532779
research-article2014
AORXXX10.1177/0003489414532779Annals of Otology, Rhinology & LaryngologyMiguel et al
Article
The Effect of Insomnia on Tinnitus
Annals of Otology, Rhinology & Laryngology 2014, Vol. 123(10) 696–700 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0003489414532779 aor.sagepub.com
George S. Miguel, DO1, Kathleen Yaremchuk, MD1, Thomas Roth, PhD2, and Ed Peterson, PhD3
Abstract Objective: The objective is to see how chronic tinnitus sufferers who are unmanageable to maximized medical therapy can benefit by decreasing their subjective complaints from a sleep evaluation and treatment. However, the proper identification of these particular patients has not been described well in the literature when attempting to correlate these 2 diagnoses. Thus, tinnitus patients with and without insomnia, based on ICD-9 diagnosis, were evaluated using the Tinnitus Reaction Questionnaire and Insomnia Severity Index to determine correlations between insomnia and tinnitus. Methods: Patients with a diagnosis of tinnitus and tinnitus along with insomnia who were treated at our institution from 2009 to 2011 were identified. Tinnitus Reaction Questionnaire and Insomnia Severity Index responses were obtained through written and telephone interviews. A Pearson product moment correlation was used to determine the effect of insomnia on tinnitus. Additional analyses identified whether Tinnitus Reaction Questionnaire scores were associated with a possible benefit from an evaluation for insomnia in tinnitus patients. Results: A total of 117 patients met inclusion criteria. A significant correlation was found between the Insomnia Severity Index score and Tinnitus Reaction Questionnaire severity (r = 0.64; P = .001). Tinnitus Reaction Questionnaire severity was shown to be a good predictor of sleep disturbance and good in predicting group association, especially the “emotional” subscore component (sensitivity 96.9% and specificity 55.3% for identifying tinnitus patients with insomnia). The greater the insomnia disability as exhibited by an elevated Insomnia Severity Index score, the more severe the patient’s complaints were regarding the tinnitus. Conclusion: Results suggest that if the emotional score on the Tinnitus Reaction Questionnaire is ≥ 15, the Insomnia Severity Index may be useful to identify patients who may benefit from further treatment and evaluation of insomnia. The robust correlation between the Tinnitus Reaction Questionnaire and Insomnia Severity Index objectively showed that patients with insomnia have an increased emotional distress associated with their tinnitus. Both questionnaires can be used together with a high degree of specificity and sensitivity in predicting tinnitus patients with an underlying sleep disturbance. Keywords insomnia, tinnitus
Introduction More than 36 million patients are plagued with tinnitus.1 Evidence-based research has shown a strong correlation between tinnitus and other psychological disturbances. Researchers have long been challenged regarding the effects of tinnitus on other comorbidities and few systematic studies have been done. Tinnitus may be associated with worsening of anxiety, depression, irritability, and sleep disturbances. Of these, the relationship between depression and tinnitus has been well documented through the validated Tinnitus Handicap Index in relation to the Beck Depression Inventory questionnaire.2 Recent epidemiological studies found an association between sleep disturbance (ie, insomnia) and a decreased capacity for dealing with stress due to pain stimuli.1 The
studies concluded that sleep disturbance may be linked to an increased severity of the previously mentioned psychological disturbances. Few reports have documented the relationship between tinnitus and sleep. Such studies have been 1
Department of Otolaryngology–Head and Neck Surgery, Henry Ford Hospital, Detroit, Michigan, USA 2 Department of Pulmonary and Critical Care Medicine, Section of Sleep Medicine, Henry Ford Hospital, Detroit, Michigan, USA 3 Department of Public Health Sciences, Henry Ford Hospital, Detroit, Michigan, USA Corresponding Author: George S. Miguel, DO, Department of Otolaryngology–Head and Neck Surgery, Henry Ford Hospital, 2799 West Grand Boulevard, Suite K-8, Detroit, MI 48202, USA. Email: [email protected]
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
697
Miguel et al done in the geriatric population, correlating chronic tinnitus with sleep disturbances based on initiation of sleep and sleep maintenance.3 However, most studies are mainly suggestive that sleep difficulties are related to increased tinnitus severity. Moreover, most reports are brief and include only a few questions on sleep without statistical significance. Folmer and Griest4 concluded that their findings underscored the importance of proper identification and successful treatment of patients with tinnitus and insomnia. An obstacle has been the ability to properly identify patients whose sleep disturbance may be worsening their perception of tinnitus. It is important to distinguish and be aware of these 2 clinical entities in order to correctly identify these patients so that proper treatment can be initiated earlier, resulting in higher patient satisfaction. Roth and AncoliIsrael1 concluded that chronic pain sufferers have a decreased tolerability to other chronic conditions such as tinnitus or insomnia. It is unfortunate that it is difficult to discern whether sleep makes tinnitus worse or the reverse; however, we can accept the fact that they work synergistically together. The accepted and validated questionnaire indexes for tinnitus5 and insomnia6 patients, respectively, have not been statistically correlated or compared in these patient groups. This study uses these validated questionnaires in 2 groups of tinnitus patients (tinnitus and tinnitus with insomnia) to determine the correlation between tinnitus and insomnia. In doing so, we intend to show that understanding the relationship between tinnitus and insomnia is important in the management and proper identification of these patients early in the treatment process. It is the hope that by showing correlation between these 2 validated questionnaires, a clinician can evaluate and treat patients with tinnitus from a different angle once all other reasons have been exhausted.
