Auditory threshold shifts after glycerol administration to patients with suspected Menière's disease: a retrospective...

Nearly half a century ago, administration of glycerol was shown to temporarily improve the threshold of hearing in patients with suspected Menière's disease (glycerol test). Although a positive test result provides strong evidence of Menière's disease, the test has not gained widespread acceptance. A probable reason is that there is no consensus as to the definition of positive. Moreover, a negative test result is of little diagnostic value because Menière's disease cannot be excluded. By reanalyzing archived data, the authors sought to understand the test in light of signal detection theory. Moreover, they explored the possibility of estimating the probability of a positive test result from the pretest audiogram.

The study is based on audiograms from 347 patients (356 ears) who underwent a glycerol test to corroborate a suspected diagnosis of Menière's disease. Subsequent to an initial pure-tone audiogram, glycerol (1.2 mL/kg body weight) was orally administered; follow-up audiograms were obtained after 1, 2, 3, and 4 hr. Transcription of the audiograms into a computer-readable form made them available for automated reanalysis.

Averaged difference audiograms provided detailed insight into the frequency dependence and the temporal dynamics of the glycerol-induced threshold reduction. The strongest threshold reduction was observed 4 hr after glycerol intake, although nearly the same effect was already found after 3 hr. Strong overall threshold reductions were associated with a pronounced maximum at approximately 1000 Hz; weaker effects were associated with a plateau between 125 and 1000 Hz and a rapid decrease toward higher frequencies. To date, criteria suggested for a positive test result vastly differ in both sensitivity (with regard to the detection of a threshold reduction) and specificity (1 minus false-positive rate). Here, a criterion based on the aggregate threshold reduction in adjacent audiometric frequencies is suggested. This approach does not only seem to be more robust but also permits to freely adjust the false-positive rate. A positive test result is particularly likely when the mean low-frequency hearing loss is approximately 60 dB and the mean high-frequency hearing loss does not exceed 50 dB.

If the pretest audiogram does not render a positive test result unlikely, a state-of-the-art implementation of the glycerol test is a competitive method for corroborating a suspected diagnosis of Menière's disease.