Study setting
Patients will be recruited at the tertiary tinnitus clinic of the Antwerp University
Hospital (UZA). When tinnitus patients present themselves at the ear-nose-throat department,
a thorough diagnostic evaluation takes place. A systematic medical history is obtained and coexistent symptoms (e.g. hearing loss, sleeping disorders, cervical tension
or pain, or temporomandibular joint dysfunction) are enquired. Apart from the medical history, psychoacoustic characteristics of the tinnitus sound are quantified and the impact
of tinnitus on quality of life is estimated via questionnaires. A thorough clinical
exam containing micro-otoscopy, pure-tone audiometry, and psychoacoustic measures
is performed. Together, these examinations will allow the evaluator to determine whether
the patient meets the inclusion criteria for the study (Table 1). In case inclusion
criteria are met, patients will be informed about the clinical trial and asked to
participate.
Eligibility criteria
Adult patients (male/female) with chronic (> 6 months), non-pulsatile subjective tinnitus
will be included in the study (Table 1). Confounding effects of anxiety and depression
will be controlled by using the Hospital Anxiety and Depression Scale (HADS). In case
of uncertainty concerning the mental state or possible active psychiatric disorders,
the patient will first be evaluated by the psychiatric diagnosis team of the Antwerp
University Hospital.
Table 1. Inclusion and exclusion criteria.
Interventions
Participants are invited for a first clinical testing at the site (baseline measurement,
T0). This testing consists of pure tone audiometry, psychoacoustic tinnitus analysis
(loudness and pitch matching of the tinnitus), speech-in-quiet (SPIQ) and speech-in-noise
(SPIN) measurements, questionnaires, and explorative research tests (Repeatable Battery
for the Assessment of Neuropsychological Status, adjusted for Hearing impaired individuals
(RBANS-H) and Cortical Auditory Evoked Potentials (CAEP)). The entire test protocol
will take approximately 2.5 - 3 hours. The patient visit and follow-up schedule is
illustrated in Fig. 1. After the baseline measurements patients are randomized into
the HD-tDCS group or the sham control group. Fig. 2 represents an overview of all
interventions and outcome measures in accordance with the Standard Protocol Items:
Recommendations for Interventional Trials (SPIRIT) 2013 guidelines [29].
When patients are enrolled in the study, 6 sessions of sequential dual-site HD-tDCS
(or sham stimulation for the control group) are planned within 3 weeks’ time (2 times
per week, with at least one day between sessions: Ta - Tf). Stimulation (or sham) is provided for 30 minutes each session. The electrodes are
positioned at the rDLPFC and the LTA. The positioning is according to the 10/20 international
system for EEG electrode placement. For the rDLPFC stimulation, the central anode
will be placed at F4 and the adjoining cathodes at F2, F6, FC4, and AF4 [24]. For
the LTA stimulation, the central anode is positioned between at CP5 with adjoining cathodes at C5, TP7, CP3, and P5. A constant current of 2 mA is applied
for 15 minutes at each site with a fade-in and fade-out of 20s, resulting in a total
stimulation duration of 30 minutes. For the sham stimulation, constant current will
only be applied for 20s as previously described [30]. The direct current is transmitted
by means of five sintered silver/silver chloride (Ag/AgCl) ring electrodes with inner
radius of 6 mm and outer radius of 12 mm and delivered by a battery-driven Soterix
Medical 1x1 tDCS low-intensity stimulator and 4x1 multichannel stimulation adaptor
from Soterix Medical Inc, New York, NY, with a maximum output of 2 mA. The ring electrodes
are stabilized using HD-Electrode holders hooked in the Soterix Medical HD-cap and
filled with EEG electrode gel (Neurax, Bonheiden, Belgium), following the guidelines
for 4x1 HD tDCS stimulation [22].
