Using temporal bone computed tomography to predict sensorineural hearing loss in otic capsule-sparing temporal bone fracture.

OBJECTIVES: The purpose of this study was to investigate the efficacy of otic capsule-sparing (OCS) length for the prediction of sensorineural hearing loss (SNHL) in patients with OCS temporal bone fracture.

METHODS: Thirty-four patients with OCS temporal bone fracture were enrolled, and their temporal bone computed tomography (TBCT), audiogram, and medical records were reviewed. The TBCT measured the shortest length between the otic capsule and fracture line. This length was referred to as the OCS length and was used to predict SNHL. Ossicular dislocation and fracture were also evaluated. Patients were divided into two groups according to the presence of SNHL. Univariate and multivariate analyses were performed for age, gender, brain hemorrhage, mean bone conduction threshold on the contralesional side, OCS length, and ossicular dislocation and fracture on TBCT. A receiver operating characteristic (ROC) curve was produced to evaluate the efficacy of OCS length for the prediction of SNHL. To determine an association between degree of SNHL and OCS length, regression analysis was conducted in the SNHL group.

RESULTS: The mean OCS lengths of the SNHL and non-SNHL groups were 4.42±1.67mm and 8.00±5.71mm, respectively. In the univariate analysis, a relatively significant association (P <0.20) was found between SNHL and age, brain hemorrhage, mean bone conduction threshold on the contralesional side, OCS length, and incus dislocation. Multivariate analysis was performed using these factors. On multivariate analysis, OCS length (P=0.030, odds ratio=0.598; 95% confidence interval 0.375-0.952) was the only independent factor associated with SNHL. The area under the curve (AUC) was 0.747. When the cut-off value of OCS length was 5.27mm, the sensitivity and specificity for the prediction of SNHL were 71.4% and 69.2%, respectively. In the regression analysis, OCS length showed a significant association with degree of SNHL in the SNHL group (P=0.025, β=-12.822, SE=5.282).

CONCLUSIONS: The major finding of our study was that OCS length was significantly associated with SNHL in patients with OCS temporal bone fracture. Such patients with a short OCS length had a higher likelihood of SNHL.