Abstract
Background Tinnitus can be related to many different etiologies, such as hearing loss or a noise trauma, but it also can be related to the somatosensory system of the cervical spine. The diagnosis of cervicogenic somatic tinnitus (CST) is made when the predominant feature is the temporal coincidence of appearance or increase of both neck pain and tinnitus.
Objective The aim of this study was to assess the diagnostic value of clinical cervical spine tests in people with CST.
Design A cross-sectional study was conducted.
Setting The study was conducted at a tertiary referral center.
Patients Consecutive adult patients with chronic subjective nonpulsatile tinnitus were included. Exclusion criteria were vertigo, Ménière disease, middle ear pathology, intracranial pathology, cervical spine surgery, whiplash trauma, and temporomandibular dysfunction.
Measurements A full ear, nose, and throat examination was conducted to classify patients into CST and non-CST groups. The physical therapist examination included completion of the Neck Bournemouth Questionnaire (NBQ) and the following clinical cervical spine tests: manual rotation test, adapted Spurling test (AST), trigger point tests, and tests for strength and endurance of the deep neck flexors.
Results Eighty-seven patients with tinnitus were included, of whom 37 (43%) were diagnosed with CST. The diagnosis of CST becomes less likely with NBQ scores of <14 points (sensitivity of 80%, likelihood ratio [LR] of 0.3, and posttest probability of 19%). Absence of trigger points corresponded to an LR of 0.3, a sensitivity of 82%, and a posttest probability of 22%. A positive manual rotation test and AST indicate a higher probability of CST (LR of 5, specificity of 90%, and posttest probability of 78%).
Limitations A limited number of clinical cervical spine tests were used in this study. Although tests with good validity and reliability were included, additional tests could provide more information on cervical spine dysfunction in patients with CST.
Conclusions Clinical cervical spine tests can support the diagnostic process for CST. An NBQ score of <14 points and the absence of trigger points can help to exclude CST. In contrast, a positive manual rotation test and AST can help to include CST. In future studies, these tests should be included in a multidisciplinary assessment of patients with suspected CST.
Tinnitus is the phantom sensation of sound in the absence of overt acoustic stimulation.1 It occurs in 10% to 15% of the adult population.2 Tinnitus can be related to many different etiologies, such as hearing loss or a noise trauma. Within the group of patients with chronic subjective nonpulsatile tinnitus, a subgroup can be defined where the tinnitus is related to the somatosensory system of the cervical spine.3 This type of tinnitus is named cervicogenic somatic tinnitus (CST).
The existence of CST is supported by several animal studies, which have shown connections between the dorsal column of the spinal cord and the cochlear nuclei.4,5 The axons of the dorsal column originate from the C1–C8 dorsal roots of the spinal cord. Stimulation of the C2 dorsal root ganglion in particular generates responses from cells in the cochlear nuclei.6 Additionally, Matsushima et al7 demonstrated that tinnitus improved in 52% of the patients after an occipital nerve block. Other recent studies in humans showed that, in some patients, tinnitus could be evoked or modulated by input from the somatic system, for instance, by forceful muscle contractions of the head, neck, and limbs and pressure on myofacial trigger points.8–10 These findings might explain the ability of some patients with tinnitus to modulate their tinnitus by certain head or neck movements. For example, some patients indicate that their tinnitus worsens when performing a combined cervical spine extension and rotation.11
As tinnitus often co-occurs with neck complaints,11 patients regularly seek help through cervical spine treatment. In these cases, physical therapy treatment can be considered as a treatment option, as suggested in several case studies.3 In previous research,11 a group of patients with CST was characterized by: a positive score on the Neck Bournemouth Questionnaire (NBQ), a positive manual rotation test and adapted Spurling test (AST), and the presence of sensitive trigger points. To date, however, little information is available on the diagnostic value of these tests and the causal relationship with tinnitus.
