Systematic Review of Extratympanic Electrocochleography in Ménière's Disease Diagnosis (PDF)

International Journal of Audiology ISSN: 1499-2027 (Print) 1708-8186 (Online) Journal homepage: https://www.tandfonline.com/loi/iija20 A systematic review and meta-analysis of extratympanic electrocochleography in Ménière’s disease diagnosis Aysha Ayub, Li Qi & Desmond A. Nunez To ci...

International Journal of Audiology ISSN: 1499-2027 (Print) 1708-8186 (Online) Journal homepage: https://www.tandfonline.com/loi/iija20 A systematic review and meta-analysis of extratympanic electrocochleography in Ménière’s disease diagnosis Aysha Ayub, Li Qi & Desmond A. Nunez To cite this article: Aysha Ayub, Li Qi & Desmond A. Nunez (2019): A systematic review and meta-analysis of extratympanic electrocochleography in Ménière’s disease diagnosis , International Journal of Audiology, DOI: 10.1080/14992027.2019.1606947 To link to this article: https://doi.org/10.1080/14992027.2019.1606947 Published online: 08 May 2019. Submit your article to this journal Article views: 60 View Crossmark data Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=iija20 INTERNATIONAL JOURNAL OF AUDIOLOGY https://doi.org/10.1080/14992027.2019.1606947 REVIEW ARTICLE A systematic review and meta-analysis of extratympanic electrocochleography in Me
nie re’s disease diagnosis Aysha Ayuba,b, Li Qib,c,d and Desmond A. Nunezb a Faculty of Science, University of British Columbia, Vancouver, Canada; bDivision of Otolaryngology Head and Neck Surgery, Department of Surgery, University of British Columbia, Vancouver, Canada; cNeuro-Otology Clinic, Vancouver General Hospital, Vancouver, Canada; dSchool of Audiology and Speech Sciences, University of British Columbia, Vancouver, Canada ABSTRACT ARTICLE HISTORY Objective: Determine whether a combination of electrocochleography determined summating/action Received 27 April 2018 potential (SP/AP) ratio and other audiological measurements has greater sensitivity and specificity than Revised 1 April 2019 that achieved with electrocochleography SP/AP ratio alone in diagnosing definite M
eniere’s Disease. Accepted 3 April 2019 Design: Systematic review and meta-analysis. KEYWORDS Study sample: Pubmed, Cochrane Library, and Web of Science were searched using search terms M
eniere’s Disease; “electrocochleography”, “ECochG, ,“ECoG”, “M
eniere’s Disease”, and “Idiopathic Endolymphatic Hydrops”. electrocochleography; elec- Inclusion criteria were extratympanic electrocochleography methodology, English language publication trocochleography (ECochG); between January 2002 and December 2017, and the 1995 American Academy of Otolaryngology and electrocochleography Head and Neck Surgery M
eniere’s disease diagnostic criteria. Five articles satisfied inclusion criteria and (ECoG); idiopathic were sufficiently detailed for aggregate quantitative analysis of SP/AP ratio (315 subjects) and combin- endolymphatic hydrops; ation audiological measures (113 subjects). electrophysiology; condi- Results: The diagnostic sensitivity and specificity of the SP/AP amplitude ratio was 47.6% and 83.8% and tions/pathology/disorder of combination diagnostic measures 63.5% and 89.3%, respectively. Point estimates of sensitivity (p ¼ 0.248) and specificity (p ¼ 0.969) and the summary Receiver Operator Characteristic Curve (p ¼ 0.407) were not statistically significant. Conclusion: Statistically, combination diagnostic measures do not result in greater accuracy of definite Me
niere’s disease diagnosis compared to the SP/AP amplitude ratio alone. However, given the small sam- ple size further studies are recommended to arrive at a definitive conclusion. Introduction in individuals without a diagnosis of M
eniere’s Disease and failed to find ELH in others diagnosed with M
eniere’s Disease (Foster M
eniere’s Disease, also called idiopathic endolymphatic hydrops and Breeze 2013). (ELH), was first described by a French otologist Prosper M
eniere The diagnosis of M
eniere’s Disease is normally made on the in 1861 (Baloh 2001). In 1972, the Subcommittee on Equilibrium basis of symptoms and pure tone audiometry (Zhang et al. and Its Measurements of the American Academy of 2016). According to the AAO-HNS (1995) guideline the diagno- Otolaryngology and Head and Neck Surgery (AAO-HNS) sis of M
eniere’s Disease is based on a selection of clinical symp- defined M
eniere’s disease as a disorder of hearing loss, and ver- toms, signs and exclusion of other identifiable causes tigo, with hydropic distention of the endolymphatic system as a (Committee on Hearing and Equilibrium 1995). However, the pathological correlate (AAOO and AAO-HNS 1992). In 1995, diagnosis of M
eniere’s Disease is still a challenging task for clini- aiming to standardise the diagnosis, the AAO-HNS subcommit- cians due to a wide and varied spectrum of clinical presentations tee published new clinical criteria for M
eniere’s Disease that clas- and the fluctuating nature of M
eniere’s Disease. Therefore, a sified the strength of diagnosis of M
eniere’s disease as certain, quantitative objective diagnostic test in addition to the pure tone definite, probable, or possible (Monsell et al. 1995). audiogram to aid accurate diagnosis is desirable. The reported prevalence of M
eniere’s Disease varies Electrocochleography (ECochG) was introduced to aid the between 20 and 200 per 100,000 persons globally (Lempert diagnose of M
eniere’s Disease since the 1970s (Eggermont 2017). 2012). M