Immediately after the last HD-tDCS session (Tf) , the patient is instructed to fill in all questionnaires. A follow-up (T1) at three months after the baseline measurements will contain the full clinical study
protocol: pure tone audiometry, tinnitus analysis, SPIQ/SPIN, questionnaires, RBANS-H,
and CAEP. A second follow-up (T2) is planned at six months after baseline measurements, comprising only the complete
set of questionnaires.
[Insert Figure 1 about here]
Except for mild itching or tingling sensations, headaches, or fatigues, no adverse
effects of HD-tDCS treatment are expected. In case patients do report serious side
effects, HD-tDCS treatment will be discontinued.
No concomitant tinnitus treatment will be allowed between baseline measurement and
the three-month follow up. After the three-month time point, participants will be
allowed to undergo additional treatment. In case participants were allocated to the
sham group, they will have the right to undergo an actual HD-tDCS treatment free of
cost after the three-month time point.
[Insert Figure 2 about here]
Outcomes
Primary outcome measure
The primary outcome is the change in the Tinnitus Functional Index (TFI) questionnaire
from baseline to follow-up at three months. The TFI is a self-reported questionnaire
assessing the impact of tinnitus on quality of life [31, 32]. The patient must answer
25 questions on a Likert scale ranging from 0 to 10. For questions 1 and 3, which
are expressed in percentages, the Likert scale is transformed to a 0-10 scale post hoc. The total score is then calculated as the mean of all questions multiplied by 10
and expressed as a number between 0 and 100. Results of the TFI include the total
score and the score on eight subscales: intrusiveness, sense of control, cognition,
sleep disturbance, auditory difficulties, relaxation, quality of life, and emotional
distress. A decrease of 13 points on the total TFI score is considered as a clinically
relevant difference [32]. Based on previous studies, we expect mean TFI scores at baseline of approximately
50 [28].
Secondary outcome measures
Tinnitus Questionnaire (TQ)
The Dutch validated version of the TQ is used to differentiate between emotional and
cognitive distress, auditory difficulties, and self-experienced intrusiveness caused
by tinnitus [33]. The total score can range from 0 to 84, assigning a subject to a
distress category: slight (score = 0 – 30, grade 1), moderate (score = 31 – 46, grade
2), severe (score = 47 – 59, grade 3) and very severe (score = 60 – 84, grade 4) [34].
Speech, Spatial and Qualities of Hearing Scale-12 questionnaire (SSQ12)
The SSQ12 is a short form of the Speech, Spatial and Qualities of Hearing scale [35,
36]. The questionnaire is used in clinical research to measure several aspects of
hearing ability, such as speech comprehension in quiet and noise, localization of
sound, distance and movement, segregation, and listening effort. Responders rate their
ability to do or experience the situation described in each question by marking a
1-10 scale (1 = not at all, 10 = perfectly). Scores of all 12 questions are averaged
to obtain a global SSQ-12 score.
Hospital Anxiety and Depression Scale (HADS)
The Hospital Anxiety and Depression Scale (HADS) is used to detect states of anxiety
or depression [37]. Patients must answer a total of 14 questions, of which 7 belong
to the subscale ‘depression’ and 7 pertain to the subscale ‘anxiety’. Scores of 7
or less on each subscale indicate non-cases. Scores of 8-10 are borderline abnormal
(borderline case), while scores of 11-21 are abnormal (case).
Visual Analogue Scale (VAS)
The VAS enquires the mean loudness and maximum loudness of the tinnitus. The patient
must score their tinnitus loudness on a scale of 0 (absence of tinnitus) to 100 (cannot
be any louder).
Hyperacusis Questionnaire (HQ)
The patient’s hypersensitivity to sound is investigated using the 14-item HQ [38].
The 14 questions assess three dimensions (attentional, social, and emotional). The
answer categories are: "no" (score of 0 points), "yes, a little" (1 point), "yes,
quite a lot" (2 points), and "yes, a lot" (3 points). A total score of 28 or more
indicates clinically significant hyperacusis.
