Limited Audiological Assessment Results in Children with Otitis Media with Effusion

Abstract

Objective.

Clinical practice guidelines predicate the need for evaluation of hearing in children with otitis media with effusion (OME). The objective of this work was to characterize the completeness of hearing assessment results in children with OME.

Design.

40 participants with OME completed two full audiological assessments, one in a clinical setting and a second in a research setting. An additional 14 participants without OME completed a single audiological assessment in the research setting as a control group. The success of various behavioral and objective audiometric tests completed in each setting was quantified and evaluated.

Results.

Findings indicate that ear-specific behavioral audiometric information is substantially limited in children with OME, particularly in clinical settings. In contrast, objective testing including tympanometry and otoacoustic emission testing was largely successful.

Conclusions.

Ear-specific behavioral audiometric information is limited in children with OME and, consequently, consideration of these data for use as part of clinical decision making is also limited. Objective tests were more successful but are not direct measures of hearing.

INTRODUCTION

Otitis media with effusion (OME) is one of the most common childhood illnesses in the United States, affecting more than 90% of children by age 10 (Paradise et al., 1997). It is also the most common cause of acquired hearing loss in childhood (Coleman & Cervin, 2019). Children with OME generally present in clinic relatively asymptomatic with subtle otoscopy findings (e.g., no infection, perforation). OME is often accompanied by a fluctuating and transient conductive hearing loss (CHL) with thresholds that can range from 0 to 55 dB HL (Al-Salim et al., 2021; Gravel et al., 2006), and can last for variable lengths of time (Marchant et al., 1984). The CHL that results from OME causes auditory deprivation during critical developmental periods in childhood that can lead to long-term deficits in speech, language, and auditory processing (Whitton & Polley, 2011). Management of OME includes watchful waiting and surgical placement of tympanostomy tubes. While there are risks associated with surgery in young children, waiting for OME to resolve spontaneously may also carry risks, including those associated with the auditory deprivation that can occur when CHL is present.

OME and tympanostomy tube clinical practice guidelines recommend that a behavioral hearing assessment be completed in all children where OME has persisted for at least 3 months and that hearing status should be considered as part of the decision as to whether to proceed with surgical placement of tubes or watchful waiting (Rosenfeld et al., 2016, 2022). Unfortunately, despite these guideline recommendations, failure to obtain audiometric testing pre-operatively prior to tube placement is common (Gisselsson-Solen, 2018; Sajisevi et al., 2017). Sajisevi and colleagues (2017) found that only 13.5% of children with OME persisting 3 months or greater had a hearing assessment and Gisselsson-Solen and colleagues (2018) found that only 42% of children undergoing tube placement had a documented pre-operative hearing test.

There are known challenges and clinical constraints (e.g., child state, availability of a testing assistant, pediatric expertise, physical setup of testing space, etc.) associated with testing the hearing of young children (Widen et al., 2000). Determination of hearing status in children with OME generally needs to occur in a single clinical visit, both due to the fluctuating nature of the disease as well as the general course of clinical care for these patients. This contrasts with children with stable and permanent hearing loss, where documentation of hearing can often be achieved by combining information across several clinical assessment visits. Additionally, while several studies have documented reasonable success of audiometric testing in the sound field for children with OME (e.g., Sidell et al., 2014), and there is no question that sound field information is of value in the clinical management of these children, ear-specific hearing information is also important. OME status can vary markedly between the two ears in the same patient, and it has been shown that there is no association between the pathological findings in an ear with OME and its contralateral ear (Raghavan et al., 2023). Sound field testing would only provide information on the better hearing ear and could, for example, erroneously suggest that a child has normal hearing despite having a moderate conductive loss in one ear due to differences in OME status between the two ears. Additionally, asymmetries between the two ears may put children at even higher risk for known sequalae of OME, like binaural processing deficits, given how this may alter binaural level and timing cues (Thornton et al., 2021), further emphasizing the value of ear-specific audiometric information in this population.

