Main finding
We found that singing produced changes in physiological parameters including VO2, end tidal CO2, METs, heart rate and minute ventilation, comparable with those seen when walking at a moderate to brisk pace, consistent with the changes in these parameters seen during moderate intensity physical activity.
Research regarding the oxygen cost of singing by people with limited singing experience is limited. Sliiden et al present data from 20 professional musical theatre performers which suggest similar physiological responses to the current study when singing compared with rest, including heart rate, VO2, minute ventilation and breath volume.35 Another study of nine final-year musical theatre students compared cardiorespiratory parameters while singing and dancing together, with dancing alone. The study found significantly lower breathing frequency and higher lactate when singing and dancing together, compared with dancing alone, but other parameters including VO2 and heart rate did not differ significantly.36 However, singing alone (without dancing) was not compared with rest which limits comparisons. Regarding ventilatory volumes, our findings support previous research that suggests increases during singing and speech compared with spontaneous breathing.7 37–41 However, much of the previous research concerns speech alone, and where singing has been investigated, the studies have largely focused on professional singers, or employed limited protocols that do not fully represent the range of activities engaged in during a community singing group. As such, application of previous research findings to the most common contexts in which people sing is challenging. To our knowledge, this is the first study to systematically assess physiological parameters in people who do not sing regularly, including pulmonary ventilation volumes, during the various singing activities commonly found in amateur community singing groups. As such, our findings build on those of other studies by demonstrating comparable physiological responses related to singing in non-singers, and by comparing singing with a standardised form of physical activity in the form of treadmill walking. Of note, the relative increases above baseline in ventilatory parameters may be of importance when considering aerosol transmission of infectious agents, including SARS-CoV-2.42
An important consideration when interpreting our findings is that the extent to which people are moving is also likely to be a major factor in determining the physiological demands of the activity. Though completely static singing is unrealistic, we should consider that different types of singing encourage different levels of body movement, gesture and dance like movements, in addition to voice production. A further point for consideration is the extent to which changes in the physiological parameters assessed result from physical exertion, or a degree of relative hyperventilation required for vocalisation. For example, one might expect to see larger ventilatory volumes, and possibly heart rate, because of the air flow velocity and volume requirements for vocalisation. However, the pattern of end tidal CO2 during singing, compared with walking, suggests that hyperventilation alone does not account for the changes in the other parameters seen during the singing component. Furthermore, while minute ventilation approximately doubles from baseline, VO2 approximately quadruples, suggestive of an important contribution from higher cardiac output, respiratory muscle oxygen extraction and skeletal muscles involved in movement, however the relative contribution of these factors has not been investigated here.
It is also useful to consider how our findings apply to people with respiratory disease. In the current studies, participants did not have any activity-limiting illnesses and are substantially younger than many people with common long-term respiratory conditions, such as chronic obstructive pulmonary disease. People with respiratory conditions may be more restricted in their ability to engage in singing activities in general, which could influence the physiological demands experienced. However, SLH sessions are specifically designed for people with respiratory disease and personally adapted to individual participant’s abilities during sessions, to enable participation despite individual restrictions. However, the potential for physiological responses to differ by age group and the presence of respiratory disease highlights the need to evaluate potential differences in future research.
Methodological considerations
This study has multiple strengths. To our knowledge, this is the first study to compare the physiological demands of singing with walking, using measures of ventilation, VO2, end tidal CO2, and perceived effort and dyspnoea simultaneously. The focus on people who are not professional singers or performers makes the findings highly relevant for people who do not regularly engage in singing.
Certain limitations should be mentioned. First, the use of healthy, relatively young participants may limit the extent to which our findings can be extrapolated to older people, or those with significant medical conditions, such as those with chronic respiratory disease (CRD). However, individuals with CRD are likely to find activities such as singing more, rather than less physiologically demanding, as a proportion of their VO2 max.43 Therefore, one might reasonably suspect that the potential for physical benefits related to training effects would also be increased, though in what way, and to what extent, remains unclear. Additionally, this would require the individuals with CRD to engage with the activity in the same way as the healthy volunteers of the current study, which for many people with CRD would not be possible. Given the multiple uncertainties regarding specific responses in people with CRD, further research including such participants is clearly required. Second, the sample size is small; although it was sufficient to meet the aims of the study by comparing the parameters during protocol components, replication of our findings in larger samples is encouraged. Third, although we considered real-world applicability when developing the components of the protocol, the total protocol duration was approximately 25 min, while most community singing sessions are longer. As such, further studies during real-world community singing group sessions would be of interest. Lastly, though this study has demonstrated that singing induces physiological responses that are similar in magnitude to moderate intensity physical activity, this study has not assessed training effects of singing. As such we cannot draw clear conclusions from this study alone regarding impacts on physical fitness.
It is possible that given the jaw movement required for singing that a dynamic air leak could have gone unnoticed. However, this is unlikely as we tested the fit before starting. Furthermore, if there was a leak, the ventilatory values would have been underestimated, rather than overestimated. Additionally, this would not have influenced the relative proportions of gases recorded in the analysis.
This study has raised multiple directions for future research. To build on these findings, future studies could include maximal exercise tests for comparison; evaluate if training effects occur following a programme of singing; directly compare professional and amateur singers; specifically assess the impact of musical genre, volume and physical movements; and compare healthy controls with people with certain chronic diseases, in whom singing is being delivered in a therapeutic context. It would also be valuable to explore how the different session components could be adapted and varied, and how this influences outcomes.