Impact of anxiety associated with COVID 19 on tinnitus

Background We investigated how the anxiety associated with COVID 19 impacts the severity of tinnitus and the outcomes of tinnitus therapy. Methods and Findings A retrospective research design was used to compare the clinical characteristics of tinnitus between patients from March to April 2020 under pandemic pressure and those from the matching period in 2019. Tinnitus severity was evaluated using the Tinnitus Handicap Inventory (THI) questionnaire and the minimum masking level (MML) measure while anxiety was quantified using the Zung Self-rating Anxiety Scale (SAS). The assessments were repeated after a 2-month interval, in which sound therapy was applied to a subgroup of patients. In all, 94 and 70 cases were reported in the 2020 and 2019 groups, respectively. The effects of the pandemic on emotional status were evident from a higher incidence of anxiety and much higher SAS scores in the 2020 group. There was also an increase in the THI scores in the 2020 group, but the between-year difference was smaller than that of the SAS score. Moreover, there was no between-year difference in MML or the treatment effect, as measured via both THI and MML. Furthermore, the reduction in SAS score in the second assessment was significantly smaller in the 2020 group. However, a positive correlation between the initial SAS score and the improvement was seen within the 2020 group. Conclusions Anxiety increased greatly in tinnitus subjects due to the COVID 19 pandemic. However, this was not associated with an increase in tinnitus severity in 2020. Instead, there was no between-year difference in the THI score or MML or in the improvement of either measure after treatment. The smaller improvement in SAS score and the positive correlation with the initial SAS score in the 2020 group suggests that the SAS score change in this group might largely be due to the natural relief of pressure as the pandemic decelerated in China. Therefore, the anxiety change induced by the COVID 19 pandemic is not likely to have a strong impact on tinnitus.

and 76 in the same period in 2019 (Fig 1). Table 1 compares the demographics and 1 8 2 tinnitus characteristics between the 2 years' groups. The case load for tinnitus appeared to 1 8 3 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 4, 2020.
be higher in 2020 than in the same period in 2019 (99 vs. 76, or an increase of 30.3%).

8 4
Such an increase could be fully attributed to a normal fluctuation due to the temporary 1 8 5 pause on non-emergency patients in our hospital between January and February 2020.
The two groups were matched by age and sex despite their different sizes. The chi-square test was used for the between-group comparison of sex, educational background, site, 1 9 1 anxiety, and the risk factor of sensorineural hearing loss; the t-test was used for age; a Mann-Whitney rank-1 9 2 sum test was used for duration; and a Fisher's exact test was used for the risk factors of noise exposure, 1 9 3 hypertension, hyperthyroidism, and head or neck trauma. In the 2020 group, 74 out of 99 subjects had an SAS score higher than 45 (the 1 9 7 criterion for anxiety), thus the incidence of anxiety was 74.7% in this group, significantly investigate the difference in the involvement of anxiety in tinnitus, a two-way ANOVA 2 0 0 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 4, 2020. . https://doi.org/10.1101/2020.07.02.20145532 doi: medRxiv preprint was performed on the SAS scores against the factors of year group (2020 vs. 2019) and 2 0 1 anxiety (with anxiety vs. without). Overall, the group had a significant effect, with the 2 0 2 2020 group having higher SAS scores (61.9 ± 11.9) than the 2019 group (48.2 ± 8.2; 2 0 3 F 1, 171 = 74.684, p < 0.001, Fig 2A). The anxiety effect also significantly interacted with COVID-19 outbreak not only increased the incidence of anxiety but also the degree of anxiety in those who suffered from it. No significant differences in the averaged SAS 2 1 0 score was seen between the patients without anxiety in 2020 and 2019 (43.9 ± 1.5 vs.  . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 4, 2020. . https://doi.org/10.1101/2020.07.02.20145532 doi: medRxiv preprint 1 3 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

4 4
A two-way ANOVA was performed on MMLs against both year group and anxiety.

4 5
The subjects in 2020 had significantly lower MMLs (8.3 ± 3.5 dB SL) compared to those between the years and the small sample sizes in patients receiving hearing aids, we 2 6 5 analyzed only the treatment outcomes for sound therapy.  suggest that the increased anxiety probably reduced the effectiveness of the treatment.  To fully evaluate the effects of the sound treatment, the changes in SAS, THI, and 2 9 0 MML scores were compared between years, as shown in Fig 3. First, the SAS score was 2 9 1 significantly reduced in both years after treatment, as indicated by the significant 2 9 2 treatment effect in two-way repeated measure (RM) ANOVA (F 1, 94 = 71.614, p < 0.001).