While assessment of hearing is important in children with OME, it is clear that current clinical practice is not consistent with guideline recommendations. The goal of the current work is to characterize the completeness of hearing assessment results obtained in a single visit using standard audiological tests in children with OME. Further, we aim to explore the success of audiological assessment in both a clinical setting and an idealized research setting to help determine how clinical constraints may be contributing to limited assessment results. This will allow us to identify areas where improvements are needed in the assessment and management of hearing loss in children with OME.

METHODS

Population

40 participants (19 males) diagnosed with OME and who were scheduled for tympanostomy tube placement were recruited from otolaryngology clinics at Boys Town National Research Hospital (BTNRH) in Omaha, NE. Data from 14 participants (5 males) with normal hearing matched for age were also included. Diagnosis of OME was made by a BTNRH otolaryngologist based upon observation of middle ear effusion in the absence of signs of acute infection via standard or pneumatic otoscopy. Participants with OME ranged in age from 8 months to 4 years, 5 months (mean age = 22 months) and matched control participants ranged in age from 8 months to 4 years, 6 months (mean age = 24 months). Participants had no reported developmental delays or diagnoses that might affect performance on our measures and were only included if they were under the age of 5, as we would expect audiological testing to be largely successful in typically developing children aged 5 and older. Audiological testing results from both ears of each participant were analyzed, resulting in 108 ears (80 with OME, 28 control) being included in our analyses, focusing primarily on ear-specific audiological test results. Informed consent was obtained from all participants and caregivers for testing procedures approved by the BTNRH Institutional Review Board.

Procedure

Participants with OME received a full audiological assessment at two separate visits, one in the clinic and one in the research setting, while matched control participants received a single audiological assessment in the research setting.

The clinical assessment occurred on the same day as the otolaryngology medical evaluation when OME was diagnosed and took place in one of three BTNRH clinics by one of seven audiologists with pediatric expertise. The clinical tests attempted/completed were chosen at the discretion of the clinical audiologist and were completed using different equipment; however, test parameters were consistent across clinics. When available (and as needed), a trained test assistant assisted with the evaluation. Clinic visits were scheduled as 30-minute appointments. Participants were consented and enrolled into the study after their otolaryngologist diagnosed them with OME, which occurred after their clinical audiology appointment, so clinical assessment data were pulled from the medical records retrospectively.

Research assessments took place in a BTNRH laboratory. In participants with OME, this was completed within 72 hours prior to tube placement surgery. Research assessments were completed by one of two research audiologists with pediatric expertise, assisted by a trained test assistant. In contrast to the clinical visits, research visits were longer (up to 2 hours), a test assistant participated in all visits (while a test assistant only participated in 53% of the clinical visits), and all assessments were attempted in all but 4 participants (2 children with OME on whom audiometric testing was not attempted due to scheduling constraints, and 2 matched control children on whom testing was not completed in the second ear due to child compliance [n=1] or abnormal middle ear status observed on tympanometry [n=1]). For children with OME, there was an average of 18 days separating the two visits (with a range of 2–57 days).

Tympanometry

Tympanometry is a measure of middle ear pressure and mobility of the tympanic membrane. Abnormal tympanometry is often observed in the presence of OME. Tympanometry was measured using a 226 Hz probe tone and a positive to negative pressure sweep from +200 to −300 daPa. In the research visit only, wideband tympanometry was also measured using an Interacoustics Titan tympanometer and was recorded in response to a wideband click stimulus (0.226–8 kHz) with downswept pressure at a level of 96 dB peak-to-peak equivalent sound pressure level (≈ 65 dB nHL). Wideband tympanometry is similar to standard tympanometry but measures middle ear mobility across a wider range of frequencies, allowing for a more comprehensive assessment of middle ear mechanics (Merchant et al., 2022).

