| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Psychiatry (S.K.R., K.A.B., M.A.G., M.B.S.), Division of Otolaryngology–Head and Neck Surgery, Department of Surgery (E.S.V., J.P.H.), and the Department of Family & Preventive Medicine (M.B.S.), University of California, San Diego School of Medicine, San Diego, California; Veterans Administration San Diego Healthcare System, San Diego, California (S.K.R., M.B.S.).
Address correspondence and reprint requests to Shannon K. Robinson, MD, Department of Psychiatry-University of California, San Diego School of Medicine, Veterans Administration San Diego Healthcare System, 3350 La Jolla Village Dr, Mail Code 116A La Jolla, CA 92161-116A. E-mail: skrobinson{at}ucsd.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSNOTESREFERENCES |
|---|
Design: One hundred twenty tinnitus sufferers participated in a randomized double-blind placebo-controlled trial. Paroxetine or placebo was increased to a maximally tolerated dose (up to 50 mg/day), and patients were treated for a total of 31 days at the maximal dose.
Methods: Patients with chronic tinnitus were recruited from our university-based specialty clinic by referral from otolaryngologists and audiologists in the local community and by advertisement. Patients with psychotic or substance use disorders or suicidal ideation were excluded, as were those using psychoactive medications (this resulted in only 1 subject with major depression in the study) or any other medications that interact with paroxetine and those with inability to hear at one’s tinnitus sensation level. Fifty-eight percent of patients were male, 92% were Caucasian, and the average age was 57.
Outcomes Measures: Tinnitus matching, the Tinnitus Handicap Questionnaire, the question: How severe (bothered, aggravating) is your tinnitus? Quality of Well-Being and other psychological questionnaires.
Results: Paroxetine was not statistically superior to placebo on the following tinnitus measures (tinnitus matching, 5- or 10-db drop, Tinnitus Handicap Questionnaire, quality of well-being measures, how severe, how bothered, positive change). There was a significant improvement in the single item question, How aggravating is your tinnitus? for those in the paroxetine group compared with the placebo group.
Conclusions: These results suggest that the majority of individuals in this study did not benefit from paroxetine in a consistent fashion. Further work remains to be done to determine if subgroups of patients (e.g., those who tolerate higher doses, those who are depressed) may benefit.
Key Words: tinnitus • antidepressant
Abbreviations: UCSD = University of California, San Diego.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSNOTESREFERENCES |
|---|
Tinnitus has been associated with a higher lifetime prevalence of depression than is found in the general population, a higher prevalence of current major depression (10). Depressed tinnitus patients rate their tinnitus severity significantly higher than nondepressed tinnitus patients (11,12) and have more psychosocial disability (10). Additionally, insomnia and inability to concentrate (often seen as symptoms of anxiety or depression) are commonly reported to accompany tinnitus (13). Patients who exhibit decreases in depression scores exhibit greater reductions in tinnitus severity scores (14). Antidepressants are often prescribed for tinnitus, despite little evidence of their efficacy (15).
Only two double-blind studies had been completed (before this study) looking at antidepressants in tinnitus patients. Mihail and colleagues (16) studied trimipramine in 26 patients with subjective tinnitus, with each subject serving as his own control, and found no significant effects. Sullivan and colleagues (17) conducted a randomized placebo-controlled trial of nortriptyline in subjects with chronic tinnitus that was severe enough to disrupt daily activity. They showed decreased depression levels (as measured by the Hamilton Rating Scale for Depression), decreased functional disability (as measured by the Multidimensional Pain Inventory modified for tinnitus but not on other tinnitus disability measures), and decreased tinnitus loudness in the worst ear when adjusted for baseline levels. Based on the later observations, antidepressant treatment of tinnitus looked promising, but it was unclear if its effects on tinnitus were independent of its effects on depression. Additionally despite promising results this study was never replicated, nor is nortriptyline widely used (possibly because of side effects of tricyclic antidepressants).
Thus far, there had been few data to support the use of the next generation of antidepressants, the selective serotonin reuptake inhibitors, for the treatment of tinnitus. Furthermore, some data suggested that selective serotonin reuptake inhibitors and related compounds may actually cause tinnitus (18,19). However, open-label case reports of beneficial effects of treatment with fluoxetine and paroxetine have been published (20,21). Folmer and colleagues (22) reported a retrospective analysis of a group of 37 patients who started taking selective serotonin reuptake inhibitors after the onset of their tinnitus and were also in psychotherapy. Thirty of the 37 patients returned questionnaires (mean of 20 months later), and a significant improvement in total Tinnitus Severity Index was found for these patients.