Methods This retrospective observational study was approved by the Henry Ford Hospital Institutional Review Board. Patients with a diagnosis of tinnitus only (ICD-9 code 388.30, 388.31, 388.32) and those with both tinnitus and insomnia (ICD-9 code 780.52) who were seen between January 2009 and December 2011 were identified in the administrative database and placed into 2 groups. The patients were referred and seen at the Henry Ford Tinnitus Clinic. Tinnitus Reaction Questionnaire and Insomnia Severity Index responses were obtained through written and telephone interviews. Patients whose diagnosis of insomnia was made prior to a diagnosis of tinnitus were excluded from the study. This ensured that tinnitus was the initial and primary complaint of the patient rather than insomnia. Other exclusion criteria included deceased patients, those lost to follow-up, and those who refused to participate in the study. Demographics of the patient included an average age of 63
years, and the study included 72 men and 45 women. Most of these patients were referred to our tinnitus clinic after exhaustive medical treatment from the community. However, we did not analyze the treatments they received (ie, medications, behavioral, etc) with regard to their tinnitus. These 2 groups were made based on ICD-9 code only and insomnia was not ruled out for those in the tinnitus-only group. We wanted to see in the tinnitus-only group if there was, inherently, undiagnosed sleep disturbance. Those with a primary diagnosis of insomnia were not addressed because the intent was to see how tinnitus was affected by insomnia rather than how insomnia affected someone’s quality of life. The Tinnitus Reaction Questionnaire5 was developed to assess the psychological disturbance one has from tinnitus. It is a series of 26 questions looking at one’s perception in categories that can involve feelings of anger, annoyance, confusion, and so on. Each question is based on a score of 0 points (not at all) to 4 points (almost all of the time) with a maximum total of 104. The Insomnia Severity Index6 is a reliable and valid instrument to quantify perceived insomnia severity. The Insomnia Severity Index is a series of 7 questions asking about sleep patterns, behavior, and one’s perception of sleep. Each question is scored on a scale from 0 (none) to 4 (severe). Total scores are as follows: 0-7 = no clinically significant insomnia; 8-14 = subthreshold insomnia; 15-21 = clinical insomnia (moderate severity); 22-28 = clinical insomnia (severe). Statistical analysis examined the degree of linear relationship between Tinnitus Reaction Questionnaire and Insomnia Severity Index scores using the Pearson correlation coefficient. The overall score and scores from 3 subscales used in the Tinnitus Reaction Questionnaire (emotional, functional, and catastrophic) were evaluated using this coefficient. Correlation estimates were compared using a method for both independent and dependent correlations. The mean Tinnitus Reaction Questionnaire scores in the 2 groups (tinnitus alone and tinnitus and insomnia) were compared using Student t tests. This methodology was also used to compare the 3 subscales. The areas under the curve of the receiver operating characteristics curve were estimated using linear logistic regression. This score between 0 and 1, which measures the amount of predictability between the variables, was then used to find the cutoff point that best described the sensitivity and specificity. We maximized the sum of these 2. Estimates of the 2 screening characteristics and their associated 95% confidence intervals (CIs) were estimated. A P value less than .05 was considered statistically significant.
Results Of the 117 patients who met the inclusion criteria, 85 were in the tinnitus-only group and 32 were in the tinnitus +
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
698
Annals of Otology, Rhinology & Laryngology 123(10)
Table 1A. Correlation Between Insomnia Severity Index and Tinnitus Reaction Questionnaire. Comparison ISI vs TRQ ISI vs emot+ ISI vs funct+ ISI vs catas+
Population Correlation P Value* All
0.63 0.59 0.59 0.58
.001 .001 .001 .001
Group
Correlation P Value*
Tinnitus + insomnia
.001 .002 .001 .005
0.64 0.52 0.52 0.48
Group
Correlation
P Value*
Tinnitus
0.55 0.52 0.53 0.52
.001 .001 .001 .001
Abbreviations: catas, catastrophic; emot, emotional; funct, functional; ISI, Insomnia Severity Index; TRQ, Tinnitus Reaction Questionnaire. *Testing against 0. +Dependent correlation test P value, emot vs funct 0.802, emot vs catas 0.822, funct vs catas 0.985.
Table 1B. Correlation Between Insomnia Severity Index and Tinnitus Reaction Questionnaire. Comparison
Correlation
P Valuea
Tinnitus + insomnia vs tinnitus
ISI vs TRQ ISI vs emot+ vs funct+ vs catas+
.522 .985 .955 .832
Abbreviations: catas, catastrophic; emot, emotional; funct, functional; ISI, Insomnia Severity Index; TRQ, Tinnitus Reaction Questionnaire. a Independent correlation test. +Dependent correlation test P value, emot vs funct 0.802, emot vs catas 0.822, funct vs catas 0.985.
