The accuracy of tympanometry assessed by its sensitivity and specificity

Tympanometry has become a widely accepted diagnostic tool for the diagnosis of otitis media with effusion (OME). The present study was conducted to assess the accuracy of the technique by calculating its sensitivity and specificity. Forty children with chronic OME and recurrent otitis media were studied. Immediate preoperative tympanograms were compared with operative findings of tympanocentesis, and sensitivity and specificity data were calculated. Approximately one third of the B type tympanograms (effusion) changed to A (effusion free) or C type (possible effusion) during the waiting period for the operation, showing resorption of fluid from the middle ear cavity. Sensitivity score was 95% bearing out the C type tympanograms. When C type tympanograms were included into the calculation, sensitivity dropped to 73%, showing liquid frequently present in cases of C type tympanograms. Tympanometry scored as more reliable in the presence of adenoid hypertrophy, when sensitivity was 83% and specificity was 90%. The absence of adenoid hypertrophy led to a drop of sensitivity to 58% and to a drop of specificity to 71%. These results may indicate that inhalant induction of anaesthesia has less effect on the middle ear in the presence of adenoidal hypertrophy and concomitant eustachian tubal dysfunction. Tympanometry proved to be a reliable diagnostic tool for the diagnosis of otitis media with effusion. Long preoperative waiting periods and the absence of adenoid hypertrophy could be responsible for the differences between preoperative tympanograms and the middle ear status observed at tympanocentesis.