Introduction: Robust clinically relevant epidemiological and audiological data are needed to prepare for future clinical trials aiming at preventing cisplatin-induced ototoxicity in this suffering cancer population. We assessed the incidence, severity, and potential risk factors of symptomatic cisplatin-induced hearing loss in a large cohort of adults.

Methods: Retrospective cohort study at a tertiary care university hospital. The study group included consecutive patients over 18 years old treated with cisplatin-based chemotherapy without concomitant inner ear radiotherapy or other ototoxic medication. Every participant underwent baseline pretreatment audiometry and was asked for audiological symptoms (tinnitus or subjective hearing loss) during the treatment. If symptomatic, comparative standard audiometry (0.125 to 8 kHz) was performed. Hearing loss was defined by a threshold shift =15 dB HL in at least one of the tested frequencies.

Results: A total of 401 cancer patients (59% males) with a mean age of 56 years (range 18-80) were included. Eighty-one patients (20%) developed symptomatic hearing loss, predominantly affecting the high frequencies from 4 to 8 kHz. Among them, 49 (60%) experienced simultaneous new-onset tinnitus. None of the analyzed potential risk factors (age, sex, smoking, hypertension, diabetes, dyslipidemia, chemotherapeutic regimen, and cumulative cisplatin dose) was statistically correlated with hearing loss.

Discussion: At least 1 out of 5 patients treated with cisplatin developed audiological symptoms associated with audiometric hearing loss within the 0.125 to 8 kHz range, for which new-onset tinnitus is a sensitive symptom. Not all audiological symptoms are accompanied by audiometric change. No predisposing factor could be identified. Standardized audiological monitoring before and during cisplatin-based chemotherapy allows quantitative assessment of early audiometric signs of ototoxicity, offering to optimize anticancer therapy while minimizing morbidity in a multidisciplinary setting.