Design: Systematic review and meta-analysis of 10 randomised trials
Population: 284 patients receiving cisplatin
Primary outcome: Change in hearing threshold in treated vs control ears
Effect: No significant difference (prediction interval −3.77 to 3.20); one STS trial 40% vs 85% ototoxicity

A meta-analysis of randomised trials in Otology & Neurotology (21 July 2026) pooled 10 studies with 284 patients testing intratympanic dexamethasone, N-acetylcysteine or sodium thiosulfate, mostly comparing a treated ear with the patient's other ear, to prevent cisplatin ototoxicity.

Across agents and frequencies, hearing thresholds did not differ between treated and control ears. Neither dexamethasone nor N-acetylcysteine showed a benefit on its own. Sodium thiosulfate data could not be pooled, but one trial found ototoxicity in 40% of treated ears against 85% of control ears.

Cisplatin ototoxicity is permanent and common. Intravenous sodium thiosulfate has trial evidence in children with localised solid tumours, but intratympanic approaches remain experimental. Baseline and serial audiometry is the practical step available now.

Arrange baseline audiometry before cisplatin and repeat it during treatment.
Do not offer intratympanic dexamethasone or N-acetylcysteine to prevent cisplatin hearing loss outside a trial.
Treat intratympanic sodium thiosulfate as promising but unproven.
Tell oncology early when audiometry shows a threshold shift, so dose decisions can be discussed.

Why it matters

It stops an unproven procedure spreading and points to monitoring as the practical safeguard.

The statistics, in plain English

A prediction interval spanning both sides of zero (−3.77 to 3.20 dB) means a new trial could plausibly find a small benefit or a small harm. With 284 patients in total, a modest effect could still have been missed.