110 100 90 80 70 60 50 40 30 20 10 0
insomnia group. The correlation between the Tinnitus Reaction Questionnaire and Insomnia Severity Index scores was moderately high (0.62) and significant (P = .001) for all patients (Table 1). The correlation was 0.64 for the tinnitus + insomnia group and lower at 0.55 for the tinnitus-only group. The 2 correlations when compared to each other, however, were not significantly different (P = .522). The positive correlation indicates that as Insomnia Severity Index scores increase, so do Tinnitus Reaction Questionnaire scores (Figure 1). Table 1 also provides the correlations between the Tinnitus Reaction Questionnaire and Insomnia Severity Index for the 3 subscales of the Tinnitus Reaction Questionnaire. These correlations ranged around 0.5 and all were significantly different from 0. We tested them against one another and none of the comparisons were significant. The Tinnitus Reaction Questionnaire and mean Insomnia Severity Index scores were tested for differences between the 2 groups (Table 2). All comparisons were significant with higher scores seen for the tinnitus + insomnia group. The areas under the curve were estimated as 0.786 with a 95% CI of 0.69 to 0.87. Areas under the curve estimates for the 3 subscales were 0.793 for emotional, 0.720 for functional, and 0.786 for catastrophic (Figure 2). Table 3 shows that the emotional subscale had the highest area under the curve at 0.79 (95% CI, 0.71-0.88). We used the emotional subscale result to find the best cutoff point to maximize the sum of specificity
Correlation Between TRQ and ISI
trq 120
0
10
20 isi both only group correlations for both(solid)=0.64, for only(dashed)=0.55 p-value=0.522
30
Figure 1. Data plots for correlation between Tinnitus Reaction Questionnaire and Insomnia Severity Index.
and sensitivity. A score of x ≥ 15 resulted in an estimated sensitivity of 96.9% (88.8, 99.9) and a specificity of 55.3% (44.1, 66.1) for predicting the presence of insomnia. Using this criterion, 38 of the tinnitus-only patients were incorrectly classified. Of these 38, 29 (76.3%) reported a sleep disturbance score of ≥ 3 despite not having insomnia according to the Insomnia Severity Index (Table 4).
Discussion Assessing the severity of tinnitus in patients can be difficult for a variety of reasons. A chronic tinnitus patient presents a multidimensional clinical picture consisting of anxiety, depression, annoyance, or self-reported emotional distress. In 1983, Tyler and Baker7 asked patients to list the difficulties experienced as a result of tinnitus, and the most frequent complaint reported was “getting to sleep.” The association between tinnitus and depression is well known, and studies have shown a high correlation between
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
699
Miguel et al Table 2. Comparison Between Tinnitus + Insomnia Versus Tinnitus Group Scales and Insomnia Severity Index. Tinnitus
Tinnitus + Insomnia
Scale
N
Mean
SD
n
Mean
SD
P Value
TRQ overall TRQ emotional TRQ functional TRQ catastrophic ISI
85 85 85 85 85
42.4 15.5 17.3 9.7 12.1
25.3 9.6 11.9 8.3 7.2
32 32 32 32 32
71.6 26.6 26.1 18.8 17.5
20.3 9 9.4 7 5.1
.001 .001 .001 .001 .001
Abbreviations: ISI, Insomnia Severity Index; TRQ, Tinnitus Reaction Questionnaire.
Table 3. Area Under the Curve. ROC Curve for TRQ Factors tinnitus plus insomnia
Sensitivity
1.0 0.9 0.8 0.7
Scale
AUC (95% CI)
TRQ overall TRQ emotional TRQ functional TRQ catastrophic
0.79 (0.70-0.87) 0.79 (0.71-0.88) 0.72 (0.62-0.82) 0.79 (0.69-0.88)
0.6
Abbreviations: AUC, area under the curve; CI, confidence interval; TRQ, Tinnitus Reaction Questionnaire.
0.5 0.4 0.3
Table 4. Classification Rule in Predicting Insomnia With Tinnitus Reaction Questionnaire and Insomnia Severity Index.a
0.2 0.1 0.0 0.0
0.1
0.2 curve
0.3
0.4 0.5 0.6 1 - Specificity Catastrophic Emotional
0.7
0.8
0.9
Tinnitus + Insomnia Tinnitus
1.0
Functional
area under curve: Cata=0.786, Emot=0.793, Func=0.720
Figure 2. Data plots for the area under the curve.
Predict Patient to have tinnitus + insomnia (emotional score ≥ 15) Patient to have tinnitus only (emotional < 15) Total N
31
38
1
47
32
85
a
depressive symptoms and tinnitus severity. In 1993, Alster et al8 were the first to evaluate the possibility of sleep disturbance affecting the severity of tinnitus. Their data suggested that those who sought sleep evaluations and treatment noted a decrease in their tinnitus complaint. In daily clinical practice, it is uncommon for physicians to ask specifically about sleep disturbances in tinnitus patients and for patients with insomnia to be evaluated for tinnitus. This study supports previous conclusions that increased sleep disturbance correlates with an increase in tinnitus severity.3,7-9 The Tinnitus Reaction Questionnaire can be used initially to determine the severity of a patient’s perception of tinnitus. This study supports the use of the Insomnia Severity Index to help identify patients recalcitrant to optimized medical therapy and who may benefit from an evaluation of their sleep complaints. There was no single correlation between the Tinnitus Reaction Questionnaire and Insomnia Severity Index that showed a statistical significance in differentiating the 2 groups of patients (Table 1). Because the tinnitus + insomnia group had a diagnosis of
Of the 38 tinnitus-only with emotional ≥ 15, 76.3% have sleep score ≥ 3. Sensitivity = 96.9%. Specificity = 55.3%.
