3.2 | Hearing Outcome
Figure 2 shows hearing outcome between SBASSR and WASSR groups. A two-way ANOVA was conducted that examined the effects of ASSR grouping and follow-up assessment on hearing outcome. There was a statistically significant interaction between the effects of ASSR grouping and follow-up assessment on mean pure-tone threshold (F(2, 153) = 3.299, p = 0.040) and WRS (F(2, 100) = 3.7, p = 0.028). Post hoc testing using the Sidak’s multiple comparisons test was conducted to compare the hearing outcome between two ASSR groups. The last mean pure-tone thresholds was significantly better in the SBASSR group compared to the WASSR group (36.2 ± 21.1 dB HL vs 53.9 ± 21.4 dB HL, respectively, t (153) = 2.727, p = 0.021). However, the SBASSR group elicited a slight better last WRS compared to the WASSR group (77.6 ± 31.8% vs 58.8 ± 34.6%, respectively), which was not statistically significant (t (150) = 2.053, p = 0.120). The hearing outcome at one-week follow-up assessment was not significantly difference between two ASSR groups on mean pure-tone threshold (t (153) = 0.705, p = 0.861) and WRS (t (150) = 0.554, p = 0.926). The mean pure-tone thresholds at one-week follow-up assessment were 50.3 ± 19.9 dB HL for SBASSR group and 54.1 ± 23.4 dB HL for WASSR group. The WRS at one-week follow-up assessment were 60.1 ± 31.3% for SBASSR group and 60.4 ± 36.0% for WASSR group.
Table 2 shows the clinical factors related to hearing recovery. Twenty-eight patients in hearing improved group showed 40.6 ± 13.6 dB HL of hearing gain, but 25 patients in non-improved group showed 1.6 ± 9.1 dB HL of hearing gain. The proportion of patients from the SBASSR group who had hearing improvement was 68.8% whereas the proportion from the WASSR group who had hearing improvement was only 28.6%. The difference in proportions is significant (χ²(1, N = 53) = 8.214,p = 0.004). The side of affected ear and method of steroid treatment were also significantly associated with hearing recovery (allp < 0.05). A logistic regression was performed to ascertain the effects of ASSR grouping on the likelihood of hearing recovery. Variables (ASSR group, age, side, tinnitus, and steroid treatment) found to have possible association in univariate analysis (p < 0.20) were entered into the multivariate analysis model. However, age was removed in multivariate linear regression model with backward selection. The final logistic regression model with backward selection was statistically significant, χ2(4) = 21.782, p< 0.001. The model explained 45% (Nagelkerke R2) of the variance in hearing recovery and correctly classified 69.8% of cases. Patients in SBASSR group were 5.718 times more likely to have hearing recovery than patients in WASSR group. Patients without tinnitus were 17.631 times more likely to have hearing recovery than patients with tinnitus.