IDENTIFICATION OF THE RESPONSE

The technical and interpretive expertise required for ERA tests is greater than for conventional audiometry, Although they are objective tests, the judgement of the presence of the response and measurement of the diagnostic parameters are subjective. Technical experience is essential, especially as some of the tests are more difficult than others, and failure to recognize the responses may lead to inadequate evaluation of the results.

Automatic response recognition is desirable and is possible within certain limitations, but, ultimately, there is currently no substitute for an experienced tester. Furthermore, poor-quality recording is unlikely to be compensated for by automatic signal processing.

Auditory-evoked potentials are the averages from ongoing bio-electrical activity and ambient electrical noise. The signal-to-noise ratio is one of the most important factors for good-quality recordings, and it is up to the tester to decide the intensity of the stimuli, number of sweeps, number of replications, artefact rejection limits, and other technical recording factors, although some laboratory systems continue averaging until a certain S/N ratio is attained. Both physiologic and non-physiologic electrical noise can have an effect on test duration, threshold sensitivity, and threshold variability. Nevertheless, even in non-ideal conditions, an adequate response may be obtained (Fig. 11.1).

Passive co-operation by the patient is adequate for most non-invasive ERA tests. However, natural sleep in babies, sedation, or general anaesthesia may be required to keep the patient still and reduce excessive myogenic activity. In many anxious patients, an oral dose of diazepam reduces muscle activity in the scalp, around the jaws, and in the neck.

Electrical interference can considerably reduce the quality of the recording. It may be a problem when the tests are carried out in neonatal units or operating theatres. All non-essential electrical equipment, especially fluorescent lights, should be switched off.

Transtympanic electrocochleography has the best signal-to-noise ratio of the early responses, and detection and interpretation are comparatively easier than in other evoked potentials. Transtympanic ECochG can be carried out with only local anaesthesia in a co-operative adult, but in a child it requires general anaesthesia. Fortunately, sedation and general anaesthesia do not degrade the response.

Interpretation of the SVR may be more difficult, as there may be myogenic artefacts, as well as contamination by prominent alpha and theta rhythmical activity on the EEC The subject’s eyes should be open in order to reduce alpha rhythm. The SVR is known to be a stimulus-orientated response, and is usually recorded easily in a patient who anticipates the stimulus. Natural sleep and general anaesthesia have an effect on both EEG and SVR, making identification of the response at the threshold more difficult.

Interpretative skills for MLR in conscious co-operative patients are similar to those needed for SVR. The response is also affected by general anaesthesia. A clearer response can be achieved by stimulating at a rate of 40 stimuli/s, which is called the 40 Hz MLR response.

At present, ABR is used more often than other ERA tests in infants and in neuro-otologic patients. ABR is not affected by sedation or general anaesthesia, and myogenic activity can thus be reduced to improve the quality of the recording. Accurate measurement of the latencies of the peaks of the response is required for neuro-otologic testing, and the latencies should be compared with normative data, using similar stimulation and recording parameters.

Myogenic auditory-evoked potentials, such as the postauricular muscle response (PAM), vary with the state of muscular tonus at a particular time, and, therefore, are less suitable for audiometric assessment than neurogenic auditory-evoked potentials.

CLINICAL CONSIDERATIONS OF THE PROCEDURE

ECochG requires the placement of the electrode by an otologist through the tympanic membrane. In any patient (i.e. children, the mentally handicapped, etc.) who is incapable of lying still or who cannot be trusted to avoid turning onto the electrode, general anaesthesia is required.

For all other test situations, ERA can be performed by the technical staff in conscious patients. A relaxed state is required for all tests except the PAM, and there must be minimal body movement, scalp or neck muscle movement, and jaw clenching. The best recordings are obtained in relaxed patients in a supine position. Eye closure, natural sleep, and general anaesthesia improve recordings of ABR, but may have an adverse effect on the detection of SVR, which is thought to be of cortical origin. Therefore, for SVR, the patient should be awake with his eyes open.

If the patient is hyperactive and not testable with SVR because of gross myogenic electrical artefacts, it is unlikely that good responses will be obtained by trying to record ABR or MLR in similar conditions, although PAM recording may be successful.

Natural sleep after a feed is an ideal condition for testing neonates with ABR and MLR. In older infants, sedation may be required, which can be administered orally or per rectum. For children who are difficult to test, including those that are hyperactive or have brain damage, mental handicap, or behavioural problems, sedation may not be adequate, and they may need light general anaesthesia in order to achieve good recording conditions. It is unnecessary to use muscle relaxants, although they can improve the recordings further. Some clinics have reported that transtympanic ECochG in adults may cause momentary discomfort when placing the electrode, but most patients tolerate transtympanic needle ECochG well, even without local anaesthesia. Local anaesthetic sprays and iontophoresis have been used to anaesthetize the tympanic membrane, but their effectiveness is variable.

The duration of testing, especially for threshold assessment, is an important factor in the clinical application of auditory-evoked potentials. ECochG using broad-band clicks requires the least testing time. Click-evoked ABR requires more time than does ECochG, but the testing time can be reduced by increasing the stimulation rate to 30/s or more, without compromising threshold estimation. However, the waveform of the response may be affected. For example, the screening time for the threshold using click-evoked ABR in a sleeping neonate typically takes about 20 minutes. Frequency-specific responses from ECochG and ABR can be obtained, but it is a very time-consuming procedure and may not be practical in every circumstance. A frequency-specific air-conduction SVR audiogram, at four frequencies in both ears, in an adult or adolescent can be completed in about an hour. The 40 Hz MLR compares similarly in testing time with SVR; however, there has been less validation in cases of hearing loss.

The risks associated with ERA should be considered. Complications may occur if general anaesthesia is used, and those compromising the laryngeal airway in children are serious. (Crowley et al 1976) reported that 3 out of 603 patients known to have had general anaesthesia experienced laryngospasm. ECochG results from different centres have been assessed in 2256 patients, of whom 13 who had local or general anaesthesia experienced nausea, vomiting, and vertigo. Acute otitis media was reported in 2 out of 1875 patients following transtympanic ECochG, which is a risk of about 0.1% (Crowley et al 1976). Out of more than 3000 transtympanic ECochG reported by Gibson et al (1983), 1 patient complained that his trigeminal neuralgia was worse for 2 months following the test, and another patient complained that his tinnitus had become louder.