NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Llewellyn A, Norman G, Harden M, et al. Interventions for adult Eustachian tube dysfunction: a systematic review. Southampton (UK): NIHR Journals Library; 2014 Jul. (Health Technology Assessment, No. 18.46.)

Cover of Interventions for adult Eustachian tube dysfunction: a systematic review

Interventions for adult Eustachian tube dysfunction: a systematic review.

Show details

Chapter 3Results

Study selection

A total of 3022 records were identified from the searches of databases and other sources including trial registers. There were 196 records identified as potentially relevant and ordered as full papers. Reference checking identified a further 17 records, bringing the total number of ordered papers to 213 (Figure 1). These included seven studies which were reported in languages other than English. Consultation with readers of these languages indicated that they were, at best, small uncontrolled studies. A list of papers excluded is provided in Appendix 3 .

FIGURE 1. Flow chart of study selection.


Flow chart of study selection.

Twenty-three records representing 19 studies met the inclusion criteria (see Figure 1 ). This included seven records related to six comparative studies (three RCTs, two non-RCTs and one retrospective controlled before-and-after study)24,25,40,41,5153 and 13 case series (reported across 16 records).18,2023,34,5463

Fourteen of the studies were of adults only: all 13 of the case series and one comparative study. Four comparative studies were identified in the searches that included mixed populations of adults and children,24,25,41,52 and one did not specify whether or not the whole population were adults.40 In each case, it appeared from the information in the publication that a majority of the patients were adults. These were included following a protocol amendment (see Study selection and Chapter 2, Inclusion and exclusion criteria). Attempts to contact the study authors for clarification of whether or not studies were conducted in adults and/or to obtain separate data for adults in a mixed adult/paediatric population were unsuccessful in all except the one surgical comparative study and two of the case series, where author contact established that all patients were adults.18,51,62

There were a further 19 records representing 15 uncontrolled observational studies (see Appendix 3 ) that met all of the inclusion criteria except that the study was of a mixed adult/child population or a mixed population of patients with ETD and other middle ear problems, but data on adults with ETD were not reported separately.45,57,6480 We attempted to contact the authors but were unsuccessful in obtaining separate adult data and/or data on patients identified as having ETD. These studies were, therefore, excluded; a list and further details are provided in Appendix 3 .

Two ongoing RCTs and one ongoing uncontrolled trial were identified; no outcome data were available for these.43,81,82 Details of the study characteristics of these trials are provided in Appendix 4 .

In Patient characteristics we provide an overview of the characteristics of the included studies. Further details about the included studies, their results and quality assessment of the included studies are reported in Chapter 4 (see Quality of included studies), subgrouped by type of intervention (non-surgical and surgical interventions).

Overview of study characteristics

The included studies were very diverse. The interventions evaluated in the primary studies, the range of outcomes assessed and the measures used varied considerably across the studies. Variation in diagnostic and other inclusion criteria also contributed to substantial heterogeneity in the patient populations (see Patient characteristics).

The interventions evaluated by the comparative studies were, with one exception, pharmacological treatments or mechanical devices (Table 2). Only three RCTs were identified and these were of non-surgical interventions: there were no RCTs of surgical interventions (see Table 2 ). Two of the RCTs assessed different pharmacological interventions25,52 and one evaluated a mechanical device.40 This was a small study and was reported only as a letter to an editor.40 All of these studies used a control of either placebo or no active treatment. A single retrospective controlled before-and-after study evaluating a surgical intervention was identified.51 All other studies assessing surgical treatments were case series.



Overview of interventions assessed in the review

None of the studies was conducted in the UK. Studies of pharmacological agents were carried out in the USA, Denmark and Sweden, studies of mechanical devices in the USA and Italy. Where reported, studies of surgical interventions were conducted in the USA, Germany, Denmark, Finland, Russia and Mexico (see Table 2 ). Publication dates ranged from 1976 to 2011 for studies of pharmacological interventions, were 1999 and 2008 for mechanical devices and ranged from 2005 to 2012 for studies of surgical interventions.

Patient characteristics

Four of the studies of non-surgical interventions reported data on a mixed adult/paediatric population. In the case of Gluth et al.,52 the proportion of adults in the trial (63%) was reported and some separate data on the primary outcome were reported for the paediatric subgroup (aged 6 to 17 years), allowing results for adult patients to be calculated.52 In the three other studies, the age ranges indicated that all patients were adolescents or adults, with age ranges of 14–66 years,24 12–75 years25 and 16–76 years.41 In the fourth study, there was some uncertainty as to whether or not all patients were adults, but the mean age was 39.2 years, suggesting that, unless the range was very wide, the probability was that a majority of individuals were adults.40

The mean ages of patients undergoing a surgical intervention ranged from 40 to 63 years and age ranges extended from 18 to 90 years where reported. The single controlled before-and-after study evaluating surgery had an age range of 21 to 56 years (mean age not reported).51 All of the included studies which reported information on patient sex had approximately equal numbers of male and female patients.

There was notable variation in the inclusion criteria used by individual studies, and, consequently, in the characteristics of the patient populations. In particular, there was variation in the length of time patients were required to have had ETD symptoms, the severity of the symptoms and the previous treatments attempted. While two studies of surgical intervention required that patients had a history of 5 years of ETD symptoms,20,23 other studies required merely that patients experienced ongoing ETD symptoms following an aeroplane flight,41,54 while others specified that the symptoms be ‘chronic’18,56,61 or that participants were required to have a long history of symptoms.21 Although the symptoms considered to be associated with ETD were specified as part of the inclusion criteria in many of the studies, ETD itself was rarely defined. None of the studies of pharmacological agents or mechanical devices specified a minimum duration of symptoms, although one required that symptoms be ‘persistent’.40 Where symptoms were assessed and reported using a scale to quantify severity, it appeared that patients were in considerable discomfort at baseline.40,55

Notably, three of the surgical studies included only patients with a diagnosis of otitis media with effusion with an aetiology indicating that ETD was responsible for the condition.22,23,51 In these and four other studies, including two evaluations of pharmacological treatments, patients with a perforated tympanic membrane or current pressure equalisation tubes were included either as an identified subgroup or together with patients with an intact eardrum.20,21,24,25

Previous treatment histories also varied widely both among and in some cases within studies. Poor reporting was an issue, particularly in the studies of pharmacological interventions. One study of a mechanical device reported previous therapies of nasal decongestants, non-steroidal anti-inflammatory drugs (NSAIDs), antihistamines and antibiotics,40 while the other did not report treatment history.41 Of the studies of pharmacological interventions, two did not report treatment histories25,52 while the third noted only one previous pressure equalisation tube.24

Where treatment histories were documented for surgical patients, they indicated considerable variation, with patients in some studies having failed multiple therapies including both medical and surgical interventions, while patients in others had no surgical history or no prior treatment history. In four studies of surgical intervention, failure of conservative or medical management was a criterion for enrolment in the study.20,23,54,61 In four further surgical studies, repeated courses of medical therapy were documented for all patients.56,60,63,83 Three studies required prior ventilation tubes.22,23,61 One surgical study reported no prior treatments.58 Within some studies, there was variation in the reported treatment histories, with three studies reporting prior surgeries in some but not all of the patients.56,63,83 Three studies of surgical intervention did not report treatment histories18,21,62 and one noted only that most patients had had previous ETD treatment (Professor Maria Boboshko, St. Petersburg Pavlov State Medical University, 2012, personal communication).51

There were also differences in the approach to comorbidities such as gastric reflux and rhinosinusitis. In five studies, the incidences of comorbidities were not reported.18,55,58,62,63 Other studies documented substantial minorities of patients with these and related conditions,22,23,56,60,61 while some used their presence as an exclusion criterion.20,21,54 One study of tuboplasty included only patients with a diagnosis of rhinosinusitis.56


As can be seen from Table 2 , the interventions represented in the review covered a range of pharmacological, surgical and mechanical devices.

