The effect of singing on snoring and daytime somnolence
- First Online:
- Cite this article as:
- Pai, I., Lo, S., Wolf, D. et al. Sleep Breath (2008) 12: 265. doi:10.1007/s11325-007-0159-1
The objective of the study is to compare the prevalence and severity of snoring and daytime somnolence amongst semiprofessional choir singers and non-singers. It is a cross-sectional comparative study and the setting is at a tertiary otorhinolaryngology referral centre. Adult singers were recruited from two mixed-gender choirs in London. The control group consisted of healthy volunteers who do not sing. The weight and height of all participants were measured by a single investigator. A questionnaire was completed by each subject, and the snoring habit section completed by their spouses or partners. The snoring scale score (SSS) and Epworth Sleepiness Scale (ESS) were utilised to assess the severity of snoring and daytime somnolence, respectively. The mean age of the singers was 46.3 years (20:32, males to females) and the control group 43.3 years (23:32, males to females). There was no difference in body mass index (BMI; p = 0.180) and ESS score (p = 0.770) between singers and non-singers. Regression analysis showed no significant relationship between the number of years of singing and ESS score (p = 0.390) although there was a linear relationship between age and SSS for both singers (R2 = 0.11; p = 0.02) and non-singers (R2 = 0.20; p = 0.01). Based on the general linear model, singers have significantly lower SSSs compared to non-singers when adjusted for age (p = 0.0147), BMI (p = 0.0389) and both age and BMI (p = 0.0153). Singing practice may have a role in the treatment of snoring but does not appear to influence daytime somnolence.
KeywordsSleep-disordered breathingSnoringDaytime somnolenceSinging
Sleep-disordered breathing is a condition that many respiratory physicians and otolaryngologists commonly encounter. Symptoms may range from simple palatal snoring to obstructive sleep apnoea (OSA) with oxygen desaturations and frequent nocturnal arousals. As well as causing social embarrassment to affected individuals and annoyance to bed partners, it has been suggested that sleep-disordered breathing, in particular OSA–hypopnoea syndrome, is an independent risk factor for cardiovascular disease, hypertension, type II diabetes mellitus and occupational and road traffic accidents [1–7]. In a recent cross-sectional study involving patients in the primary care setting in the United States, Germany and Spain, snoring was shown to be a significant problem in over 40% of these subjects and approximately one third were considered to be at high risk of OSA .
Current treatment modalities for snoring and OSA range from conservative lifestyle modifications to invasive surgery, depending on the nature and severity of the problem. Surgical interventions, which may be considered when less invasive measures have failed, carry the risk of potential complications, and the overall long-term success rate is no more than 50% [9, 10]. For many patients with sleep-disordered breathing, especially those with significant simple snoring (apnoea–hypopnoea index <10), there exists no treatment option that is effective, acceptable and justifiable. For this reason, there is considerable interest in the development of alternative, noninvasive treatment modalities.
In an interesting pilot study, Ojay and Ernst  hypothesised that vocal training may reduce snoring by improving the pharyngeal muscle tone. The study involved 20 chronic snorers (snoring every night, aged between 18 and 65 years) who received instructions on singing techniques and singing exercises. After 3 months, the candidates were reassessed using voice-activated tape machines recording the episodes of snoring, and non-overweight patients were found to experience a significant reduction in snoring. In view of this finding, our group conducted a cross-sectional study to compare the prevalence and severity of snoring and daytime somnolence between semiprofessional singers and non-singers.
Materials and methods
The singer group consisted of 52 individuals (20:32, males to females) with a mean age of 46.3 years (range = 26–70). The mean number of years of choir singing was 30.9 years (range = 5–62). The non-singer group consisted of 55 individuals (23:32, males to females) with a mean age of 43.3 years (range = 16–74). None of the subjects in this group sang regularly.
