Minding the Gap in Vertebroplasty: Vertebral Body Fracture Clefts and Cement Nonunion.

Background: Vertebroplasty and kyphoplasty are leading treatments for patients with vertebral body compression fractures. Although cement augmentation has been shown to help relieve pain and instability from fractures containing a cleft, there is some controversy in the literature regarding the procedure's efficacy in these cases. Additionally, some of the literature blurs the distinction between clefts and cement patterns (including cement nonunion and cement fill pattern). Both clefts and cement patterns have been mentioned in the literature as risks for poorer outcomes following cement augmentation, which can result in complications such as cement migration.

Objectives: This study aims to identify the prevalence of fracture clefts and cement nonunion, the relationship between them as well as to cement fill pattern, and their association with demographics and other variables related to technique and outcomes.

Study Design: Retrospective cohort study.

Setting: Interventional radiology department at a single site university hospital.

Methods: This retrospective cohort study assessed 295 vertebroplasties/kyphoplasties performed at the University of Colorado Hospital from 2008 to 2018. Vertebral fracture cleft and cement nonunion were the main variables of interest. Presence and characterization of a fracture cleft was determined on pre-procedural imaging, defined as an air or fluid filled cavity within the fractured vertebral body on magnetic resonance or computed tomography. Cement nonunion was evaluated on post-procedural imaging, defined as air or fluid surrounding the cement bolus on magnetic resonance or computed tomography or imaging evidence of cement migration. Cement fill pattern was assessed on procedural and/or post-procedural imaging. Pain improvement scores were based on a visual analog score immediately prior to the procedure and during clinical visits in the short-term follow-up period. Additional patient demographics, medical history, and procedure details were obtained from electronic medical chart review.

Results: Pre-procedural vertebral fracture clefts were demonstrated in 29.8% of our cases. Increasing age, secondary osteoporosis, and thoracolumbar junction location were associated with increased odds of clefts. There was no significant difference in pain improvement outcomes in patients following cement augmentation between clefted and non-clefted compression fractures. Clefts, especially large clefts, and cleft-only fill pattern were associated with increased odds of cement nonunion. Procedure techniques (vertebroplasty, curette, and balloon kyphoplasty) demonstrated similar proportion of cement nonunion and distribution of cement fill pattern.

Limitations: Cement nonunion was observed in only 6.8% of cases. Due to this low proportion, statistical inference tends to have low power. Multiple levels were treated in nearly half of the study's patients undergoing a single vertebroplasty/kyphoplasty session; in these cases, each level was treated as independent rather than spatially correlated within the same study patient.

Conclusions: Vertebral body fracture clefts are not uncommon and are related to (but distinct from) cement nonunion and cement fill patterns. Our study shows that, although patients with clefts will benefit from cement augmentation just as much as patients without a cleft, the performing provider should take note of cement fill and take extra steps to ensure optimal cement fill. These providers should also identify cement nonunion and associated complications (such as cement migration) on follow-up imaging.