Three studies, including two RCTs and one non-RCT, assessed pharmacological interventions. These were a nasal steroid spray,52 a topical administration of decongestant,25 and oral administration of a combination of antihistamine and ephedrine.24

Two studies (one small RCT and one non-RCT) assessed pressure equalisation devices.40,41

Surgical interventions assessed were various techniques for Eustachian tuboplasty, balloon dilatation, thermal or laser-assisted myringotomy, and laser coagulation of the Eustachian tube pharyngeal opening. One case series assessed dexamethasone administered via micro-wick following myringotomy.61 Seven case series assessed forms of tuboplasty;20,21,23,54,56,62,63 three case series assessed balloon dilatation;18,22,55 and two assessed myringotomy.58,60 For other interventions, a single retrospective controlled before-and-after study51 or case series were available.61

Seven of the surgical studies reported that concomitant treatments were administered in addition to the intervention that was the focus of the assessment. These included a range of additional surgical interventions and/or pharmacological treatments.2023,55,56,61

The interventions assessed by the included studies did not include several methods which may be used in primary care, including active observation, supportive care, antibiotics, leukotriene receptor antagonists (LTRAs) or nasal douching. There were also no included studies assessing the most common surgical intervention for ETD: insertion of pressure equalisation tubes. For many of the eligible interventions, there was, therefore, no evidence which met the review inclusion criteria. For most interventions where evidence was available, there was no randomised and often no controlled evidence. The number of studies identified for each of the eligible interventions is reported in Table 3 , indicating where gaps in the evidence remain.



Overview of gaps in the evidence identified


Eleven studies, comprising one of the three pharmacological studies,52 one of the two studies of mechanical devices40 and nine studies of 13 studies of surgical interventions,18,20,51,55,56,58,6163 assessed the primary outcome. Of the studies which did assess symptoms, only four reported using a systematic method to quantify improvement or change in symptomatology. One used scales specific to the condition for which at least some information is available on reliability and validity (the ETDQ-7 and SNOT-22).55 One used a symptom questionnaire which appeared to be a modified version of the ETDQ-7, and two used a visual analogue scale (VAS).20,40,52

All other studies which reported symptom change restricted reporting to the number of patients reporting improvement in either global ETD symptoms,55,56,63 specific symptoms such as tinnitus or aural fullness51,58,61 or non-specific ‘improvement’18 or resolution of symptoms.62

A considerable range of outcomes specified as relevant in the review protocol were assessed (Table 4), although none of the studies reported quality of life, and early tube extrusion was not a relevant outcome as no studies of pressure equalisation tubes were included.



Reporting of key outcomes

Hearing was assessed in seven of the studies of surgical interventions20,23,51,56,60,61,63 and in one of the mechanical device studies.41 Various audiometry measures were reported, of which the most common was the average pure-tone threshold.23,56,61,63 Air–bone gap was also commonly reported20,41,51,61 and one study also reported speech discrimination score,61 while another reported bone conduction in addition to air conduction.20 One study reported only that a patient reported a change in hearing.60

Most studies reported some measure of middle ear function but the measure used varied considerably. Conversion to a type A tympanogram was reported by one study of a mechanical device,41 one study of a pharmacological intervention52 and eight studies of surgical interventions.18,2023,51,54,55,61 Other measures of tympanometry were reported as alternative or additional outcomes. These included ‘improvement to normal or more normal tracing’56 and ‘improvement’.63

Alternative measures of middle ear function which included outcomes related to pressure equalisation (e.g. tubotympanometry,40 Valsalva manoeuvre,20,21,23,25,84 passive opening pressure20,21,24) and measures based on the appearance of the middle ear on clinical examination22,23,55,57 were used. Multiple outcomes were reported by several studies for both middle ear function and hearing.20,55,56,84

Two surgical studies reported on clearance of middle ear effusion.23,51

All three of the pharmacological studies24,25,52 and one of the two studies of mechanical devices41 measured treatment success and provided a definition of successful treatment. Eight of the surgical studies measured treatment success; however, in three of these, the authors did not state how they defined treatment success.18,51,62 Where provided, the definitions of treatment success varied widely and often included multiple criteria. For example, Gluth et al. defined treatment success as a normal (type A) tympanogram in both ears or as meeting this criterion and not requiring/using additional specified treatments during the study, and reported results according to both criteria.52 In other studies, different definitions of success were used dependent on patient characteristics; for example, one case series required an improvement in tympanometry or hearing impairment as well as symptoms but an improvement only in symptoms for patients who had normal tympanometry and hearing at baseline.55

Reporting of safety data was not consistent between studies and for the majority of studies there was insufficient information to establish how systematically the data were collected. Two of the three pharmacological studies reported adverse events information, and 11 of the studies of surgical interventions reported some information.

Finally, there were considerable differences in the length of follow-up, both between studies and within studies for different outcome measures. Where reported, length of follow-up ranged from the extremely short term (30 minutes or up to 3 hours), which may have very limited clinical relevance,24,25 to the more typical short term of between 1 and 8 weeks,40,41,52,54,55,58 and 1 to 2 years or longer in some of the studies assessing surgical interventions.20,21,23,51 Length of follow-up was not clear in some studies, with assessments being noted as, for example, ‘post-operative’56 or ‘study completion’62 or only a mean duration of follow-up being reported.18,61

Quality of included studies

Quality of non-surgical studies

Of the three studies reported as being randomised, Gluth et al. 52 was considered to be at low risk of bias. Jensen et al. 25 had an unclear risk of bias primarily due to multiple gaps in reporting. The third RCT, Alpini and Mattei,40 was considered at high risk of bias due to the lack of blinding of outcome assessors, personnel and participants (Table 5). The two non-randomised studies24,41 were considered at high risk of bias due to the lack of randomisation and the lack of clarity around potential differences between groups at baseline (see Table 5 ).



Quality assessment of RCTs (non-surgical studies): risk of bias summary

Gluth et al. 52 was the only trial to describe appropriate randomisation methods and adequate allocation concealment methods. Three studies24,25,52 blinded their participants (placebo-controlled studies) but none clearly reported blinding study personnel, and therefore none of the studies were considered to have a low risk of performance bias (i.e. a risk of differences between groups in the care received or in exposure to factors other than the interventions).48

Detection bias was unclear in all except two studies24,52 which clearly stated blinding outcome assessors. Attrition bias was considered low in all studies. Only Gluth et al. 52 reported sufficient data to rule out selective reporting of outcomes. Two trials25,52 reported a power calculation, but only Jensen et al. 25 reported sufficient power to detect a significant treatment effect. Follow-up duration was considered sufficient to detect a short-term impact in only two studies.41,52 Further details are reported in Tables 5 and 6 .



