Predictors of outcome following cochlear implantation in adults

Abstract

Open-set speech recognition without the use of visual clues is common following multichannel cochlear implantation in adults. However, large variations exist among the observed audiological outcomes. This research sought to identify preoperative and operative predictors of postoperative implant performance.The study took the form of a retrospective case review with statistical modelling. Subjects were 117 post-lingually deafened adult recipients of multichannel cochlear implants with a mean age of 52 years (range 18-79) and a mean duration of deafness of 16 years (range 0.125-48). Outcomes were open-set Bench-Kowal-Bamford (BKB) sentence test scores achieved after nine months of device use. Factors tested for predictive ability included historical variables: duration of profound deafness, aetiology of deafness and age at implantation; preoperative audiological variables: pre-operative speech recognition scores; electrode variables: percentage of active electrode array inserted; device variables: implant type and speech-processing strategy.BKB outcome scores demonstrated significant variability and were skewed towards higher outcome scores (mean 56% words correct). A multivariate ordinal regression analysis was therefore employed to identify independent predictors of implant performance using outcomes of poor, average and good performance based on word recognition cut-offs of 0-30, 30-70 and 70%+ respectively. Multivariate ordinal regression analysis demonstrated that duration of hearing loss prior to implantation was the only independent predictor of implant performance. The effect on implant performance of each additional year of deafness was small (9% greater chance of being in a poorer outcome group). This suggests that there are other factors influencing outcome that we were unable to identify in this study.