eniere’s Disease typically starts between the ages of 30 ECochG can be recorded with transtympanic (TT) or extratym- and 50 years, and in most cases only one ear is affected (Lempert panic (ET) electrodes (Eggermont 2017). TT and ET approaches 2012; Lacour et al. 2007). have different advantages and disadvantages. Although the TT The aetiology of M
eniere’s Disease remains elusive. method has a lower signal-to-noise ratio and larger and easier to Histological studies of temporal bones of patients with M
eniere’s read potentials compared to the ET method, the TT method is Disease suggested that endolymphatic hydrops may be respon- invasive and requires an otologist (Sass, Densert, and Arlinger sible for M
eniere’s Disease; however, studies have identified ELH 1998, Ferraro 2010). This article is restricted to ET ECochG test CONTACT Desmond A Nunez [email protected] Division of Otolaryngology Head and Neck Surgery, Diamond Health Care Centre, University of British Columbia, 2775 Laurel Street, Vancouver V5Z 1M9, Canada Supplemental data for this article can be accessed here. ß 2019 British Society of Audiology, International Society of Audiology, and Nordic Audiological Society 2 A. AYUB ET AL. because it is noninvasive, can be performed without physician Exclusion criteria assistance, and is anecdotally more widely used. In an ET ECochG test, a sound stimulus is used to generate a Single case reports, reviews, correspondences, intratympanic elec- potential response, which typically occurs within 5 milliseconds trocochleography studies, or studies that did not provide suffi- of the stimulus onset (early-latency auditory evoked potential) cient details to calculate the sensitivity and specificity of the tests (Minaya and Atcherson 2015). Three elements are generated in reported were excluded. ECochG recordings: the cochlear microphonic (CM), the coch- lear summating potential (SP) and the auditory nerve action Appraisal tools potential (AP) (Minaya and Atcherson 2015). A 2009 survey of primarily United States based otologists and The Oxford Centre for Evidence-Based Medicine (March 2009) neurotologists demonstrated the low uptake of ECochG testing levels of evidence were adopted to classify the level of evidence with only 17.5% of otologists and neurotologists using ECochG in the studies reviewed (CEBM 2009). Two authors (LQ and routinely, whilst 45.5% did not use it at all (Nguygen, Harris, DAN) independently assessed the risk of bias in the included and Nguygen 2010). This suggests that the clinical usefulness of studies using the Quality criteria for diagnostic accuracy studies ECochG tests is in doubt. Discrepancies in the reported sensitiv- QUADAS-2 tool (Sch€ unemann et al. 2013). The lowest risk of ity of ECochG tests contribute to the dilemma. The variation is bias rank assigned by either author was adopted for the purposes in part due to intrasubject variation in recorded measurements of this review. A study was judged as high quality if it demon- related to the fluctuating nature of M
eniere’s Disease with results strated low risk of bias in the majority of domains assessed with during symptomatic periods varying from those in nonsympto- the risk of bias tool. matic periods; interstudy differences in diagnostic criteria, and recording techniques; intersubject variation in level (e.g. possible or definite M
eniere’s Disease) and stage (early or late) of Extracted data M
eniere’s Disease. A summating potential response can be The following data were extracted from each article: study absent in up to 35.5% of normal hearing ears without M
eniere’s design, sample size, AAO-HNS classified level of M
eniere’s Disease further reducing the utility of ECochG testing (Grasel Disease diagnostic certainty in the patient group, ECochG et al. 2017). A number of researchers have sought to improve the accuracy recording technique, the diagnostic criteria and the number of of ECochG testing by adding different test criteria, or combining true positive (TP), false negative (FN), true negative (TN), and SP/AP amplitude ratio testing with other audiological tests. This false positive (FP) results in each study (Table 1, Figure 1). The study reviews the recent evidence to determine if, in definite number of TP, FN, TN, and FP results were extracted independ- M
eniere’s Disease patients, combination test strategies consisting ently by two authors (DAN and AA). Disagreements were of electrocochleographically determined summating/action poten- resolved by discussion. The TP, FN, TN, and FP were taken dir- tial (SP/AP) ratio and other audiological measurements are more ectly from the studies; however, if only sensitivity and specificity accurate than the electrocochleographic summating/action poten- were available, these values were imputed from the number of tial (SP/AP) ratio test alone. M
eniere’s Disease patients and controls. Methods Analysis The Pubmed, Cochrane library, and Web of Science databases The AAO-HNS M
eniere’s Disease criteria were used to assign were searched using Keywords: “Electrocochleography”, patients M
eniere’s Disease status dichotomously for analysis pur- “ECochG”, “ECoG”, “M
eniere’s Disease”, and “Idiopathic poses- either definite M
eniere’s Disease (disease present), or con- Endolymphatic Hydrops”. The Boolean operator AND was trols (disease absent) if no definite M
eniere’s Disease diagnosis included between each search term for a more focussed search. or other M
eniere’s Disease diagnoses (including possible and Examples of searches include “Electrocochleography AND probable M