Health Utilities Index (HUI23)
The HUI23 is a family of generic health profiles and preference-based systems for
the purposes of measuring health status, reporting health-related quality of life
(QoL), and producing utility scores [39]. The HUI23 comprises a 15-item questionnaire.
The resulting total health-related QoL score ranges from 0.00 (dead) to 1.00 (perfect
health).
Pure-tone audiometry
Pure tone linear audiometry will be performed according to current clinical standards
(International Organization for Standardization (ISO) 8253-1:2010), using a two-channel
AC-40 audiometer (Interacoustics, Assens, Denmark) in a soundproof booth. Air conduction
thresholds will be measured at 125 Hz, 250 Hz, 500 Hz, 1 kHz, 2 kHz, 3 kHz, 4 kHz,
6 kHz, and 8 kHz using headphones. When air conduction thresholds between 250 Hz and
4 kHz exceed normality levels of 20 dB HL, the bone conduction threshold will be measured
at 250 Hz, 500 Hz, 1 kHz, 2 kHz, 3 kHz, and 4 kHz in order to distinguish between
conductive and sensorineural hearing loss.
Tinnitus analysis
Participants are asked whether they perceive the tinnitus unilaterally, bilaterally,
or centrally, and whether the tinnitus sound is a pure tone, a noise, or a mixture
of different sounds (polyphonic). Psychoacoustic characteristics of the tinnitus sound
are assessed in a soundproof booth. As a psychoacoustic equivalent of frequency, the
tinnitus pitch is obtained by use of a pitch matching technique. A two-alternative
forced choice procedure will be employed using the contralateral ear as the reference
ear. In cases where tinnitus is perceived bilaterally, the choice of ear is made arbitrarily.
Using this technique, an attempt is made to identify the center pitch of the tinnitus.
When multiple tinnitus sounds are perceived, patients are asked to focus on the most
troublesome tinnitus sound. Pairs of pure tones (or noises in case of noise-like tinnitus),
differing by one or more octaves, are presented to the subject, who has to indicate
which of the stimuli best resembles the tinnitus sound. This procedure is repeated
and finer adjustments are made to obtain the closest possible match of the tinnitus
pitch.
Loudness is the perceptual correlate of sound intensity. The tone or noise defined
as the pitch match is presented to the ipsilateral ear (when appropriate) and a loudness
match is made by use of an alternating procedure. Because of compressed dynamic range
frequently present at the tinnitus frequency, final loudness measurements are made
with 1 dB steps. The absolute level of tinnitus loudness is measured in dB hearing
level (dB HL). In addition, a calculation is made to provide a measurement of relative
loudness expressed in dB sensation level (dB SL), which is defined as the level of
the loudness match minus the auditory threshold at the tinnitus frequency.
Changes in speech-in-quiet (SPIQ) and speech-in-noise (SPIN) understanding
Speech comprehension in quiet is investigated using the Dutch NVA lists, which were
developed by the Dutch Society for Audiology (Nederlandse Vereniging voor Audiologie,
NVA) [40]. Each of the four lists consists of twelve monosyllabic words (consonant-vowel-consonant),
of which the first one is used for training. The percentage of correctly identified
items determines the speech recognition score. The lists are presented through headphones.
The Leuven Intelligibility Sentences Test (LIST) is used to measure speech reception
in noise [41]. The long-term average frequency spectrum of the speech signal matches
the spectrum of the noise. The noise level is constant at 65 dB SPL, while the sound
level of the speech signal is changed according to the subject’s response. Lists of
ten sentences are presented through headphones. If the subject correctly repeats the
keywords of a sentence, the sound level of the next sentence is decreased by 2 dB
SPL. In case of an incorrect response, the level is increased by 2 dB SPL. The levels
of the last five sentences of the list and the imaginary 11th sentence are averaged to acquire the speech reception threshold (SRT).