insomnia, there was an assumption that the correlation between the Tinnitus Reaction Questionnaire and Insomnia Severity Index would be higher compared to those without a diagnosis of insomnia. If the correlations between these 2 groups were close and significant, this might indicate that the Insomnia Severity Index is diagnostic. For example, the tinnitus-only group might suffer from inherent sleep problems. However, this was not the case. Rather, the predictability of inherent sleep problems from the subscale of the Tinnitus Reaction Questionnaire score was examined with use of the area under the curve. This showed a statistical significance for all 3 subscales with the emotional component having the highest effect. Table 4 shows the rule in predicting a patient to have inherent sleep disturbance, that is, insomnia, if the combined scores are ≥ 15. As mentioned previously, 29 (76.3%) of 38 tinnitus-only patients were incorrectly classified; these patients reported a sleep disturbance score of ≥ 3
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
700
Annals of Otology, Rhinology & Laryngology 123(10)
despite not having insomnia according to the Insomnia Severity Index. Correct classification could be done by looking at their summarized emotional subscore, which was > 15. This is a good example that 1 question regarding sleep in questionnaires is not a sufficient indicator for the presence or absence of insomnia. The classification rule shown in Table 4 may guide a clinician to approach these tinnitus patients with an Insomnia Severity Index questionnaire and further sleep evaluation. Limitations of our study include the difficulty to assess whether tinnitus or insomnia contributed to the patients’ initial complaint. We attempted to control this variable by excluding those who were diagnosed with insomnia before being seen for tinnitus. The group recruited was based on ICD-9 codes and that insomnia was not completely ruled out in the tinnitusonly group. As seen, sleep disturbance is underdiagnosed and not identified routinely. However, we should also note that there is the possibility that sleep disturbance may have been addressed in these patients and the diagnosis of insomnia was never coded. Furthermore, the size of the cohort has its limitations, and the strength of the correlations might be improved with a larger size population. Stouffer et al9 asked 528 patients to rate different conditions that reduced the severity of their tinnitus, and 26% responded sleep. This article has shown that the Tinnitus Reaction Questionnaire and Insomnia Severity Index can be used together in guiding a clinician to identify and treat those who have sleep disturbance. Different treatment regimens may consist of treating their sleep disturbance with medications or cognitive behavioral therapy. Melatonin has been shown to be useful in decreasing the severity of tinnitus in this subgroup of patients.10 This study also found an improvement in symptoms with placebo. Tinnitus is a process involving cognitive, emotional, and psychophysiological components. Such factors increase distress in patients, and studies have reported that tinnitus sufferers show maladaptive behaviors when exposed to a stress/pain stimulus.11 Thus, sleep complaint in these patients can be the consequence of a combined effect of a decrease in their tolerance to tinnitus. As shown, the results support the notion that sleep disturbance is underdiagnosed in this population and that treating their sleep disturbance may help reduce their symptomatology. This study demonstrates the use of a validated questionnaire in identifying patients whose sleep complaints may be responsible for a worsening of their perception of tinnitus. Further studies looking at this cohort after a proper sleep evaluation and treatment may be done to better understand how these 2 clinical entities relate.
Conclusion The identification of patients with insomnia with worsening tinnitus symptoms may assist in medical management. The
study results are consistent with those of previous investigations that sleep disturbance affects a patient’s perception of tinnitus severity. These questionnaires give objective measures in classifying tinnitus sufferers and their insomnia complaints. The Tinnitus Reaction Questionnaire and Insomnia Severity Index may be beneficial when used together in classifying this patient subgroup that has concomitant insomnia problems with a sensitivity of 96.9% and specificity of 55.3%. Authors’ Note This article was originally presented as a poster presentation at the Triological Society Meeting in San Diego, California, on April 21, 2012.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References 1. Roth T, Ancoli-Israel S. Daytime consequences and correlates of insomnia in the United States: results of the 1991 National Sleep Foundation Survey. Sleep. 1999;2(suppl 22):S354-S358. 2. Budd RJ, Pugh R. The relationship between locus of control, tinnitus severity, and emotional distress in a group of tinnitus sufferers. J Psychosom Res. 1995;39:1015-1018. 3. Asplund R. Sleepiness and sleep in elderly persons with tinnitus. Arch Gerontol Geriatr. 2003;37:139-145. 4. Folmer R, Griest S. Tinnitus and insomnia. Am J Otolaryngol. 2000;21:287-293. 5. Wilson P, Henry J, Bowen M, Haralambous G. Tinnitus Reaction Questionnaire: psychometric properties of a measure of distress associated with tinnitus. J Speech Hearing Res. 1991;34:197-201. 6. Bastien CH, Vallières A, Morin C. Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Med. 2001;2:297-307. 7. Tyler RS, Baker LJ. Difficulties experienced by tinnitus sufferers. J Speech Hear Disord. 1983;48:150-154. 8. Alster J, Shemesh Z, Ornan M, Attias J. Sleep disturbance associated with chronic tinnitus. Biol Psychiatry. 1993;34: 84-90. 9. Stouffer JL, Tyler RS, Kileny PR, Dalzell LE. Tinnitus as a function of duration and etiology: counselling implications. Am J Otol. 1991;12:188-194. 10. Rosenberg SI, Silverstein H, Rowan PT, Olds MJ. Effect of melatonin on tinnitus. Laryngoscope. 1998;108:305-310. 11. Heinecke K, Weise C, Schwarz K, Rief W. Physiological and psychological stress reactivity in chronic tinnitus. J Behav Med. 2008;31:179-188.