eniere’s Disease). Data from individual articles were M
eniere’s Disease” and “ECochG AND Idiopathic aggregated except for articles without contemporaneous controls. Endolymphatic Hydrops”. No additional filters were added. SAS 9.4 (SAS Institute Inc., Cary, NC) was used for bivariate Articles quoted by included studies were also screened. and summary Receiver Operator Characteristic Curve (sROC) comparison of the distribution of test positive definite M
eniere’s Disease patients identified by SP/AP ratio alone with the distri- Inclusion criteria bution identified by a combination of ECochG test parameters or SP/AP ratio combined with other audiological tests. The test Cross-sectional, retrospective and prospective human studies; uti- negative results were similarly compared. lising the 1995 American Academy of Otolaryngology-Head and Review Manager 5.3 was used to create a forest plot, a hier- Neck Surgery M
eniere’s disease diagnostic criteria; with extra- archical bivariate model and a sROC to graphically display test tympanic electrocochleography testing; and published between January 1, 2002 and December 31, 2017 in the English language performance. were included. This time period was chosen to capture the stud- ies that began to appear in the early 21st century suggesting that Results new measurements such as the area curve ratio or combination audiological criteria improved ECochG diagnostic accuracy Five articles published between 2002 and 2017 satisfied the study (Devaiah et al. 2003). Restricting the search to the last 20 years inclusion and exclusion criteria and were detailed enough for also improves the chance that comparable ECochG equipment analysis (Figure 2, Table 2). Several other articles identified by and techniques were used thus decreasing interstudy heterogen- the literature search were excluded for some of the following rea- eity in ECochG test results. sons. Specificity and/or sensitivity data was not reported, and  INTERNATIONAL JOURNAL OF AUDIOLOGY 3 Table 1. Extracted study characteristics. M
eniere’s Disease Group (size, mean age, Control group (size, age/range, Diagnostic criteria (SP/AP Recording Author Study design ± standard deviation) mean ± standard deviation) amplitude ratio) approach (ECochG) Gerenton et al. (2015) Prospective Definite (73, 49.9 ± 12.5 y) Healthy (30, 20–30 y) 0.4 clicks 17/sec, 95 dB nHL (85 dB nHL if subject complained too intense) Kim et al. (2005) Retrospective Definite (60, 49.9 ± 11.0 y) Possible (32, 46.5 ± 12.5 y) and 0.4 clicks 7.1/sec, 90 dB nHL Probable M
eniere’s Disease (5, 56.8 ± 12.0 y) Lamounier et al. (2017) Prospective Definite (12, 48.41 y) Healthy (12, matched to 0.5 clicks 11.3/sec, 99 dB HL M
eniere’s Disease group) Lee et al. (2011) Prospective Definite (47, 45.7 ± 15.3 y) Non M
eniere’s Disease (other 0.4 clicks 7.1/sec, 90 dB HL vestibular conditions) (41, 41.5 ± 15.1 y) Mammarella et al. (2017) Retrospective Definite (7, not available) Probable M
eniere’s Disease and 0.37 clicks 7.1/sec, 110 dB Non M
eniere’s Disease (other conditions) (27, not available) Figure 1. Forest plot indicating sensitivity and specificity of SP/AP amplitude ratio and combination tests. could not be extracted from 2 papers, Oh et al. (2014) and Satar consisting of an ECochG SP/AP ratio cut-off and other audiometric et al. (2008). B€uki et al. (2010), B€ uki, J€ unger, and Avan (2012), criteria namely Distortion Product Otoacoustic Emission (DPOAE) Martin-Sanz et al. (2013, 2015), Zack-Williams, Angelo, and Yue posture induced shift (Gerenton et al. 2015) and Vestibular Evoked (2012), and Moon et al. (2012) either lacked contemporaneous Myogenic Potentials (VEMPs) (Lamounier et al. 2017). controls or used the contralateral ear as controls. Contralateral Kim et al. (2005) evaluated ECochG SP/AP amplitude ratio ears in patients with M
eniere’s Disease were not accepted as con- findings on 97 M
eniere’s Disease patients classified by AAO- trol group data in this study because as highlighted by B€ uki HNS M
eniere’s Disease diagnostic criteria. The current review’s et al.’s (2012) work some contralateral ‘control’ ears of patients methodology allocated the patients as 60 Definite M
eniere’s with diagnosed unilateral M
eniere’s Disease on review have fea- Disease patients and 37 controls. The SP/AP amplitude ratio was tures of M
eniere’s Disease on history and pure tone audiometry. positive (above 0.4) in 41 and 17, M
eniere’s Disease and control Data from studies such as, Devaiah et al. (2003) and Al- subjects, respectively. The ratio was negative (below 0.4) in 19 Momani et al. (2009), that had no patients classified as definite and 20, M
eniere’s Disease and control subjects, respectively. M
eniere’s Disease were excluded from further analysis since it Mammarella et al.’s (2017) study described the ECochG find- was not possible to determine intrastudy sensitivity and specifi- ings, with a 0.37 SP/AP amplitude cut-off, in 7 definite city of the diagnostic tests for definite M
eniere’s Disease. The M
eniere’s Disease patients, 21 probable M
eniere’s Disease and 6 findings in the definite M
eniere’s Disease group could not be patients with other causes for their ear symptoms including auto- extracted from the combined M
eniere’s Disease group data in immune, sudden sensorineural hearing losses and basilar some articles, such as Hwang et al. (2008). Others included a migraine. All of the nondefinite M
eniere’s Disease patients were number of patients with bilateral M
eniere’s Disease but allocated categorised as controls for the current review. Six M