Quality assessment of non-randomised controlled studies (non-surgical studies)

Quality of surgical studies

The single controlled before-and-after study assessing a surgical intervention had adequately reported eligibility criteria, determined through author contact (Professor Maria Boboshko, St. Petersburg Pavlov State Medical University, 2012, personal communication), and had an adequate length of follow-up.51 However, details of the intervention received by the control group were limited, and the representativeness of the sample and adequacy of participation rates were unclear (Table 7).



Quality assessment of controlled before-and-after study (surgical study)

The quality of the included case series was variable (Table 8). Nine of the 13 studies were prospective and three reported consecutive recruitment of patients.22,56,63 All except one study had adequately reported eligibility criteria and there was at least 80% follow-up from baseline in all except one study where this was not clear.18 It was not clear in any study whether or not the patients assessed constituted a representative sample of those eligible. Even in studies using prospective and consecutive recruitment of patients, the fact that it was unclear whether or not patients were a representative sample meant that these case series, and the controlled before-and-after study, were still at high risk of selection bias, in which patients considered likely to demonstrate a good outcome were preferentially enrolled. This risk is increased for studies in which enrolment was not consecutive and particularly so where recruitment was not prospective. Only six studies used an appropriate statistical analysis with reporting of tests to assess the significance of differences between baseline and follow-up,2023,54,56 and only four reported appropriate measures of variability such as standard deviations (SDs) for assessments at baseline and follow-up.20,22,54,55



Quality assessment of case series (surgical studies)

Six studies also reported both potential confounding and prognostic factors,2023,55,56 while two reported confounding but not prognostic factors18,60 and one reported prognostic but not confounding factors.54 Even an ideal case series is subject to the limitation that it is impossible to assess what the outcome would have been for patients had they not been given the intervention assessed. The lack of a control group means that it is unclear how much of the observed benefit(s) may be attributed to this therapy and how much may be a consequence of spontaneous recovery and/or regression to the mean. The natural history of ETD means that this is of particular relevance, as symptoms may be expected to change in intensity and frequency and, in some cases, to resolve without intervention.

In the case series included in this review, an additional consideration is the administration of concomitant therapy. The seven studies reporting confounding factors all documented additional treatments other than the intervention assessed. In some cases, all patients received additional surgical or pharmacological therapies, while in others a proportion of patients were given concomitant treatment according to perceived clinical need resulting from their ETD and/or comorbidity.

Non-surgical studies

Study characteristics

The included studies evaluated a diverse range of interventions. Only single studies were identified for each intervention: there was one randomised placebo-controlled trial for each of a pressure equalisation method,40 nasal steroids52 and a topical decongestant,25 and a non-RCT of a combination of antihistamine and ephedrine24 and a politzerisation method.41 A description of the interventions is provided in Table 2 . All the studies were small: the number of participants ranged from 20 to 91. Follow-up duration was short, ranging from 30 minutes up to 10 weeks after treatment initiation.

Only one study reported the setting and context in which interventions were delivered (by an ENT specialist in a private otology practice).41 Antihistamine, ephedrine and topical decongestants were administered as a single dose.24,25 None of the studies reported on treatment adherence and compliance. One trial reported on the use of concomitant treatments (antibiotics and/or oral decongestants) for 14 (15%) patients.52 Although the requirement for participants to have ETD was clearly stated in all studies, the condition was not explicitly defined in any of them. Selection criteria varied across the studies. Two studies included patients with a perforated eardrum,24,25 while they were excluded from one.52 Two studies excluded patients with comorbidities, such as craniofacial syndromes and cleft palate,52 active cholesteatoma, or upper respiratory tract infection.25 Diagnostic methods of ETD varied. All studies used a combination of at least two diagnostic tools, with tympanometry being the most frequent. Further details on selection criteria are reported in Table 9 .



Patient inclusion criteria and baseline characteristics (non-surgical studies)

Where reported, baseline symptoms, related conditions and previous treatments varied between the studies. Three studies reported that the included participants had symptoms associated with ETD, such as fullness, ear pain, plugged sensation, popping sensation, dampened or loss of hearing.40,41,52 However, only Alpini and Mattei reported an objective measurement of symptom severity at baseline (VAS score for fullness in the ear).40 Gluth et al. used a non-validated symptoms questionnaire but the baseline results were not reported,52 and patients in Silman and Arick41 all reported ETD symptoms during aeroplane travel but the study did not report measuring their severity. None of the studies reported on the duration or persistence of ETD symptoms before treatment.

Three studies reported using tympanometry at baseline and found results indicating abnormal middle ear function in most evaluated ears.24,41,52 Participants in Alpini and Mattei were reported to have residual middle ear effusion, although they showed normal audiometric and tympanometric results at baseline.40 All patients in Jensen et al. 23 and about one-third of participants in Holmquist and Larsson had eardrum perforations.24,25 Prevalence of eardrum perforations was not reported in the other studies. Two studies did not report any related conditions.40,41

Only two studies reported on the use of previous treatments.24,40 All patients in Alpini and Mattei had received pharmacological treatment for otitis media, such as nasal decongestants, NSAIDs, antihistamines and antibiotics.40 Holmquist and Larsson was the only study to report on the number of patients who had had ventilation tubes placed.24 Further details on patient characteristics are reported in Table 9 .


The results from individual non-pharmacological studies are reported below, grouped by type of outcome. Given that there was only a single study for each intervention, a quantitative synthesis was not possible. Indeed, due to the poor reporting of results for most outcomes, it was not always possible to extract standard data such as baseline and follow-up values with associated SDs to allow calculation of the 95% CIs for the treatment effect.

Change in severity or frequency of symptoms

Two RCTs reported a measure of symptom frequency or severity (Table 10).40,52 Gluth et al. 52 used a Eustachian Tube Dysfunction Subject Questionnaire of unknown reliability and validity. The questionnaire assessed frequency and severity of fullness in the ear, pain, plugged sensation, popping sensation, and dampened hearing, all assessed using a five-point Likert scale. The study reported the number of patients who had a positive, a negative or no change in individual symptoms following treatment. Alpini and Mattei used a VAS ranging from 0 to 10 to measure the severity of fullness in the ear of patients.40



Change in symptoms (non-surgical studies)

The total number of participants for the analysis was unclear in the Alpini and Mattei40 study and the total symptom scores results in Gluth et al. 52 were not reported. Analyses were conducted on a per-patient basis.

Alpini and Mattei38 reported a difference of four points in favour of a pressure equalisation device compared with placebo on a 0–10-point VAS at 1-week follow-up. A measure of variance was not reported and it was not possible to calculate a 95% CI for the between-group difference. The study reported the statistical significance of the change from baseline for each of the groups separately, but not for the difference between groups at follow-up.

Gluth et al. 52 stated there was a non-statistically significant difference between nasal steroids and placebo in overall symptom score at 6 weeks’ follow-up, which favoured the placebo group (p = 0.07), though the actual data were not reported. An additional analysis (analysis of covariance), which adjusted for difference in symptoms at baseline, showed no difference (p = 0.27) between the two groups at follow-up. Analysis of individual symptoms showed a significant between-group difference in plugged sensation in the ear, which was more severe (p = 0.03) and more frequent (p = 0.02) for those receiving nasal steroids. The data underpinning these analyses were not reported.