There are a number of possible mechanisms by which antidepressants may improve tinnitus. Some of these proposed mechanisms have to do with nonselective binding of antidepressants, such as anticholinergic and antihistaminic binding (16). To date, these have few to no data to support them. However, there is growing evidence to support the theory originally postulated by Shea and colleagues (23) that inhibiting serotonin reuptake inhibits the transmission of the tinnitus impulse. This proposed mechanism is supported by data showing the auditory cortex to be rich in serotonin receptors and that alteration of these receptors can alter auditory evoked potentials (5,24). In addition to a direct effect on serotonin receptors in the auditory cortex, another possible mechanism of improvement is the indirect benefit on tinnitus that may occur with the treatment of depression and anxiety disorders, even in subsyndromal forms.
Our clinical objectives were to decrease tinnitus aggravation and improve quality of well-being. Using a double-blind placebo-controlled trial, we hypothesized that the selective serotonin reuptake inhibitor paroxetine would (1) decrease the perceived loudness of tinnitus, that is, produce a drop in tinnitus sensation level, (2) decrease participants’ disability associated with their tinnitus, and (3) increase participants’ health-related quality-of-life. We also hypothesized that (4) paroxetine’s effect on tinnitus and tinnitus-related disability would occur independent of participants’ history of depressive or anxiety disorders, and (5) any tinnitus-related improvements would occur independent of changes in depressive or anxious symptomatology.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSNOTESREFERENCES |
|---|
Intake Assessment
The intake assessment included a history and physical examination performed by a neuro-otologist, an audiogram with tinnitus matching, and a psychiatric assessment.
Audiometric testing included a baseline assessment of hearing sensitivity in both ears, including pure-tone hearing thresholds, speech reception thresholds, and word discrimination scores. Impedance testing was performed only if needed as determined by the audiologist and evaluation/presentation of patient’s symptoms. Subjects with asymmetric findings (e.g., >10-db pure tone hearing threshold) were investigated with brain stem auditory evoked responses or magnetic resonance imaging. If a retrocochlear lesion was ruled out, then the participant was included.
The Structured Clinical Interview for Diagnostic and Statistical Manual, Fourth Edition, a standardized psychiatric interview, was administered by a psychiatrist or a master’s-level clinician to determine the presence of psychiatric disorders, including major depressive disorder, dysthymia, mania, panic disorder, social or specific phobia, generalized anxiety disorder, obsessive compulsive disorder, and posttraumatic stress disorder (25). The Alcohol Use Disorders Identification Test (a self-report of alcohol use) was used to help screen out patients with active alcohol dependence (26).
Treatment Protocol
Treatment began at 10 mg of paroxetine (or placebo) per day for the first week. A 10-mg starting dose was chosen because this has been shown to be a well-tolerated dose (27). The dose was then increased to 20 mg per day for 2 weeks. Based on participant response and tolerance to side effects (adverse events), the dose was increased in 10-mg increments every 2 weeks to a maximum of 50 mg per day. A maximum of 50 mg was used because this is the maximum dose approved for use in depressed patients by the US Food and Drug Administration (27). Total duration of treatment was 100 days. If a participant dropped out of the study, the participant had his or her last observation carried forward as long as he or she took at least 10 mg of the medication for at least 1 week and returned for at least 1 visit while on medication. Participants began tapering off the medication after follow-up testing of tinnitus and psychological variables.
Assignment and Masking
Participants (n = 120) were randomized to treatment with either paroxetine (Paxil) or placebo in identical capsules. Randomization was done in blocks of 10 subjects in a 1 paroxetine to 1 placebo ratio. Originally, there was a randomization in blocks of 10 for participants with major depression present at time of initial assessment and participants without major depression present. Although this was an ideal study design, our recruitment efforts unfortunately led to only 1 participant with major depression, and therefore this participant’s data were not stratified or separated out from the other participants. Participants were allocated to paroxetine or placebo after completing initial assessments. Neither participants nor study personnel were aware of which treatment the participant was receiving. Randomization schemes were provided by the pharmacy and were kept in the pharmacy until all 120 participants had completed their last visit, i.e., had tapered off study medication and come in for a follow-up visit 1 month later.