Although animal studies have demonstrated the existence of neural connections that can cause CST, the diagnosis is challenging. The diagnosis of CST is made when at least one of the following occurrences prior to the onset of tinnitus is present: (1) evident history of head or neck trauma; (2) tinnitus association with movement of or pressure on the cervical spine; (3) recurrent pain episodes in the head, neck, or shoulder girdle; (4) temporal coincidence of appearance or increase of both pain and tinnitus; and (5) increase of tinnitus during inadequate postures during rest, walking, working, or sleeping.3 The most important diagnostic criterion is the temporal coincidence of onset or increase of both neck pain and tinnitus. As these diagnostic criteria are based on medical history, the accuracy of the diagnosis depends partly on the memory of the patient. Therefore, all other causes of tinnitus need to be ruled out before CST can be diagnosed, which is very time-consuming and expensive. Thus, a set of clinical cervical spine tests that can be used in the CST diagnosis in addition to the other diagnostic criteria would be very useful.
Therefore, the aim of this study was to determine the diagnostic value of a set of clinical cervical spine tests in diagnosing CST.
Materials and Method
Participants
A prospective cohort study of patients with chronic subjective tinnitus was performed in Antwerp University Hospital. During a 6-month period, all consecutive patients with tinnitus who presented themselves at the tinnitus clinic were asked to participate in the study.
Patients were included if they had chronic subjective tinnitus (existing for >3 months), regardless of the presence of neck complaints. Patients with vertigo, Ménière disease, middle ear pathology, or intracranial pathology were excluded because these conditions result in tinnitus subtypes that are easily diagnosed. Therefore, these tinnitus subtypes can easily be distinguished from CST. Patients also were excluded after cervical spine surgery, as some of the cervical spine tests in the examination might be contraindicated in these patients. Patients with whiplash-associated disorders also were excluded because of the heterogeneity of the cervical spine complaints and the risk of influence due to central sensitization in these patients. Finally, patients with a temporomandibular dysfunction were excluded, because these patients compose a different somatic tinnitus subtype that was beyond the interests of this study.
Diagnosis
All patients were investigated by means of medical history; complete ear, nose, and throat (ENT) examination with microscopic otoscopy and audiometry (comprising standard hearing tests and tests to objectify the loudness, pitch, and unilateral, bilateral, or central character of the tinnitus); tinnitus assessment (comprising assessment of tinnitus loudness using a visual analog scale [VAS] and of tinnitus annoyance using the tinnitus questionnaire); temporomandibular and cervical spine investigation; and brain magnetic resonance imaging (MRI) to obtain a tinnitus workup. Based on this complete ENT and audiologic examination and the brain MRI, the ENT physician diagnoses the tinnitus subtype. The diagnosis of CST is made when the predominant feature is the temporal coincidence of onset or increase of both neck pain and tinnitus.
Physical Examination
The examination of the cervical spine was performed by a physical therapist with master's degrees in physical therapy and manual therapy and 9 years of experience in musculoskeletal assessment and treatment. At all times, the physical therapist and the participants remained blinded to the ENT diagnosis. The physical examination consisted of a set of clinical tests of the cervical spine and the NBQ.
First, the left and right passive rotation movements of the cervical spine was investigated using the manual rotation test.12 Using this test, the quality of the passive rotation movement on C0–C2 and C2–C7 levels was rated based on 3 parameters: range of motion (hypermobility/normal/hypomobility), end-feel (hard/normal/soft/empty), and pain provocation (VAS score >2 cm). The test was considered positive when, for at least one movement, 2 out of 3 parameters were aberrant. Normal mobility at the C0–C2 segments was set at an estimated 45 degrees of rotation for participants under 40 years of age and at an estimated 35 degrees of rotation for those older than 40 years. At the C2–C7 segments, normal rotation was set at an estimated 25 to 30 degrees of rotation.12
Second, the AST, a segmental provocation test using a combination of cervical extension, lateral flexion, and rotation, was used (Fig. 1). This test was considered positive when, at least on one level, pain was provoked with a VAS score >2 cm.12 Both tests have shown high sensitivity (77.8) and specificity (77.3) in discriminating patients with neck dysfunction from asymptomatic controls.12
Adapted Spurling test.