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
research-article2014
AORXXX10.1177/0003489414532779Annals of Otology, Rhinology & LaryngologyMiguel et al
Article
The Effect of Insomnia on Tinnitus
Annals of Otology, Rhinology & Laryngology 2014, Vol. 123(10) 696–700 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0003489414532779 aor.sagepub.com
George S. Miguel, DO1, Kathleen Yaremchuk, MD1, Thomas Roth, PhD2, and Ed Peterson, PhD3
Abstract Objective: The objective is to see how chronic tinnitus sufferers who are unmanageable to maximized medical therapy can benefit by decreasing their subjective complaints from a sleep evaluation and treatment. However, the proper identification of these particular patients has not been described well in the literature when attempting to correlate these 2 diagnoses. Thus, tinnitus patients with and without insomnia, based on ICD-9 diagnosis, were evaluated using the Tinnitus Reaction Questionnaire and Insomnia Severity Index to determine correlations between insomnia and tinnitus. Methods: Patients with a diagnosis of tinnitus and tinnitus along with insomnia who were treated at our institution from 2009 to 2011 were identified. Tinnitus Reaction Questionnaire and Insomnia Severity Index responses were obtained through written and telephone interviews. A Pearson product moment correlation was used to determine the effect of insomnia on tinnitus. Additional analyses identified whether Tinnitus Reaction Questionnaire scores were associated with a possible benefit from an evaluation for insomnia in tinnitus patients. Results: A total of 117 patients met inclusion criteria. A significant correlation was found between the Insomnia Severity Index score and Tinnitus Reaction Questionnaire severity (r = 0.64; P = .001). Tinnitus Reaction Questionnaire severity was shown to be a good predictor of sleep disturbance and good in predicting group association, especially the “emotional” subscore component (sensitivity 96.9% and specificity 55.3% for identifying tinnitus patients with insomnia). The greater the insomnia disability as exhibited by an elevated Insomnia Severity Index score, the more severe the patient’s complaints were regarding the tinnitus. Conclusion: Results suggest that if the emotional score on the Tinnitus Reaction Questionnaire is ≥ 15, the Insomnia Severity Index may be useful to identify patients who may benefit from further treatment and evaluation of insomnia. The robust correlation between the Tinnitus Reaction Questionnaire and Insomnia Severity Index objectively showed that patients with insomnia have an increased emotional distress associated with their tinnitus. Both questionnaires can be used together with a high degree of specificity and sensitivity in predicting tinnitus patients with an underlying sleep disturbance. Keywords insomnia, tinnitus
Introduction More than 36 million patients are plagued with tinnitus.1 Evidence-based research has shown a strong correlation between tinnitus and other psychological disturbances. Researchers have long been challenged regarding the effects of tinnitus on other comorbidities and few systematic studies have been done. Tinnitus may be associated with worsening of anxiety, depression, irritability, and sleep disturbances. Of these, the relationship between depression and tinnitus has been well documented through the validated Tinnitus Handicap Index in relation to the Beck Depression Inventory questionnaire.2 Recent epidemiological studies found an association between sleep disturbance (ie, insomnia) and a decreased capacity for dealing with stress due to pain stimuli.1 The
studies concluded that sleep disturbance may be linked to an increased severity of the previously mentioned psychological disturbances. Few reports have documented the relationship between tinnitus and sleep. Such studies have been 1
Department of Otolaryngology–Head and Neck Surgery, Henry Ford Hospital, Detroit, Michigan, USA 2 Department of Pulmonary and Critical Care Medicine, Section of Sleep Medicine, Henry Ford Hospital, Detroit, Michigan, USA 3 Department of Public Health Sciences, Henry Ford Hospital, Detroit, Michigan, USA Corresponding Author: George S. Miguel, DO, Department of Otolaryngology–Head and Neck Surgery, Henry Ford Hospital, 2799 West Grand Boulevard, Suite K-8, Detroit, MI 48202, USA. Email: [email protected]
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
697
Miguel et al done in the geriatric population, correlating chronic tinnitus with sleep disturbances based on initiation of sleep and sleep maintenance.3 However, most studies are mainly suggestive that sleep difficulties are related to increased tinnitus severity. Moreover, most reports are brief and include only a few questions on sleep without statistical significance. Folmer and Griest4 concluded that their findings underscored the importance of proper identification and successful treatment of patients with tinnitus and insomnia. An obstacle has been the ability to properly identify patients whose sleep disturbance may be worsening their perception of tinnitus. It is important to distinguish and be aware of these 2 clinical entities in order to correctly identify these patients so that proper treatment can be initiated earlier, resulting in higher patient satisfaction. Roth and AncoliIsrael1 concluded that chronic pain sufferers have a decreased tolerability to other chronic conditions such as tinnitus or insomnia. It is unfortunate that it is difficult to discern whether sleep makes tinnitus worse or the reverse; however, we can accept the fact that they work synergistically together. The accepted and validated questionnaire indexes for tinnitus5 and insomnia6 patients, respectively, have not been statistically correlated or compared in these patient groups. This study uses these validated questionnaires in 2 groups of tinnitus patients (tinnitus and tinnitus with insomnia) to determine the correlation between tinnitus and insomnia. In doing so, we intend to show that understanding the relationship between tinnitus and insomnia is important in the management and proper identification of these patients early in the treatment process. It is the hope that by showing correlation between these 2 validated questionnaires, a clinician can evaluate and treat patients with tinnitus from a different angle once all other reasons have been exhausted.