eniere’s M
eniere’s Disease status by ear rather than by patient, such that Disease patients tested positive and one negative. One control an individual patient could have both definite and possible tested positive and 26 tested negative. M
eniere’s Disease at the same time, including Chung et al. Lee et al.’s (2011) study compared the diagnostic accuracy of (2004). Data from these types of studies were excluded. an ECochG 0.4 SP/AP amplitude ratio with that of a test of Therefore, 5 studies were used to compare the utility of a single CHAMP in MD and Non-MD patients. The CHAMP results ECochG cut-off SP/AP amplitude ratio test (Kim et al. 2005; Lee were excluded for the purposes of this review. The ECochG et al. 2011; Gerenton et al. 2015; Lamounier et al. 2017; SP/AP amplitude ratio alone results of 7 M
eniere’s Disease Mammarella et al. 2017) against combination test strategies patients that tested positive, 26 M
eniere’s Disease patients that 4 A. AYUB ET AL. Figure 2. Systematic review PRISMA study flow chart. Table 2. Test sensitivity and specificity results comparing ECochG SP/AP amplitude cut-off and ECochG in combination tests in definite M
eniere’s Disease patients and controls (amalgamated results of Kim et al. 2005; Lee et al. 2011; Gerenton et al. 2015; Lamounier et al. 2017, and Mammarella et al. 2017). ECochG SP/APa amplitude Test type ratio alone Combination test protocol Total Test result Definite Controls Definite Controls Definite Controls M
eniere’s M
eniere’s M
eniere’s Disease Disease Disease Patients Patients Patients Positive (True positive for 88 21 54 3 142 24 definite M
eniere’s Disease patients, false positive for controls)b Negative (False negative for 97 109 31 25 128 134 definite M
eniere’s Disease patients, true negative for controls)b Total 185 130 85 28 270 158 Sensitivity 47.6% 63.5% Specificity 83.8% 89.3% a Summating potential/action potential. b True Positive and False negative based on AAO-HNS Definite M
eniere’s Disease diagnostic criteria. tested negative, 1 control patient that tested positive, and 37 con- potentials (VEMP) in a series of 12 M
eniere’s Disease patients trol patients that tested negative were amalgamated with the and 12 healthy controls. VEMP tests were classified as positive if other ECochG SP/AP amplitude ratio studies. there were no reproducible waves or there was a greater than Lamounier et al. (2017) evaluated ECochG with a 0.50 SP/AP 34% interaural asymmetry in the amplitude of the response. amplitude ratio cut-off and Cervical Vestibular Evoked Myogenic Lamounier et al. (2017) presented their results by ear but these  INTERNATIONAL JOURNAL OF AUDIOLOGY 5 were analysed by subject, and where a patient had bilateral and flow and timing of the study in the majority (60%) of the M
eniere’s Disease a positive response in either ear was inter- sample (Figure 3). preted as a positive ECochG result and likewise for the control Only English language articles were included. The search subjects. In this study, the SP/AP amplitude ratio was positive in strategy design precluded the identification of articles published 8 and 1 M
eniere’s Disease and control subjects, respectively. The in languages other than English, therefore, the review is at high ratio was negative in 4 and 11, M
eniere’s Disease and control risk of publication language bias. Trial registers were not subjects, respectively. With the addition of VEMP, the SP/AP searched to mitigate this risk. amplitude ratio was positive in 12 and 1 M
eniere’s Disease and control subjects, respectively. The ratio was negative in 0 and 11 M
eniere’s Disease and control subjects, respectively. Heterogeneity Gerenton et al. (2015) compared SP/AP amplitude thresholds Clinical heterogeneity  using a 0.4 cut-off and DPOAE shift with postural change [–18 ,  þ34 ] in 73 definite M
eniere’s Disease patients with 30 controls. Heterogeneity may arise due to differences in M
eniere’s Disease The SP/AP amplitude ratio results in controls were 1 positive, 15 subjects. For the purposes of this review only patients classified negative, and 14 no result stated. The corresponding results in using AAOHNS criteria as having definite M
eniere’s Disease the M
eniere’s Disease subjects were 26 positive, and 47 negative. patients were categorised as such. However, intersubject variation The DPOAE findings are extrapolated from the controls used for in M
eniere’s Disease stage, duration, acuity, and severity of establishing a 95% confidence interval normative range as 5% of symptoms is expected to affect diagnostic results (Takeda and 30 being abnormal or 1.5 subjects, which is rounded up to 2. Kakigi 2010). Since only 16 control subjects had SP/AP results, it is likely that Kim et al. (2005) classified definite M
eniere’s Disease patients 1 of these 16 subjects had a positive DPOAE shift. Therefore, the by disease stage based on audiometric results and found that patients in stage 2 were the most likely to have abnormal SP/AP combined DPOAE and M
eniere’s Disease positive and negative ratios (Kim et al. 2005). Mammarella et al. (2017) also classified results are 2, and 14 in the controls; and 42, and 31 in the patients based on audiogram shape but did not make any find- M
eniere’s Disease subjects, respectively. ings related to this classification. The remaining studies reviewed did not indicate M
