Discussion
ABR is an objective hearing measurement rarely affected by anesthesia, sleep and environmental noises. Blackman-gated tone-burst ABR had been reported to predict the PTA thresholds for a wide range of frequencies while click evoked ABR only focused on 2kHz to 4kHz PTA thresholds. Compared to click-ABR, Blackman-gated 1kHz tone-pip ABR worked better in assessing low-to-mid hearing frequencies which are mainly impaired in CHL patients. The pre- and intra-fs-ABR only prolonged the surgery for 10 to 15 minutes. Instead of insert earplug, loudspeaker was better in maintaining sound intensity and aseptic principle.
In terms of different stimuli, both S1 and S3 worked well in OR. The wave V latency of S3 was longer than that of S1 because longer rise phase activates less nerve fiber compared to shorter rise phase. In S1, fs-ABRI and PTAI showed better correlation and in S3, the absolute intra-fs-ABR value was better in predicting the long-term HI.
Background noise of the air conditioner and anesthesia machine, magnetic and electronic influence of various machines might affect the monitoring. Therefore, only the anesthesia machine was left open while testing. We noticed that the mean hearing threshold of fs-ABR of NH in OR was 6.08dB higher than that in SPC, but it was not clinically significant, and would pose little impact on the efficiency of prediction.
Additionally, the mean bias between post-surgery PTA and intra-fs-ABR was smaller than that of the pre-PTA and pre-fs-ABR, and the difference between the two biases were 6.024dB and 14.09dB for S1 and S3 respectively. That meant intra-fs-ABR underestimated the post-PTA for 6.024 and 14.09 in average for S1 and S3 respectively which might be caused by the following factors: firstly, the revised tympanic membrane and the underneath cartilage would increase the mass of acoustic transducer system. It usually took 3-6 months for the cartilage to be absorbed which was unavoidable. Secondly, blood and serum would pile up in the external canal during monitoring, which would cause artificial conductive hearing loss. Necessary steps had been taken to minimize the impact of local errhysis, like using gelatin sponge with hemostatic at the external ear canal before testing.
ETD was a known etiology for CHL with a prevalence of 1% in adult and presented in 70% of patients undergoing tympanoplasty caused by chronic otitis media or cholesteatoma(10). Most otologists agree that better ETF is critical for the better long-term HI of middle ear surgery(11). We also observed that ETD affected the HI in a negative way. ETS-7 had been demonstrated to be valuable as a diagnostic follow-up method(12). Subjects with ETS≥7 showed better HI outcome and the coefficient factor between post-surgery PTA and intra-fs-ABR rose to 0.8, higher than that of whole S1 (r=0.5). We didn’t find any statistically significant correlation among parameters in Sub1(ETS<7). Therefore, it is recommended to pay attention to ETF to optimize the HI.
However, there are several conditions should be taken into consideration in future work. Firstly, more subjects should be enrolled to test the feasibility of this method among different kinds of ossiculoplasty. Secondly, the impact of errhysis on the threshold shift should also be studied in animal models. Thirdly, an automatic threshold determination method is now under study and we are trying to further minimize the testing time to optimize patients’ benefit.