Methods This retrospective observational study was approved by the Henry Ford Hospital Institutional Review Board. Patients with a diagnosis of tinnitus only (ICD-9 code 388.30, 388.31, 388.32) and those with both tinnitus and insomnia (ICD-9 code 780.52) who were seen between January 2009 and December 2011 were identified in the administrative database and placed into 2 groups. The patients were referred and seen at the Henry Ford Tinnitus Clinic. Tinnitus Reaction Questionnaire and Insomnia Severity Index responses were obtained through written and telephone interviews. Patients whose diagnosis of insomnia was made prior to a diagnosis of tinnitus were excluded from the study. This ensured that tinnitus was the initial and primary complaint of the patient rather than insomnia. Other exclusion criteria included deceased patients, those lost to follow-up, and those who refused to participate in the study. Demographics of the patient included an average age of 63
years, and the study included 72 men and 45 women. Most of these patients were referred to our tinnitus clinic after exhaustive medical treatment from the community. However, we did not analyze the treatments they received (ie, medications, behavioral, etc) with regard to their tinnitus. These 2 groups were made based on ICD-9 code only and insomnia was not ruled out for those in the tinnitus-only group. We wanted to see in the tinnitus-only group if there was, inherently, undiagnosed sleep disturbance. Those with a primary diagnosis of insomnia were not addressed because the intent was to see how tinnitus was affected by insomnia rather than how insomnia affected someone’s quality of life. The Tinnitus Reaction Questionnaire5 was developed to assess the psychological disturbance one has from tinnitus. It is a series of 26 questions looking at one’s perception in categories that can involve feelings of anger, annoyance, confusion, and so on. Each question is based on a score of 0 points (not at all) to 4 points (almost all of the time) with a maximum total of 104. The Insomnia Severity Index6 is a reliable and valid instrument to quantify perceived insomnia severity. The Insomnia Severity Index is a series of 7 questions asking about sleep patterns, behavior, and one’s perception of sleep. Each question is scored on a scale from 0 (none) to 4 (severe). Total scores are as follows: 0-7 = no clinically significant insomnia; 8-14 = subthreshold insomnia; 15-21 = clinical insomnia (moderate severity); 22-28 = clinical insomnia (severe). Statistical analysis examined the degree of linear relationship between Tinnitus Reaction Questionnaire and Insomnia Severity Index scores using the Pearson correlation coefficient. The overall score and scores from 3 subscales used in the Tinnitus Reaction Questionnaire (emotional, functional, and catastrophic) were evaluated using this coefficient. Correlation estimates were compared using a method for both independent and dependent correlations. The mean Tinnitus Reaction Questionnaire scores in the 2 groups (tinnitus alone and tinnitus and insomnia) were compared using Student t tests. This methodology was also used to compare the 3 subscales. The areas under the curve of the receiver operating characteristics curve were estimated using linear logistic regression. This score between 0 and 1, which measures the amount of predictability between the variables, was then used to find the cutoff point that best described the sensitivity and specificity. We maximized the sum of these 2. Estimates of the 2 screening characteristics and their associated 95% confidence intervals (CIs) were estimated. A P value less than .05 was considered statistically significant.
Results Of the 117 patients who met the inclusion criteria, 85 were in the tinnitus-only group and 32 were in the tinnitus +
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
698
Annals of Otology, Rhinology & Laryngology 123(10)
Table 1A. Correlation Between Insomnia Severity Index and Tinnitus Reaction Questionnaire. Comparison ISI vs TRQ ISI vs emot+ ISI vs funct+ ISI vs catas+
Population Correlation P Value* All
0.63 0.59 0.59 0.58
.001 .001 .001 .001
Group
Correlation P Value*
Tinnitus + insomnia
.001 .002 .001 .005
0.64 0.52 0.52 0.48
Group
Correlation
P Value*
Tinnitus
0.55 0.52 0.53 0.52
.001 .001 .001 .001
Abbreviations: catas, catastrophic; emot, emotional; funct, functional; ISI, Insomnia Severity Index; TRQ, Tinnitus Reaction Questionnaire. *Testing against 0. +Dependent correlation test P value, emot vs funct 0.802, emot vs catas 0.822, funct vs catas 0.985.
Table 1B. Correlation Between Insomnia Severity Index and Tinnitus Reaction Questionnaire. Comparison
Correlation
P Valuea
Tinnitus + insomnia vs tinnitus
ISI vs TRQ ISI vs emot+ vs funct+ vs catas+
.522 .985 .955 .832
Abbreviations: catas, catastrophic; emot, emotional; funct, functional; ISI, Insomnia Severity Index; TRQ, Tinnitus Reaction Questionnaire. a Independent correlation test. +Dependent correlation test P value, emot vs funct 0.802, emot vs catas 0.822, funct vs catas 0.985.
