Sagittal decompensation after corrective osteotomy for lumbar degenerative kyphosis: classification and risk factors

A retrospective study

To classify the types and identify related factors on sagittal decompensation after corrective osteotomy for lumbar degenerative kyphosis (LDK).

There has been a skeptical view of surgical treatment of LDK owing to loss of sagittal balance even after correction of kyphosis. However, there had been no report on the classification and risk factors of sagittal decompensation.

A total of 23 LDK patients who had undergone corrective osteotomy were enrolled. The mean follow-up period was 45.7 months. Radiographic parameters including sagittal balance, the cross-sectional area of paravertebral muscles, were analyzed. We classified the type of sagittal decompensation into thoracic (Group T) and lumbar decompensation (Group L) with a reference line from the posterosuperior corner of the sacrum to the center of the T12-L1 disc. The type of sagittal decompensation was defined with the location of T1 and the reference line at the last follow-up radiographs.

The mean number of fusion segments was 7.7. Sagittal balance improved from 26.4 cm to 4 cm immediately after operation but deteriorated to 11.2 cm at the last follow-up. The decompensation was greater in Group T (11 cases) than in Group L (12 cases) (9.1 cm vs. 5.2 cm, P = 0.03). The comparative analysis showed significant differences between groups T and L in thoracic kyphosis at the last follow-up (Group T:L = 40.5°:27.5°, P = 0.04), preoperative thoracic kyphotic angle (Group T:L = 19.6°:-1°, P = 0.01), mean ratio of cross-sectional area of paravertebral muscles to intervertebral disc in T12-L1, and incidence of the preoperative compensatory thoracic lordosis (Group T:L = 27.3%:100%, P = 0).

The mean sagittal decompensation after corrective osteotomy for LDK was 38.3%. The etiology was loss of lumbosacral lordosis in Group L and progression of kyphosis at the proximal unfused segments in addition to lumbosacral loss in Group T. The decompensation was greater in the thoracic type than in the lumbar type and was considered relevant to a large preoperative thoracic kyphotic angle, absence of compensatory thoracic lordosis, and atrophy of paravertebral muscles.