Distortion Product Otoacoustic Emissions (DPOAEs)

Distortion product otoacoustic emissions (DPOAEs) are a measure of outer hair cell function. While often used as an assessment of inner ear function, DPOAEs can be influenced by pathologies of the middle ear as the eliciting stimulus has to travel through the middle ear and the resulting emissions must travel backward from the inner ear through the middle ear to be recorded in the ear canal. DPOAEs are often used in clinic to get a general pass/fail sense of how a child may be hearing, though they are not a direct measure of hearing. DPOAEs were measured in a quiet room at ambient pressure at L1/L2 levels of 65/55 dB SPL from 1–8 kHz during the clinical visit and 1–10 kHz during the research visit. During the clinical visit, DPOAEs were generally only measured when ear-specific behavioral audiometric thresholds could not be obtained, while they were measured for all participants during the research visit.

Behavioral Audiometry

Depending on the child’s age, visual reinforcement audiometry (VRA), conditioned play audiometry (CPA), or conventional audiometry was used to attempt to obtain air conduction audiometric thresholds at octave frequencies 0.25–8 kHz and bone conduction audiometric thresholds at octave frequencies 0.25–4 kHz, if indicated. Procedures were based on published guidelines from the American Academy of Audiology (AAA, 2020). Bone conduction thresholds were measured with masking used when an air-bone gap > 10 dB was present at a given frequency. When a young child was observed to be fatiguing to the task, air conduction and unspecified bone conduction thresholds at octave frequencies from 0.5–4 kHz were prioritized. Lateralization was not required to indicate a reliable response to VRA. Data collection in the research setting focused on ear-specific data collection, so sound field outcomes were not quantified. All behavioral audiometric assessments were completed in a double walled sound booth.

RESULTS

Tympanometry

Table 1 summarizes the completed testing for both tympanometry and DPOAEs. For children with OME, standard 226 Hz tympanometry was successfully completed on 98% of ears (78/80) during the clinical visit and 100% of ears (80/80) during the research visit. For matched control participants, who only participated in the research visit, standard tympanometry was completed on 100% of ears (28/28). Wideband tympanometry (not shown in the table) was completed at the research visit only and was successfully completed on 100% of ears with OME (80/80) and 93% of matched control ears (26/28).

Table 1.

Completed tympanometry and distortion product otoacoustic emission (DPOAE) testing by ear across the clinic and research visits. Measurement of DPOAEs was considered complete when at least 50% of DPOAE frequencies attempted were obtained. Only ears on which a test was attempted are included in these calculations. n indicates number of ears.

	Tympanometry		DPOAEs
Visit	n	Completed	n	Completed
Clinic	80	98%	52	90%
Research	80	100%	79	98%
Control	28	100%	26	89%

Distortion Product Otoacoustic Emissions (DPOAEs)

Measurement of DPOAEs was generally successful during both the clinical and research visits, where DPOAEs were successfully measured in 90% of ears (47/52) with OME in the clinic and 99% of ears (78/79) during the research visit. DPOAEs were also successful in 89% of matched control ears (24/27) during their research visit. Measurement of DPOAEs was considered complete when at least 50% of DPOAE frequencies attempted were measured, regardless of whether the DPOAE was present or absent. Only ears on which a test was attempted are included in these calculations (i.e., for DPOAEs in the clinic, tests were completed on 90% of the 52 ears DPOAEs were attempted on). The lower number of ears on which DPOAEs were attempted in the clinic (52 vs. 79 ears, Table 1) can be attributed to the clinical protocol where DPOAEs are typically not attempted if any ear-specific behavioral audiometric information is successfully obtained.

