Validation of Roussouly classification in predicting the occurrence of adjacent segment disease after short-level lumbar fusion surgery

Abstract Background Recent studies demonstrated that restoring sagittal alignment to the original Roussouly type can remarkably reduce complication rates after adult spinal deformity surgery. However, there is still no data proving the benefit of maintaining ideal Roussouly shape in the lumbar degenerative diseases and its association with the development of adjacent segment disease (ASD). Thus, this study was performed to validate the usefulness of Roussouly classification to predict the occurrence of ASD after lumbar fusion surgery. Materials and Methods This study retrospectively reviewed 234 consecutive patients with lumbar degenerative diseases who underwent 1- or 2-level fusion surgery. Demographic and radiographic data were compared between ASD and non-ASD groups. The patients were classified by both “theoretical” [based on pelvic incidence (PI)] and “current” (based on sacral slope) Roussouly types. The patients were defined as “matched” if their “current” shapes matched the “theoretical” types and otherwise as “mismatched”. The logistic regression analysis was performed to identify the factors associated with ASD. Finally, clinical data and spinopelvic parameters of “theoretical” and “current” types were compared. Results With a mean follow-up duration of 70.6 months, evidence of ASD was found in the 68 cases. Postoperatively, ASD group had more “current” shapes classified as type 1 or 2 and fewer as type 3 than the non-ASD group ( p < 0.001), but the distribution of “theoretical” types was similar between groups. Moreover, 80.9% (55/68) of patients with ASD were mismatched, while 48.2% (80/166) of patients without ASD were mismatched ( p < 0.001). A multivariate analysis identified age [odds ratio (OR) = 1.058)], 2-level fusion (OR = 2.9830), postoperative distal lordosis (DL, OR = 0.949) and mismatched Roussouly type (OR = 4.629) as independent risk factors of ASD. Among the four "theoretical" types, type 2 had the lowest lumbar lordosis, DL, and segmental lordosis. When considering the "current" types, current type 2 was associated with higher rates of 2-level fusion, worse DL, and greater pelvic tilt compared with other current types. Conclusions DL loss and mismatched Roussouly type were significant risk factors of ASD. To decrease the incidence of ASD, an appropriate value of DL should be achieved to restore sagittal alignment back to the ideal Roussouly type. Level of Evidence : Level 4.