Number (males to females)
Mean age (years)
Three singers were known to have OSA, two of whom were receiving continuous positive airway pressure (CPAP) treatment. They had ESS scores of 0, 4 and 6 and SSSs of 0, 5 and 8, respectively. Seven singers suffered from chronic nasal obstruction, including the three subjects with OSA. Only one patient with nasal obstruction, who did not have OSA, required nasal steroid spray treatment. However, even after excluding all the singers suffering from nasal obstruction and OSA, the difference in SSS between singers and non-singers remained statistically significant (Wilcoxon rank-sum test; p = 0.002). Amongst the non-singer group, none of the subjects suffered from nasal obstruction or OSA.
One of the key aspects of learning to sing is to master correct breathing control, which is aided by various breathing techniques. It has been suggested that singing exercises may therefore lead to an improvement in the control of soft palate and pharyngeal muscles . We therefore hypothesised that trained, regular singers, in the absence of other co-morbidities, are less likely to snore compared to non-singers. Semiprofessional singers were recruited into the current study, all of whom had received formal singing training. The least number of years of choir singing was 5 years, significantly longer than the 3 months of singing prescribed by Ojay and Ernst . Our results show that choir singers have significantly lower SSS values compared to non-singers when adjusted for age and BMI. However, ESS values are not statistically different between the two groups, suggesting that daytime somnolence is not affected by singing. ESS scores and SSSs are not directly comparable as the former test assesses daytime somnolence and the latter one tests loudness, periodicity and frequency of snoring. Although three subjects in the singer group had OSA, their ESS scores are not particularly high, possibly due to the use of CPAP. Based on the current study model, it is not possible to determine whether or not regular singing practice has had any effects on their OSA. However, even after excluding subjects with OSA and nasal obstruction, non-singers were found to have higher SSS values compared to singers. In our study, there appeared to be no relationship between the number of years singing and daytime somnolence.
Our results show that with increasing age there is an increase in SSS values amongst both singers and non-singers. This may be due to an increase in soft tissue laxity in the oropharynx and surrounding structures. Singers, nonetheless, have lower SSS values compared to non-singers when matched for age, supporting our hypothesis that singing practice may have a beneficial effect on snoring by improving oropharyngeal muscle tone. As illustrated in Fig. 2, the difference in SSS values between singers and non-singers increases with increasing age. It is therefore possible that older snorers, who are more likely to have significant oropharyngeal laxity, would benefit more from singing practice.
In this study, the mean BMI values in both singer and non-singer groups are within the normal ranges and many of the subjects would be considered to be at low risk for snoring. The results, however, show a positive linear relationship between BMI and SSS values in both groups (Fig. 3). This is consistent with previous findings in the medical literature, which highlight high BMI as an important risk factor for sleep-disordered breathing [14–16]. As was the case with age, singers were found to have significantly lower SSS values compared to non-singers when matched for BMI. Interestingly, as BMI approaches 40, the difference in SSS values between the two groups decreases, suggesting that snorers with normal BMI are more likely to benefit singing practice. This is in keeping with the results of the pilot study by Ojay and Enrst, in which normal BMI was shown to be a good prognostic factor, together with compliance with singing practice and onset of snoring in adulthood.
Our study does have a number of limitations. First of all, SSS values are overall low in both singer and non-singer groups and it may be argued that many subjects in this study are unlikely to have problems with snoring anyway. Secondly, although there is a statistically significant difference between the two groups in SSS values, the magnitude of the difference is small and it is not possible to determine whether or not this small difference is clinically relevant. Perhaps, most importantly, no conclusion can be drawn about cause and effect as this is purely an observation from a population study.
Our results suggest that singing practice may have a role in the treatment of simple snoring. Although firm conclusions cannot be drawn from the data, our study does provide support for conducting further research into this noninvasive and novel approach to the management of a common yet difficult problem. An ideal study would be a prospective controlled trial involving patients with simple snoring, randomized to regular singing practice or no intervention. Our study also suggests that patient selection is important for an accurate evaluation. Based on current evidence, studies should initially involve patients who are not clinically obese and do not have OSA or anatomical risk factors warranting surgical intervention.