OBJECTIVE Using an appropriate dose of an aminoglycoside antibiotic cotreated with a loop diuretic a guinea pig model of high-frequency loss can be obtained mimicking cochlear implant candidates with low-frequency residual hearing. We examined the stability of this model over time. BACKGROUND A well-established method to create an animal model for profound deafness is cotreatment with an aminoglycoside antibiotic and a loop diuretic. Recent data indicated that reduction of the aminoglycoside dose might yield selective high-frequency hearing loss. Such a model is relevant for studies related to hybrid cochlear implant devices, for example, with respect to preservation of residual hearing. METHODS Guinea pigs received an electrode for chronic recording of compound action potentials to tones to assess thresholds. They were treated with a coadministration of kanamycin (200 mg/kg) and furosemide (100 mg/kg), after which, the animals were sacrificed for histologic analysis at 2, 4, or 7 weeks. RESULTS After 2 to 7 weeks threshold shifts were greater than 50 dB for 8 to 16 kHz in 15 of 17 animals, whereas threshold shifts at 2 kHz or lower were less than 50 dB in 13 animals. Major threshold shifts occurred the first 2 to 4 days; subsequently, some spontaneous recovery occurred and, after 2-3 weeks thresholds, remained stable. Inner hair cell loss still progressed between 2 and 4 weeks in the most basal cochlear region; thereafter, hair cell loss was stable. CONCLUSIONS An appropriate animal model for selective severe high-frequency hearing loss was obtained, which is stable at 4 weeks after ototoxic treatment.