9 3
However, the SAS score was generally higher in 2020 than in 2019, as indicated by the 2 9 4 significant year effect in the ANOVA (F 1, 94 = 61.19, p < 0.001; Fig 3A). There was no 2 9 5 significant interaction between the factors of year and treatment. The SAS score was 2 9 6 reduced in both years in post hoc tests for the effects of the treatment (q = 6.911, p < 2 9 7 0.001 in 2020 and q = 10.48, p < 0.001 in 2019). Interestingly, however, the SAS score of 2 9 8 the 2020 group after being reduced by the treatment (58.0 ± 10.6) was significantly 2 9 9 higher than the before-treatment SAS score in the 2019 group (48.3 ± 8.5) according to a 3 0 0 Mann-Whitney rank-sum test (U = 534, p < 0.001). These results suggest that the anxiety 3 0 1 associated with COVID-19 had not been counteracted by the treatment.
3 0 2 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 4, 2020. To further identify the between-year difference and the effects of anxiety on the SAS 3 1 2 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

1 8
A significant interaction between the two factors was also seen (F 1, 92 = 9.037, p = 3 1 9 0.002). Correspondingly, the amount of the SAS score reduction from the treatment in the 3 2 0 patients without anxiety was significantly smaller (-1.5 ± 7.2) than those with anxiety in 3 2 1 2020 (7.0 ± 8.0, q = 5.364, p < 0.001) and those without anxiety in 2019 (6.0 ± 3.5, 3 2 2 q = 4.636, p = 0.002; Fig 3D). That is, the patients without anxiety in 2020 had less 3 2 3 improvement in SAS score. However, no difference was seen in SAS score improvement 3 2 4 in 2019 between the subjects with and without anxiety. The sound therapy appeared to reduce the THI scores in both years, as indicated by 3 2 6 the significant treatment effect in the two-way RM ANOVA (F 1, 94 = 58.405, p < 0.001). A 3 2 7 significant effect of the year was evidenced by an overall higher THI in the subjects, 3 2 8 overall, in the 2020 group (40.7 ± 6.7) than for those in the 2019 group (32.7 ± 8.3, 3 2 9 improvements were observed in both the 2020 and 2019 groups (from 9 ± 4.4 dB SL to 3 4 0 7.3 ± 4.2 dB SL and from 10.0 ± 3.8 dB SL to 7.9 ± 3.8 dB SL respectively, Fig 3C).

4 5
To further evaluate whether the initial anxiety was a factor that impacted on the 3 4 6 treatment, the Pearson product moment correlation was calculated between the initial 3 4 7 SAS score and the changes in the SAS, THI, and MML scores. There was a weak, tinnitus and those without. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 4, 2020. anxiety probably related to the special population investigated in this study, because the 4 3 3 incidence has been found to be much lower in other studies of the general population. For It is also noticeable that, in most of the previous studies, the goal was to evaluate 4 3 9 how tinnitus worked as a promoting factor for anxiety [40] , or to simply established 4 4 0 whether there was a link between anxiety and tinnitus [40,50,51,53,54] . In the present 4 4 1 study, we evaluated whether anxiety plays a causative role in promoting the development that the between-year differences in the tinnitus clinic afforded a good chance to verify 4 4 9 whether anxiety plays a causative or promotive role for tinnitus. Unfortunately, the level of anxiety was also evaluated with different methods in these 4 5 4 studies (ICD-9 code, HADS, and symptoms of depression and anxiety, respectively). In because the SAS has good psychometric credentials and continues to be widely used in Therefore, the anxiety increase caused by the COVID-19 pandemic seems to have 4 9 0 had a limited impact on tinnitus itself. This idea is further supported by the outcomes of underlying stress in association with the deceleration of the pandemic during the two-4 9 9 month interval between the two evaluations in the present study. With this idea, it 5 0 0 becomes understandable that those who had a higher SAS score might achieve more  This idea is further supported by the fact that there was no significant between-year 5 0 5 difference in the improvements in the THI score and MML in the present study. Our reductions were reported to be larger in subjects with anxiety (e.g., by higher HADS 5 0 9 scores in [52]) and furthermore, the reduction in anxiety was found to parallel the 5 1 0 improvement in THI by sound therapy [58,59]. In the present study, however, the 5 1 1 improvements in both THI score and MML were not larger in the 2020 group, in which 5 1 2 anxiety was much more severe. This conflict can also be explained by an assumption that severity, and therefore no impact on the treatment outcome. There were several limitations to our study. This was a retrospective study in which 5 1 7 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 4, 2020. data that allowed a comparison between years and the number of subjects using this counseling alone for their treatment, over the other two, probably due to financial 5 2 2 constraints. The combination of counseling with another therapy has been suggested by 5 2 3 others to achieve better outcomes [60][61][62][63]. In addition, the overall sample size in the 5 2 4 present study was small as the data were collected only from one hospital within a limited 5 2 5 period. Although the data and conclusion are solid in the present study, further 5 2 6 investigation would be helpful to verify the conclusion with a larger sample.

2 7
Conclusion 5 2 8 There was a substantial increase in anxiety of tinnitus subjects in 2020 under the 5 2 9 pressures produced by the COVID-19 pandemic. However, this increase was not matched 5 3 0 by the between-year difference in tinnitus loudness but matched better the change in THI score. The efficacy of sound therapy on tinnitus itself showed no significant difference 5 3 2 between years. While the post-treatment SAS score was still much higher in the 2020 5 3 3 group, there was a positive correlation between the SAS score improvement and the 5 3 4 initial SAS score in that group. This might have resulted from the natural reduction in by the pandemic is unlikely to be a promoting or enhancing factor for tinnitus. The authors acknowledge the colleagues for participating in this study. We are deeply 5 4 0 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 4, 2020. . https://doi.org/10. 1101/2020 indebted to the families who participated in the study. Competing interests 5 4