Outcome Measures
Audiometric Measures
An assessment of tinnitus in the affected ear(s) was performed at baseline and each follow-up visit by matching reported tinnitus to externally presented sounds. The matching was done using a two-sample forced-choice method to determine the frequency and intensity level after characterizing the type of tinnitus sound (e.g., pure tone, narrow band noise, warbled tone). Tinnitus was measured in db hearing level and was then converted to db sensation level by subtracting the patient’s hearing threshold at the tinnitus apparent frequency (measured at 1/24 octaves) from the measured intensity level (measured in 1-db steps) of the tinnitus. At follow-up, the audiologist asked participants if they had observed any change in their tinnitus since the baseline visit. Responses were coded as positive change or no positive change. Due to the well-known fact that some of these patients may experience the effects of recruitment and/or hyperacusis, the actual loudness of the tinnitus was expected to vary between patients; however, this study’s focus was primarily investigating change in the tinnitus within the individual. Although not in the original design of the study, most participants were also asked by the audiologist at baseline and follow-up visits if they experienced more sensitivity to moderately loud and loud sounds than others around them (hyperacusis: yes or no) and were asked to rate their aggravation level of their tinnitus on an 8-point scale (0–7).
Psychological Symptom Measures
Depressive symptoms were assessed using the Hamilton Rating Scale of Depression, 17-item version (observer-rated), and the Beck Depression Inventory (self-rated) (28,29). Anxiety symptoms were assessed using the Hamilton Rating Scale for Anxiety (observer-rated) and the Beck Anxiety Inventory (self-rated) (30,31). The Hamilton Rating Scale for Depression and the Hamilton Rating Scale for Anxiety were administered by a psychiatrist or master’s-level clinician. The Symptom Checklist-90-Revised (self-rated) was administered to provide a more broad-based index of mental and behavioral symptoms that may not have been captured by the other measures (32). Sleep problems were assessed using the Pittsburgh Sleep Quality Index (33).
Disability and Quality of Life
Disability was assessed with specific reference to tinnitus using a specialized measure, the Tinnitus Handicap Questionnaire (self-rated) (34). Factor analysis has revealed 3 subscales: (1) the physical, emotional, and social consequences of tinnitus, (2) hearing ability of the patient, and (3) the patient’s view of tinnitus (34). In order to further understand the disability associated with tinnitus, participants answered the following questions on an 8-point Likert scale: "How severe is your tinnitus?" with 0 = not at all and 7 = loudest sound imaginable, and "How bothered are you by your tinnitus?" with 0 = not at all and 7 = very, very much. A more generic measure of disability was used as well, the World Health Organization’s 5-item Disability Inventory (self-rated) (35). Health-related quality of life was assessed using the Quality of Well-Being Scale, a widely used quality of life measure (36).
Adverse Events
Participants completed a adverse events questionnaire at every visit in order to assess the severity (none, mild, moderate, severe) of the following: headache, gastrointestinal distress, sexual dysfunction, insomnia, drowsiness, dry mouth, sweating, tremor, and any other side effects they may experience.
Sample Size Projection
Based on the literature on tricyclic antidepressant treatment of tinnitus we expected a 60% response rate in our paroxetine-treated group and a 30% response rate in our placebo-treated group. "Response" was defined by our primary outcome measure as a 10-db drop in perceived tinnitus volume. In order to detect this moderate-sized effect with a power of 80% and a p = .05, our calculations indicated we would need 100 subjects to complete the study. In order to compensate for early dropouts, we enrolled 120 subjects, based on dropout rates in paroxetine trials for depression and anxiety disorders and tinnitus medication trials.
Statistical Analysis
Outcomes were tested using repeated-measures analyses of variance with assessments (baseline and follow-up) as the within-subject variables and treatment group (paroxetine versus placebo) as the between-subject variable for continuous variables and 
2 for noncontinuous variables. For subjects who terminated from the study prematurely, follow-up measures consisted of the last observation carried forward. Relationships between outcome measures were examined using Spearman correlation coefficients. Statistical tests were two-tailed, and p values <.05 were considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSNOTESREFERENCES |
|---|
|
|
|
Participant Flow
There were no significant differences between the paroxetine and the placebo groups on any of the baseline measures. Twenty-six participants dropped out of the trial, but only five of these 26 participants were excluded from follow-up analyses because no data were available beyond the baseline assessment. In other words, 21 participants had their last observation carried forward, resulting in a total of 115 participants with follow-up data. These 115 participants were used in our intent-to-treat analysis. Five participants did not take 10 mg daily for 1 week and return for 1 follow-up visit, and therefore, they were eliminated from analysis. Four of these participants dropped out of the study because of adverse events. One participant thought his tinnitus was worse. None of these subjects returned for further evaluation of their tinnitus. Among the 26 participants who dropped out of the study (including 21 participants who had their last observation carried forward and five participants who were dropped from analyses), most participants dropped out because of adverse events. These included those assessed with the side effect questionnaire, as well as a perceived increase in tinnitus. Significantly more participants in the paroxetine group (n = 17) dropped out because of adverse events than those in the placebo group (n = 5) (2 [1, n = 118] = 7.54, p < .05). Among the 115 participants who completed the trial and the participants who had their last observation carried forward, significantly more participants in the paroxetine group reported moderate or severe sexual dysfunction, drowsiness, and dry mouth than in the placebo group at follow-up. See Table 4.
