Determinants of Hearing Recovery in Multidrug-Resistant Tuberculosis Patients with Ototoxic Hearing Loss: A Prospective Observational Cohort Study in Kinshasa, Democratic Republic of the Congo

Purpose. The ototoxic hearing loss (HL) during the treatment of multidrug-resistant tuberculosis (MDR-TB) remains a public health challenge, especially in resource-limited settings such as the Democratic Republic of the Congo (DRC). This study seeks to identify the risk factors associated with the hearing recovery (HR) in MDR-TB patients in order to implement effective preventive strategies. Methodology. This prospective observational and analytic study included patients with MDR-TB. It was carried out in Kinshasa (Democratic Republic of the Congo) between February 15, 2020 and February 14, 2021. Sociodemographic, clinical, biological, therapeutic, and audiometric data were obtained and analyzed using SPSS for Windows software version 21. The hearing recovery (HR) was considered good if the final hearing loss (HL) was < 45 dB and poor if it was ≥ 45 dB. Pearson’s Chi-square or Fisher’s exact tests, as appropriate, were used to compare proportions. A multivariate logistic regression model was used to identify independent determinants of the HR. Odds ratios were provided with their 95% confidence intervals (Cis). Statistical significance was set at p < 0.05. Findings. Of the 310 patients were examined (mean age 34.5 ± 14 years, men 64.8%), 245 patients (79%) received only an aminoglycosides-based regimen, 31 (10%) received only bedaquiline (BDQ), and 34 (11%) received aminoglycosides (AG) followed by BDQ. Twenty-eight patients (9%) had a positive HIV serology, two-thirds were anemic (67.4%) and one-third (32.9%) had hypoalbuminemia. According to the final HL at the three months of treatment, 89 patients (28.7%) had experienced a good HR. Age under 40 years (p =0.002), the BDQ regimen (p < 0.001), normal hearing and mild HL (p < 0.001), and the type A and B audiometric curve (p < 0.001) were significantly associated with a favorable HR. Whereas, the presence of tinnitus (p = 0.031), hypoacusis (p < 0.001), hypoalbuminemia (p = 0.046), anemia (p = 0.044), and chronic renal failure (p = 0.040) were associated with a poor HR. In the multivariable logistic regression analysis, a moderate HL (OR 3.71; 95% CI [1.71 – 8.28]; p =0.001) and a type B audiometric curve (OR 2.27; 95% CI [1.72 -7.8]; p =0.001) emerged as the main determinants of a good HR. Unique contributor to theory, policy and practice: The hearing recovery in MDR-TB patients is poor during the treatment. This requires a regular audiometric monitoring to detect the early hearing loss. To reduce the risk of an ototoxic hearing impairment, it is recommended to prioritize the use of bedaquiline and to address adverse prognostic factors such as anemia, hypoalbuminemia, and impaired renal function. Hence, the need for large prospective and interventional studies to confirm these results.