Sources of variability in wideband energy reflectance measurements in adults

Wideband acoustic immittance measurements of the middle ear, such as wideband energy reflectance (ER), can provide information about how the middle ear functions across the traditional audiometric frequency range. These measurements are being investigated as a new means of evaluating conductive hearing disorders, and studies have been reported on a number of middle-ear disorders. However, the normative database for wideband ER is still being developed, and more information is needed about sources of test variability.

The purpose of the present study was to evaluate sources of variability in wideband ER measurements at baseline and across annual tests for up to 5 yr in subjects with normal hearing.

The main group consisted of 112 subjects (187 ears), 24 females and 88 males, with normal hearing and normal 0.226-kHz admittance tympanometry. An additional 24 adults with abnormal 0.226-kHz tympanometry provided baseline comparison data.

A longitudinal design was used in obtaining annual measurements of audiometry, tympanometry, and wideband ER at ambient pressure in adults.

Clinical audiometry and tympanometry data and 1/3-octave wideband ER measurements were obtained at baseline and annually for up to four additional tests. Descriptive statistics and t-tests were used to explore differences in 1/3-octave baseline ER measures in terms of subject age, test ear, sex, and clinical tympanometry. Longitudinal mixed-effects linear regression models at 1.0, 2.0, and 4.0 kHz were used to examine the different sources of variance affecting ER over time.

There were small but statistically significant mean differences in ER for baseline measurements as a function of ear, sex, and age. Compared with these results, data for 29 ears with abnormal 0.226-kHz tympanometry differed from mean normal data across a broad frequency range by as much as 20%. ER varied as a function of peak compensated static acoustic admittance (Ytm) for measures at 1.0 kHz but was unrelated to Ytm at 2.0 and 4.0 kHz. ER also varied as a function of the test ear, with significantly higher ER on the left at 1.0 and 2.0 kHz, but was not significantly related to the test ear at 4.0 kHz. The standard deviation for test-retest variability was about 0.1 at each frequency, which is consistent with previous studies.

Mean wideband ER at baseline showed small but significant differences related to sex, ear, and age. ER was significantly related to Ytm at 1.0 kHz in the longitudinal data but not at 2.0 or 4.0 kHz and to the test ear at 1.0 and 2.0 kHz but not at 4.0 kHz. When evaluated at ambient pressure, ER for ears with negative middle-ear pressure was similar to that of ears with abnormally low Ytm. Therefore it might be necessary to evaluate wideband acoustic immittance compensated for middle-ear pressure by using tympanometry to obtain an effective differential diagnosis of middle-ear disorders in adults.