Behavioral Audiometry

Behavioral audiometry was attempted on most participants during both the clinical (88%; 70/80) and the research visit (95%; 76/80) in ears with OME and in matched control ears (96%; 27/28). However, successful measurement of ear-specific behavioral audiometric thresholds was considerably more limited compared to tympanometry and DPOAEs in both settings. A detailed presentation of the limited results obtained during behavioral audiometric threshold testing is depicted in Table 2. Of particular note, in children with OME, while at least a single sound field sound awareness threshold was obtained in 80% of children during the clinical visit (not shown), zero ear-specific audiometric thresholds could be measured on 74% of ears (52/70) in the clinic and on 53% of ears (40/76) at the research visit on which ear-specific behavioral audiometry was attempted. No ear-specific audiometric thresholds could be measured on 70% of the matched control ears (19/27). Of the 26% of ears where some audiometric information was obtained in clinic (for ears where behavioral assessment was attempted), a four-frequency pure-tone average (PTA) (i.e., mean air conduction thresholds from 0.5–4 kHz) was obtained in only 17% of ears (12/70). While outcomes were better during the research visit, a four-frequency PTA was obtained in only 36% of ears with OME (27/76) and 30% (8/27) of control ears. The success of bone conduction testing was more limited than air conduction testing and was thus only broken down by whether 1 or more unmasked or masked bone conduction thresholds were measured, with almost no ear specific masked bone conduction successfully measured in the clinic (3%) and limited masked bone conduction (21%) obtained during the research visit.

Table 2.

Comparisons of behavioral audiometric testing success for the entire study population (top section) as well as broken down by test type: visual reinforcement audiometry (VRA) versus conditioned play audiometry (CPA) (bottom two sections). Testing success is broken down by transducer type (air or bone) and number of thresholds obtained. Only ears on which a test was attempted are included in these calculations. n indicates number of ears. Note that one child (2 ears) tested in the clinic and 1 child (2 ears) tested in the lab (control group) were tested via conventional audiometric techniques and is therefore represented in the top section but not in either bottom section as VRA and CPA were not used. Bone conduction outcomes are not presented for control ears as they were rarely, if ever, tested, due to thresholds within-normal-limits.

	Ear-Specific Audiometry: All Methods
Visit	n	Age (months)	Attempted, 0 Thresholds	≥ 1 Air Threshold	≥ 4 Air Thresholds	≥ 1 Unmasked Bone Threshold	≥ 1 Masked Bone Threshold
Clinic	70	23	74%	26%	17%	13%	3%
Research	76	23	53%	47%	36%	4%	21%
Control	27	24	70%	30%	30%
	Audiometry: VRA
Visit	n	Age (months)	Attempted, 0 Thresholds	≥ 1 Air Threshold	≥ 4 Air Thresholds	≥ 1 Unmasked Bone Threshold	≥ 1 Masked Bone Threshold
Clinic	48	15	100%	0%	0%	4%	0%
Research	56	15	70%	30%	14%	4%	13%
Control	15	14	100%	0%	0%
	Audiometry: CPA
Visit	n	Age (months)	Attempted, 0 Thresholds	≥ 1 Air Threshold	≥ 4 Air Thresholds	≥ 1 Unmasked Bone Threshold	≥ 1 Masked Bone Threshold
Clinic	20	39	20%	80%	50%	10%	35%
Research	20	41	5%	95%	95%	5%	45%
Control	10	41	20%	80%	80%

Table 2 further breaks down the success of audiometric testing by testing technique, comparing outcomes for VRA, generally used in younger children under the age of 2 ½ years, and CPA generally used in children older than 2 ½ years. Findings demonstrate that no ear-specific audiometric information was obtained when VRA was used during the clinic visit. While some ear-specific information was obtained using VRA during the research visit, it was limited, with nothing obtained on 70% of ears with OME, at least one threshold obtained on 30% of ears with OME, and a four-frequency PTA obtained on only 14% of ears with OME. No thresholds were obtained using VRA in the matched control ears. CPA achieved greater proportions of success during both the clinical and research visits, with at least one threshold obtained in 80% of ears with OME in the clinic and 95% of ears during the research visit, and 80% of matched control ears. Further, a four-frequency PTA was obtained in all 95% of ears with OME during the research visits, 80% of matched control ears, and in 50% of ears with OME during the clinic visit.

Effect of Age

In the clinic, all ears on which no thresholds could be measured were from children 48 months and younger, while in the research lab the children on whom thresholds could not be measured were all younger than 34 months. With the exception of one child in the clinic, some ear-specific audiometric thresholds were obtained on all children older than 36 months. There were no significant differences (based on t-tests) in age between the groups of any of the comparisons made.