110 100 90 80 70 60 50 40 30 20 10 0
insomnia group. The correlation between the Tinnitus Reaction Questionnaire and Insomnia Severity Index scores was moderately high (0.62) and significant (P = .001) for all patients (Table 1). The correlation was 0.64 for the tinnitus + insomnia group and lower at 0.55 for the tinnitus-only group. The 2 correlations when compared to each other, however, were not significantly different (P = .522). The positive correlation indicates that as Insomnia Severity Index scores increase, so do Tinnitus Reaction Questionnaire scores (Figure 1). Table 1 also provides the correlations between the Tinnitus Reaction Questionnaire and Insomnia Severity Index for the 3 subscales of the Tinnitus Reaction Questionnaire. These correlations ranged around 0.5 and all were significantly different from 0. We tested them against one another and none of the comparisons were significant. The Tinnitus Reaction Questionnaire and mean Insomnia Severity Index scores were tested for differences between the 2 groups (Table 2). All comparisons were significant with higher scores seen for the tinnitus + insomnia group. The areas under the curve were estimated as 0.786 with a 95% CI of 0.69 to 0.87. Areas under the curve estimates for the 3 subscales were 0.793 for emotional, 0.720 for functional, and 0.786 for catastrophic (Figure 2). Table 3 shows that the emotional subscale had the highest area under the curve at 0.79 (95% CI, 0.71-0.88). We used the emotional subscale result to find the best cutoff point to maximize the sum of specificity
Correlation Between TRQ and ISI
trq 120
0
10
20 isi both only group correlations for both(solid)=0.64, for only(dashed)=0.55 p-value=0.522
30
Figure 1. Data plots for correlation between Tinnitus Reaction Questionnaire and Insomnia Severity Index.
and sensitivity. A score of x ≥ 15 resulted in an estimated sensitivity of 96.9% (88.8, 99.9) and a specificity of 55.3% (44.1, 66.1) for predicting the presence of insomnia. Using this criterion, 38 of the tinnitus-only patients were incorrectly classified. Of these 38, 29 (76.3%) reported a sleep disturbance score of ≥ 3 despite not having insomnia according to the Insomnia Severity Index (Table 4).
Discussion Assessing the severity of tinnitus in patients can be difficult for a variety of reasons. A chronic tinnitus patient presents a multidimensional clinical picture consisting of anxiety, depression, annoyance, or self-reported emotional distress. In 1983, Tyler and Baker7 asked patients to list the difficulties experienced as a result of tinnitus, and the most frequent complaint reported was “getting to sleep.” The association between tinnitus and depression is well known, and studies have shown a high correlation between
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
699
Miguel et al Table 2. Comparison Between Tinnitus + Insomnia Versus Tinnitus Group Scales and Insomnia Severity Index. Tinnitus
Tinnitus + Insomnia
Scale
N
Mean
SD
n
Mean
SD
P Value
TRQ overall TRQ emotional TRQ functional TRQ catastrophic ISI
85 85 85 85 85
42.4 15.5 17.3 9.7 12.1
25.3 9.6 11.9 8.3 7.2
32 32 32 32 32
71.6 26.6 26.1 18.8 17.5
20.3 9 9.4 7 5.1
.001 .001 .001 .001 .001
Abbreviations: ISI, Insomnia Severity Index; TRQ, Tinnitus Reaction Questionnaire.
Table 3. Area Under the Curve. ROC Curve for TRQ Factors tinnitus plus insomnia
Sensitivity
1.0 0.9 0.8 0.7
Scale
AUC (95% CI)
TRQ overall TRQ emotional TRQ functional TRQ catastrophic
0.79 (0.70-0.87) 0.79 (0.71-0.88) 0.72 (0.62-0.82) 0.79 (0.69-0.88)
0.6
Abbreviations: AUC, area under the curve; CI, confidence interval; TRQ, Tinnitus Reaction Questionnaire.
0.5 0.4 0.3
Table 4. Classification Rule in Predicting Insomnia With Tinnitus Reaction Questionnaire and Insomnia Severity Index.a
0.2 0.1 0.0 0.0
0.1
0.2 curve
0.3
0.4 0.5 0.6 1 - Specificity Catastrophic Emotional
0.7
0.8
0.9
Tinnitus + Insomnia Tinnitus
1.0
Functional
area under curve: Cata=0.786, Emot=0.793, Func=0.720
Figure 2. Data plots for the area under the curve.
Predict Patient to have tinnitus + insomnia (emotional score ≥ 15) Patient to have tinnitus only (emotional < 15) Total N
31
38
1
47
32
85
a
depressive symptoms and tinnitus severity. In 1993, Alster et al8 were the first to evaluate the possibility of sleep disturbance affecting the severity of tinnitus. Their data suggested that those who sought sleep evaluations and treatment noted a decrease in their tinnitus complaint. In daily clinical practice, it is uncommon for physicians to ask specifically about sleep disturbances in tinnitus patients and for patients with insomnia to be evaluated for tinnitus. This study supports previous conclusions that increased sleep disturbance correlates with an increase in tinnitus severity.3,7-9 The Tinnitus Reaction Questionnaire can be used initially to determine the severity of a patient’s perception of tinnitus. This study supports the use of the Insomnia Severity Index to help identify patients recalcitrant to optimized medical therapy and who may benefit from an evaluation of their sleep complaints. There was no single correlation between the Tinnitus Reaction Questionnaire and Insomnia Severity Index that showed a statistical significance in differentiating the 2 groups of patients (Table 1). Because the tinnitus + insomnia group had a diagnosis of
Of the 38 tinnitus-only with emotional ≥ 15, 76.3% have sleep score ≥ 3. Sensitivity = 96.9%. Specificity = 55.3%.