Quality of life

None of the studies reported quality of life outcomes.


One non-RCT reported the effect of treatment on hearing.41 All patients underwent a complete audiological evaluation (including pure-tone air and bone conduction thresholds, speech recognition thresholds, suprathresholds, speech recognition score for monosyllabic words, and tympanometry) at baseline and follow-up. However, only data on air–bone gap (difference between the threshold for hearing acuity by bone conduction and by air conduction measured by pure-tone audiometry) were reported. Tympanometric data are reported in the following subsection. Mean (SD) air–bone gap results were reported for each group on a per-patient basis. A mean air–bone gap of > 10 dB was considered significant. All 28 patients included in the study were analysed. It was unclear how many ears were evaluated in each patient and in total.

There was a statistically significant difference in air–bone gap favouring modified politzerisation at 3 to 4 weeks following 6 weeks of treatment (MD 12.90 dB; 95% CI 2.85 dB to 22.95 dB). Air–bone gap increased (indicating worsening in hearing) in both arms of the study: in the intervention group (by a mean 0.6 dB, SD 8.7 dB) and in the control group at follow-up (by a mean 13.5 dB, SD 17.1 dB), although the reasons for this deterioration are unclear. Further details are reported in Table 11 .



Hearing outcomes (non-surgical studies)

Middle ear function

All five non-surgical studies reported outcome measures of middle ear function. Two assessed outcome based on tympanometry,41,52 and three used other outcome measurement methods.24,25,40

Middle ear function (tympanogram normalisation)

Two studies assessed resolution of abnormal tympanometric results (from B or C to A) from baseline.41,52 Follow-up was short, ranging from 6 weeks to 9–10 weeks after treatment initiation.

Silman and Arick41 used an acoustic immitance device, with an air pressure ranging from 200 daPa to –400 daPa, with a 50 daPa/second rate of air-pressure change. Mean (SD) tympanometric peak results were reported for on a per-patient basis. Normalisation was defined as tympanogram peak pressure ≥ –100 daPa at follow-up (9–10 weeks after initiation of a 6-week treatment). All 28 patients included in the study were analysed. It was unclear how many ears were evaluated in each patient and in total. It was also unclear if one or both ears were required to have normal tympanometric peak pressure at follow-up for the treatment to be considered successful, and it was unclear if one or both ears had abnormal tympanometric peak pressure at baseline.

Gluth et al. 52 used tympanograms (with external auditory canal volume measurements) to evaluate middle ear function. Results were reported per patient (primary analysis) and per ear (secondary analysis) for patients with follow-up data. Treatment success was tympanogram normalisation, defined as a change from abnormal tympanogram (type B or C) at baseline to normal tympanogram (type A) at 6 weeks’ follow-up. Seventy-four patients (including 44 adults) were included in the analyses. All patients at baseline had an abnormal tympanogram in at least one ear (47% of patients in the intervention group had both ears with abnormal tympanogram at baseline vs. 54% in the placebo arm).

Silman and Arick41 found that modified politseratisation was associated with a significantly reduced risk of having an abnormal tympanometric peak pressure at follow-up compared with no intervention [risk ratio (RR) 0.36; 95% CI 0.15 to 0.87]. Mean tympanometric peak pressure improved in the treatment group (decrease of 182.9 daPa, SD 153.0) and deteriorated in the control group (increase of 18.1 daPA, SD 182.4). The difference between the two arms significantly favoured patients undergoing modified politzerisation (MD –201.00 daPa; 95% CI –325.71 daPa to –76.29 daPa).

Overall, there was no statistically significant difference in the percentage of patients with tympanogram normalisation between nasal steroids treatment (7 out of 37: 19%) and the placebo group (12 out of 37: 32%), (RR 1.20; 95% CI 0.91 to 1.58) in Gluth et al. 52 When considering patients with treatment failure (those who took an additional treatment such as antibiotics, oral decongestant or nasal spray in the placebo group) as having incomplete resolution, rates of resolution were slightly lower in the intervention group (5 out of 37: 14%) and the placebo arm (9 out of 37: 24%), although the difference between the two group remained non-statistically significant (RR 1.14; 95% CI 0.91 to 1.43).

In Gluth et al.,52 it was also possible to calculate results on a per-patient basis for the subgroup of adults with follow-up data (22 patients in each arm). No statistically significant differences were found between intervention and placebo in this subgroup, including when accounting for treatment failures. Per-ear analyses were consistent with these findings. Further details are reported in Appendix 5 .

Middle ear function (other outcome measurement methods)

Three additional non-surgical studies measured changes in middle ear function, all significantly favouring treatment. Follow-up duration was short, ranging from 30 minutes to 1 week.24,25,40

Alpini and Mattei40 used tubotympanometry to assess Eustachian tube function in patients who had recovered from otitis media. The test evaluates the impedance of the eardrum during Valsalva’s manoeuvre and swallowing, by recording the inflow and outflow of air through the tube as a pattern of impedance change which is classified as normal, obstructive, or patent.86,87 The study used tubotympanometry to assess ‘residual Eustachian dysfunction’, for which no definition was provided. A normal result indicated normal Eustachian tube function, but it was unclear whether abnormal tubotympanometry indicated abnormal Eustachian tube occlusion, patency or both [the test was used in conjunction with vestibular evoked myogenic potential (VEMPs), a neurophysiological assessment technique used to determine the function of nerves and organs (utricle and saccule) within the inner ear. VEMPs results were not reported as they focused on the inner ear and were therefore considered to be beyond the scope of the review]. All 20 patients in the study except one in the control group had abnormal tubotympanometry at baseline. It is unclear whether all patients or only those with abnormal tubotympanometry (n = 19) were included in the analysis.

Holmquist and Larsson24 measured middle ear function using an air pressure equalisation technique (pressure regulator and manometry) in ears with eardrum perforations and tympanometry for ears with intact eardrum. Improvement in middle ear function was evaluated in terms of number of ears with reduction in opening pressure of ≥ 100 mmH2O (patients with perforated eardrums) or pressure change in normalising direction of ≥ 100 mmH2O (patients with intact eardrums). The study reported that Eustachian tube function was measured three or four times within 3 hours of treatment intake. All 39 treated ears (of 32 patients) were analysed, although it is unclear which of these measurements were taken into account to assess treatment success.

Jensen et al. 25 assessed Eustachian tube function using the Valsalva manoeuvre and the aspiration/deflation tests (using an initial pressure of ±200 mmH2O). Valsalva results were considered positive if the test resulted in an audible passage of air at least once in five Valsalva manoeuvres (listening test). Aspiration/deflation tests results were positive if the aspiration test showed a residual pressure of –100 mmH2O or more, or if the deflation test showed a residual pressure of +100 mmH2O or less. Positive treatment effect was defined as a change from pathological to normal measurement in Valsalva, aspiration or deflation tests in at least one occurrence. In addition, results from the three tests (Valsalva, aspiration and deflation tests) were reported separately. Patients were analysed for each test if they had a pathological measurement at baseline (35 patients with negative Valsalva, 35 with pathological aspiration test, 28 with pathological deflation test). All 36 patients had at least one pathological measurement at baseline, and all were analysed for the treatment success outcome.

