Computed Tomography Hounsfield Units as a Predictor of Reoperation and Graft Subsidence After Standalone and Multilevel Lateral Lumbar Interbody Fusion

Abstract

Standalone single and multilevel lateral lumbar interbody fusion (LLIF) have been increasingly applied to treat degenerative spinal conditions in a less invasive fashion. Graft subsidence following LLIF is a known complication and has been associated with poor bone mineral density (BMD). Previous research has demonstrated the utility of computed tomography (CT) Hounsfield units (HUs) as a surrogate for BMD. In the present study, we investigated the relationship between the CT HUs and subsidence and reoperation after standalone and multilevel LLIF. A prospectively maintained single-institution database was retrospectively reviewed for LLIF patients from 2017 to 2020, including single and multilevel standalone cases with and without supplemental posterior fixation. Data on demographics, graft parameters, BMD determined by dual-energy x-ray absorptiometry, preoperative mean segmental CT HUs, and postoperative subsidence and reoperation were collected. We used 36-in. standing radiographs to measure the preoperative global sagittal alignment and disc height and subsidence at last follow-up. Subsidence was classified using the Marchi grading system corresponding to disc height loss: grade 0, 0%-24%; grade I, 25%-49%; grade II, 50%-74%; and grade III, 75%-100%. A total of 89 LLIF patients had met the study criteria, with a mean follow-up of 19.9 ± 13.9 months. Of the 54 patients who had undergone single-level LLIF, the mean segmental HUs were 152.0 ± 8.7 for 39 patients with grade 0 subsidence, 136.7 ± 10.4 for 9 with grade I subsidence, 133.9 ± 23.1 for 3 with grade II subsidence, and 119.9 ± 30.9 for 3 with grade III subsidence (P= 0.032). Of the 96 instrumented levels in the 35 patients who had undergone multilevel LLIF, 85, 9, 1, and 1 level had had grade 0, grade I, grade II, and grade III subsidence, with no differences in the HU levels. On multivariate logistic regression, increased CT HU levels were independently associated with a decreased risk of reoperation after both single-level and multilevel LLIF (odds ratio, 0.98; 95% confidence interval, 0.97-0.99; P= 0.044; and odds ratio, 0.97; 95% confidence interval, 0.94-0.99; P= 0.017, respectively). Overall, the BMD determined using dual-energy x-ray absorptiometry was not associated with graft subsidence or reoperation. Using a receiver operating characteristic curve to separate the patients who had and had not required reoperation, the threshold HU level determined for single-level and multilevel LLIF was 131.4 (sensitivity, 0.62; specificity 0.65) and 131.0 (sensitivity, 0.67; specificity, 0.63), respectively. Lower CT HUs were independently associated with an increased risk of graft subsidence after single-level LLIF. In addition, lower CT HUs significantly increased the risk of reoperation after both single and multilevel LLIF with a critical threshold of 131 HUs. The determination of the preoperative CT HUs might provide a more robust gauge of local bone quality and the likelihood of graft subsidence requiring reoperation following LLIF than overall BMD.