Relation between otitis media and sensorineural hearing loss: a systematic review

• Published in: BMJ open Vol. 11; no. 8; p. e050108
• Main Authors: Elzinga, Hester Beatrice Emilie, van Oorschot, Hanneke Doremiek, Stegeman, Inge, Smit, Adriana L
• Format: Journal Article
• Language: English
• Published: England BMJ Publishing Group LTD 12.08.2021; BMJ Publishing Group
• Series: Original research
• Subjects: Adult; Bias; Bone Conduction; Case-Control Studies; Ear diseases; Ear, Nose and Throat/Otolaryngology; Hearing loss; Hearing Loss, Sensorineural - etiology; Humans; Otitis Media - complications; Otitis Media, Suppurative - complications; Otolaryngology; Patients; Pediatrics; Systematic review; Young Adult; paediatric otolaryngology; audiology; otolaryngology
• ISSN: 2044-6055; 2044-6055
• DOI: 10.1136/bmjopen-2021-050108

ObjectivesThis systematic review summarises the evidence on the correlation between recurrent acute otitis media (rAOM) or chronic suppurative otitis media (CSOM) and sensorineural hearing loss (SNHL).Research methodsPubMed, Embase and Cochrane Library databases were searched from inception to 15 January 2021. Two authors independently identified articles, extracted data and performed quality assessment for included studies. Studies comparing the sensorineural hearing levels of patients with a history of rAOM/CSOM for >3 months to a control group were included.ResultsScreening of 4168 articles lead to inclusion of two case-control studies (control-group: patients non-OM) and seven cohort-studies (control group: contralateral ear). Quality assessment indicated considerable risk of bias in all studies. Reported populations varied (sample size 13–607, mean age 22–41.5 years, mean duration of disease 6.1–12.4 years). The OR for SNHL in the OM-group was 3.30–7.86 (95% CI 1.16 to 9.40, p<0.05) in cohort studies (n=2), and 0.05 (95% CI 0 to 0.78, p<0.05) in a case-control study. Mean/median bone conduction thresholds were respectively 1.19–32.21/0–10 dB higher on all frequencies (0.5–4 kHz) for the OM-group in four cohort studies (p<0.05). Two other studies reported no statistical test outcomes.ConclusionDue to the high risk of bias of included studies, effect estimates heterogeneity and suboptimal research designs, no conclusion on the correlation between OM and SNHL can be made. It emphasises the need for future prognostic studies.