Diagnostic accuracy of CE Chirp.

Objective: There has been an increase in the use of the CE-Chirp stimulus in automated auditory brainstem response (AABR) equipment for neonatal hearing screening. The purpose of this study is to evaluate the diagnostic accuracy of the LS CE-Chirp-evoked auditory brainstem response (ABR) compared to the click-evoked ABR for the identification of different degrees and configurations of sensorineural (SNHL) hearing loss.

Method: 49 ears with mild to moderate SNHL were assessed: 16 ears with rising SNHL and 33 ears with sloping high frequency SNHL. Behavioural pure tone thresholds were obtained at 125-8000 Hz and ABR thresholds were measured using the click and LS CE-Chirp stimuli respectively. Click- and LS CE-Chirp-evoked thresholds were compared with each other and with behavioural pure tone average at 500, 1000, 2000 Hz (PTA), high frequency average at 2000, 4000, 8000 Hz (HFA) and low frequency average at 250, 500, 1000 Hz (LFA). Diagnostic accuracy of the two ABR stimuli was also compared by using ROC curves.

Results: Differences between click- and LS CE Chirp-evoked ABR, and behavioural thresholds were not statistically significant (p > 0.05). The highest significant correlations for ABR using clicks to behavioural thresholds was found at 2000 and 4000 Hz, whereas, the highest correlation for LS CE-Chirp ABRs to behavioural thresholds was found at 1000, 2000 and 4000 Hz (r > 0.7, p < 0.001). A very strong, positive correlation was found between both click (r = 0.805, p < 0.001) and LS CE-Chirp (r = 0.825, p < 0.001) and the behavioural PTA. LS CE-Chirp ABR thresholds were closer to mid and low frequency thresholds than the click ABR while the click-evoked thresholds were in closer proximity to HFA. Sensitivity and specificity and false negative rates were identical. Diagnostic accuracy of the LS CE-Chirp ABR was equal to or better than click for low (area under the curve (AUC) = 0.83), mid (AUC = 0.89) and high frequency hearing losses (AUC = 0.73). However, scatterplots indicated more frequent underestimation of behavioural pure tone thresholds at mid and high frequencies with the LS CE-Chirp than for the click ABR.

Conclusion: The diagnostic accuracy of the LS CE Chirp-evoked ABR is equivalent or better than the click-evoked ABR. The importance of ongoing surveillance and consideration of ABR screening protocols is consequently emphasized.