In Alpini and Mattei,40 tubotympanometry became normal in 9 out of 10 patients using a pressure equalising device, compared with 3 out of 10 patients in the control group after 1 week of treatment. The improvement in the treatment arm was statistically significant compared with control (RR 0.13; 95% CI 0.02 to 0.85).

In Holmquist and Larsson,24 a positive improvement in middle ear function was recorded in 11 out of 19 ears receiving antihistamine–ephedrine combination, compared with 2 out of 20 ears receiving placebo at up to 3 hours after receiving the intervention. The difference between the groups significantly favoured the treatment arm (RR 0.47; 95% CI 0.27 to 0.81).

In Jensen et al.,25 treatment effect was positive for 12 out of 19 patients undergoing direct application of nasal decongestant, compared with 7 out of 17 patients in the placebo group. The difference was not statistically significant (RR 0.63; 95% CI 0.31 to 1.27). Separate results of each test showed, that compared with placebo, patients receiving topical decongestants had significantly improved Valsalva results 30 minutes following treatment (RR 0.47; 95% CI 0.28 to 0.80). However, no significant effect was demonstrated by the aspiration test (RR 0.94; 95% CI 0.60 to 1.48) or the deflation test (RR 0.80; 95% CI 0.40 to 1.58). The authors concluded that decongestants only had a positive effect on Eustachian tube function at unphysiologically high pressure increases (as with the Valsalva manoeuvre), but not in the case of the smaller and more physiological pressures generated by the aspiration/deflation test.

Clearance of middle ear effusion

No non-surgical study reported clearance of middle ear effusion as a treatment outcome.

Need for additional treatment

Need for additional treatment was assessed in one pharmacological study.52 Gluth et al. reported that 7 out of 37 patients (19%) receiving nasal steroids required antibiotics or oral decongestants while enrolled in the study. In the placebo group, 7 out of 37 (19%) received antibiotics, oral decongestants and/or nasal spray. There was no statistically significant difference between the groups (RR 1.00; 95% CI 0.39 to 2.57).

Adverse events, discontinuation

Two out of five non-surgical studies stated that they addressed adverse events of treatments, although it was not clear how data on adverse events were elicited.25,52 Jensen et al. 23 reported no adverse events following treatment with topical decongestant (follow-up duration unclear). Gluth et al. 52 reported that minor events (coughs and nosebleeds) occurred in both arms of the study during a 6-week course of nasal steroids, which did not lead to any treatment discontinuation. However, the number of adverse events was not reported. No discontinuations due to lack of treatment effectiveness were reported.

Key findings of non-surgical studies

There were five studies included that evaluated a variety of non-surgical interventions, namely pharmacological treatments and manual pressure equalisation device.24,25,40,41,52 None of the interventions was evaluated by more than one study. All of the non-surgical studies included adults, and all except one explicitly reported including a minority of children; therefore, the generalisability of the findings to the broader population of adults with ETD is uncertain. Nearly all patients had a diagnosis of ETD, although the studies were rarely explicit about how they had defined ETD. Diagnostic methods varied, but nearly all studies used a combination of at least two methods, such as tympanometry and audiometry. A single RCT was identified that was assessed as at low risk of bias, though the data available for some outcomes were sparse in this study.52

One RCT,52 which was the only study with a low risk of bias, showed no evidence that nasal steroids were effective at improving the severity and frequency of ETD symptoms among patients with otitis media with effusion and/or negative middle ear pressure by the end of 6 weeks of treatment. Minor adverse events were reported (coughs and nosebleeds), although there were no significant differences between the two arms of the study.

One RCT25 reported some evidence of improvement in middle ear function for patients with a history of chronic otitis media 30 minutes after receiving direct application of a topical decongestant on the pharyngeal opening of the Eustachian tube. The trial suggested that treatment only improved middle ear function when patients were subject to unphysiologically high pressure changes. The study reported no adverse events. However, the internal validity of this study is unclear, notably due to multiple gaps in reporting of design characteristics and very short-term follow-up. This is also likely to be an unrealistic treatment for primary care; in secondary care, endoscopic guidance would be required.

One non-RCT24 found a significant improvement in middle ear function (significant reduction in opening pressure for patients with eardrum perforation or pressure change in normalising direction for patients with intact eardrums) for patients receiving a single dose of antihistamine and ephedrine compared with placebo.24 However, the reliability of these findings is uncertain, notably due to a high risk of selection bias and very short follow-up duration (3 hours). All three pharmacological studies included very few patients, and, of the two studies that reported a power calculation,25,52 only one reported sufficient power to detect a significant treatment effect.24

Two studies40,41 evaluated the use of two different manual pressure equalisation devices, both of which had a high risk of bias. One RCT40 found that self-administration of a manual device applying mild negative pressure to the external ear canal three times a day for 1 week was associated with a significant reduction in severity of fullness in the ear and middle ear function (measured by tubotympanometry) for ETD in patients with residual middle ear effusion by the end of treatment. A non-RCT41 found a statistically significant difference in middle ear function (tympanometric peak pressure) and in hearing (air–bone gap) 9–10 weeks following initiation, which favoured the use of modified politzerisation twice-weekly for 2 weeks. However, the difference in hearing reflected an unexplained deterioration in the control group rather than an improvement in those who received the intervention. Neither study reported data on adverse events of the interventions, making the safety of these interventions uncertain.

Some studies reported some or all data on a per-ear basis, where some patients received bilateral treatment; where this was the case, it was unclear if any statistical analysis undertaken incorporated the within-patient correlation of outcome data.

Overall, the evidence from non-surgical studies is limited due to the small number and size of the studies, poor reporting of study design characteristics, definition of ETD and diagnostic criteria, patient characteristics and outcome data, and limited follow-up duration. The evidence available for any single intervention was extremely sparse. This precludes any definitive conclusions on the effectiveness and safety of non-surgical interventions for ETD in adults.

Surgical studies

Study characteristics

With the exception of Boboshko et al.,51 all of the studies of surgical interventions were case series. Silverstein et al. (n = 11) used surgery to permit the topical application of steroids (see Table 2 ).61 The other interventions assessed were procedural variations of laser Eustachian tuboplasty (seven studies, n = 182);20,21,23,54,56,62,63 balloon dilatation of the Eustachian tube (three studies, n = 103);18,22,55 and myringotomy (two studies; n = 121).58,60 General anaesthesia was used in five studies,22,5456,63 local in seven studies,18,20,21,51,58,60,61 a combination of the two in one study,23 and one study did not report which method was employed.62

In most of the studies, all patients were described as having a diagnosis of ETD, although in two these patients were a defined subgroup for whom separate data were reported.21,60 In three studies, the diagnosis was otitis media with effusion due to ETD;22,23,51 these patients would be expected to have a worse prognosis than those in other studies without otitis media with effusion. Three studies excluded patients with comorbidity such as allergies or reflux disease;20,21,54 conversely, one study required a diagnosis of rhinosinusitis severe enough to warrant surgery.56 Some studies included patients with a perforated eardrum as a separate subgroup,20,21 while others did not report separate data for these patients, although they were identifiable from their baseline tympanometry.22,23 Baseline tympanometry data are discussed together with the tympanometric response to therapy in Results, below.

