XRays in standing, sitting and supine position; measurements of pelvic parameters and lumbar lordosis
Patients and Methods
We included healthy volunteers aged 18–50 years, non-lumbalgic, with no history of surgery, infection or pelvic or spinal tumoural pathology. This study was approved by a local ethics committee.
Standing: erect and relaxed, hands on the clavicles, horizontal gaze.
Sitting position: patient in a comfortable, natural position, with variable height seat, horizontal thighs, knee flexion at 90°. The upper limbs also relaxed, hands placed and crossed on the thighs to allow visualization of the lumbar spine.
Supine position: strict supine position on horizontal radiology table. Hands crossed on the chest, in a comfortable and natural position.
Radiological and Statistical Analyses
Pelvic parameters and lumbar lordosis from L1 to S1 were calculated using Surgimap 2.1.1. In the supine position, the sacral slope was calculated with respect to the vertical and the pelvic tilt in relation to the horizontal. Statistical analysis was performed using SPSS software. For each parameter we calculated the means with standard deviation. Normality of distribution of values was verified by the Shapiro–Wilk test. The correlation test used was the Spearman test.
Radiographs of 15 patients, five women and ten men, with an average age of 42.9 years were analysed.
Mean values of spinopelvic parameters (in degrees)
The majority of studies on spinopelvic balance are done in a static erect position. These studies have shown strong correlations between pelvic incidence and sacral slope, pelvic incidence and lumbar lordosis or lumbar lordosis and thoracic kyphosis. In fact, in real life, when standing, the human is most often in motion. He adopts other static postures in daily life, but most often in a sitting or lying position. The time spent during the day while sitting or lying down varies according to the individual, which depend in particular on socio-economic factors and age . Some elderly people spend <10% of a day while standing . In addition, disability may limit a patient to only seated and lying positions. So it seems restrictive to refer exclusively to the erect position for the study of the relationships and correlations between the pelvis and the lumbar spine.
Nevertheless, there has never been a study into the correlation between lumbar lordosis and pelvic incidence in other positions of daily life, particularly sitting or lying. However, it is important to know if this correlation persists in order to confirm the importance of pelvic incidence in posture, whether it is standing, sitting or lying down. It therefore seemed necessary to know if there is a correlation between pelvic parameters and lumbar lordosis in the sitting or lying position. A lack of correlation could suggest that much too much attention has been paid to the PI, for example in the planning of lumbar spinal fusion.
PI, a parameter specific to an individual , is nevertheless used in the calculation of theoretical lumbar lordosis when planning lumbar arthrodesis. Its importance is recognized: the suboptimal adjustment of lumbar arthrodesis can accelerate the degeneration of the adjacent level  and cause muscle pain. The lack of approximation between postoperative lumbar lordosis and pelvic incidence is correlated with poor clinical outcomes [6, 7].
In a standing position, the values of the pelvic and spinal parameters in our study are comparable to large data from the literature . Similarly, the correlation coefficients PI/LL, LL/SS and PI/SS are superimposable on the coefficients found in the literature. The correlations found in the literature between PI/SS and LL/SS are statistically stronger, i.e. with a coefficient greater than 0.7 than the LL/PI correlation [2, 9–15]. Our population is therefore representative of the general population concerning the lumbopelvic complex.
In the sitting position, the coupled movements of the coxofemoral joints and the lumbar spine have been well described in the literature and specified by Lazennec : hip flexion is accompanied by flexion of the lumbar spine, pelvic retroversion (increased PT) and decreased SS. These changes in pelvic parameters and lumbar lordosis are well documented in our study. These modifications are done harmoniously, with persistence of a strong correlation between PI/LL and SS/LL. The pelvic floor therefore also conditions the posture while seated.
In the supine position, we find a strong correlation PI/LL and very strong LL/SS. It is in a supine position that the PI/LL, LL/SS and LL/PT correlations are the strongest. Presumably, the suppression of gravity excludes the mechanisms of pelvic adaptation to gravity which appear when standing or sitting and must distort the excellent correlation found in the supine position. Nevertheless, surprisingly, we find an increase in the sacral slope and a decrease in PT and LL, which does not correspond to the coupled movements previously described. Perhaps while lying on the X-ray table, there is a discreet extension of the coxofemoral joints that has not been studied in this article. This hyperextension of the hips, which induces an increase of SS by coupled movement, cannot possibly be coupled with hyperextension of the lumbar spine because the back of the patient rests on the table. We had not considered this hypothesis, which should be confirmed by the measurement of the pelvifemoral angle. In our study, this angle could not be measured because the supine radiographs did not include the upper third of the femurs. The other potential explanation for this asymmetric variation of PT and LL in the supine position is secondary to the suppression of gravity constraints on the lumbopelvifemoral complex.
The PI is therefore an essential parameter not only in the erect position but also in sitting or lying. It is therefore essential to regulate lumbar arthrodesis according to the value of PI [1, 17– 19], even in elderly patients with reduced physical activity.