Effect of Support of a Test Assistant

A test assistant was available and utilized for all research study visits, while a test assistant was utilized as needed and as available for clinical visits. A test assistant was present for 67% of the clinical visits where VRA was utilized, while a single audiologist completed testing in the other 33% of cases. Zero ear specific thresholds were obtained whether the test assistant was present or not. In clinical visits where CPA was utilized, a test assistant was present for 30% of visits and a four-frequency PTA was obtained in all cases. In contrast, a single tester was present for the other 70% of CPA visits, and a four-frequency PTA was obtained in only a third of those cases (with 1–3 thresholds obtained in half, and nothing obtained in the remaining ears). This increased success in CPA with a second tester present is also supported by the fact that a four-frequency PTA was obtained in 80% of matched control ears tested via CPA with a testing assistant in the research setting versus 50% in the clinic. The average age of those tested via CPA with two testers was 36.6 months in contrast to 40.5 months for those tested with a single tester, suggesting a lower success rate was not due to younger age, though this age difference was not statistically significantly different.

Overall Success of Audiometric Testing

Given the limited ear-specific behavioral audiometric testing results coupled to guideline recommendations, we wanted to summarize what types of audiometric information were available in this population for clinical decision making based on data available during the clinic visit. This summary is shown in Table 3. Any sort of ear-specific behavioral audiometric information was available in only 22% of ears, with a four-frequency PTA available in only 15% of ears. In 75% of ears, the only ear-specific information that was available from the clinical audiology visit was the tympanogram and/or DPOAEs, which are not direct measures of hearing.

Table 3.

Summary of ear-specific audiologic testing information available after the clinical visit for clinical decision making for all ears (n=80), including tympanometry, distortion product otoacoustic emissions (DPOAEs), some audiometric information (1–3 thresholds), or a four-frequency pure-tone average (PTA).

Ear-Specific Information Available for Clinical Decision Making
No Audiological Information	3%
Tympanometry Only	25%
Tympanometry & DPOAEs Only	50%
Tympanometry & 1–3 Audiometric Thresholds	7%
Tympanometry & PTA (4 Audiometric Thresholds)	15%

DISCUSSION

Hearing assessment in children with OME is important and indicated based on clinical practice guidelines (Rosenfeld et al., 2016, 2022), particularly for those undergoing tympanostomy tube placement surgery. However, the results of this study demonstrate that audiometric assessment in children with OME is limited. Findings indicate that any ear-specific behavioral audiometric information is obtained in only 22% of ears in children with OME under the age of 5 scheduled for tympanostomy tube placement, and a four-frequency PTA available in only 15% of ears. Audiological assessment results obtained in an idealized research setting were more complete than the clinical results. This may be due to greater access to testing resources and supports such as longer visit times and the consistent availability of a testing assistant but could also be due to learning effects given the research visit occurred second, or a combination of these factors. Data from the age-matched control group demonstrated similar success rates to the clinic in some respects (percentages where no thresholds were able to be obtained), while higher success rates in others with more similar performance to participants with OME in the research setting (number of thresholds obtained). However, even in the idealized research setting, many children were still not able to complete the full audiological testing battery, with zero ear-specific thresholds obtained in 53% of ears and a four-frequency PTA obtained in only 36% of ears.

Poor audiometric test success makes it challenging to factor hearing status into clinical decision making as recommended by the published guidelines (Rosenfeld et al., 2016, 2022). Although not specifically investigated here, these results would imply a lack of adherence to these guidelines (as has been demonstrated previously, i.e., Gisselsson-Solen, 2018; Sajisevi et al., 2017). This is an important consideration given the high incidence and relevance of this condition and management decisions which often include surgical tympanostomy tube placement.