Diagnostic methods varied between studies, but all (except one which did not report these) used more than one method. In all except two studies, the diagnostic process included tympanometry. Although the requirement for participants to have ETD was clearly stated by authors, as with the non-surgical studies, the condition was rarely defined. Most commonly, some of the symptoms characteristic of patients with a diagnosis of ETD such as otalgia during pressure change and aural fullness were cited,18,21,51,5456,61,63 together with requirements for patients to have an abnormal tympanogram or abnormal appearance on clinical examination.20,21,23,5456,61,63 In some cases, ETD was cited as an inclusion criterion with no further information.62 Full details of the range of criteria used are shown in Table 12 .



Patient inclusion criteria and baseline characteristics (surgical studies)

Patients in most of the studies had extensive histories of previous treatment for ETD and/or related conditions. These included multiple antibiotic and steroid treatments as well as previous aural and nasal surgeries (see Table 2 ). In five studies, failure of one or more medical therapies or previous surgical intervention (e.g. pressure equalisation tubes) was a criterion for inclusion in the study,20,22,23,54,61 while in nine studies, all patients had failed one or more previous treatment.20,22,23,5456,60,61,63


Severity and frequency of symptoms

Nine studies assessed change in symptoms of ETD. All reported the number of patients who experienced improvement in symptoms, although the definitions for the outcome varied in terms of symptoms specified; two studies did not specify any symptoms (Table 13).56,63 In the majority of studies, the methods for outcome assessment were not reported; the authors simply referred to the number of patients who showed improvement or resolution of symptoms. Where this was the case, it appeared that the assessment was based on patient report, although one study included the appearance of the tympanic membrane at examination in the criteria, together with symptomatology.18 The presence or intensity of subjective symptoms was assessed in some studies but methods of determining whether a patient had improved or resolved symptoms and thresholds for these improvements/resolutions were not always specified. One study specified that the outcome was assessed by ‘patient response’55 and another specified that the assessment was subjective.56 One study used a VAS but reporting of outcome data was limited.20 Another study reported mean scores on the ETDQ-7 and the SNOT-22 scales.55 Length of follow-up ranged from 2 months to 2 years, where reported.



Change in symptoms (surgical studies)

Four of the studies assessed various techniques for tuboplasty. The study by Caffier et al. only reported the numbers of patients showing improvement for the outcome of tinnitus, but did show patient-reported data using a VAS for overall satisfaction and improvement of ETD, aural fullness and hearing loss.20 Scores of between 5 and 7 on the VAS were reported for all outcomes at 1 year; these were described by the authors as high, though the possible score range on the scale was not reported.20 Of the other three studies, two reported success rates of > 90%62,63 and one reported a success rate of 70% based on resolution of symptoms;56 symptoms were specified as ear blockage, ear pain, hypoacusis and autophony in one study but were not specified in the other two.

Two studies assessing balloon dilatation reported symptom resolution.18,55 One reported improvement in 92% of ears at 6 months’ follow-up;55 the second reported improvement in 71% of ears after a mean of 30 weeks’ follow-up.18 McCoul et al. assessed symptoms using the ETDQ-7 and the SNOT-22; data were reported for durations of follow-up ranging from 3 weeks to 6 months.55 ETDQ-7 score improved by a mean of 1.8 points (SD 1.2; 22 ears) at 6-month follow-up from a mean baseline score of 4.5 (SD 1.2; 31 ears) (p = 0.001). The possible score range on the ETDQ-7 is 1.0 to 7.0, with higher scores denoting more severe symptoms.26 The clinical relevance of changes in ETDQ-7 score was not interpreted in the study, and no data on minimally important clinical difference were found; therefore, the clinical relevance of the change in ETDQ-7 is unclear. The SNOT-22 score improved by a mean of 23.3 (SD 19.6; 21 ears) points at 6-month follow-up from a mean baseline score of 51.4 (SD 21.1; 33 ears) (p = 0.001). The possible score range on the SNOT-22 is 0 to 110.27 The minimally important difference in SNOT-22 score was estimated to be 8.9;27 therefore, reported changes in SNOT-22 from baseline were likely to be clinically relevant. Both measures showed statistically significant improvements from baseline at all time points.

Studies of laser-assisted tympanostomy (myringotomy) and topical steroid application reported symptom resolution or improvement in 79% and 72% of patients, respectively.58,61 The controlled before-and-after study found that all 25 patients treated with point laser coagulation reported disappearance or reduction of unpleasant feeling and noise in the ear; results for the 15 patients in the control group were not reported.51

Across all of the studies, the proportion of patients showing improvement in ETD symptoms, either not further defined by the authors or defined as multiple symptoms (as opposed to a single symptom of tinnitus20 or otitis media with effusion resolution23), was between 70% and 100%. Given the small sample sizes, this suggests some consistency across interventions, despite wide variations in patient characteristics, treatments assessed and use of concomitant therapies.

Quality of life

None of the studies assessed quality of life.


Seven studies reported some information on changes in hearing following the intervention (Table 14),20,23,51,56,60,61,63 although one study of myringotomy noted only that 1 out of 13 patients experienced an unspecified change in hearing.60



Hearing outcomes (surgical studies)

Length of follow-up ranged from 4 months60 to 2 years23,51 where specified; one study reported only that the assessment was ‘postoperative’.56 Five studies reported pure-tone averages, measured in DBs across a range of speech frequencies. Other outcomes including the air–bone gap (also measured in DBs) were reported by some studies as additional61 or alternative20,51 measures of hearing.

Change from baseline to follow-up in pure-tone ranged from a mean improvement of 6 dB to 10.7 dB. The change in air–bone gap ranged from a mean improvement of –6 dB to –12.3 dB. The statistical significance of change from baseline to follow-up was not reported in all studies.

Four studies that evaluated a form of tuboplasty assessed changes in hearing.20,23,56,63 Three of these reported pure-tone averages with improvements of between –6 dB and –10 dB at time points between ‘postoperative’ and 2 years’ follow-up.23,56,63 The fourth study reported a statistically significant decrease in the air–bone gap of –12.3 dB, although patients with perforated tympanic membranes showed larger improvements than those with intact membranes.20 This study also reported improvements in both air and bone conduction and found a significant improvement in air–bone gap at 1 year in patients with perforated (p < 0.001) and in patients with intact ear drums at baseline (p < 0.05).20

None of the studies of balloon dilatation reported data on hearing.

Two studies of other interventions reported a hearing outcome. The study which assessed topical steroid application by MicroWick (Silverstein MicroWick™, Anthony Products, Indianapolis, IN, USA) reported statistically non-significant improvements of 6 dB in the mean pure-tone average, a statistically non-significant change in the air–bone gap of –6 dB at a mean of 7 months’ follow-up, and a slight improvement in mean speech discrimination scores from 94% to 97%.61

Mean air–bone gap was significantly smaller in patients receiving laser coagulation surgery than in controls at 1-year follow-up [MD –6.70 (95% CI –9.91 to –3.49)].51 Twenty-four of the 31 ears (77.4%) that received laser surgery had an air–bone gap of < 10 db at follow-up, compared with 4 out of 15 ears (26.7%) in the control group at 1 year. This difference was statistically significant (RR 0.31; 95% CI 0.15 to 0.63). Data on the numbers of patients with pure-tone average thresholds between 0 dB and 10 dB, 11–20 dB, 21–30 dB and > 30 dB were also reported and indicated greater improvement in the intervention group.

