Our data-driven approach assessed the relationship between the hearing-related QoL, the level of tinnitus annoyance, and also the perceived change in the level of tinnitus annoyance since cochlear implantation in 2322 adult CI recipients. The analysis demonstrated a statistically significant association between hearing-related QoL assessed by the SSQ12 and both tinnitus annoyance and perceived change in tinnitus annoyance at least one year after implantation. Overall, results suggested that CI recipients who experienced bothersome tinnitus showed worse subjective hearing-related QoL; whereas, CI recipients reporting less bothersome tinnitus since implantation had better hearing-related QoL. The SSQ12 total scores were also statistically significantly associated with the age of a recipient and presence of unilateral versus bilateral implants (Table 5). Coefficients of determination (R2) showed that only 8.4% of the variance in SSQ12 scores was explained by age, unilateral versus bilateral implantation and tinnitus annoyance levels. Therefore, it is highly likely that other important factors might contribute to the variations in SSQ12 scores.
Tinnitus prevalence was 33.9% in our cohort population, which is lower than the expected 40–54% prevalence found in previous studies on adult CI users 9,18,24. The variability in tinnitus prevalence in CI users can be explained by a non-response bias in previous studies and variations in tinnitus definition and heterogeneous assessment methods to identify the presence of tinnitus 2,25. This prevalence significantly varies with age (Table 4), as reported by other authors 2,26−28. Based on the age group classification used in this study, tinnitus prevalence increases with age, with up to 38.6% for the recipients aged between 54 and 65 years old, and thereafter declines for older age group. The same trend was reported by previous studies using different cut offs and age groupings 2,26.
Unilateral CI recipients reported significantly more tinnitus than bilateral CI recipients (Table 2). This is to be expected due to greater benefit with two implants compared with one 29. However, levels of tinnitus annoyance do not significantly differ between unilateral and bilateral implants. This might suggest that two implants do not bring significant benefit in terms of tinnitus annoyance compared to a single implant. These findings, however, remain hard to interpret without detailed information about the tinnitus percept (e.g. location) and complementary information on recipients’ hearing loss profiles.
In our study, we used a validated multi-item questionnaire, the SSQ12, to measure hearing-related QoL 22. The SSQ12 has no question related to tinnitus. The total score ranges from 0 to 10, with 10 indicating a perfect hearing-related QoL. We found that SSQ12 total scores significantly decreased with increasing tinnitus annoyance. A mean difference of 2.36 points in SSQ12 total scores was found between CI recipients reporting their tinnitus as not at all bothersome and recipients with extremely bothersome tinnitus, which is more than double the clinically significant change of 1.0 SSQ points suggested by the SSQ developers 21. A mean difference in SSQ12 scores of the same range (2.55 SSQ12 points) was reported in a study investigating the difference between normal-hearing and hearing-loss groups 30. Likewise, Wyss et al. showed a statistically significant improvement of 2.2 SSQ12 points at one-year post-implant in 1013 auditory implant recipients 31. Hence, the difference observed between the two extreme levels in tinnitus annoyance post-implantation showed the same magnitude of difference reported between pre-implantation and one-year post-implantation. The high mean differences found between different levels of tinnitus annoyance raises questions about the importance and the impact of tinnitus on hearing-related QoL in CI recipients, which may need further focus in clinics and exploration in future studies.
This study suggested a negative association between tinnitus and hearing-related QoL. When controlling for age and unilateral versus bilateral implants, mean SSQ12 scores were significantly lower in adult CI recipients with tinnitus than in CI recipients without tinnitus. Furthermore, tinnitus annoyance was also negatively associated with hearing-related QoL. The demonstrated association corresponds with the findings of previous studies investigating perceived tinnitus distress in CI recipients 9,16−18. In a study from Opperman et al., an increase in perceived tinnitus distress was correlated with a decrease in hearing-related QoL based on the Abbreviated Profile of Hearing Aid Benefit (APHAB) scores 18. This is in line with our findings on perceived changes in tinnitus annoyance where CI recipients who experienced less bothersome tinnitus showed better subjective hearing-related QoL. This might be related to the impact of tinnitus on psychological distress such as stress, coping strategies and depressive and anxiety symptoms 32,33, but also its impact on speech perception 34; that is, both psychological and perceptual factors can affect hearing-related QoL. Moreover, SSQ12 scores were not significantly different between adult CI recipients experiencing tinnitus as not at all bothersome and CI recipients without tinnitus, highlighting the importance of tinnitus-related distress over the presence of tinnitus. Further research is needed to fully understand the factors involved in this relation and its implications on CI outcomes.