Measurement of ear-specific behavioral audiometric thresholds was particularly limited in children under the age of 3 years, the age range at which OME is most common (with a peak incidence around age 2, Atkinson et al., 2015). Degree of hearing loss, typically determined by the PTA obtained from the audiogram, is what is referenced in the clinical practice guidelines for use in clinical decision making in the management of children with OME (Rosenfeld et al., 2022). While the OME guidelines present VRA as a testing method that provides results that reliably indicate hearing status, the studies cited in the guidelines to support VRA as a reliable method often excluded children with OME, allowed for testing over multiple sessions, and/or used a screening method that did not require threshold searches that would extend test time (Widen et al., 2000). These are all procedures that are not consistent with the hearing assessment and standard course of clinical care of children with OME. As our results demonstrate, the success of obtaining ear-specific behavioral audiometric information in the clinic in a single visit in children with OME using VRA is limited at best.

Tympanometry and DPOAE results were much easier to obtain than audiometry in children with OME, with success in 85–100% of ears. However, while ear specific, neither of these are direct measures of hearing. Standard tympanometry is only weakly associated with presence and degree of CHL (Al-Salim et al., 2021; Margolis et al., 1994). DPOAEs may provide a general idea about hearing in children with normal middle-ear status, however, they are not likely to provide a reliable estimation of hearing status in children with middle-ear effusion because DPOAEs rely on both forward and reverse transmission. That is, in ears with OME, the stimulus that evokes the DPOAE is reduced on the way in and then the DPOAE itself is reduced on the way out because of the middle-ear dysfunction. Hearing only relies on forward transmission, so the reduction in the DPOAE during reverse transmission would not be reflective of how a child is hearing. Indeed, previous work in children with OME has shown that tympanograms can be abnormal and DPOAEs can be reduced and/or absent even in ears with audiometric thresholds within the normal range (Al-Salim et al., 2021).

Overall, these findings demonstrate that, while most children (80%) did have at least a sound awareness threshold obtained in the sound field, ear-specific behavioral audiometric information is substantially limited in children with OME. Ear-specific hearing status is therefore likely not being factored into clinical decision making regarding OME management on a regular basis. It should also be noted the clinical results reported here are from an urban specialty hospital setting with highly trained clinicians who specialize in testing pediatric populations and have significant clinical resources at their disposal. Outcomes may be poorer in settings without these advantages and resources.

Tympanostomy tube guidelines do acknowledge that audiologic testing isn’t always obtainable and, in such cases, recommend use of a screening questionnaire to gauge hearing difficulty (Rosenfeld et al., 2022). However, the screening questions ask parents to report behaviors that are not necessarily applicable to very young children, where OME is common. As such there remains a lack of options available to clinicians to ascertain hearing status when audiometric testing cannot be completed on a child with OME.

While sound field information is certainly of clinical value in the management of this population, particularly when combined with objective measures (Baldwin et al., 2010), and behavioral hearing thresholds should always be attempted and remain the gold standard, these results underscore the need for an alternative means to estimate ear-specific hearing levels of children with OME when behavioral thresholds cannot be obtained, ideally one that is objective, quick, and non-invasive (Rosenfeld et al., 2016). A recent study demonstrated that wideband tympanometry, coupled to an automated computational model, can be used to estimate CHL (the four-frequency PTA) within a clinically-meaningful margin of error (Merchant & Neely, 2022). Wideband tympanometry is a quick, easy, and non-invasive test and is well tolerated similar to standard tympanometry, as demonstrated by 100% test success of both wideband tympanometry and standard tympanometry during the research visit in this study. As such, wideband tympanometry would be a useful addition to the audiological test battery in children with OME, particularly in conjunction with other standard objective testing methods with high success rates shown to be diagnostically beneficial in this population (e.g., standard tympanometry and DPOAEs, Blankenship et al., 2017).

CONCLUSIONS

This work demonstrates poor success of ear-specific behavioral audiometric testing in a single visit in children with OME. This limitation makes it difficult to use hearing status as part of OME clinical decision making, despite guideline recommendations. Objective testing (tympanometry and DPOAEs) is more successful but does not provide a clear indication of how a child is hearing. This work underscores the need for an alternative method to estimate hearing status in children with OME.