The limitation of this study concerns especially the absence of analysis of the subpelvic sector, in particular the pelvifemoral angle which could have helped us in the understanding of certain angular variations.
Neurovascular Risks During the Insertion of the S1 Screw: An Anatomical Study
The point of entry of the S1 screw is located at the lateral portion of the lower edge of the upper articular facet of the sacrum [20–26]. The aim is classically convergent from 30° to 40°, targeting the anterosuperior corner of the sacrum. According to a cadaveric study , the ideal convergence would be 35° ± 4. Roy-Camille  initially described screw insertion parallel to the vertebral endplate of S1 in 1983. De Peretti  describes a better bone fixation if the screw is upward of 10°.
Nevertheless, the behaviour of these screws depends on the quality of the bone, as the quality may be poor in osteoporotic patients. In addition, it has been shown that taking the anterior sacral cortex (bicortical screw) significantly enhances this sacral screw strength. But these bicortical screws expose the risk of damage to neurological and vascular structures located in the inner pelvis [29–31]. Also, we wanted to define the safety zone for these sacral screws and measure the distance between an ideally placed screw and the vascular and nerve structures that could potentially be injured by the screw.
In Vivo CT Measurements
In order to evaluate the risk of neurological or vascular injury during S1 screw insertion, we simulated the positioning of an S1 screw on ten pelvic CT scanners. Measurements were performed on CT scans of patients with no traumatic, tumour or infectious condition on the L5, S1 and S2 vertebrae. There were 6 men and 4 women with a mean age of 56 years.
Measurement of the distance of the S1 screw from the first sacral foramen. We simulated the implantation of two S1 right and left screws of 6.5 mm diameter on each scanner, for a total of 20 screws, using the Osirix Viewer® software . The entry point of the screw was classically located at the lateral aspect of the base of the upper articular facet of S1. The angle of convergence of the screw was determined in the axial plane with respect to the axis of the pedicle of S1. Then, in the sagittal reconstruction plane, this angle of convergence was reported to lie exactly in the plane of the pedicle. The path of the screw was then simulated by a 6.5-mm-thick line. We then measured the distance separating the line simulating the screw of the first sacral hole.
Measurement of the distance of the S1 screw with respect to the lumbosacral trunk. On these same scanners we simulated by a line of 6.5- mm-thick with the implantation of two S1 screws on the axial section passing below the upper plate of S1. We then measured the distance separating the end of the line simulating the head of the lumbosacral trunk screw.
Measuring the distance of the screw S1 from the iliac vessels. On these same scanners we simulated a line of 6.5-mm-thick for the implantation of two screws S1 on the axial section passing below the upper plate of S1. We then measured the distance between the end of the line and the common iliac artery screw on the right and the common iliac vein on the left.
The measurement for the medial sacral artery was hardly feasible since it is too small a vessel to be viewed on a standard scanner.
On average, the S1 screw goes to 5.2 mm from the first sacral foramen (±0.75 mm). It can be considered that screw insertion of S1 is a low risk procedure for the S1 root if it is inserted appropriately. In our experiment, we recommend performing S1 screw under fluoroscopy magnification in order to target the anterosuperior corner of the vertebral body of S1.
The distance between a bicortical S1 screw and the iliac vessels is on average 22 mm ± 2 mm. The average distance between a bicortical S1 screw and the lumbosacral trunk is on average 8.5 mm ± 0.8 mm.
The common iliac vessels can be reached if the screw in S1 is insufficiently convergent. On the right, the common iliac artery is in contact with the bone surface. The vein is located just in front of the artery. On the left, it is the common iliac vein which is located against the bone and which could therefore be reached. In another cadaveric study of 30 cadavers, screwing S1 with the entry point at the foot of the articular facet of S1 and a median convergence of about 10° did not produce injury of the iliac vessels. In this study, the medial sacral artery was reached in 4 out of 30 cases.
In our study, we were able to measure the distance between an ideally located S1 screw and the iliac vessels and the lumbosacral trunk. In case of a bicortical screw, the risk of vascular injury is very low, the distance between a converging screw of about 30° and the iliac vessels being >20 mm. On the other hand, the distance between the tip of the screw and the lumbosacral trunk is much smaller, 8.5 mm. It is therefore sufficient that the aim is not sufficiently convergent for a S1 screw to cause conflict with the lumbosacral trunk. The lumbosacral trunk is bulky, 8 mm wide, applied to the anterior surface of the sacral ala (wing). In addition to its volume, its vulnerability is increased by the fact that it is adherent to the bone to which it is attached by fibrous tissue .
Finally, the distance between the S1 screw and the S1 foramen had never been measured. Presumably, the breach of the sacral foramen by the S1 screw is rare; the S1 foramen can be visualized on the posterior aspect of the sacrum during the placement of the screw. Nevertheless, the distance between the ideally positioned S1 screw and the S1 foramen is small, of the order of 5 mm.
One must therefore be particularly careful when inserting a screw at S1 not to have a point of insertion that is too low and/or a downward trajectory that would risk compression of the S1 root.