Sleep Medicine 10 (2009) 657660

Contents lists available at ScienceDirect

Sleep Medicine
journal homepage: www.elsevier.com/locate/sleep

Original Article

Risk of sleep apnea in orchestra members

Devin L. Brown a,*, Darin B. Zahuranec a, Jennifer J. Majersik a, Patricia A. Wren c, Kirsten L. Gruis d, Michael Zupancic d, Lynda D. Lisabeth a,b
a

Stroke Program, University of Michigan Medical School, The Cardiovascular Center Stroke Program, 1500 E. Medical Center Drive - SPC#5855, Ann Arbor, MI 48109-5855, USA Department of Epidemiology, University of Michigan School of Public Health, University of Michigan, Ann Arbor, MI 48109, USA c School of Health Sciences, Oakland University, Rochester, MI 48309, USA d Department of Neurology, University of Michigan, 1500 East Medical Center Drive, Ann Arbor, MI 48109, USA
b

a r t i c l e

i n f o

a b s t r a c t
Background: Obstructive sleep apnea (OSA) is a common condition with substantial health consequences. A recent randomized trial found that playing the didgeridoo improved both subjective and objective sleep measures. We undertook a cross-sectional survey of professional orchestra players to test the hypothesis that playing a wind instrument would be associated with a lower risk of OSA. Methods: An anonymous internet-based survey of professional orchestra members assessed risk of sleep apnea using the Berlin questionnaire. Multivariable logistic regression was used to test the association between playing a wind instrument and having a high risk score on the Berlin questionnaire, both unadjusted and adjusted for age, body mass index, and gender. Results: A total of 1,111 orchestra members responded, including 369 (33%) wind instrument players. Wind players were more often male and had a higher body mass index than non-wind players. Of all musicians, 348 (31%) had a high risk of sleep apnea. Wind players were more likely than non-wind players to be at high risk in unadjusted analysis (Odds ratio = 1.47, 95% CI 1.13, 1.91), though this association was not signicant in adjusted analysis (Odds ratio = 1.12 (0.82, 1.54)). Conclusion: Playing a wind instrument was not associated with a lower risk of OSA. 2008 Elsevier B.V. All rights reserved.

Article history: Received 27 December 2007 Received in revised form 5 April 2008 Accepted 11 May 2008 Available online 17 November 2008 Keywords: Sleep apnea obstructive Questionnaires Sleep apnea syndromes Epidemiology Sleep Risk factors

1. Introduction Obstructive sleep apnea (OSA) is an important medical condition that predisposes patients to cerebrovascular and cardiovascular disease [13]. Although the standard treatment is continuous positive airway pressure, compliance is often poor, suggesting the need for alternative therapies [4]. Recently, a randomized trial showed that playing the didgeridoo, a wind instrument of the indigenous Australians, was associated with a reduction in sleepdisordered breathing events and daytime sleepiness [5]. Playing the didgeridoo requires the use of circular breathing, a technique used to produce a continuous tone without break, accomplished by the use of the cheeks as a reservoir of air while breathing through the nose rather than the mouth. It has been suggested that practicing this wind instrument may train airway muscles leading to less collapse of oropharygeal muscles at night, resulting in its benecial effect on sleep apnea. To our knowledge, no epidemiological study has tested for an association between OSA, or its associated features, and frequent use of wind instruments. We hypothesized that playing a wind

instrument would be associated with a reduced risk of OSA. To test this hypothesis, we undertook a cross-sectional survey of professional orchestra players.

2. Methods We conducted a web-based survey of professional orchestra members. This study was granted an exemption by the University of Michigan Institutional Review Board. 2.1. Subjects The International Conference of Symphony and Opera Musicians (ICSOM) is a professional organization that represents over 4000 orchestral musicians from the top 52 American Federation of Musician orchestras in the United States and Puerto Rico. Orchestras range in size, the largest ones having over 100 musicians. They are typically comprised of four main sections: strings, winds, percussion, and keyboard. Wind instrument players typically represent approximately 30% of the orchestra and are almost equally divided into two groups: woodwinds and brass. ICSOM has email addresses on 3665 of their approximately 4300 active musician members.