Third, the strength and endurance of the deep neck flexor muscles were measured using the craniocervical flexion test (CCFT).13 The CCFT is performed with the patient in a supine crook-lying position with the neck in a neutral position such that the line of the face is horizontal and a line bisecting the neck longitudinally is horizontal to the testing surface. An uninflated pressure sensor is placed behind the neck so that it abuts the occiput and is inflated to a baseline pressure of 20 mm Hg. The craniocervical flexion movement is performed gently and slowly as a head-nodding action (as if saying “yes”). The CCFT tests the activation and endurance of the deep cervical flexors in progressive inner range positions as the patient attempts to sequentially target five 2-mm Hg progressive pressure increases from the baseline of 20 mm Hg to a maximum of 30 mm Hg and maintain an isometric contraction at the progressive pressures as an endurance task.13 The strength of the deep neck flexor muscles was scored as aberrant when the participant could not maintain the isometric contraction at all 5 pressures for 2 to 3 seconds without compensations or abnormal movement patterns. For the endurance test, a minimal isometric contraction of 10 seconds at all 5 pressures was used as a cutoff point. This cutoff point was not determined using a receiver operating characteristic curve.
Finally, the tenderness of 16 myofascial trigger points was tested by applying manual pressure. A trigger point was identified as positive when the participant scored more than 2 cm on a VAS for pain. The test was considered positive when at least one trigger point was found to be positive. The locations of the trigger points were determined according to the findings of Teachey et al14 and are displayed in Figures 2 and 3.
Locations of trigger points in the levator scapulae muscle (left) and splenius capitis muscle (right).
Locations of trigger points in the upper trapezius muscle (left) and sternocleidomastoid muscle (right).
In addition to these clinical tests, the NBQ was used to assess self-reported pain intensity, limitations in activities of daily living, depression, and self-control. The NBQ has shown high sensitivity (83.3) and specificity (90.9) in identifying patients with neck disorders.12 The NBQ was considered positive when a participant scored >14/70.
All tests were performed in the sequence as mentioned above, starting with the manual rotation test and ending with the trigger point tests. All of the tests were scored dichotomously as negative or positive, based on the thresholds mentioned above. Additionally, the provocation of changes in tinnitus loudness was queried after each test.
Data Analysis
The data were analyzed using 2 × 2 contingency tables. From the 2 × 2 contingency tables, sensitivity, specificity, likelihood ratio (LR), and pretest and posttest probability values were calculated to discriminate participants with CST from those without CST. The pretest probability is calculated for each test, as due to missing data, this probability can vary slightly. The pretest probability is the probability of the CST diagnosis without the extra information of the cervical spine tests. The posttest probability is the probability of the CST diagnosis taking into account a positive or negative result of a test.
We used IBM SPSS Statistics for Windows (version 22.0, IBM Corp, Armonk, New York) for obtaining the contingency tables. The sensitivity, specificity, LR, and pretest and posttest probability values were calculated using Microsoft Excel for Windows (Microsoft Corp, Redmond, Washington).
Role of the Funding Source
This study was funded by a research grant from the University of Antwerp.
Results
In total, 87 adult patients with tinnitus, with a mean age of 49.7 years (SD=14.4), were included in the study between August 2012 and January 2013. No significant differences in sex (P=.729) or age distribution (P=.07) were found between the CST and non-CST groups. However, a tendency of the participants in the CST group to be slightly older (mean age=53 years, SD=13) than those in the non-CST group (mean age=47 years, SD=15) was noted. The demographic features are shown in Table 1.
Demographic Characteristicsa
In our study sample, 43% of the participants were diagnosed with CST, and 46% complained of a certain degree of neck pain. The average tinnitus loudness was 51 mm (SD=25) on a VAS, the average tinnitus questionnaire score was 41 (SD=17) out of 82 points, and the average hyperacusis questionnaire score was 18 points (SD=8) (Tab. 2). Tinnitus was unilateral in 36% of the participants, bilateral in 48%, and central in 3%. Tinnitus had a high frequency >2,000 Hz) in 86% of the participants, hyperacusis was reported in 56%, difficulties concentrating were reported in 78%, and 60% complained of sleeping problems. No significant differences between the CST and non-CST groups were found regarding the tinnitus characteristics.
Tinnitus Characteristics: Average (SD)a
To quantify the diagnostic value of the clinical cervical spine tests and the NBQ, sensitivity, specificity, LR, and pretest and posttest probability were calculated. An overview of the clinimetric properties of the tests is presented in Table 3. The relevant items are mentioned below.