|
Outcomes
The 115 participants who were included in the follow-up analyses took their maximum tolerated dose of medication for an average of 31 days, ranging between 17 to 57 days, following a conservative up-titration schedule. The total duration of treatment from start to finish on average was 100 days.
There were no differences in baseline scores on the Beck Depression Inventory, Beck Anxiety Inventory, Hamilton Rating Scale for Depression, Hamilton Rating Scale for Anxiety, Symptom Checklist-90-Revised, World Health Organization Disability Inventory, Quality of Well-Being Scale, Pittsburgh Sleep Quality Index, or Tinnitus Handicap Questionnaire between paroxetine and placebo when the 115 participants used in the intent to treat analysis were looked at. Of these 115 participants, 58 received placebo and 57 received paroxetine.
The pure tone average (average of 500-, 1000-, and 2000-Hz hearing thresholds) was calculated pre- and posttreatment for both ears for the placebo and paroxetine groups. The placebo group pure tone average went from 21.9 ± 1.7 db to 20.7 ± 2.1 db and the paroxetine group went from 21.6 ± 2.0 db to 23.3 ± 2.1 db. The 0.8-db improvement in the placebo group and 1.8-db decrement in the paroxetine group were statistically different (p < .05), but are clearly below clinically significant change levels. Thus it was concluded that there was no substantial hearing change with treatment and therefore little possibility for development of recruitment changes affecting tinnitus sensation levels.
Intent-to-treat analyses revealed a significant treatment effect of paroxetine on participants’ report of aggravation caused by tinnitus to the study’s audiologist but on none of the other outcome measures (Table 5). No treatment effects of paroxetine were found on psychological measures (Table 6).
|
|
Table 7 shows that changes in some tinnitus symptoms were correlated with changes in depressive or anxious symptomatology. Correlations were not, however, in the expected direction. For example, changes in depression levels, as measures by the Beck Depression Inventory, were inversely correlated with changes in tinnitus severity.
|
Because some patients had reported to the treating clinician that their tinnitus was much improved, we did post hoc analyses in order to determine if significant treatment effects were found in a subgroup of participants. There was no evidence from our analyses that any subgroup we looked at, other than those subjects who reached a dose of 50 mg per day, benefited consistently (on a number of measures, both objective and subjective). Group differences were examined for the participants who reached a 50 mg/d dosage in the paroxetine group (n = 27) compared with all participants in the placebo group. Participants on paroxetine at 50 mg daily benefited significantly compared with all participants who received placebo on the following: the Tinnitus Handicap Questionnaire subscale 2 (which measures perceived hearing ability of the patient, F(1,76) = 8.36, p = .005) and the questions how severe is your tinnitus (F(1,39) = 5.46, p = .025), how bothered are you by your tinnitus (F(1,40) = 4.82, p = .034) and how aggravated are you by your tinnitus (F(1,55) = 8.01, p = .006). Additionally, subjects in the 50-mg paroxetine group were more likely to have a 10-db drop in tinnitus in either ear than were participants in the placebo group, 2 (1, n = 73) = 5.28, p = .027.