Repeatable Battery for the Assessment of Neuropsychological status, adjusted for Hearing
impaired individuals (RBANS-H)
Cognitive functioning will be evaluated by use of the RBANS [42], adjusted to test
Hearing impaired individuals (RBANS-H) [18]. The RBANS assesses five cognitive domains,
i.e. Immediate Memory, Visuospatial / constructional, Language, Attention, and Delayed Memory, and consists of twelve subtests.
The domain Immediate Memory consists of two subtests. In the subtest List Learning, the same list of 10 words is presented over four trials, whereas the subtest Story Memory consists of a twelve-item short story that is presented twice. After each presentation,
the subject must recall as much of the words or the story as possible. Two subtests
belong to the domain Visuospatial / constructional: a Figure Copy test, in which the subject is asked to copy a geometric figure, and the subtest Line Orientation, during which the subject needs to match two lines according to their orientation.
The domain Language includes the subtests Picture Naming and Semantic Fluency. In the first subtest, ten line drawings are to be named by the subject. In the second
subtest, the subject is given one minute to generate as many examples as possible
from a certain semantic category. The subtests Digit Span and Coding contribute to the domain Attention. In the first subtest, a string of digits is presented, after which the subject is
asked to repeat the digits in the correct order. In the subtest Coding, the subject is asked to complete a page of symbols with the corresponding digits
according to a key on top of the page. The last domain, Delayed Memory, enquires how many items the subject recalls from the subtests List Learning, Story Memory, and Figure Copy.
The RBANS-H is developed especially for the purpose of examining the cognitive function
of individuals with hearing impairment. To reach this purpose a number of adjustments
to the original RBANS have been made. By means of an accompanying PowerPoint presentation,
written explanations are given to support the oral instructions and ascertain that
the participant understands the instruction. In addition, all relevant stimuli are
not only presented orally but also visually. All adjustments were made in accordance
to the RBANS guidelines [43]. The total score pre- and post-therapy will be compared in order to reveal any changes
in different aspects of cognition due to HD-tDCS.
Cortical Auditory Evoked Potentials (CAEP)
Exploratory research is performed towards CAEP as a potential objective measure of
therapy effect. Brain potentials will be elicited using an oddball paradigm during
which the patient is instructed to press a button every time a rare stimulus (2 kHz)
is presented in between frequent stimuli (1 kHz) through shielded headphones (Audio Technica ATH M30x Refaeds). Rare and frequent stimuli will occur with a probability of respectively 20% and 80%
and the rise and fall time of the presented stimuli are both 5 ms. Stimuli are delivered by use of the Software
Presentation TM (Neurobehavioral Systems, Inc). During this task, EEG is recorded
(Micromed TM SD LTM 64 Express) using the interface ‘Gilat Medical TM Event Related
Potentials system’. In total, 31 silver/silver chloride (Ag/AgCl) electrodes are placed
according to the 10-20 Standard International Electrode System referenced to a chin
electrode, with the ground electrode placed on the right mastoid. Vertical electrooculogram
(EOG) is recorded using one electrode located below the right eye. After recording
the EEG is sampled at 1024 Hz with 22-bit A/D resolution and band passed between 0.02 Hz
– 450 Hz (Micromed TM SD LTM 64 Express).
All data will be analysed by one researcher using Gilat Medical TM analysis software
(Karkur, Israel). Recordings will be segmented into time epochs of 2s which are time-locked
to the stimuli. Baseline correction for each trial will be performed using the average
of the 200 ms prior to stimuli onset for each channel separately. For each patient,
all trials will be averaged according to the condition (target and non-target). An
Independent Component Analysis (ICA) algorithm will be used to correct for external
artefacts, such as eye blinks.
The analysis software will be used to calculate the area below the curve and the centre
of gravity of peaks within specific time windows in the average trace for the target
condition. The time window for the different components is determined according to
visual inspection of the average trace over all electrodes. Furthermore, latency average
of the correct response is calculated as the time from stimulus onset to the time
of the button press. Finally, a group average for the target and non-target records
of the baseline measurement and the follow-up measurement is performed in order to
compare the averaged response at baseline with the averaged response at the follow-up
at 3 months.