Clinimetric Propertiesa
For a negative NBQ result (<14/70), the probability of diagnosing a patient with CST decreased from 41% to 19% (negative LR=0.3). Moreover, the sensitivity of the NBQ for diagnosing CST was high (80%), indicating a low risk of false-negative results. On the other hand, the specificity of the NBQ for diagnosing CST was only moderate (62%), indicating a high risk of false-positive results. Likewise, the absence of trigger points reduced the probability of the CST diagnosis from 43% to 22% (negative LR=0.3). A high sensitivity (82%) also was found for this test.
A positive manual rotation test in combination with a positive AST increased the probability of diagnosing a patient with CST from 43% to 78% (positive LR=5). In addition, the specificity of the combination of manual rotation test and AST was high (90%), indicating a low risk of false-positive results. A positive AST, without a positive manual rotation test, still increased the probability of a CST diagnosis to 65% (positive LR=3) with a specificity of 84%. In case one of the clinical tests was able to provoke the tinnitus, the probability of a CST diagnosis increased from 43% to 64%, with a specificity of 90%.
The CCFT strength test showed an acceptable specificity of 74%, but the positive and negative LR values were too close to 1 for the test to be helpful for diagnosing CST. The CCFT endurance test showed overall low sensitivity, specificity, LR, and posttest probability values.
Discussion
The aim of this study was to assess the diagnostic value of clinical cervical spine tests in people with CST. We found that several tests can be useful in the diagnostic process for CST. In patients with suspected CST, the NBQ can be administered first. In case of an NBQ score of <14 points, the diagnosis of CST becomes less likely, as is shown in the negative posttest probability, sensitivity, and negative LR values. Similar results were found for using the NBQ to diagnose patients with neck pain.12 The co-occurrence of tinnitus and cervical spine dysfunction can easily lead to suspected causality. In this case, a substantial cervical spine dysfunction is needed. The NBQ allows the therapist to objectify the cervical spine complaint. As described in a previous study, an NBQ score of >14 points is needed for a substantial cervical spine complaint.12 An NBQ score of <14 points indicates that the cervical spine dysfunction is not substantial enough, and causality can be excluded.
The trigger point test can be added to the NBQ in the diagnostic evaluation of patients with suspected CST. The absence of trigger points will reduce the probability of CST, as is shown in the negative posttest probability, sensitivity, and negative LR values.
When there is a positive NBQ score, the risk of false positives is too high, so a highly specific test, such as the combination of manual rotation test and AST, is necessary to decrease the risk of false positives. When both tests are positive, the diagnosis of CST becomes more likely, as is shown in the positive posttest probability, specificity, and positive LR values. In case of a positive AST, without a positive manual rotation test, the probability of CST still increases, as is confirmed in the positive posttest probability, specificity, and positive LR values. Similar results were previously found for using the manual rotation test and AST to diagnose patients with neck pain and cervical facet joint pain.12,15 Using a combination of tests will increase the diagnostic value, as was demonstrated in the study by Schneider et al,15 who investigated cluster diagnostics in patients with cervical facet joint pain.
The CCFT was not useful in diagnosing patients with CST. This finding is in accordance with the findings of a recent study11 that showed no differences in CCFT scores comparing neck-related symptoms in patients with and without CST. No information is available on differences in CCFT scores between patients with CST and asymptomatic controls, which could be interesting because in other populations with neck-related symptoms, such as patients with cervicogenic headache, a significant difference in CCFT scores could be found between patients and asymptomatic controls.16 Previous research has shown that impairments in patients with neck pain are mainly observed at the lower levels of the test (22, 24, and 26 mm Hg),17 and we also used the higher levels. This use of the higher levels of the test might have caused the absence of differences between the CST and non-CST groups. In future research, 26 mm Hg should be used as a cutoff for normal.
To date, no clinical tests have been described in the literature for diagnosing CST. Therefore, the clinical diagnosis of CST is currently based on the patient's medical history. As several clinical tests have proven to be useful in the diagnostic process for CST, we advise adding the NBQ to the anamnesis. For an NBQ score of <14 points, the CST diagnosis becomes very unlikely, and further cervical spine tests will not be necessary. In case of an NBQ score of >14 points, the ENT can refer the patient to the physical therapist, who can perform the manual rotation test and AST to rule out false-positive CST diagnoses. In this way, the physical therapist can cooperate with the ENT during the diagnostic process.