insomnia, there was an assumption that the correlation between the Tinnitus Reaction Questionnaire and Insomnia Severity Index would be higher compared to those without a diagnosis of insomnia. If the correlations between these 2 groups were close and significant, this might indicate that the Insomnia Severity Index is diagnostic. For example, the tinnitus-only group might suffer from inherent sleep problems. However, this was not the case. Rather, the predictability of inherent sleep problems from the subscale of the Tinnitus Reaction Questionnaire score was examined with use of the area under the curve. This showed a statistical significance for all 3 subscales with the emotional component having the highest effect. Table 4 shows the rule in predicting a patient to have inherent sleep disturbance, that is, insomnia, if the combined scores are ≥ 15. As mentioned previously, 29 (76.3%) of 38 tinnitus-only patients were incorrectly classified; these patients reported a sleep disturbance score of ≥ 3
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
700
Annals of Otology, Rhinology & Laryngology 123(10)
despite not having insomnia according to the Insomnia Severity Index. Correct classification could be done by looking at their summarized emotional subscore, which was > 15. This is a good example that 1 question regarding sleep in questionnaires is not a sufficient indicator for the presence or absence of insomnia. The classification rule shown in Table 4 may guide a clinician to approach these tinnitus patients with an Insomnia Severity Index questionnaire and further sleep evaluation. Limitations of our study include the difficulty to assess whether tinnitus or insomnia contributed to the patients’ initial complaint. We attempted to control this variable by excluding those who were diagnosed with insomnia before being seen for tinnitus. The group recruited was based on ICD-9 codes and that insomnia was not completely ruled out in the tinnitusonly group. As seen, sleep disturbance is underdiagnosed and not identified routinely. However, we should also note that there is the possibility that sleep disturbance may have been addressed in these patients and the diagnosis of insomnia was never coded. Furthermore, the size of the cohort has its limitations, and the strength of the correlations might be improved with a larger size population. Stouffer et al9 asked 528 patients to rate different conditions that reduced the severity of their tinnitus, and 26% responded sleep. This article has shown that the Tinnitus Reaction Questionnaire and Insomnia Severity Index can be used together in guiding a clinician to identify and treat those who have sleep disturbance. Different treatment regimens may consist of treating their sleep disturbance with medications or cognitive behavioral therapy. Melatonin has been shown to be useful in decreasing the severity of tinnitus in this subgroup of patients.10 This study also found an improvement in symptoms with placebo. Tinnitus is a process involving cognitive, emotional, and psychophysiological components. Such factors increase distress in patients, and studies have reported that tinnitus sufferers show maladaptive behaviors when exposed to a stress/pain stimulus.11 Thus, sleep complaint in these patients can be the consequence of a combined effect of a decrease in their tolerance to tinnitus. As shown, the results support the notion that sleep disturbance is underdiagnosed in this population and that treating their sleep disturbance may help reduce their symptomatology. This study demonstrates the use of a validated questionnaire in identifying patients whose sleep complaints may be responsible for a worsening of their perception of tinnitus. Further studies looking at this cohort after a proper sleep evaluation and treatment may be done to better understand how these 2 clinical entities relate.
Conclusion The identification of patients with insomnia with worsening tinnitus symptoms may assist in medical management. The
study results are consistent with those of previous investigations that sleep disturbance affects a patient’s perception of tinnitus severity. These questionnaires give objective measures in classifying tinnitus sufferers and their insomnia complaints. The Tinnitus Reaction Questionnaire and Insomnia Severity Index may be beneficial when used together in classifying this patient subgroup that has concomitant insomnia problems with a sensitivity of 96.9% and specificity of 55.3%. Authors’ Note This article was originally presented as a poster presentation at the Triological Society Meeting in San Diego, California, on April 21, 2012.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References 1. Roth T, Ancoli-Israel S. Daytime consequences and correlates of insomnia in the United States: results of the 1991 National Sleep Foundation Survey. Sleep. 1999;2(suppl 22):S354-S358. 2. Budd RJ, Pugh R. The relationship between locus of control, tinnitus severity, and emotional distress in a group of tinnitus sufferers. J Psychosom Res. 1995;39:1015-1018. 3. Asplund R. Sleepiness and sleep in elderly persons with tinnitus. Arch Gerontol Geriatr. 2003;37:139-145. 4. Folmer R, Griest S. Tinnitus and insomnia. Am J Otolaryngol. 2000;21:287-293. 5. Wilson P, Henry J, Bowen M, Haralambous G. Tinnitus Reaction Questionnaire: psychometric properties of a measure of distress associated with tinnitus. J Speech Hearing Res. 1991;34:197-201. 6. Bastien CH, Vallières A, Morin C. Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Med. 2001;2:297-307. 7. Tyler RS, Baker LJ. Difficulties experienced by tinnitus sufferers. J Speech Hear Disord. 1983;48:150-154. 8. Alster J, Shemesh Z, Ornan M, Attias J. Sleep disturbance associated with chronic tinnitus. Biol Psychiatry. 1993;34: 84-90. 9. Stouffer JL, Tyler RS, Kileny PR, Dalzell LE. Tinnitus as a function of duration and etiology: counselling implications. Am J Otol. 1991;12:188-194. 10. Rosenberg SI, Silverstein H, Rowan PT, Olds MJ. Effect of melatonin on tinnitus. Laryngoscope. 1998;108:305-310. 11. Heinecke K, Weise C, Schwarz K, Rief W. Physiological and psychological stress reactivity in chronic tinnitus. J Behav Med. 2008;31:179-188.
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 28, 2015
