongly correlated with the sacral slope and lumbar lordosis, and ensures the individual an economical standing position. It is important for determining the sagittal curve of the spine. The angle of incidence has also been shown to depend partly on the sagittal anatomy of sacrum, which is established in childhood while learning to stand and walk. The purpose of this study was (1) to define the relationship between the sacrum and the angle of incidence, and (2) to compare these parameters in three populations: young adults, infants before walking, and patients with spondylolisthesis. Forty-four normal young adults, 32 infants not yet walking and 39 patients with spondylolisthesis due to isthmic spondylolysis underwent a sagittal full-spine radiography. A graphic table and the software for bidimensional study of the sacrum developed by J. Hecquet were used to determine various anatomic and positional parameters. Comparison tests of means, and multiple and partial correlation tests were used. A study of

No abstract