eniere’s Disease audiometric classified stage Level and quality of evidence and no study clearly addressed the impact of variation in dur- ation and severity of symptoms on the tests. The five studies analysed were all individual cohort studies and The results of all assessment techniques including all tests satisfied the 2 b level of evidence of the Oxford Centre for reviewed in this study depend on individual test subjects’ ves- Evidence-Based Medicine (CEBM 2009). Lee et al.’s (2011), tibular and cochlear function which will, therefore, reflect inter- Lamounier et al.’s (2017), and Mammarella et al.’s (2017) studies subject heterogeneity in the severity and extent of M
eniere’s provided high quality evidence whereas Kim et al. (2005) and inner ear involvement. Gerenton et al. (2015) were of low quality due to their high risk of bias based on the QUADAS-2 tool (Sch€ unemann et al. 2013). Details are provided in Supplemental Material Table 1. Methodological heterogeneity Varying diagnostic cut-off values affect sensitivity and specificity. Bias In this review, three studies (Kim et al. 2005; Lee et al. 2011; and Gerenton et al. 2015) used 0.4 as the SP/AP amplitude cut-off The majority of data is drawn from studies at high risk of bias value for a positive diagnosis. Lamounier et al. (2017) and comprising 275 of the 428 subjects reported in the review. The Mammarella et al. (2017) used 0.5 and 0.37 as the SP/AP ampli- risk of bias in patient selection was either unclear or high in all tude cut-off, respectively. Lower cut-off values are associated studies. The reference standard was at low risk of bias in 40%, with higher sensitivities and lower specificities, while higher cut- unclear risk of bias in 40%, and high risk of bias in 20% of the off values are associated with higher specificities and lower sensi- sample, respectively. There was low risk of bias in the index test tivities (Habibzadeh, Habibzadeh, and Yadollahie 2016). Since Figure 3. Risk of bias summary findings for all included studies. 6 A. AYUB ET AL. the studies used different cut-off values, a subject in one study may have tested positive according to criteria used in that study, but would have tested negative if enrolled in another study using a higher cut off value. This variation is a source of methodo- logical heterogeneity. Heterogeneity in diagnostic test results may also arise due to differences in the time when the test was conducted relative to the subjects last major episode of vertigo. For example, Gerenton et al. (2015) distinguished between symptomatic and asymptom- atic definite M
eniere’s Disease patients. Mammarella et al. (2017) performed ECochG within 24 h of making an initial diagnosis of M
eniere’s Disease though they did not state how this corre- sponded to the time of the subjects last major vertigo episode. Other studies did not specify when the tests were conducted. On first principles the sensitivity of diagnostic M
eniere’s Disease tests will vary with the degree of endolymphatic hydrops and the latter in turn with the severity of M
eniere’s Disease symptoms hence test protocols that vary with respect to the presence of symptoms will yield heterogeneous results. There is also heterogeneity in the types of tests included in the combination category for this review as one study investi- gated DPOAE and the other VEMP. DPOAE and VEMP tests like ECochG are not only used for diagnosing M
eniere’s Disease, therefore, the presence of other hearing and balance disorders can confound the results (McMahon et al. 2008). This review Figure 4. Hierarchical bivariate model. Sensitivity and specificity summary point does not seek to distinguish between or identify which combin- estimate are plotted as the solid dots. 95% confidence interval cannot be shown ation of tests are superior. for the combination studies due to the limited number of studies. Statistical heterogeneity Statistical heterogeneity cannot be calculated in a meta-analysis of diagnostic test accuracy, as there is currently no I2 counterpart (Bossuyt et al. 2013). Heterogeneity will be overestimated and differences caused by diagnostic cut-off effects will not be explained if I2 values are calculated independently for sensitivity and specificity (Bossuyt et al. 2013). Hence we are unable to comment on the effect of statistical heterogeneity in this review. Analysis The between-study variability was assumed to be not dependent on the individual study protocols for the bivariate model sensi- tivity and specificity analysis, since the between-study variability could not be estimated for the only two combination studies included. The 80% [95% confidence interval (CI); 28%, 97%] sensitivity point estimate for combination test strategies whilst greater was not statistically significantly different to the 52% [95% CI; 23%, 80%] sensitivity point estimate for single ECochG cut-off SP/AP amplitude ratio tests studies (p ¼ 0.248). The spe- cificity point estimates for combination studies and single SP/AP ratio studies were similar at 91% [95% CI; 67%, 98%] and 91% [95% CI; 43%, 99%], respectively (p ¼ 0.969). The result is illus- trated in Figure 4. The bivariate model is best used when the Figure 5. sROC curve. cut-off values are the same for all studies; however, two studies used a different SP/AP amplitude ratio cut-off value to the other three, limiting the precision of these findings. demonstrates improved discrimination between diseased and The hierarchical sROC (HSROC) model can account for dif- non-diseased subjects (Macaskill et al. 2010). While both the ferent study cut-off values. HSROC analysis like the bivariate combination of tests and