Identification of neonatal hearing impairment: experimental protocol and database management

Ear-canal acoustic admittance and reflectance measurements in human neonates. II. Predictions of middle-ear in dysfunction and sensorineural hearing loss

This report describes relationships between middle-ear measurements of acoustic admittance and energy reflectance (YR) and measurements of hearing status using visual reinforcement audiometry in a neonatal hearing-screening population. Analyses were performed on 2638 ears in which combined measurements were obtained [Norton et al., Ear Hear. 21, 348-356 (2000)]. The measurements included distortion-product otoacoustic emissions (DPOAE), transient evoked otoacoustic emissions (TEOAE), and auditory brainstem responses (ABR). Models to predict hearing status using DPOAEs, TEOAEs, or ABRs were each improved by the addition of the YR factors as interactions, in which factors were calculated using factor loadings from Keefe et al. [J. Acoust. Soc. Am. 113, 389-406 (2003)]. This result suggests that information on middle-ear status improves the ability to predict hearing status. The YR factors were used to construct a middle-ear dysfunction test on 1027 normal-hearing ears in which DPOAE and TEOAE responses were either both present or both absent, the latter condition being viewed as indicative of middle-ear dysfunction. The middle-ear dysfunction test classified these ears with a nonparametric area (A) under the relative operating characteristic curve of A = 0.86, and classified normal-hearing ears that failed two-stage hearing-screening tests with areas A = 0.84 for DPOAE/ABR, and A = 0.81 for TEOAE/ABR tests. The middle-ear dysfunction test adequately generalized to a new sample population (A = 0.82).

Ear-canal acoustic admittance and reflectance effects in human neonates. I. Predictions of otoacoustic emission and auditory brainstem responses

This report describes the extent to which ear-canal acoustic admittance and energy reflectance (YR) in human neonates (1) predict otoacoustic emission (OAE) levels and auditory brainstem response (ABR) latencies, and (2) classify OAE and ABR responses as present or absent. Analyses are reported on a subset of ears in which hearing screening measurements were obtained previously [Norton et al., Ear. Hear. 21, 348-356 (2000a)]. Tests on 1405 ears included YR, distortion-product OAEs, transient-evoked OAEs, and ABR. Principal components analysis reduced the 33 YR variables to 5-7 factors. OAE levels decreased and ABR latencies increased with increasing high-frequency energy reflectance. Up to 28% of the variance in OAE levels and 12% of the variance in ABR wave-V latencies were explained by these factors. Thus, the YR response indirectly encodes information on inter-ear variations in forward and reverse middle-ear transmission. The YR factors classify OAEs with an area under the relative operating characteristic (ROC) curve as high as 0.79, suggesting that middle-ear dysfunction is partly responsible for the inability to record OAEs in some ears. The YR factors classified ABR responses less well, with ROC areas of 0.64 for predicting wave-V latency and 0.56 for predicting Fsp.

Newborn screening: rationale for a comprehensive, fully integrated public health system

Newborn screening has existed for approximately four decades. During that period of time, newborn screening has evolved conceptually from a laboratory test for a single disorder, phenylketonuria (PKU), to a multi-part public health system involving education, screening, diagnostic follow-up, treatment/management, and system evaluation. At a time when newborn screening is recognized as a model for predictive medicine, it also faces critical challenges that will determine its future credibility and viability. In order to understand these challenges, it is helpful to review briefly the history of newborn screening.

Identification of neonatal hearing impairment: ear-canal measurements of acoustic admittance and reflectance in neonates

OBJECTIVES

1) To describe broad bandwidth measurements of acoustic admittance (Y) and energy reflectance (R) in the ear canals of neonates. 2) To describe a means for evaluating when a YR response is valid. 3) To describe the relations between these YR measurements and age, gender, left/right ear, and selected risk factors.

DESIGN

YR responses were obtained at four test sites in well babies without risk indicators, well babies with at least one risk indicator, and graduates of neonatal intensive care units. YR responses were measured using a chirp stimulus at moderate levels over a frequency range from 250 to 8000 Hz. The system was calibrated based on measurements in a set of cylindrical tubes. The probe assembly was inserted in the ear canal of the neonate, and customized software was used for data acquisition.

RESULTS

YR responses were measured in over 4000 ears, and half of the responses were used in exploratory data analyses. The particular YR variables chosen for analysis were energy reflectance, equivalent volume and acoustic conductance. Based on the view that unduly large negative equivalent volumes at low frequencies were physically impossible, it was concluded that approximately 13% of the YR responses showed evidence of improper probe seal in the ear canal. To test how these outliers influenced the overall pattern of YR responses, analyses were conducted both on the full data set (N = 2081) and the data set excluding outliers (N = 1825). The YR responses averaged over frequency varied with conceptional age (conception to date of test), gender, left/right ear, and selected risk factors; in all cases, significant effects were observed more frequently in the data set excluding outliers. After excluding outliers and controlling for conceptional age effects, the dichotomous risk factors accounting for the greatest variance in the YR responses were, in rank order, cleft lip and palate, aminoglycoside therapy, low birth weight, history of ventilation, and low APGAR scores. In separate analyses, YR responses varied in the first few days after birth. An analysis showed that the use of a YR test criterion to assess the quality of probe seal may help control the false-positive rate in evoked otoacoustic emission testing.

