Intraoperative round window electrocochleography and speech perception outcomes in pediatric cochlear implant recipients.

OBJECTIVES: The goal was to measure the magnitude of cochlear responses to sound in pediatric cochlear implant recipients at the time of implantation and to correlate this magnitude with subsequent speech perception outcomes.
DESIGN: A longitudinal cohort study of pediatric cochlear implant recipients was undertaken. Intraoperative electrocochleographic (ECoG) recordings were obtained from the round window in response to a frequency series at 90 dB nHL in 77 children totaling 89 ears (12 were second side surgeries) just before device insertion. The increase in intraoperative time was approximately 10 min. An ECoG "total response" metric was derived from the summed magnitudes of significant responses to the first, second, and third harmonics across a series of frequencies. A subset of these children reached at least 9 months of implant use and were old enough for the phonetically balanced kindergarten (PB-k) word test to be administered (n = 26 subjects and 28 ears). PB-k scores were compared to the ECoG total response and other biologic and audiologic variables using univariate and multiple linear regression analyses.
RESULTS: ECoG responses were measurable in almost all ears (87 of 89). The range of ECoG total response covered about 60 dB (from ~0.05 to 50 μV). Analyzing individual ECoG recordings in bilaterally implanted children revealed poor concordance between the measured response in the first versus second ear implanted (r = 0.21; p = 0.13; n = 12). In a univariate linear regression, the ECoG total response was significantly correlated with PB-k scores in the subset of 26 subjects who were able to be tested and accounted for 32% of the variance (p = 0.002, n = 28). Preoperative pure-tone average (PTA) accounted for slightly more of the variance (r = 0.37, p = 0.001). However, ECoG total response and PTA were significantly but only weakly correlated (r = 0.14, p = 0.001). Other significant predictors of speech performance included hearing stability (stable versus progressive) and age at testing (22 and 16% of the variance, respectively). In multivariate analyses with these four factors, the ECoG accounted for the most weight (β = 0.36), followed by PTA (β = 0.26). In a hierarchical multiple regression analysis, the most parsimonious models that best predicted speech perception outcomes included three variables: ECoG total response, and any two of preoperative PTA, age at testing, or hearing stability. The various three factor models each predicted approximately 50% of the variance in word scores. Without the ECoG total response, the other three factors predicted 36% of variance.
CONCLUSIONS: Intraoperative round window ECoG recordings are reliably and easily obtained in pediatric cochlear implant recipients. The ECoG total response is significantly correlated with speech perception outcomes in pediatric implant recipients and can account for a comparable or greater proportion of variance in speech perception than other bio-audiologic factors. Intraoperative recordings can potentially provide useful prognostic information about acquisition of open set speech perception in implanted children.