Sample size and power
The primary outcome is the change in TFI from baseline to follow-up at three months.
A TFI decrease of 13 points is considered a minimal clinically important difference.
Assuming a standard deviation of 20 points (SD found in international literature [44]
as well as in a previous study of our research group [45]) and a significance level
of 0.05 an achieved sample size of 39 per group is required to detect an effect of
13 points with 80% power using a two-sample t-test. Taking into account a possible
20% drop-out, 50 patients will be recruited into each group. Hence, a total of 100
patients will be recruited for the study.
Assignment of interventions
Allocation
Participants will be randomized into the treatment group or the placebo group in a
1:1 ratio. The randomized list will be generated by an independent researcher using
QMinim Online Minimization®. The allocation sequence will depend on the date on which
participants are referred to the trial. Stratified randomization according to the
grade of tinnitus severity on the TFI and gender will be performed. The TFI categorizes
patients from grade 1 (slight tinnitus distress) to grade 5 (severe tinnitus distress).
As patients with grade 1 are generally not in need of intervention, they will be excluded
from the study (inclusion criterion: TFI > 24). Therefore, a total of 8 strata will
be defined by TFI grade and gender. Patient enrollment and allocation will be performed
by an independent researcher.
In order to avoid any test bias the researcher performing the baseline and follow-up
measurements is blinded to the allocation of the patient. The investigator will be unblinded by an independent researcher after the three-month
time point (T1). The independent researcher will be able to access the randomization sequence via
QMinim, an online portal. For technical reasons, it is not possible to blind the investigator administering
the HD-tDCS. Therefore, for each patient, the investigator administering the HD-tDCS
will differ from the researcher performing the baseline and follow-up measurements.
Blinding
Both the patient and the investigator performing baseline and three-month follow-up
measurements will be blinded up until and including the three-month time point (T1). Patients will be unblinded by an independent clinician at a separate hospital visit
following the three-month follow-up time point.
Data collection, management, and analysis
Data collection and management
OpenClinica LLC is used to enter and store data in a clean, secure, and efficient
manner. This software package is developed especially for electronic data management
in clinical research. Only the principal investigators have access to this password-protected
database. Validation checks such as range checks for data values are programmed so
that the number of mistakes is minimized. The information collected in this study
is kept strictly confidential. Individual information and results are coded, with
only the researcher knowing which code was assigned to each participant. The data
are kept for 20 years.
Statistical methods
Data will be analyzed using SPSS statistical software version 25 (SPSS Inc., Chicago,
IL, USA). The primary outcome is the change in TFI (Tinnitus Functional Index) from
baseline to 3 months after the treatment. A linear regression model using delta TFI
(pre minus post therapy) as an outcome and treatment as a predictor with correction
for age, gender, hearing loss, HADS score, and TFI grade will be used as primary outcome
analysis. All patients will be used in an intention-to-treat (ITT) analysis.
Per protocol analyses will be performed for all secondary outcome measurements to
evaluate the treatment effect for the compliers. The change in secondary outcomes (which are all continuous variables) from baseline
to follow-up at 3 months will be analyzed with a linear regression model with treatment
as predictor controlling for age, gender, hearing loss, HADS score, and TFI grade.
The questionnaires that are administered repeatedly over time will be modelled by
a linear mixed model using patient as a random intercept to correct for repeated measures
coming from the same patient. The Holm-Bonferroni method will be used to correct for multiple testing.
Study monitoring
An independent Data and Safety Monitoring Board (DSMB) at the University Hospital
of Antwerp will monitor the clinical trial for adverse events, adherence to the study
protocol, and potential early stopping. The DSMB will meet twice a year and will biannually receive reports from the research team on patient safety, protocol deviations, data
completion, and enrollment.