The study sample was recruited in a tertiary referral center for patients with tinnitus, where all patients underwent a thorough ENT examination for diagnosing the tinnitus subtype. The diagnosis of CST was made when the predominant feature was the temporal coincidence of onset or increase of both neck pain and tinnitus. This diagnosis is more limited than the original criteria for somatic tinnitus of Sanchez and Rocha,3 where the diagnosis is made when at least one of the following occurrences prior to the onset of tinnitus is present: (1) evident history of head or neck trauma; (2) tinnitus association with some manipulation of the teeth, jaw, or cervical spine; (3) recurrent pain episodes in the head, neck, or shoulder girdle; (4) temporal coincidence of appearance or increase of both pain and tinnitus; (5) increase of tinnitus during inadequate postures during rest, walking, working, or sleeping; and (6) intense periods of bruxism during the day or night.
In our study, we decided to exclude patients with whiplash trauma because of the complexity of the pathology and the heterogeneity of the clinical presentation. We also excluded patients with temporomandibular disorders, as the population of interest in this study was the CST population. For the above-mentioned reasons, criteria 1, 6, and part of 2 were not used for the diagnosis of CST in this study. As we wanted to minimize the risk of including patients with tinnitus and cervical spine disorders, but without any relation between the tinnitus and the neck complaints in the CST group, we decided to primarily use criterion 4 for the diagnosis of CST. However, we recognize the possibility that in our non-CST group, some of the patients with significant neck dysfunction actually were patients with CST who did not notice or comment on a coincidental onset of neck pain and tinnitus symptoms.
In our study, the diagnosis of CST was made in 43% of the patients. This number corresponds with the percentages reported by Abel et al.18 They found that 36.7% of the patients complained of neck, jaw, or facial pain. This finding, which was recently confirmed by Ostermann et al19 and Fabijanska et al,20 implies that our study sample was representative for the entire population. The pretest probability to diagnose a patient with CST was generally 43%, but small differences can be noted due to missing data for some of the clinical tests.
No significant differences were found in tinnitus loudness, tinnitus questionnaire scores, hyperacusis questionnaire scores, or hearing impairment between the CST and non-CST groups. These results correspond to the findings of Vielsmeier et al21 in patients with and without temporomandibular joint–related tinnitus and outline the necessity for cervical spine investigation, as the tinnitus characteristics alone cannot differentiate between both populations.
A limited number of clinical tests were used in this study. The manual rotation test and AST were included because they have been proven to have good sensitivity and specificity in the diagnosis of nonspecific neck pain. The CCFT was included because it has proven good reliability22 for measuring deep neck flexor endurance. Additionally, differences in CCFT scores were found when comparing patients with cervicogenic headache and healthy controls.16 The trigger point test was added because this test has proven to be positive in patients with tinnitus.14 The reliability of this test, however, can be questioned. To eliminate the risk of interrater pressure differences in this study, all tests were performed by the same therapist, who had 9 years of experience in clinical examination of the cervical spine. The use of pressure algometry, however, would increase the reliability of this measure, but then the resemblance to daily clinical practice would be reduced. In future research, other highly reliable clinical cervical spine tests should be tested for validity in diagnosing CST, and the effect of cervical spine treatment also should be investigated.
Clinical tests of the cervical spine can support the diagnostic process of patients with CST. An NBQ score of <14 points and the absence of trigger points can help to exclude CST. In contrast, a positive manual rotation test and AST can help to include CST. We advise that these tests should be included in a multidisciplinary assessment of patients with suspected CST.
Footnotes
All authors provided concept/idea/research design and data analysis. Mrs Michiels, Dr Van de Heyning, and Professor De Hertogh provided writing. Mrs Michiels provided data collection. Mrs Michiels, Dr Van de Heyning, and Dr Truijen provided project management. Dr Van de Heyning, Dr Truijen, and Professor De Hertogh provided fund procurement and consultation (including review of manuscript before submission). Dr Van de Heyning provided participants. Dr Van de Heyning and Dr Truijen provided facilities/equipment and institutional liaisons.
Ethical approval for the study was obtained from the local ethics comity.
This study was funded by a research grant from University of Antwerp.
- Received October 13, 2014.
- Accepted May 21, 2015.
- © 2015 American Physical Therapy Association