Repair of subtotal tympanic membrane perforations: A temporal bone study of several tympanoplasty materials

The aim of this project was to investigate the effects of different types of graft material, and different remaining segments of the native TM on its motion. In twelve human temporal bones, controlled TM perforations were made to simulate three different conditions. (1) Central perforation leaving both annular and umbo rims of native TM. (2) Central perforation leaving only a malleal rim of native TM. (3) Central perforation leaving only an annular rim of native TM. Five different graft materials (1) perichondrium (2) silastic (3) thin cartilage (4) thick cartilage (5) Lotriderm(®) cream were used to reconstruct each perforation condition. Umbo and stapes vibrations to acoustic stimuli from 250 to 6349 Hz were measured using a scanning laser Doppler vibrometer. Results showed that at low frequencies: in the Two Rims condition, all grafting materials except thick cartilage and Lotriderm cream showed no significant difference in umbo velocity from the Normal TM, while only Lotriderm cream showed a significant decrease in stapes velocity; in the Malleal Rim condition, all materials showed a significant decrease in both umbo and stapes velocities; in the Annular Rim condition, all grafting materials except Lotriderm and perichondrium showed no significant difference from the Normal TM in stapes velocity. Umbo data might not be reliable in some conditions because of coverage by the graft. At middle and high frequencies: all materials showed a significant difference from the Normal TM in both umbo and stapes velocities for all perforation conditions except in the Annular Rim condition, in which silastic and perichondrium showed no significant difference from the Normal TM at umbo velocity in the middle frequencies. In the low frequencies, the choice of repair material does not seem to have a large effect on sound transfer. Our data also suggests that the annular rim could be important for low frequency sound transfer.