ABSTRACT
<p><b>OBJECTIVE</b>To investigate the sensitive factors which were used in routine audiological tests to find out otitis media with effusion (OME) in newborn infants.</p><p><b>METHODS</b>Subjects of this study were 48 infants, including 31 males and 17 females, who failed in the universal newborn hearing screening. The age ranged from 1.5 to 12 months with the average age of 4.3 months. All subjects accepted temporal bone CT and routine audiological assessments, including air-conduction and bone-conduction auditory brainstem response (ABR), 40 Hz-auditory event related potential (40 Hz-AERP), distortion-product otoacoustic emission (DPOAE), acoustic reflex, tympanometries using 226 Hz and 1000 Hz probe tone. Nine factors were statistically analyzed using Kappa test, Univariate chi(2) test and multivariate condition Logistic stepwise regression analysis, which included the results of acoustic immittance, the air-conduction and bone-conduction ABR thresholds, the difference between air-conduction and bone-conduction ABR thresholds, the latency of ABR wave I, duration between ABR wave I and V, 40 Hz-AERP thresholds, amplitudes and thresholds of DPOAE, and acoustic reflex thresholds (ART).</p><p><b>RESULTS</b>Seventy-seven ears were diagnosed with OME, and 19 ears were normal. CT scan of temporal bone was set as a comparative standard. Kappa test indicated that the results of tympanometry with 1000 Hz probe tone (Kappa = 0.745, P < 0.001), the air-conduction ABR threshold (Kappa = 0.453, P < 0.001), the latency of ABR wave I (Kappa = 0.430, P < 0.001), the threshold of 40 Hz-AERP (Kappa = 0.582, P < 0.001), and DPOAE (Kappa = 0.495, P < 0.001) had agreement with the results of temporal bone CT on evaluating the function of middle ear. Univariate analysis indicated that sensitive factors of middle ear function in newborn infants were tympanometry with 1000 Hz probe tone (P < 0.001), ART (P < 0.001), the air-conduction ABR threshold (P < 0.001), the difference between air-conduction and bone-conduction ABR thresholds (P < 0.001), the latency of ABR wave I (P < 0.001), the threshold of 40 Hz-AERP (P < 0.001) and DPOAE (P < 0.001). And multivariate conditional Logistic stepwise regression model showed that tympanometry with 1000 Hz probe tone (P < 0.001) and 40 Hz-AERP threshold (P = 0.004) can be substituted into Logistic stepwise regression equation.</p><p><b>CONCLUSIONS</b>Tympanometry with 1000 Hz probe tone and are sensitive factors to find out OME in newborn infants. The air conduction ABR threshold, ABR wave I latency, 40 Hz-AERP threshold and DPOAE could reflect the middle ear function of newborn infants effectively.</p>