Middle ear function

Nine studies reported on the numbers of patients displaying a type A tympanogram following treatment (Table 15). The proportion of patients with an abnormal (non-type A) tympanogram at baseline ranged from 0% to 66%, although tracings in at least one study, while meeting criteria for type A, were nevertheless considered to be abnormal by the authors.55 Length of follow-up ranged from 2 weeks to 2 years.



Middle ear function: tympanogram type A (normal) (surgical studies)

In Caffier et al. and Sedlmaier et al., tympanograms were described as normal rather than as type A, with other tympanogram tracings being classified as ‘flat’, ‘flatter’ or ‘negative’.20,21 These two studies also did not report the number of patients who experienced change from baseline from an abnormal to a type A tympanogram; instead, they reported the number of type A tympanograms at baseline and follow-up. The overall number of patients with type A tympanograms increased in both studies. However, because it is possible for treatment to have a negative impact on tympanogram results in some patients, this does not preclude some individuals having converted from type A to B/C.

Four studies assessed types of tuboplasty.20,21,23,54 Caffier et al. reported that four patients (26%) had a normal tracing at follow-up compared with two (13%) patients at baseline.20 Poe et al. 23 reported conversion to type A in 15% of patients. A third study found an increase of 13% in the number of patients with a normal pattern, assessed 8 weeks after surgery.21 The fourth study assessing tuboplasty which reported change from baseline found normalisation in 36% patients.54

Three of the studies assessed balloon dilatation. Rates of conversion to type A tympanogram were 36%,22 71%18 and 96%.55 The lower conversion rate of 36% may be due, at least in part, to the fact that patients in this study had a diagnosis of chronic otitis media with effusion. However, both of the studies reporting higher conversion rates also administered concomitant surgical treatment, either to all or to a majority of patients, meaning that these studies may overestimate the effect of the treatment.

In studies assessing other interventions, Silverstein reported that 50% of the patients with an abnormal baseline tracing had converted to a type A pattern after a mean of 8 months’ follow-up after MicroWick application of dexamethasone.61 One study of laser coagulation reported no type A tympanograms at baseline and improvement to type A in 97% of affected ears 2 weeks after surgery.51

Seven studies also reported a range of other measures of middle ear function. These included ability to perform the Valsalva manoeuvre,2022,54 passive tubal opening,21 response to pressure testing (including Eustachian tube opening and closing pressures)20,54 changes in the waveform of the tympanogram other than a shift from types B or C to type A,56,63 mucosal inflammation score,22 Eustachian tube endoscopy scores,23 and results of clinical examination of the tympanic membrane.55 Multiple measures were reported by five studies, all of which also reported data on type A tympanograms (see Table 15 ).2023,54

Changes in tympanogram other than the presence of a type A pattern were reported in two studies of tuboplasty. One reported that 96% (24 out of 25) of patients with an abnormal tympanogram (defined as non-type A) demonstrated improvement (undefined).63 Another reported the development of a normal or more normal tracing in 65% (11 out of 17 patients) who showed baseline abnormality (not defined).56

Changes in the number of patients with a positive Valsalva manoeuvre were reported by three studies of tuboplasty;20,21,54 and one of balloon dilatation in patients with otitis media with effusion.22 All studies showed an increase in the proportion of patients, ranging from 61% to 100%, with a positive Valsalva manoeuvre (see data extraction tables in Appendix 5 for full details).

Eustachian tube endoscopy scores which assessed valve dilatation, mucosal swelling and function of the levator veli and tensor veli were reported by the tuboplasty study in otitis media with effusion patients and did not show significant changes from baseline,23 while muscosal inflammation scores were reported in the balloon dilatation study in otitis media with effusion patients and did show a statistically significant improvement.22 The other balloon dilatation study assessed tympanic membrane retraction and found an improvement in all affected patients.55 Full details are reported in the data extraction tables (see Appendix 5 ).

Measures of response to pressure were reported by one tuboplasty study, which assessed the closing and opening Eustachian tube pressures and the presence of a blocked pattern using a pressure chamber. As with the tympanometry results, this showed a statistically significant improvement.54 Normal passive tubal opening was assessed by two other studies of tuboplasty.20,21 Interepretation of graphical data indicated improvement in both studies in the number of patients showing normal response but the statistical significance was unclear. Full details are reported in the data extraction tables (see Appendix 5 ).

Clearance of middle ear effusion

Clearance of middle ear effusion was reported by two studies.23,51 In one study of patients with chronic otitis media with effusion,23 there was resolution of the symptom in 4 of 11 patients at 6 months following tuboplasty. The method for assessing resolution appeared to be clinical examination. The second study reported that 6% of ears treated with point laser coagulation experienced otitis media with effusion recurrence at 9 to 11 months’ follow-up, compared with 40% of ears in the control group at 1 to 6 months’ follow-up.51

Need for additional treatment

Eight studies reported data on the need for additional treatment: three of tuboplasty,23,56,63 three of balloon dilatation18,22,55 and one each of myringotomy60 and MicroWick application of steroids.61 The additional treatment documented was either a repeat of the original procedure18,55,60 or the insertion or removal of pressure equalisation tubes or myringoplasty for persistent perforation.60,61 One study documented repeated treatments only in those patients with a good response to initial therapy.18 The duration of follow-up varied from 4 months to 2 years, where reported.

In the three studies assessing forms of tuboplasty, one reported that no additional treatment (pressure equalisation tubes) was required.63 One study reported that 2 of the 20 patients required pressure equalisation tubes following tuboplasty,56 while the study conducted in patients with chronic otitis media with effusion reported that two of the eight patients on whom follow-up data were available for this outcome required them.23

One study of balloon dilatation reported repeat procedures in 2 of the 22 patients.55 The second reported data only for ears which had shown an initial benefit of treatment (71 of 100 ears); seven of these required a repeat dilatation.18 The third study reported only concomitant treatments of insertion or removal of pressure equalisation tubes.22

One study of myringtotomy in 13 patients documented one repeat procedure and two myringoplasties for persistent perforation.60 Myringoplasties were also required in 3 of 11 patients who underwent MicroWick application of dexamethasone.61

Adverse events and complications of Eustachian tube dysfunction

Information on adverse events was reported in all except three studies.58,62,63 Duration of follow-up varied from 1 week51 to between 1 and 2 years where it was reported;23 in two of the studies, it was clear that a post-surgical assessment was conducted but no assessment of long-term adverse events was carried out.21,54 The adverse events documented were generally minor, consisting of discomfort, minor lacerations, formation of adhesions (synechia), and granulomas. Individual cases of bleeding and radiculopathy were documented following balloon dilatations.

