**4. Discussion**

The present systematic review describes the effect of cochlear implantation on tinnitus in patients with SSD and disabling tinnitus. Reduction of tinnitus was reported in a relatively high number of studies (31 studies, 479 patients). No aggravation of tinnitus was reported in any patient. When compared to no treatment, CI was associated with better tinnitus suppression scores. These findings are encouraging in considering CI for SSD patients with disabling tinnitus, more specifically when conventional treatments fail to relieve the tinnitus. Although results are promising so far, the indication of CI for these patients is not yet widespread.

Most studies included in the present review assessed tinnitus using subjective tools; these are available in different languages, are not time consuming, and provide validated scores. VAS and THI were the most frequently used, followed by TQ. Although it could seem advantageous to not to be limited to a single tool, particularly since not all tools are validated for all languages, and some are more difficult to use than others, the heterogeneity of tools employed hampers comparison between studies. It is of note that objective tools were less likely to be used, which is possibly related to the difficulty of access to equipment required for electrophysiological and radiological assessments but also to the lack of available personnel with the skills to perform the assessments and interpret the results. These tools are, however, interesting in further understanding the mechanism of tinnitus reduction as well as the anatomical areas intervening in this process. It may also be helpful in identifying parameters that can predict prognosis. More generally, further research is needed to objectively assess treatment related physiological processes.

All SSD patients included in this review had disabling tinnitus, but the characteristics of their deafness were variable in terms of interval between onset and cochlear implantation, aetiology, and type of CI device. This makes it difficult to compare studies, but suggests treatment is successful independent of these factors. The risk of bias assessment showed a lack of precise inclusion criteria as well as a definition of the intervention in many studies. This emphasizes the need for a randomized clinical trial with clearly defined inclusion criteria and standard and clear intervention and rehabilitation protocols.

Whether tinnitus was evaluated as a primary complaint or not, CI succeeded to alleviate tinnitus. Studies in which tinnitus was evaluated as a primary complaint discussed several tinnitus characteristics including residual inhibition and recurrence of tinnitus after deactivation of implant. These studies were less likely to discuss hearing aspects or psychosocial benefits compared to studies where tinnitus was not the primary complaint.

Our review included all studies until December 2021. The present systematic review differs from previous published reviews in several ways. First, and to the best of our knowledge, this is the only review in which the listed studies have been divided into two groups depending on whether or not tinnitus was the primary complaint. The latter division permits reducing the risk of false-positive and false-negative errors. Second, the present review is not limited by the type of questionnaires used to assess tinnitus [50]: all validated multi-item questionnaires have been considered. In addition, studies using subjective assessment tools and studies using objective assessment tools were included. Audiological and neurophysiological levels of evidence were simultaneously considered when available. Last but not least, data on short- and long- term tinnitus suppression were analysed. The improvement of tinnitus, reflected by a significant reduction in various validated multi-item questionnaire scores, should strengthen considering CI in SSD with disabling tinnitus when conventional treatments are insufficient.

The present study has certain limitations; similar to the previously published systematic reviews, studies were mostly observational, and there was wide heterogeneity of tools used and a small sample size. This may preclude generalization of the results to a wider more heterogeneous population. Further studies with larger samples are needed to develop prediction models of tinnitus outcomes after cochlear implantation, where objective methods of tinnitus could be of interest.
