A controlled study of the reliability of pure tone audiometry for the detection of gentamicin auditory toxicity.

Hearing changes in patients treated with gentamicin have been prospectively investigated. A group of patients who had a similar spectrum of clinical conditions but did not receive gentamicin have been studied as a control group. The audiometric technique was first validated by recording audiograms from healthy volunteers, once in an anechoic chamber and twice in a ward. It was concluded that changes in threshold less than 20 db. should not be regarded as significant when measured in a ward. In the gentamicin-treated group an initial audiogram was not a reliable measure of the patients' normal hearing because of the influence of their clinical condition. Twenty-one of the control patients were elective admissions to hospital and had their hearing first tested when they were clinically well. There was a 24 per cent incidence of unexplained hearing abnormalities--similar to the incidence of hearing loss in the gentamicin-treated group at follow-up. It was concluded that aminoglycoside auditory toxicity should not be assessed retrospectively. The incidence of hearing changes in the control and gentamicin-treated groups were similar. Possible causes of hearing loss in surgical patients are discussed and recommendations are made for the conduct of future studies of drug-induced auditory toxicity.