the SP/AP ratio amplitude ratio model analysis found that the combination of tests increases ECochG test are unsurprisingly better than an uninformative diagnostic accuracy, though not statistically significantly test, the combination of tests curve is further toward the upper (p ¼ 0.407) (Figure 5). In a sROC plot, the diagonal dashed left (Figure 4). However, HSROC analysis is limited by only two line indicates an uninformative test (Macaskill et al. 2010). As a combination studies being included. The most flexible HSROC curve moves toward the upper left corner of the plot, the test model cannot be determined with certainty and in the present  INTERNATIONAL JOURNAL OF AUDIOLOGY 7 analysis assumes that the shape parameter of the ROC curve is suggesting that both ECochG SP/AP amplitude ratio tests alone the same for both single ECochG cut-off SP/AP amplitude ratio and combination test protocols are superior to an uninformative studies and combination test studies, and that between-study test in their ability to identify patients with definite variability is not dependent on the test study protocols utilised. M
eniere’s Disease. The limitations of this review include the small number of studies (n ¼ 5), interstudy diagnostic protocol heterogeneity, and Discussion the high or unclear risks of bias in subject selection and study An increase in the SP/AP amplitude ratio above a set cut-off standards in published studies. These make it challenging to value is the established ECochG M
eniere’s Disease diagnostic apply standard meta-analytic techniques or alternative statistical parameter. However, this test is not in wide clinical use partly methods to the data aggregated from several studies. For because of low accuracy with reported sensitivity and specificity example, the bivariate model is best suited for the analysis of ranges of, 21–96% (Lee et al. 2011; Chung et al. 2004) and studies that have the same diagnostic cut-off value. This was true 71.2–97% (Hwang et al. 2008; Lee et al. 2011), respectively. for the three studies that used an SP/AP ratio diagnostic value of Lower cut off values generally result in a higher sensitivity and 0.4, but Lamounier et al. (2017) used a 0.5 diagnostic ratio and lower specificity, whilst higher cut off values have the reverse Mammarella et al. (2017) used 0.37. In addition, no attempt was effect (Habibzadeh et al. 2016). Anecdotally difficulty in eliciting made at subgroup analysis of individual test strategies as the an ECochG wave form in some otherwise normal ears also advantage of the large sample size achieved by combining several detracts from the usefulness of the test. studies would be lost. The English language restriction adopted Recently a number of authors have tried to improve ECochG in the current review’s literature search may also have reduced diagnostic accuracy by combining the SP/AP amplitude ratio the yield of studies. This can also be addressed in future reviews. with other ECochG or audiological measurements. The work on There is insufficient evidence to conclude that ET ECochG combining the SP/AP amplitude ratio with other audiological SP/AP amplitude ratio criteria when combined with other tests has yielded early promising results. Gerenton et al. (2015) ECochG and/or audiological measurements achieve a higher reported a sensitivity of 94% and a specificity of 81% when diagnostic accuracy for definite M
eniere’s Disease compared to a ECochG SP/AP amplitude thresholds and DPOAE postural single SP/AP amplitude ratio measurement. More studies are rec- threshold shifts are used to diagnose M
eniere’s Disease in a ommended to arrive at a definitive conclusion. Future studies group of definite M
eniere’s Disease patients in an active phase of the disease. These results are not generalisable to the more com- should continue to investigate differences in the diagnostic mon practice where investigations are not necessarily undertaken accuracy of single ET ECochG SP/AP amplitude ratio tests com- in close proximity to major vertigo attacks and Gerenton et al.’s pared with combined ECochG and/or other audiological meas- (2015) test accuracy was poorer in the group not assessed urements. These studies ideally will detail the approach to acutely. Lamounier et al. (2017) reported a high specificity of subject recruitment, utilise a 0.4 SP/AP ECochG amplitude ratio 84.6% and 100% using VEMP and ECochG tests separately for diagnosis in large samples of definite M
eniere’s Disease though the sensitivity of the individual tests was not high. patients and contemporaneous controls. The problems of bias The differing evidence and results obtained in individual can be addressed by future researchers registering diagnostic studies make it challenging to determine if the combination study trial protocols with registries such as ClinicalTrials.gov. ECochG test protocols are truly diagnostically more accurate The trial protocols should address blinding, randomisation and than the single SP/AP amplitude ratio in M
eniere’s Disease espe- clinical heterogeneity biases. The latter will likely require multi- cially as there are interstudy variations in M
eniere’s Disease diag- institutional recruitment. The accrual of subject data from such nostic criteria used. Hence this review amalgamated the results high quality studies may allow sub-group analysis to determine if of five studies reporting AAO-HNS criteria determined definite a specific combination of ECochG SP/AP amplitude ratio and M