* Corresponding author. Tel.: +1 734 936 9075; fax: +1 734 232 4447. E-mail address: devinb@umich.edu (D.L. Brown). 1389-9457/$ - see front matter 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.sleep.2008.05.013

658

D.L. Brown et al. / Sleep Medicine 10 (2009) 657660

2.2. Sample size and power We anticipated roughly 1000 respondents and made the following assumptions: type 1 error of 0.05, 80% power, 26% prevalence of high risk for sleep apnea in non-wind instrument players [6], and a ratio of wind to non-wind player respondents of 1:2. Given these assumptions, the sample size of 1000 would allow for a detectable difference in prevalence of high risk for sleep apnea between non-wind and wind instrument players of 8% or, alternatively, an odds ratio of 0.64 [7]. 2.3. Study procedures ICSOM sent an email to all active members for whom they had a registered email address that introduced the study and offered anonymous participation through a hyperlink to the survey. Two follow-up emails were sent approximately 1 week apart to remind subjects of the survey. 2.4. Survey instrument The questions asked in the survey included the Berlin questionnaire, a validated tool that predicts sleep apnea risk [8]. Berlin questionnaire results were scored in the standard fashion resulting in a dichotomized high risk for OSA and lower risk for OSA. Subjects who earned high risk scores for at least two of the three symptom categories were considered to be at high risk for OSA [8]. Additional questions were asked about demographics, medical history including diagnosis of OSA by a physician, and other baseline characteristics including height and weight. Subjects were asked to identify the primary instrument played in the orchestra, number of hours per week the instrument was typically played, and the age at which playing the instrument was begun. To assess for misclassication of exposure to playing wind instruments, subjects were asked about any other instrument that they routinely played more than 1 h per week aside from the primary instrument played in the orchestra. Use of circular breathing was also assessed in a dichotomous fashion. 2.5. Statistical analysis Demographics and baseline characteristics were assessed using descriptive statistics. Body mass index (BMI) was calculated from the self-reported height and weight using the formula [(weight in pounds/height in inches2) 703]. Comparisons were made between those at high risk and those at lower risk for sleep apnea based on their Berlin scores using chi square tests for dichotomous variables and t-tests for continuous variables. To test the primary hypothesis, logistic regression was used to assess the association between wind instrument playing (compared with all other instrument groups combined) and being at high risk for sleep apnea. Adjustment for age, BMI, and gender were pre-specied before data collection. Age and BMI were treated continuously in the model. Odds ratios (OR) with 95% condence intervals (CI) were calculated. In exploratory analyses, wind instrument players were further divided into woodwinds and brass. Logistic regression was used to assess the association between woodwind and brass instrument playing and being at high risk for sleep apnea. In this model, woodwind and brass instrument playing were modeled as dummy variables with all other instruments as the referent group. The association between high sleep apnea risk and use of circular breathing was then tested by comparing those wind players who use circular breathing with all other instrument players. The association between a physician diagnosis of OSA and (1) playing a wind instrument compared with all other instruments and (2)

