Objective: To identify whether patients with diabetes have a higher incidence of sensorineural hearing loss than the general population and examine whether control of diabetes is related to severity of hearing loss.
Study Design: Retrospective database review; complete data mining of electronic medical record from 1989 to present.
Setting: Tertiary referral center.
Patients: Electronic medical records from 53,461 nondiabetic age-matched patients and 12,575 diabetic patients were reviewed.
Main Outcome Measures: Presence or absence of diabetes and/or sensorineural hearing loss, serum creatinine, pure tone hearing (dB), speech discrimination (%), serum cholesterol, and triglycerides.
Results: Sensorineural hearing loss was more common in the diabetic patients than in age0matched nondiabetic patients from the same institutions. Poor control of diabetes, as measured by increasing serum creatinine, but not apparent in hemoglobin A1C laboratory data, correlated with worsening hearing in patients with diabetes who had sensorineural hearing loss.
Conclusions: Sensorineural hearing loss was more common in patients with diabetes than in the control nondiabetic patients, and severity of hearing loss seemed to correlate with progression of disease as reflected in serum creatinine. This may have been due to microangiopathic disease in the inner ear.
The relationship between diabetes mellitus and hearing loss has been debated for many years. Jordao (1) in 1857 published a case report of a diabetic patient with hearing loss. Edgar (2) in 1915 was the first to report a high-frequency sensorineural hearing loss (SNHL) in a diabetic patient. Some authors conclude that there is no relationship between hyperglycemia and hearing loss; however, the bulk of the literature supports a poorly defined association.
The link between diabetes and SNHL makes intuitive sense, given the documented neuropathic and microvascular complications of diabetes and the complex blood supply of the inner ear. Most audiometric studies of hearing in patients with diabetes show a mild to moderate high-frequency SNHL (3), although Celik et al. (4) noted higher thresholds in diabetic patients at all frequencies tested. Cullen and Cinnamond (3) showed no difference in speech discrimination scores among diabetic patients and a normal population. The effects of different variables such as duration of diabetes, blood sugar control, and presence of end-organ damage on hearing loss have not yet been clarified, despite several studies of this topic. Part of the difficulty in identifying the effects of diabetes on hearing is the presence of comorbidities, such as hypertension and atherosclerosis, which could potentially affect hearing. Furthermore, the largest studies in this area have examined several hundred patients, which limit the conclusions that these studies can draw because of insufficient statistical power.
Although many reports in the literature have evaluated the relationship between these two diseases, to our knowledge no large-scale study of the laboratory and audiometric data in diabetic patients with SNHL has been published. This study is a retrospective analysis of laboratory and audiometric data of diabetic patients with SNHL, using a large computer database at the Veterans Affairs (VA) Maryland Health Care System, which is composed of three hospitals and several free-standing clinic sites. By combining sophisticated search engines with archival electronic records, we were able to study large populations and examine the effect of diabetes on hearing.