eniere’s Disease patients in comparison with controls. The other audiological test is superior to other combinations. controls consisted of patients with normal ears, other audio-ves- The clinically important conclusion of this review is that both tibular diseases, possible and probable M
eniere’s Disease which ECochG and combination tests are better than noninformative reflects the range of normal and pathological cases that the diagnostic tests in identifying patients with M
eniere’s disease and ECochG test is required to assess to identify true M
eniere’s thus it is reasonable to continue to use extratympanic EchoG for Disease sufferers from in practice. M
eniere’s disease diagnosis. The inability to conclude albeit sub- Three of the five included studies were of high quality reflect- ject to limitations that audiological tests in combination with ET ing some of the best currently available evidence relevant to the ECochG SP/AP amplitude ratio criteria improve the diagnostic review’s question, despite the generally high risk of bias in stud- accuracy of ET ECochG, means that clinicians are advised to use ies in this area. other criteria beyond improving test accuracy to justify multitest The bivariate and sROC analysis of the combined data from protocols. Additionally those centres that do or are contemplat- the five studies independently arrived at the same conclusion ing adopting multitest protocols should consider doing so as part supporting its validity. Specifically, the combination of tests of a diagnostic clinical trial of multitest protocols compared to record higher diagnostic sensitivities, but these are not statistic- ally significantly different to that achieved by SP/AP ECochG stand alone extratympanic ECochG. In this way, sufficient robust amplitude ratio tests alone in the diagnosis of definite M
eniere’s evidence will be obtained to determine if combinations of tests Disease. Additionally, the combination of diagnostic tests did not confer an advantage in the diagnosis of M
eniere’s Disease. result in the anticipated loss of specificity. Indeed, measured spe- cificity at 89.3% was higher though not statistically different from that with the single SP/AP amplitude ratio measurement. Disclosure statement However, the sROC analysis of the amalgamated data from the five studies contrasts with Oh et al.’s (2014) conclusion in No potential conflict of interest was reported by the authors. 8 A. AYUB ET AL. Funding Lacour, M., P. H. van de Heyning, M. Novotny, and B. Tighilet. 2007. “Betahistine in the Treatment of M
eniere’s disease.” Neuropsychiatric University of British Columbia: access to databases and journals. Disease and Treatment 3 (4): 429–440. Lamounier, P., T. S. de Souza, D. A. Gobbo, and F. Bahmad. Jr 2017. “Evaluation of Vestibular Evoked Myogenic Potentials (VEMP) and Electrocochleography for the Diagnosis of M
eniere’s Disease.” Brazilian Journal of Otorhinolaryngology 83 (4): 394–403. 2016.04.021 doi:10.1016/ References j.bjorl. Lee, J. B., S. J. Choi, K. Park, H. Y. Park, J. J. Hong, E. Hwang, H. J. Lee, H. AAOO and AAO-HNS. 1992. “VII: The AAOO 1972 and AAO-HNS 1985 Guidelines for Reporting Treatment Results in M
eniere’s Disease.” Acta Chan, and Y. H. Choung. 2011. “Diagnostic Efficiency of the Cochlear Otolaryngologica 112 (497): 141–152. doi:10.3109/00016489209136849. Hydrops Analysis Masking Procedure in M
eniere’s Disease.” Otology and Al-Momani, M. O., J. A. Ferraro, B. J. Gajewski, and A. Gregory. 2009. Neurotology 31 32 (9): 1486–1491. doi:10.1097/MAO.0b013e318235586c. “Improved Sensitivity of Electrocochleography in the Diagnosis of Lempert, T. 2012. “Recurrent Spontaneous Attacks of Dizziness.” Continuum M
eniere’s Disease.” International Journal of Audiology 48 (11): 811–819. 18 (5): 1086–1101. doi:10.1212/01.CON.0000421620.10783.ac. doi:10.3109/14992020903019338. Macaskill, P., C. Gatsonis, J. J. Deeks, R. M. Harbord, and Y. Takwoingi. Baloh, R. W. 2001. “Prosper M
eniere and His Disease.” Archives of Neurology 2010. “Cochrane Handbook for Systematic Reviews of Diagnostic Test 58 (7): 1151–1156. doi:10.1001/archneur.58.7.1151. Accuracy. Version 1.0. Chapter 10”. The Cochrane Collaboration. Bossuyt, P., C. Davenport, J. Deeks, C. Hyde, M. Leeflang, and R. Scholten. Accessed May 18 2018. http://methods.cochrane.org/sdt/ 2013. “Cochrane Handbook for Systematic Reviews of Diagnostic Test Mammarella, F., M. Zelli, T. Varakliotis, A. Eibenstein, C. M. Pianura, and Accuracy Version 0.9”. Chapter 11. The Cochrane Collaboration. G. Bellocchi. 2017. “Is Electrocochleography Still Helpful in Early Accessed May 18 2018. http://methods.cochrane.org/sdt/ Diagnosis of M
eniere’s Disease?” Journal of Audiology and Otology 21 (2): B€ uki, B., H. J€ unger, and P. Avan. 2012. “Cochlear Function in M
eniere’s 72–76. doi:10.7874/jao.2017.21.2.72. Disease.” International Journal of Audiology 51 (5): 373–379. doi:10.3109/ Martin-Sanz, E., J. Esteban-Sanchez, L. Rodriganez-Riesco, and R. Sanz- 14992027.2011.645508. Fernandez. 2015. “Transitory Effect on Endolymphatic Hydrops of the B€ uki, B., M. Platz, T. Haslwanter, H. J€ unger, and P. Avan. 2010. “Results of Intratympanic Steroids for M
eniere’s Disease.” The Laryngoscope 125 (5): Electrocochleography in M
eniere’s Disease after Successful Vertigo 1183–1188. doi:10.1002/lary.25057. Control by Single Intratympanic Gentamicin Injection.” Audiology and Martin-Sanz, E., C. Z. Luzardo, L. Rodriganez-Riesco, T. M. Patino, and R. Neurotology 16 (1): 49–54. doi:10.1159/000312538. Sanz. 2013. “The Use of Electrocochleography to Monitor the Response of CEBM Center for Evidence-Based Medicine. 2009. Oxford Centre for M
