We retrospectively reviewed patients who underwent surgical treatment for bimalleolus or trimalleolus ankle fractures between January 2016 and December 2021. Inclusion criteria were as follows: (1) age ≥65 years old; (2) low energy trauma; (3) preoperative CT image; and (4) BMD measurement using DXA within one year before injury. Patients who had high energy trauma, polytrauma, or open fracture (N=3), neurologic impairment (N=1), or follow-up periods of <1 year (N=5) were excluded. Finally, 51 patients were enrolled in the study (Fig. 1). Demographic characteristics, such as, age at operation, sex, BMI, smoking, diabetes, mechanism of injury, and follow-up periods were evaluated. The bimalleolus or trimalleolus ankle fractures were classified according to the Lauge-Hansen classification or Denis-Weber classification.

The patients were placed in the supine position with fully extended knee joints and neutrally positioned ankle joints for preoperative CT using a 64-layer CT scanner (Siemens, Erlangen, Germany). Scanning parameters were 1.0 mm layer thickness and 0.6 mm reconstruction interval. Image files were obtained using the INFINIT program (INFINITT, Seoul, Korea) from a high-resolution medial picture archiving communication system (PACS; IMPAX; Agfa Healthcare, Mortsel, Belgium). The HU was measured at the distal tibia metaphyseal region approximately 1 cm proximal to the plafond on axial images. If the epiphyseal plate was located at the level of 1 cm above the plafond or the fracture fragment was so large that it was impossible to measure the HU at the level of 1 cm above the plafond, the HU was measured above the level of 1 cm above the plafond. By raising one or two cut on the CT axial image, the measurement of HU was possible without the influence of the fracture fragment and epiphyseal plate. The largest possible elliptical region of the distal tibia metaphysis was drawn on the PACS system, except the fracture area and cortical margin to prevent volume averaging (Fig. 2). Three axial images were obtained from the lowest CT axial image without the influence of the fracture fragment and epiphyseal plate by raising one or two cut on the CT axial image. The HU from three axial images were averaged to generate a mean HU value. Inter- and intra-observer reliabilities were assessed using the intraclass correlation coefficient (ICC) of the radiographic measurements, and an agreement of 0.75 was considered excellent.[21]

DXA scans were conducted at the proximal femur and lumbar spine using the Horizon-W DXA scanner (Hologic Inc., Bedford, MA, USA). Bone mass density (g/cm²), T score, and Z score were obtained using DXA in all the patients. To evaluate the correlations between DXA and HU, the patients were required to have a complete DXA scan within one year before ankle CT. According to the T score, patients with T score −2.5 or less were classified into an osteoporosis group, and other patients were classified into a non-osteoporosis group.

All of the surgeries were performed under spinal anesthesia. All the fractures were treated using an injury-specific method according to the injury pattern. Associated lateral, medial, and/or posterior malleolus was treated with open reduction and internal fixation, according to standard principles, which included internal fixation of the fibular and posterior malleolus fractures using the posterolateral approach and a separate medial approach for the medial malleolus fracture. Postoperatively, posterior short-leg splints were applied for the first two weeks. Thereafter, the splint was removed, and controlled ankle motion boots were applied to encourage an early range of motion for four more weeks. Weight-bearing was not allowed during this period. At six weeks, when the fracture appeared stable and any associated fractures were healed, the boots were removed and progressive rehabilitation (progressive weight bearing and ankle range of motion exercise) was allowed. The patients had outpatient visits for the first two weeks, four weeks, three months, six months, and then every six months postoperatively.

The clinical outcome was evaluated based on the Foot and Ankle Outcome Score (FAOS). The FAOS is a subjective foot and ankle clinical outcome score for patient-relevant outcomes after a foot and ankle injury.[22] The FAOS consists of 42 items grouped into five subscales: pain (9 items), other symptoms (7 items), activities of daily living (ADL; 17 items), sport and recreational function (sport/recreation, 5 items), and foot and ankle related quality of life (QOL; 4 items). Each question of the item was scored on a five-point Likert scale between 0-4 (non, mild, moderate, severe, and extreme problems). Scores were calculated by summing the total items score of the sub-scale and dividing it by the maximum subscale score. The final score was transformed into a scale of 0-100 (0=extreme problem, 100=no problem). The FAOS was evaluated at one year postoperatively.

Inter- and intra-observer reliabilities of HU measurements were assessed using ICC of the radiographic measurements, and an agreement of 0.75 was considered excellent.[21] Values are presented as mean and standard deviation for continuous data and number or percentage for categorical data. The relationship between HU and BMD measurement using DXA were evaluated using the Pearson’s correlation coefficient. Correlations between HU and the FAOS were determined using the Pearson’s correlation coefficient, followed by a linear regression model, which was adjusted for age and BMI, as both are potential confounder. To compare the osteoporosis and non-osteoporosis groups, a nonparametric Mann-Whitney test and Fisher’s exact test were used for continuous and categorical data, respectively. All analyses were performed using SPSS version 24.0 (IBM Corp., Armonk, NY, USA), and P value of less than 0.05 was considered significant.