wind players who use circular breathing compared with all others was also tested. Exploratory analyses were conducted with and without adjustment for age, BMI, and gender. All analyses were conducted using S-plus 7.0 for Windows. 3. Results There were a total of 1111 survey respondents (30% response rate). Demographic characteristics are found in Table 1. Six respondents did not indicate that any instrument was played, leaving 1105 for analysis. There were a total of 369 wind players: 175 (47%) played a brass instrument, while 194 (53%) played a woodwind. Fifty-ve (15%) of the wind players used circular breathing. Of the non-wind players, 670 played a string instrument (bass (n = 92, 14%), cello (127, 19%), viola (146, 22%), violin (305, 46%)), 55 played percussion (harp (16, 29%), percussion (30, 55%), timpani (9, 16%)), and 12 played a keyboard instrument. On average, wind players played their instrument 25.8 h per week (SD = 8.8); while non-wind players played an average of 27.3 h per week (SD = 8.6). Only one non-wind player played a wind instrument (ute) more than 1 h per week. On average, wind players started playing their instrument at age 11.3 (SD = 3.2); while non-wind players began playing their instrument at age 9.0 (SD = 4.6). Of the 1105 respondents, 348 (31%) had high risk for sleep apnea determined by their Berlin questionnaire scores. In bivariate analysis, playing a wind instrument was positively associated with a high sleep apnea risk (OR = 1.47 (95% CI: 1.13, 1.91)), but after adjusting for age, gender, and BMI, this association was not significant (OR = 1.12 (95% CI: 0.82, 1.54)). Age, BMI, and male gender were positively associated with high sleep apnea risk. Results of the multivariable model are found in Table 2. When assessing brass players alone (referent group all non-wind instrument players), results were similar, with a signicant association in bivariate analysis (OR = 2.24 (95% CI: 1.60, 3.14), but a nonsignicant result (OR = 1.36 (95% CI: 0.91, 2.03)) after adjusting for age, gender, and BMI. When assessing woodwind players only (referent group all non-wind instrument players), the association was not signicant in either bivariate (OR = 0.96 (95% CI: 0.67, 1.37)) or adjusted analysis (OR = 0.94 (95% CI: 0.62, 1.40)). Among all instrument groups, circular breathing was not associated with a high risk of sleep apnea in bivariate (OR = 0.98 (95% CI: 0.54, 1.76)) or adjusted (OR = 0.71 (95% CI: 0.34, 1.46)) analyses. There was no association between playing a wind instrument and having a physician diagnosis of OSA in bivariate (1.60 (95%

Table 1 Demographics by instrument played Wind (n = 369) N (% or 95% CI) Age Weight Male Non-hispanic Race Caucasian Asian Other BMI Hypertension Smoking Current Previous Never Known OSA Tonsillectomy High risk Berlin score
*

Non-wind (n = 736) N (% or 95% CI) 46.8 (45.9, 47.7) 162.0 (159.3, 164.6) 330 (48) 651 (97) 616 (91) 48 (7) 11 (1) 24.7 (24.4, 25.0) 94 (14) 40 (6) 172 (25) 472 (69) 35 (5) 205 (30) 211 (29)

p-value 0.8358 <0.001 <0.001 0.41 0.0039

46.7 (45.4, 47.9) 176.5 (171.9, 181.0) 233 (68) 328 (98) 331 (97) 5 (1) 5 (1) 26.1 (25.5, 26.6) 61 (18) 11 (3) 67 (20) 263 (77) 27 (8) 107 (31) 137 (37)

<0.001 0.21 0.017

0.075 0.655 0.0043

Numbers do not all sum to total n due to missing responses.

D.L. Brown et al. / Sleep Medicine 10 (2009) 657660 Table 2 Results of the multivariable model predicting a high risk of OSA based on the Berlin questionnaire Odds ratio (95% Condence interval) Age BMI Femalea Wind playerb
a b

659

1.39 3.24 0.56 1.12

(1.09, (2.60, (0.40, (0.82,

1.78) 4.03) 0.77) 1.54)

Referent group: male. Referent group: all non-wind players (strings, percussion, keyboard).

Table 3 Results of the multivariable model predicting a physician diagnosis of OSA Odds ratio (95% Condence interval) Age BMI Femalea Wind playerb
a b

1.86 2.78 0.28 0.97

(1.15, (2.08, (0.13, (0.54,

2.99) 3.72) 0.61) 1.73)

Referent group: male. Referent group: all non-wind players (strings, percussion, keyboard).