Of the tuboplasty studies, one reported that there were no adverse events in the 20 operated patients;56 one reported no acute or long-term complications but no further information was provided;54 one reported two patients with types of synechiae and two with granuloma in the resected area but no significant surgical complications;23 and one reported discomfort which was relieved by additional anaesthesia in three patients and one case of an adhesion.20 Three other tuboplasty studies did not report data on adverse events.58,62,63

Of the two studies of balloon dilatation, one reported minor mucosal lacerations to the lumen of the Eustachian tube in 5 of the 11 patients and a contralateral radiculopathy (C6–7 disc space) which showed full recovery; no further complications were reported.22 The other study reported that one patient (of 22) experienced bleeding, which resolved after myringotomy.55

The study assessing bilateral thermal myringotomy reported no adverse effects other than persistent bilateral perforations in 2 of 13 patients.60 A study of laser ablation of the epipharyngeal Eustachian tube reported that one synechia was the only complication.21 A study of laser point coagulation reported that there were no adverse events.

One study reported complications of ETD: one patient developed profound sensorineural hearing loss as a consequence of severe otitis media following MicroWick application of topical steroids; no negative effects of treatment were documented.61

Discontinuations of treatment and other losses to follow-up

All of the studies, except that by Silverstein et al.,61 were of surgical interventions, so discontinuation of therapy was not a relevant outcome for these studies. Silverstein et al. reported one discontinuation of therapy (from 11 patients) as a consequence of the ETD complication described above.61 Similarly, early tube extrusion was not relevant, as pressure equalising tubes were not the intervention assessed in any of the included studies, although in three studies some patients had them placed before, during or after the assessed therapy.22,23,55,56

Losses to follow-up were reported by three studies: one of balloon dilatation55 and two of tuboplasty.21,23 Data from Poe et al. were complicated by missing data from different individual patients at multiple follow-up points,23 while McCoul et al. reported cumulative losses55 and Sedlmaier et al. a single follow-up duration.21 The fact that other studies did not report losses to follow-up is likely to be due to the fact that in five studies there was no prespecified study duration: these studies reported mean duration of follow-up or range of follow-up duration.18,22,51,63 In a further two studies, duration of follow-up was reported only as ‘post-surgical assessment’56 or ‘study completion’.62

Key findings of surgical studies

There were 14 included studies that evaluated a surgical intervention. These were all cases series, with the exception of a controlled before-and-after study. No RCTs were identified. Any interpretation of data from case series is limited by the uncontrolled study design: it is impossible to determine how much improvement in symptoms and other measures would have occurred in the absence of the intervention. The assessment is further limited by the fact that the recruitment of patients to the studies was not well described; although a majority of the studies were prospective, few reported consecutive recruitment of patients, and it was unclear how representative the included patients were in any of the studies. The controlled before-and-after study was also subject to potential sources of bias.51

A second major limitation was the lack of clear definitions of ETD in the included studies, which was combined with substantial variations in inclusion criteria, baseline patient characteristics and diagnostic methods. Outcome data were poorly reported in some of the studies, although losses to follow-up were generally limited. In many cases, the statistical significance of results was not reported and insufficient data were available to calculate this information. Some studies reported some or all data on a per-ear basis, where some patients received bilateral treatment; where this was the case, it was unclear if any statistical analysis undertaken incorporated the within-patient correlation of outcome data. Follow-up duration also varied substantially, and in some cases was reported only as a mean duration. Where data from multiple follow-up points were reported, these sometimes indicated continuing accrual of benefit with increased length of follow-up; this may be a consequence of the intervention or of a remission of symptoms due to the natural course of the condition of ETD. Finally, many of the studies reported the use of co-interventions for many or all of the patients; these concomitant therapies involved additional surgery as well as some pharmacological therapies. In the absence of control groups, it was not possible to separate the effects of assessed interventions, natural remission or alteration of symptoms, and co-interventions.

The greatest volume of evidence related to various techniques of Eustachian tuboplasty, with seven studies in 182 patients assessing interventions in this category.20,22,23,54,56,62,63 However, in addition to differences in the techniques used, there were wide variations between the patients in these studies, as well as differences in the outcomes reported and the measures used to assess outcomes. One study also employed sinus surgery in all patients,56 and concomitant interventions for some patients were reported in three of the other studies.20,22,23

Rates of resolution or improvement were high, at 70–90% in the three studies which assessed symptoms of ETD generally. Unsurprisingly, the study which evaluated the intervention in otitis media with effusion patients and did not systematically use cointerventions reported lower rates of symptom (middle ear effusion) resolution. Improvements in tinnitus reported by another study were also low,20 although improvements in VAS scores for other symptoms were described as significant. Where hearing was assessed, improvements in the pure-tone average were small (6 dB to 10 dB) and, therefore, potentially not clinically significant. Measures of middle ear function indicated low (15% to 40%) rates of conversion to type A tympanogram20,23,54 but higher rates (65% to 96%) of ‘normalisation of the tympanogram’.56,63 There is insufficient evidence to demonstrate the effectiveness of tuboplasty, or to determine either the details of the surgical technique which should be employed or the patients for whom it should be considered.

Of the three studies of balloon dilatation identified (n = 107), those patients who were not required to have previous surgery as a criterion for inclusion demonstrated high levels of improvement in symptoms of ear fullness, pressure, pain and tolerance to air travel or general ETD symptoms;18,55 the resolution of symptoms was not reported in the study of patients with otitis media with effusion, although tympanometry showed an improvement in 36% of patients.22 Tympanometric measurement of middle ear function in the other two studies indicated high levels of conversion to type A tracings (89% and 96%, respectively).18,55 None of the studies reported data on hearing. As with tuboplasty, there is insufficient evidence to demonstrate efficacy or to determine the population in which it should be considered, particularly when the use of cointerventions for some or all of the patients is borne in mind.

Two studies assessed methods of myringotomy.58,60 The first of these had only 13 patients, of whom 11 had an ETD diagnosis;60 these patients were treated in order to enable hyperbaric oxygen therapy for other indications. A minority of patients in the larger (n = 108) case series had a diagnosis of ETD.58 In both cases, reporting of outcome data was very limited, as was duration of follow-up. While the data suggested that there may be a benefit to myringotomy, they were too short term and poorly reported to allow any conclusions to be drawn.

The evidence base forotopical application of steroids to the Eustachian tube rests on a single case series of 11 patients.61 While the limited data available from this series suggested that there may be some benefits to the treatment, they are too limited to allow any conclusions to be drawn.

The single controlled before-and-after study evaluated laser point coagulation and indicated high levels of symptom resolution (100%) and tympanogram normalisation (97%) but did not report comparable data for control subjects. An improvement in hearing did show a statistically significant benefit compared with controls.51 Therefore, while these studies suggested there may be some benefits to these treatments, the available data were too limited to allow any conclusions to be drawn.

Overall, none of the interventions appeared to be associated with serious adverse effects, although minor complications of surgery were reported in a minority of patients in several studies. It was not clear that adverse events were systematically documented, and several studies did not report any safety data. The evidence pertaining to surgical interventions generally is of limited quantity and quality.

Copyright © Queen’s Printer and Controller of HMSO 2014. This work was produced by Llewellyn et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Included under terms of UK Non-commercial Government License.

Bookshelf ID: NBK262275


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.1M)

Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...