Purpose of the study: Ankylosing spondylarthritis (AS) can progress to stiff thoracolumbar kyphosis which has an invalidating impact. Several publications have reported results with transpedicular osteotomies using horizontal or vertical reference lines. The purpose of our work was to report our experience with a new angle allowing the evaluation of the overall kyphosis (T1-S1) before and after correction by transpedicular osteotomy. The postoperative results were compared with the pelvic incidence (PI).

Material and methods: This was a radiographic study comparing a control group (154 asymptomatic volunteers and a group of patients with AS (n=28) who underwent posterolateral fusion associated with lumbar transpedicular osteotomy. The radiographic protocol was the same for the two groups. A large view including the entire spine was obtained (lateral and AP). We measured the classic pelvic parameters (pelvic incidence and version, sacral slope), C7 tilt and the spinosacral angle (SSA). All measured were made with computer assistance by the same operator. Twelve osteotomies were performed on L4 and 16 on L3.

Results: The PI was greater in the AS group compared with controls (61 vs 51). Seven patients have a PI < 50 (46 on average) and 21 had a PI > 50 (67 on average). For the C7 tile preoperatively, the low incidences had a low sacral slope and low pelvic version and greater global kyphosis than the high incidences (90 vs 98). In the control group, the C7 tilt and the SSA were 95.4 and 135.2 respectively. In the AS group, the C7 tilt increased from 72.6 to 83.1 (p=0.0025). The SSA increased from 96.4 to 13.3 (p=0.003).

Discussion: Pelvi with a low PI have a lower sacral slope than those with a high incidence; thus they can tolerate greater kyphosis before reaching an imbalance. For the high incidences, the pelvis has to retroverse more to obtain a low sacral slope. The minimal extension of the hips can limit this mechanism. After the osteotomy, all of the radiographic parameters were improved, but the SSA remained less than in the control group. The SSA is a good indicator of global kyphosis. Insufficient correction by unique lumbar transpedicular osteotomy explains the persistent retroversion of the pelvis postoperatively. The C7 tilt is useful to assess the improvement in the sagittal balance and the SSA give a better appreciation of the kyphosis correction per se.