Purpose
The use of Hounsfield unit (HU) measurements derived from preoperative computed tomography (CT) scans has emerged as a promising surrogate for assessing bone mineral density (BMD). This systematic review aims to elucidate the role of HU in predicting spinal outcomes in patients undergoing spinal instrumentation.
Methods
A comprehensive systematic review of the English-language literature was performed across multiple databases, focusing specifically on adult patients who underwent spinal instrumentation. Adhering to PRISMA guidelines, studies published between 2017 and 2024 that investigated the correlation between HU measurements and spinal outcomes were included. Data extraction and subsequent meta-analysis were conducted utilizing a random effects model. The methodological quality of the included studies was appraised using the Newcastle-Ottawa Scale, and statistical analyses were executed with R software.
Results
A total of thirty-five retrospective studies met the inclusion criteria, encompassing a cohort of 3,927 patients. The reported rates of complications included cage subsidence at 35.5%, pedicle screw loosening at 27.9%, proximal junctional failure at 28.6%, and pseudoarthrosis at 66.7%. Notably, lower HU values were significantly associated with an increased incidence of complications across all studies. The delineated HU cutoff thresholds for predicting specific complications were as follows: 198 HU for lumbar interbody fusion cage subsidence; 116 HU and 126 HU for pedicle screw loosening in lumbar and thoracic spine fusions, respectively; 151 HU for proximal junctional failure in thoracolumbar spinal fusion; and 240–260 HU for pseudoarthrosis following anterior odontoid screw fixation. HU measurements were typically calculated as the mean HU of the vertebrae in the operational segments: cervical (C2-C7), thoracic (T1-T12), lumbar (L1-L4), or sacral (S1).
Conclusion
This meta-analysis substantiates a significant correlation between reduced HU values and spinal complications following instrumentation. Specifically, thresholds of below 130 HU for thoracic and lumbar pedicle screw loosening, below 200 HU for lumbar cage subsidence, approximately 150 HU for proximal junctional failure in the thoracolumbar region, and below 260 HU for pseudoarthrosis subsequent to anterior odontoid screw fixation were identified. The mean vertebral HU of the surgical segments serves as a reliable metric for this assessment.