CI: 0.95, 2.69)) or adjusted analysis (0.97 (95% CI: 0.54, 1.73)). Age, BMI, and male gender were positively associated with having a physician diagnosis of OSA (Table 3). Similarly, there was no association between using circular breathing and having a physician diagnosis of OSA (bivariate: 1.86 (95% CI: 0.70, 4.90); adjusted: 1.16 (95% CI: 0.37, 3.57)). 4. Discussion This cross-sectional survey of professional orchestra members found that there was an association between playing a wind instrument and having a high risk of OSA; however, this association appeared to be explained by three known confounders: age, gender, and BMI. Our a priori hypothesis that there would be an association between playing a wind instrument and having a lower risk of OSA was therefore not supported by the current data. There are several possible explanations for the differences found in the current study compared with a study that showed reduced sleep apnea severity with didgeridoo playing [5]. First, the study of didgeridoo playing may have identied a spurious association, given its small sample size. Second, the two studies addressed slightly different questions given the differences in study designs. And third, the didgeridoo studys ndings may have been due to benets of circular breathing and the current study may not have had sufcient power to identify an association between circular breathing and sleep apnea. Though circular breathing and near closure of the glottis are also used by orchestral wind instrument players, didgeridoo players use the techniques much more consistently and may achieve a greater level of oropharyngeal musculature training. In the single small trial that showed an effect of didgeridoo playing on sleep parameters, subjects practiced for 2.5 h per week. As the wind instrument players in the current study played 26 h per week, the lack of association between wind instrument playing and OSA risk in the current study is not likely due to insufcient practice time. The impetus for the study of didgeridoo playing was anecdotal evidence of OSA improvement in didgeridoo players, rather than knowledge of a strong physiological link between pharyngeal training and sleep apnea severity. In fact, there is little physiological support for the benet of pharyngeal muscle training on sleep apnea. Pharyngeal muscles of patients with OSA share some characteristics with exercise-trained skeletal muscle, such as a higher proportion of type IIa muscle bers [9]. This is thought to develop

from increased daytime muscle activity in OSA patients resulting from a compensatory mechanism [10]. Although this compensatory mechanism in OSA is not sufcient to prevent nocturnal airway collapse, perhaps dedicated pharyngeal exercises could provide a greater stimulus. There is some evidence that pharyngeal muscle in OSA is more fatigable than controls, contributing to airway collapse [11]. It could be hypothesized that exercise of pharyngeal muscle may reduce muscle fatigability through improved oxidative metabolism [12]. Despite the limited physiological substantiation, several investigators have assessed the potential benets of improving pharyngeal muscle tone through exercise/ stimulation [1315]. As an example, a randomized, controlled trial of daytime electrical stimulation of the tongue showed improvement in snoring but not in physiological sleep apnea parameters [13]. Others have investigated the effects of singing training on snoring and found a possible benecial effect [14]. The overall proportion of high risk Berlin scores in the current study was 31%. The Berlin questionnaire is a widely used, simple, validated tool for prediction of sleep apnea risk [8,16]. Prior surveys conducted either in primary care clinics or as random telephone interviews have generally found similar results, with high risk Berlin scores in approximately one-fourth to one-third of participants [6,8,17]. However, the demographics of our study population were different from these other reports. Our population was younger, which should be associated with a lower risk of OSA, and had a higher proportion of men, which should be associated with a higher OSA risk [6]. Nonetheless, it does not appear that orchestra members in general have a different risk of OSA from the population at large. Our nding that male gender, age, and BMI are associated with high risk Berlin scores is consistent with prior reports [6,17]. The under-diagnosis of OSA appears to be an issue in the orchestra community, as it does in the general population. Overall, only 6% of our study population had a physician-given diagnosis of OSA, while 31% had high risk Berlin scores, a nding even more exaggerated in the wind players. This is in keeping with research showing that the vast majority of patients with moderate to severe OSA have not been diagnosed [18]. There are several limitations of this work. We did not diagnose OSA by the gold standard, polysomnography, but used a validated scale instead. However, it is unlikely that this introduced a systematic bias that would have affected our main results. Our survey response was low and respondents may have been more likely than non-respondents to be interested in OSA, possibly due to personal symptoms of a sleep disorder. Given that the prevalence of high OSA risk in the study population was consistent with previous reports, this seems unlikely to have been a large effect. The group of orchestral musicians without email addresses ($15%) was not able to be surveyed and may be different from the group with email. As only a small percentage of our wind player respondents used circular breathing, there may have been inadequate power to detect the potential protective association of this practice. Due to the crosssectional nature of the study, the temporal relationship between playing an instrument and developing a high sleep apnea risk could not be established. However, in most subjects, exposure to the wind instrument did occur early in life. We assembled a well-characterized sample of ICSOM members who were representative of an orchestras composition. The response rate, while low in general, was sufciently well-powered to address the study hypothesis. Overall, our results dampen enthusiasm for further studies investigating the effects of oropharyngeal muscle training on sleep apnea severity. Acknowledgments The authors are grateful to The International Conference of Symphony and Opera Musicians for their collaboration. Speci-

