The Effect of Preoperative Bone Density and Subchondral Cysts on Periprosthetic Fracture Risk in Total Ankle Arthroplasty

Abstract

Category: Ankle Arthritis Introduction/Purpose: Bone density is a modifiable factor which can be addressed prior to elective surgery if necessary. However, its role pertaining to complications of total ankle arthroplasty (TAA) has not been studied. Hounsfield units (HU) can be measured on standard computed tomography (CT) imaging, and have been shown to correlate with bone mineral density measures from dual-energy X-ray absorptiometry (DEXA). There is a precedent for these measurements in the orthopedic literature: different authors have shown that patients with higher vertebral bone density on CT are at lower risk for pedicle screw loosening. We hypothesized that patients with lower bone density, as measured by HU on preoperative CT, or with large preoperative cysts, would be at greater risk for revision and periprosthetic fractures following TAA. Methods: An existing database at the authors’ institution was used to screen all patients who underwent primary TAA. Inclusion criteria included a CT scan within one year prior to surgery. Exclusion criteria included tibial or talar hardware and nonweightbearing status at the time of the CT scan. The primary outcomes were prosthetic revision and periprosthetic fracture. HU were measured on axial CT cuts at 10 mm above the tibial plafond and at 5 mm below the talar dome to approximate the location of bone cuts. HU measurements for 30 patients were made independently by two reviewers in order to establish interrater reliability. Subchondral cysts at least 5 mm in diameter were counted. Additional patient factors analyzed included age, sex, weight, body mass index (BMI), tobacco use, presence of rheumatoid arthritis, preoperative deformity =15°, and pain visual analog scale scores. Results: 198 patients with a mean 2.4 years of follow-up met the inclusion criteria. The intraclass correlation coefficients for tibial and talar HU measurements were both 0.95. Seven patients (3.5%) underwent revision, four for infection, at a mean 1.2 years postoperatively. There were seven intraoperative and nine postoperative periprosthetic fractures (3.5% and 4.5%, respectively). Neither bone density nor cysts were associated with revision (p>.05). Lower tibial and talar HU, lower weight, and lower BMI were all associated with periprosthetic fracture (Table). After controlling for age, sex, and weight, only tibial HU was significantly associated with periprosthetic fractures (p=0.018). All intraoperative fractures occurred in patients with tibial HU <200. Of patients with tibial HU <200, 10 (22%) sustained an intra- or postoperative periprosthetic fracture. Conclusion: Lower tibial and talar bone density on preoperative CT of the ankle was strongly associated with periprosthetic fracture. The low incidence of revision during the relatively short study period limited our ability to analyze effects on revision rates. In patients who have had a preoperative CT, measuring HU represents a quick, simple method of assessing bone density with excellent inter-rater reliability. In patients with tibial HU <200, surgeons may wish to consider prophylactic internal fixation of the medial malleolus.