Background: Ototoxicity is a common adverse drug reaction associated with cisplatin therapy and with radiation to the HN region. We evaluated the differential effect on hearing impairment in HNSCC patients by candidate polymorphisms of genes associated with either hearing loss or cisplatin function. Methods: In this observational study of locally-advanced HNSCC patients treated with cisplatin chemoradiation, hearing impairment attributed to treatment was defined as ≥grade 2 audiometric change from baseline to post-treatment, evaluated within 18 months of completing therapy (CTCAE v4.02). Patients were genotyped for 30 polymorphisms using Sequenom. Logistic regression evaluated associations between genetic variants and ototoxicity. Cox regression assessed relationships between genetic variants and locoregional control (LRC), distant control (DC), disease free survival (DFS) and overall survival (OS). Results: Of 246 patients who had audiometric testing pre- and post-chemoradiation, 79% were male; 76%, oropharyngeal cancers; 11%, oral cavity cancers; 8%, laryngeal cancer; 91%, stage IV; 58% had hearing loss. Two polymorphisms had significant associations with hearing loss post treatment: WFS1 rs62283056 and ABCC2 rs3740066. In an additive inheritance model, individuals with WFS1 variants had a significantly decreased risk of ototoxicity (P = 0.012; adjusted odds ratio (aOR) = 0.56; 95% CI, 0.4-0.9, per increase in one minor allele), while the minor allele of ABCC2 was associated with greater risk of ototoxicity (P = 0.016; aOR = 1.68; 95% CI, 1.1-2.6). In contrast, the same genetic variants were not associated with LRC, DC, DFS or OS in a larger cohort of 642 HNSCC patients. Conclusions: WFS1 genetic variant is associated with differential hearing loss in LA-HNSCC patients. An ABCC2 variant, involved with removal of cisplatin from cells, is associated with increased cisplatin-induced ototoxicity. The same genetic variants were not associated with any efficacy outcomes. This information could be useful in the development of predictive models for cisplatin-induced ototoxicity.