660

D.L. Brown et al. / Sleep Medicine 10 (2009) 657660 [8] Netzer NC, Stoohs RA, Netzer CM, Clark K, Strohl KP. Using the Berlin questionnaire to identify patients at risk for the sleep apnea syndrome. Ann Intern Med 1999;131:48591. [9] Series F, Cote C, Simoneau JA, Gelinas Y, St Pierre S, Leclerc J, et al. Physiologic, metabolic, and muscle ber type characteristics of musculus uvulae in sleep apnea hypopnea syndrome and in snorers. J Clin Invest 1995;95:205. [10] Mezzanotte WS, Tangel DJ, White DP. Waking genioglossal electromyogram in sleep apnea patients versus normal controls (a neuromuscular compensatory mechanism). J Clin Invest 1992;89:15719. [11] Carrera M, Barb F, Sauleda J, Toms M, Gmez C, Agust AG. Patients with obstructive sleep apnea exhibit genioglossus dysfunction that is normalized after treatment with continuous positive airway pressure. Am J Respir Crit Care Med 1999;159:19606. [12] Holloszy JO, Booth FW. Biochemical adaptations to endurance exercise in muscle. Annu Rev Physiol 1976;38:27391. [13] Randerath WJ, Galetke W, Domanski U, Weitkunat R, Ruhle KH. Tonguemuscle training by intraoral electrical neurostimulation in patients with obstructive sleep apnea. Sleep 2004;27:2549. [14] Ojay A, Ernst E. Can singing exercises reduce snoring? A pilot study. Complement Ther Med 2000;8:1516. [15] Lequeux T, Chantrain G, Bonnand M, Chelle AJ, Thill MP. Physiotherapy in obstructive sleep apnea syndrome: preliminary results. Eur Arch Otorhinolaryngol 2005;262:5013. [16] Gami AS, Pressman G, Caples SM, Kanagala R, Gard JJ, Davison DE, et al. Association of atrial brillation and obstructive sleep apnea. Circulation 2004;110:3647. [17] Netzer NC, Hoegel JJ, Loube D, Netzer CM, Hay B, Alvarez-Sala R, et al. Prevalence of symptoms and risk of sleep apnea in primary care. Chest 2003;124:140614. [18] Young T, Evans L, Finn L, Palta M. Estimation of the clinically diagnosed proportion of sleep apnea syndrome in middle-aged men and women. Sleep 1997;20:7056.

cally, this project could not have been completed without the help of Richard Levine. Dr. Lisabeth is supported by an NINDS career development award (K23 NS050161); Dr. Brown is supported by an NINDS career development award (K23 NS051202). None of the authors reports a conict of interest. References
[1] Mooe T, Franklin KA, Holmstrom K, Rabben T, Wiklund U. Sleep-disordered breathing and coronary artery disease: long-term prognosis. Am J Respir Crit Care Med 2001;164:19103. [2] Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328:12305. [3] Young T, Peppard P, Palta M, Hla KM, Finn L, Morgan B, et al. Population-based study of sleep-disordered breathing as a risk factor for hypertension. Arch Intern Med 1997;157:174652. [4] Kribbs NB, Pack AI, Kline LR, Smith PL, Schwartz AR, Schubert NM, et al. Objective measurement of patterns of nasal CPAP use by patients with obstructive sleep apnea. Am Rev Respir Dis 1993;147:88795. [5] Puhan MA, Suarez A, Lo CC, Zahn A, Heitz M, Braendli O. Didgeridoo playing as alternative treatment for obstructive sleep apnoea syndrome: randomised controlled trial. Br Med J 2006;332:26670. [6] Hiestand DM, Britz P, Goldman M, Phillips B. Prevalence of symptoms and risk of sleep apnea in the US Population: Results from the National Sleep Foundation Sleep in America 2005 Poll. Chest 2006;130:7806. [7] Fleiss JL, Tytun A, Ury SHK. A simple approximation for calculating sample sizes for comparing independent proportions. Biometrics 1980;36:3436.