CONCLUSIONS

This is the first report of wideband YR responses in neonates. Data were acquired in a few seconds, but the responses are highly sensitive to whether the probe is fully sealed in the ear canal. A real-time acoustic test of probe fit is proposed to better address the probe seal problem. The YR responses provide information on middle-ear status that varies over the neonatal age range and that is sensitive to the presence or absence of risk factors, ear, and gender differences. Thus, a YR test may have potential for use in neonatal screening tests for hearing loss.

Identification of neonatal hearing impairment: infants with hearing loss

OBJECTIVE

This article describes the audiologic findings and medical status of infants who were found to have hearing loss, detected as part of the Identification of Neonatal Hearing Impairment (INHI) project. In addition, the neonatal and maternal health variables for the group of infants who could not be tested with visual reinforcement audiometry (VRA) due to developmental and visual disability are presented.

DESIGN

The overall goal of the INHI project was to evaluate the test performance of auditory brain stem response and evoked otoacoustic emission (OAE) tests given in the newborn period. These tools were evaluated on the basis of the infants' hearing when tested behaviorally with VRA at 8 to 12 mo corrected age. The neonatal test results, VRA results, medical history information and a record of intercurrent events occurring between the neonatal period and the time of VRA were collated and reviewed. The purpose of this article is to review the characteristics of those infants who were found to have hearing loss.

RESULTS

Of 2995 infants who had VRA tests judged to be of good or fair reliability, 168 had a finding of hearing loss for at least one ear, an incidence of 5.6%. Sixty-six infants had bilateral losses, an incidence of 2%, and 22 infants had bilateral hearing losses in the moderate to profound range, an incidence 0.7%. The prevalence of middle ear problems was greater than 50% among these infants with hearing loss. From the larger group of 168 infants with hearing loss, a group of 56 infants (86 ears) was chosen as those with a low probability that the hearing loss was due to transient middle ear pathology and was more likely hearing loss of a permanent nature. These were the infants used for the analyses of neonatal test performance (Norton et al., 2000). In this selected group there were 30 infants with bilateral impairment of at least mild degree, which is an incidence of 1%. There were approximately equal numbers of ears in the mild, moderate, severe and profound range of hearing loss. Risk factors associated with hearing loss were reviewed for the total sample of infants tested with VRA and for those infants with hearing loss. A history of treatment with aminoglycosides was the risk factor most often reported in the entire sample; however, there was no difference in prevalence of this risk factor for the normal-hearing and hearing-impaired groups. The risk factor associated with the highest incidence of hearing loss was stigmata of syndromes associated with sensorineural hearing loss and other neurosensory disorders. Sixty-seven infants who returned for follow-up could not be tested with VRA due to severe developmental delay or visual disability. Many of these infants had medical histories indicating the sequelae of extreme prematurity and/or very low birthweight.

CONCLUSIONS

Most of the hearing losses found in this study were mild and, based on clinical history and tympanometry tests, many of the mild and some of the moderate impairments may have been acquired in early infancy due to middle ear effusion. In the group of infants used for determination of neonatal test performance there were approximately equal numbers of mild, moderate, severe and profound losses. Only a small percentage of infants with a conventional risk indicator for hearing loss actually had a hearing loss, and there were a significant number of infants with hearing loss who did not have a risk indicator. These findings support the need for an early identification program based on universal neonatal hearing screening rather than by targeted testing of those with risk indicators.

Identification of neonatal hearing impairment: distortion product otoacoustic emissions during the perinatal period

OBJECTIVES

1) To describe distortion product otoacoustic emission (DPOAE) levels, noise levels and signal to noise ratios (SNRs) for a wide range of frequencies and two stimulus levels in neonates and infants. 2) To describe the relations between these DPOAE measurements and age, test environment, baby state, and test time.

DESIGN

DPOAEs were measured in 2348 well babies without risk indicators, 353 well babies with at least one risk indicator, and 4478 graduates of neonatal intensive care units (NICUs). DPOAE and noise levels were measured at f2 frequencies of 1.0, 1.5, 2.0, 3.0, and 4.0 kHz, and for primary levels (L1/L2) of 65/50 dB SPL and 75/75 dB SPL. Measurement-based stopping rules were used such that a test did not terminate unless the response was at least 3 dB above the mean noise floor + 2 SDs (SNR) for at least four of five test frequencies. The test would terminate, however, if these criteria were not met after 360 sec. Baby state, test environment, and other test factors were captured at the time of each test.