eniere’s Disease Patients to Intratympanic Steroids.” Acta Evidence-Based Medicine. Levels of Evidence. Oxford, UK: CEBM. Otolaryngologica 133 (11): 1158–1164. doi:10.3109/00016489.2013.820344. Accessed January 23 2016. http://www.cebm.net/oxford-centre-evidence- McMahon, C. M., R. B. Patuzzi, W. P. Gibson, and H. Sanli. 2008. based-medicine-levels-evidence-march-2009/ “Frequency-Specific Electrocochleography Indicates That Presynaptic and Chung, W. H., D. Y. Cho, J. Y. Choi, and S. H. Hong. 2004. “Clinical Postsynaptic Mechanisms of Auditory Neuropathy Exist.” Ear and Usefulness of Extratympanic Electrocochleography in the Diagnosis of Hearing 29 (3): 314–325. doi:10.1097/AUD.0b013e3181662c2a. M
eniere’s Disease.” Otology and Neurotology 25 (2): 144–149. doi:10.1097/ Minaya, C., and S. R. Atcherson. 2015. “Simultaneous Extratympanic 00129492-200403000-00011. Committee on Hearing and Equilibrium. 1995. “Committee on Hearing and Electrocochleography and Auditory Brainstem Responses Revisited.” Equilibrium Guidelines for the Diagnosis and Evaluation of Therapy in Audiology Research 5 (1): 105. doi:10.4081/audiores.2015.105. M
eniere’s Disease.” Otolaryngology Head Neck Surgery 113 (3): 181–185. Moon, I. J., G. Y. Park, J. Choi, Y. S. Cho, S. H. Hong, and W. H. Chung. doi:10.1016/S0194-5998(95)70102-8. 2012. “Predictive Value of Electrocochleography for Determining Hearing Devaiah, A. K., K. L. Dawson, J. A. Ferraro, and G. A. Ator. 2003. “Utility of Outcomes in M
eniere’s Disease.” Otology and Neurotology 33 (2): Area Curve Ratio Electrocochleography in Early M
eniere’s Disease.” 204–210. doi:10.1097/MAO.0b013e318241b88c. Archives of Otolaryngology–Head & Neck Surgery 129 (5): 547–551. doi: Nguygen, L. T., J. P. Harris, and Q. T. Nguygen. 2010. “Clinical Utility of 10.1001/archotol.129.5.547. Electrocochleography in the Diagnosis and Management of M
eniere’s Eggermont, J. J. 2017. “Ups and Downs in 75 Years of Disease: AOS and ANS Membership Survey Data.” Otology and Electrocochleography.” Frontiers in Systems Neuroscience 11: 2. doi: Neurotology 31 (3): 455–459. doi:10.1097/MAO.0b013e3181d2779c. 10.3389/fnsys.2017.00002. Oh, K. H., K. W. Kim, J. Chang, H. S. Jun, E. H. Kwon, J. Y. Choi, G. J. Im, Ferraro, J. A. 2010. “Electrocochleography: A Review of Recoding S. W. Chae, H. H. Jung, and J. Choi. 2014. “Can we Use Approaches, Clinical Applications, and New Findings in Adults and Electrocochleography as a Clinical Tool in the Diagnosis of M
eniere’s Children.” Journal of the American Academy of Audiology 21 (3): 145–152. Disease during Early Symptomatic Period?” Acta Otolaryngologica 134 (8): doi:10.3766/jaaa.21.3.2. 771–775. doi:10.3109/00016489.2014.907500. Foster, C. A., and R. E. Breeze. 2013. “Endolymphatic Hydrops in M
eniere’s Sass, K., B. Densert, and S. Arlinger. 1998. “Recording Techniques for Disease: Cause, Consequence, or Epiphenomenon?” Otology and Transtympanic Electrocochleography in Clinical Practice.” Acta Oto- Neurotology 34 (7): 1210–1214. doi:10.1097/MAO.0b013e31829e83df. Laryngologica 118: 17–25. doi:10.1080/00016489850155071. Gerenton, G.,. F. Giraudet, I. Djennaoui, Y. Pavier, L. Gilain, T. Mom, and Satar, B., S. Karahatay, D. Sen, E. Cekin, and H. Birkent. 2008. “Analytic P. Avan. 2015. “Abnormal Fast Fluctuations of Electrocochleography and View to Concordance between Electrocochleography and Caloric Test in Otoacoustic Emissions in M
eniere’s Disease.” Hearing Research 327: M
eniere’s Disease.” European Archives of Oto-Rhino-Laryngology 265 (2): 199–208. doi:10.1016/j.heares.2015.07.016. 159–165. doi:10.1007/s00405-007-0425-7. Grasel, S. S., R. M. O. Beck, R. S. C. Loureiro, A. C. Rossi, E. R. de Almeida, Sch€unemann, G., J. Brozek, G. Guyatt, and A. Oxman, eds. 2013. GRADE and J. Ferraro. 2017. “Normative Data for TM Electrocochleography Handbook. Hamilton, Ontario, Canada: The Grade Working Group. Measures.” Journal of Otolaryngology 12 (2): 68–73. doi:10.1016/ Accessed May 13 2018: http://gdt.guidelinedevelopment.org/app/hand- j.joto.2017.04.005. Habibzadeh, F., P. Habibzadeh, and M. Yadollahie. 2016. “On Determining book/handbook.html the Most Appropriate Test Cut-off Value: The Case of Tests with Takeda, T., and A. Kakigi. 2010. “The Clinical Value of Extratympanic Continuous Results.” Biochemia Medica 26 (3): 297–307. doi:10.11613/ Electrocochleography in the Diagnosis of M
eniere’s Disease.” Journal for BM.2016.034. Oto-Rhino-Laryngology and Its Related Specialties 72 (4): 196–204. doi: Hwang, J. H., H. C. Ho, C. J. Hsu, W. S. Yang, and T. C. Liu. 2008. 10.1159/000315552. “Diagnostic Value of Combining Bilateral Electrocochleography Results Zack-Williams, D., R. M. Angelo, and Q. Yue. 2012. “A Comparison of for Unilateral M
eniere’s Disease.” Audiology and Neurotology 13 (6): Electrocochleography and High-Pass Noise Masking of Auditory 365–369. doi:10.1159/000136155. Brainstem Response for Diagnosis of M
eniere’s Disease.” International Kim, H. H., A. Kumar, R. A. Battista, and R. J. Wiet. 2005. Journal Audiology 51 (10): 783–787. doi:10.3109/14992027.2012.699199. “Electrocochleography in Patients with M
eniere’s Disease.” American Zhang, Y., B. Liu, R. Wang, R. Jia, and X. Gu. 2016. “Characteristic of the Journal of Otolaryngology 26 (2): 128–131. doi:10.1016/ Cochlear Symptoms and Functions in M
eniere’s Disease.” Chinese Medical j.amjoto.2004.11.005. Journal 129 (20): 2445–2450. doi:10.4103/0366-6999.191767.