Regarding the 27 participants on paroxetine who reached 50 mg compared with those who did not reach 50 mg of paroxetine, there were no significant baseline differences on the Beck Depression Inventory, Beck Anxiety Inventory, Hamilton Rating Scale for Depression, Hamilton Rating Scale for Anxiety, Symptom Checklist-90-Revised, World Health Organization Disability Inventory, Quality of Well-Being Scale, Pittsburgh Sleep Quality Index, or Tinnitus Handicap Questionnaire. Additionally, there was no significant worsening of tinnitus in those who received 50 mg of paroxetine versus those who did not reach 50 mg of paroxetine.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSNOTESREFERENCES |
|---|
To assess if tinnitus improvements were independent of improvements in depression or anxiety, we used a correlation matrix and found that tinnitus improvements (as measured by a decrease in sensation level, Tinnitus Handicap Questionnaire, severity/bothered/aggravation questions) were not correlated in a positive direction with decreases in depression as measured by the Beck Depression Inventory. We did find that severity and amount bothered by tinnitus were significantly correlated in a negative direction with anxiety and depression (as severity and amount bothered improved depression and anxiety worsened). It must be kept in mind that only 1 of our subjects met criteria for Major Depressive Episode, so depression and anxiety values were very low (similar to nondepressed norms in the general population), which may account for these findings. Our results of a negative correlation between depression/anxiety measures and tinnitus measures are not consistent with the majority of literature in this area that repeatedly shows a positive correlation between depression/anxiety levels and tinnitus severity (11,12,34). Given the small decrements in mood and anxiety symptoms that occurred in conjunction with small changes in tinnitus aggravation, the veracity and, certainly, the clinical significance of these observations are doubtful.
There was an 18% dropout rate in this study. This rate is comparable to other studies using antidepressants (37). This suggests that the medication was generally well tolerated. Although overall superiority of paroxetine compared with placebo could not be demonstrated, there was some evidence of benefits of the highest paroxetine dose. When we compared subjects that reached 50 mg of paroxetine to all subjects on placebo, we did find significant benefits of paroxetine in subscale 2 of the Tinnitus Handicap Questionnaire (hearing ability) and the questions how severe is your tinnitus, how bothered are you by your tinnitus, how aggravated are you by your tinnitus and a 10-db drop in tinnitus in at least 1 ear. Although it is difficult to draw conclusions from this kind of post hoc comparison, it does raise the possibility that higher doses of paroxetine (if tolerated) might be beneficial for some patients. This would need to be determined using fixed-dose study designs.
There was no indication of any effect of paroxetine on any of the psychological measures (Hamilton Rating Scale for Depression, Hamilton Rating Scale for Anxiety, Beck Depression Inventory, Beck Anxiety Inventory, Symptom Checklist-90-Revised). We theorize that this is due to a floor effect. The baseline scores on these measures (see Table 1) are all within the normal range. Therefore, there was little depressive or anxiety symptomatology in our sample at baseline and little room for improvement.
Not only are these scores low in comparison to psychiatric patients’, but they are in contrast to prior tinnitus studies. A previous study preformed in our research group revealed mean Hamilton Rating Scale for Depression score of 12.2 and Beck Depression Inventory score of 15.0 in patients seeking treatment with cognitive behavioral therapy for distressing tinnitus (38). Previous tinnitus studies using the Beck Depression Inventory have shown means ranging from 4.7 (39) to 16.1 (40–51). We believe that the very low level of psychopathology seen in our sample is due, at least in part, to our having excluded patients already taking psychotropic medication at the time of study entry. As such, this study provided a test of the hypothesis that selective serotonin reuptake inhibitors would be ineffective in patients without existing psychiatric disorders (e.g., major depression). It remains to be proven whether or not selective serotonin reuptake inhibitors might still be useful in ameliorating tinnitus in patients with comorbid mood or anxiety disorders.
Cognitive behavior therapy is generally considered the mainstay of treatment for distressed tinnitus patients. It has been shown to produce moderate to strong effect sizes on tinnitus annoyance (52). In depression and anxiety, there is some reason to believe that the best outcomes result from combining medications and psychotherapy (often cognitive behavioral therapy) (53). Following suit, Folmer and Shi (54) recommended patients with distressing tinnitus and depression be treated with antidepressant medication and therapy. Future studies into combination treatment with medications and psychotherapy would be wise.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSNOTESREFERENCES |
|---|
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSNOTESREFERENCES |
|---|
This study was funded by a grant from the Tinnitus Research Consortium. Paroxetine and matching placebo were supplied by GlaxoSmithKline. Computer support services were in part provided by the General Clinical Research Center at the University of California, San Diego, and a NIH grant (#M01RR0827).
DOI:10.1097/01.psy.0000188479.04891.74
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  W. Noble Treatments for Tinnitus Trends in Amplification, September 1, 2008; 12(3): 236 - 241. [Abstract] [PDF]
|
|
|
|
 |
|
 Minerva BMJ, December 10, 2005; 331(7529): 1416 - 1416. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