SSQ12 scores significantly decreased with age (Table 5). The age effect on SSQ or shorter forms was already observed in other studies examining minimally hearing-impaired subjects 35 or CI recipients 9. Also, SSQ12 scores were significantly higher for bilateral CI recipients compared to unilateral recipients (Table 5). This association should be further investigated to assess if it could be related to other factors such as the implant side 9,19 or patients’ hearing loss characteristics 36.
Based on the models performed, 8.4% of the variance in SSQ12 scores was explained by age, unilateral versus bilateral implants and the level of tinnitus annoyance, with the latter being the most important predictor. Tinnitus annoyance and other tinnitus related characteristics deserve further research to understand what the causal relationship of the association is. The other 92% of the variance in SSQ12 scores could be potentially explained by differences in hearing impairment 36,37, in speech perception performances 38, in implant characteristics such as implant side 19 and in cognitive-linguistic performance such as listening effort 9. The influence of non-auditory aspects, such as education level 36, socioeconomic level or additional comorbidities, should also be considered in explaining the variance in SSQ12 scores. Investigating the above characteristics and then adding the tinnitus variable would be a valuable approach to confirm or temper our findings.
The study cohort is derived from a multi-country database collected in a web-based survey platform. This unique database gathers tinnitus and individual characteristics from a large sample of 2322 Nucleus cochlear implant users. Therefore, the findings of the study are expected to be generalizable to the European adult cochlear implant population.
Some methodological issues in this study are worth considering. The first limitation is that the observational study design was not primarily aimed to study the effects of tinnitus. Due to limitations in the number of questions and length of the survey, only three questions were used to assess tinnitus characteristics. Furthermore, the question related to change in tinnitus annoyance addressed past experience, which could present a potential recall bias. In fact, CI recipients were asked to report the perceived change in tinnitus annoyance since implantation, which corresponded to a time interval of 3 years or more for 472 recipients, potentially increasing recall bias even further. Collecting prospective tinnitus outcomes pre-implantation and post-implantation would provide better insights in order to assess the change in tinnitus annoyance since implantation. In fact, the lack of longitudinal data limits the scope of our study to post-implantation tinnitus experience and prevents us from definitively estimating the effect of cochlear implant on tinnitus annoyance between pre- and post-implantation. Nevertheless, the retrospective design ensures that no adaptation process has taken place between the pre- and post-implantation periods by assessing changes at a given point in time and, thus, controlling for response shift 39. Finally, we did not fully define tinnitus and other terms in the questions and answers 25. For instance, the options related to the perceived change in tinnitus annoyance ”I did not experience it before surgery” and “I don’t recall it before surgery” could both be interpreted as reporting tinnitus newly after implantation. Therefore, these deliberations should be taken with caution since we did not have access to the pre-implantation tinnitus report to validate this interpretation.
Considering the clear association between hearing-related QoL and level of tinnitus annoyance, the identification of accompanying tinnitus should be a requirement in the standard CI candidacy evaluation. Clinicians and CI manufacturers should address tinnitus as an important factor to better manage patients’ expectation and ensure the benefit of CI. This study highlights a need for individualized tinnitus management therapies to be made available within CI counselling and rehabilitation. Further research is needed to determine the underlying mechanisms and relationships. Another aspect that will require further investigation is whether tinnitus annoyance has a direct impact on CI performance such as speech recognition.