RESULTS

DPOAE levels, noise levels and SNRs were similar for well babies without risk indicators, well babies with risk indicators, and NICU graduates. There was a tendency for larger responses at f2 frequencies of 1.5 and 2.0 Hz, compared with 3.0 and 4.0 kHz; however, the noise levels systematically decreased as frequency increased, resulting in the most favorable SNRs at 3.0 and 4.0 kHz. Response levels were least and noise levels highest for an f2 frequency of 1.0 kHz. In addition, test time to achieve automatic stopping criteria was greatest for 1.0 kHz. With the exception of "active/alert" and "crying" babies, baby state had little influence on DPOAE measurements. Additionally, test environment had little impact on these measurements, at least for the environments in which babies were tested in this study. However, the lowest SNRs were observed for infants who were tested in functioning isolettes. Finally, there were some subtle age affects on DPOAE levels, with the infants born most prematurely producing the smallest responses, regardless of age at the time of test.

CONCLUSIONS

DPOAE measurements in neonates and infants result in robust responses in the vast majority of ears for f2 frequencies of at least 2.0, 3.0 and 4.0 kHz. SNRs decrease as frequency decreases, making the measurements less reliable at 1.0 kHz. When considered along with test time, there may be little justification for including an f2 frequency at 1.0 kHz in newborn screening programs. It would appear that DPOAEs result in reliable measurements when tests are conducted in the environments in which babies typically are found. Finally, these data suggest that babies can be tested in those states of arousal that are most commonly encountered in the perinatal period.

Identification of neonatal hearing impairment: summary and recommendations

OBJECTIVES

This article summarizes the results of a multi-center study, "Identification of Neonatal Hearing Impairment," sponsored by the National Institutes of Health. The purpose of this study was to determine the performance characteristics of three measures of peripheral auditory system status, transient evoked otoacoustic emissions (TEOAEs), distortion product otoacoustic emissions (DPOAEs), and auditory brain stem responses (ABR), applied in the neonatal period in predicting hearing status at 8 to 12 mo corrected age.

DESIGN

The design and implementation of this study are described in the first two articles in this series. Seven institutions participated in this study; 7179 infants were evaluated. Graduates of the neonatal intensive care unit and well babies with one or more risk factors for hearing loss were targeted for follow-up testing using visual reinforcement audiometry (VRA) at 8 to 12 mo corrected age. Neonatal test performance was evaluated using the VRA data as the "gold standard."

RESULTS

The major results of the study are described in the nine articles preceding this summary article. TEOAEs in response to an 80 dB pSPL click, DPOAEs in response to L1 = 65 and L2 = 50 dB SPL and ABR in response to a 30 dB nHL click performed well as predictors of permanent hearing loss of 30 dB or greater at 8 to 12 mo corrected age. All measures were robust with respect to infant state, test environment and infant medical status. No test performed perfectly.

CONCLUSIONS

Based on the data from this study, the 1993 National Institutes of Health Consensus Conference-recommended protocol-an OAE test followed by an ABR test for those infants failing the OAE test-would result in low referral rate (96 to 98%). TEOAEs for 80 dB pSPL, ABR for 30 dB nHL and DPOAEs for L1 = 65 dB SPL and L2 = 50 dB SPL perform well in predicting hearing status based on the area under the relative operating characteristic curve. Accuracy for the OAE measurements are best when the speech awareness threshold or the pure-tone average for 2.0 kHz and 4 kHz are used as the gold standard. ABR accuracy varies little as a function of the frequencies included in the gold standard. In addition, 96% of those infants returning for VRA at 8 to 12 mo corrected age were able to provide reliable ear-specific behavioral thresholds using insert earphones and a rigorous psychophysical VRA protocol.

Identification of neonatal hearing impairment: a multicenter investigation

OBJECTIVES

This article describes the design of a multicenter study sponsored by the National Institutes of Health. The purpose of this study was to determine the accuracy of three measures of peripheral auditory system status (transient evoked otoacoustic emissions, distortion product otoacoustic emissions, and auditory brain stem responses) applied in the perinatal period for predicting behavioral hearing status at 8 to 12 mo corrected age. The influences of the infant's medical status, the test environment, and test and response parameters on test performance were examined.

DESIGN

Seven institutions participated in this study. There were 7179 infants evaluated in the perinatal period. All graduates of the neonatal intensive care unit (4478) and well babies with one or more risk factor for hearing loss (353) were targeted for follow-up testing using visual reinforcement audiometry (VRA) at 8 to 12 mo corrected age. Well babies without any risk indicators (N = 2348) were not targeted for follow-up VRA testing. However, 80 of these well babies did not pass the screening protocol and thus were targeted for follow-up VRA testing as well. Perinatal test performance was evaluated using the VRA data as the "gold standard."

RESULTS

The results of this study are described in a series of 11 articles following this introductory article.

CONCLUSIONS

The evaluation of newborn hearing tests required a longitudinal study in which newborn test results were compared with a gold standard based on behavioral audiometric assessment. Such an evaluation was possible because all newborns, passes as well as refers, were followed up long enough to permit reliable behavioral measurements. In addition, prenatal, perinatal, and maternal history information, test environment, and test parameter information were collected to provide data that led to a complete description of factors affecting test outcomes. All of these data were obtained in a sample of sufficient ethnic, medical, and geographic diversity in efforts to increase the generalizability of the results. Finally, the data were combined in a relational data base to examine the factors that influence test performance. Specific information related to these issues is presented in the articles that follow.