Michael Parkes
University of Birmingham, UK

https://doi.org/10.36866/pn.88.33

It is often claimed that humans – and ‘elite’ free-divers in particular – prolong breath-holding by lowering heart rate to reduce metabolic rate, and even that this might represent a harking back to some ancestral aquatic past. While the scientific literature is never straightforward, it has indicated for many years that this appears to be a myth for breath-holds in resting subjects without immersion.

Over the last decade a number of excellent television programmes about breath-holding have claimed that humans prolong breath-holding by lowering heart rate to reduce metabolic rate. These programmes are great for increasing public awareness and interest in physiology and are useful material for making undergraduate lectures interesting, topical and thought provoking. The accompanying commentaries are excellent, but have the unenviable task of converting complex ideas into a simple and intriguing message. Many viewers just enjoy the visual experience and don’t analyse the commentary closely. But if you view the films a number of times the simplifications become more noticeable. The danger of over-simplification threatens when complex physiological processes, which are incompletely studied and understood, have to be simplified to make them exciting for the viewer. Moreover each programme may repeat and so propagate the original simplification. One simplification can be the merging of breath-holds on land (visually dull) with those during immersion (visually thrilling). Another is that humans decrease heart rate during breath-holding to reduce metabolic rate and hence prolong breath-hold duration. It is valuable to highlight some of the classical physiology (Lin, 1982; Lin & Hong, 1996) and original papers indicating that this is a myth for breath-holds at rest and without immersion.

Metabolic rate does not decrease below resting levels during breath-holding

Metabolic rate (the rate of O2 consumption) is normally measured at the mouth from at least one breath, using an oxygen-filled spirometer or a Douglas bag. Expired air measurements alone are adequate to measure metabolic rate at rest or during exercise, because neither the arterial partial pressure of oxygen (PaO2) falls nor that of carbon dioxide (PaCO2) rises. Typical resting metabolic rates are ~250 ml O2,STPD min–1.

Measuring metabolic rate during breath-holding presents more of a technical challenge. Firstly, because it is only averaged over the breath-hold and secondly, because blood gases do deteriorate during breath-holding. So the true metabolic rate can only be measured from the oxygen extracted from both expired air and blood (Hong et al. 1971). The key and classic papers are more than 40 years old and so far all have found that metabolic rate in resting humans does not decrease during breath-holding.

In 1946 Stevens et al. showed in six healthy subjects that buoyancy gradually decreased during breath-holding. They proposed this was because the oxygen taken up from the gas in the unventilated lungs was not replaced by carbon dioxide gas produced as metabolism continued. Instead, while gaseous oxygen is consumed, carbon dioxide remains dissolved in the blood and tissues (because breath-holding abolishes the partial pressure gradients driving CO2 from alveolar blood to alveolar gas). The rate of change of buoyancy, i.e. the rate of decrease in lung volume, should correspond to the rate of oxygen consumption measured at the mouth. Stevens confirmed this in three subjects by showing the change in buoyancy was almost exactly equal to the rate of oxygen consumption measured by subjects breathing from an O2-filled spirometer before and after the breath-hold.

Stevens et al. used this to show that the mean rate of oxygen consumption during the entire breath-hold (the rate of change of buoyancy) was ~291 ml O2,STPD min–1, i.e. not below resting levels. (Strictly speaking, they were also immersed, but the water temperature is unknown!) The true rate of oxygen consumption must be higher than this, when also accounting for the additional fall in blood oxygen content during breath-holding. Alternatively, if the lungs are over-filled with oxygen at the start of the breath-hold (i.e. by ‘preoxygenating’ with 50–100% O2) and subjects hyperventilate so much that blood gases remain normal even at the breakpoint, oxygen consumption measured only at the mouth produces higher metabolic rate values that should be nearer to the true metabolic rate. Stevens also used breath-holds with preoxygenation (but without hyperventilation) and found that the measured mean rate of oxygen consumption was higher (369 ml O2,STPD min–1). This measure is nearer the true metabolic rate, but still fails to allow for any influences on oxygen carriage of changes in PaCO2.

In 1959 Klocke & Rahn used spirometers to measure the change in lung volume and its gas composition in six subjects during breath-holds prolonged with preoxygenation and voluntary hyperventilation. They measured no change in the volume of gaseous CO2 in the lungs during breath-holding (confirming that all metabolically produced CO2 remains dissolved). They found that the mean rate of decrease in lung volume corresponded to 300 ml O2,STPD min–1. Incidentally, the breath-hold duration of 14 minutes Rahn himself achieved with this preoxygenation and hyperventilation represents the most plausible ‘longest’ breath-hold recorded in the scientific literature (although anecdotes exist of even longer holds).

Finally, in 1971, Hong et al. attempted to derive the true rate of oxygen consumption for an entire breath-hold (but only in 2 subjects) when combining spirometry with blood gas sampling and found a mean metabolic rate of 212 ml O2,STPD min–1. Note the usual paradox in the physiology literature of there being so much variation between studies that the ‘true’ 1971 value from only 2 subjects is lower than the underestimates from the 12 subjects in 1946 and 1951!

A definitive study of validated measurements of true metabolic rate before and during breath-holding using a larger number of normal subjects would be welcome. Nevertheless, the available evidence indicates that metabolic rate does not decrease below resting levels during such breath-holds.

‘Mean’ heart rate remains above 55 beats min–1 during breath-holding

Heart rate is so easy to record during breath-holding that it is often reported, in case it might be important. There are, however, three difficulties with breath-hold studies. First, the heart rate change depends on the gases inhaled at the start of the breath-hold (heart rate does not fall when breath-holding with preoxygenation). Secondly, baseline heart rate rises in anticipation of breath-holding (especially if voluntary hyperventilation occurs) which will exaggerate subsequent ‘falls’. Thirdly, the presence of respiratory sinus arrhythmia (Cooper et al. 2003) before and during (showing that voluntary breath-holding cannot stop the central rhythm (Parkes, 2006) complicates establishing the precise heart rate changes with breath-holding.

In the best review of heart rate changes with breath-holding, Lin (1982, his Fig. 2) reports pre-breath-hold heart rates of 65–100 beats min–1 and that during breath-holding ‘mean’ heart rate always remains above 55 beats min–1. Counting the total number of heart beats during the breath-hold gives a more realistic indication of its metabolic demands (so it is better to compare mean heart rate with metabolic rate, which is itself always measured as a mean over the entire breath-hold). The overall heart rate change vs. pre (not resting conditions) therefore reported was sometimes a rise, no change or a slight fall. Subsequent literature still supports this conclusion. Furthermore, we know that in most studies measuring cardiac output during such breath-holds, it is still ~6 blood min–1.

Anecdotally, the lowest minimum heart rate value during breath-holding in individuals is ~30 beats min–1 in two subjects (Ferrigno et al. 1991, their Fig. 4). On the other hand anecdotally, the most striking example of where a heart rate fall might be expected but is not observed is during prolonged breath-holds (using preoxygenation and hyperventilation). Figure 1 shows a recent ~9 minute breath-hold in my laboratory where respiratory sinus arrhythmia persists and instantaneous heart rate never falls.

How far must heart rate fall to reduce metabolic rate?

Obviously the heart’s beating consumes oxygen. So any decrease in heart rate not accompanied by an increase in stroke volume will consume less oxygen. But the human heart (with a mean weight of only 300 g) typically consumes only ~35 ml O2,STPD min–1 at rest (Takaoka et al. 1992). In other words, even if the heart stopped beating altogether, this would reduce resting metabolic rate by only ~14%. So for resting humans breath-holding without immersion, even a large decrease in mean heart rate will cause only a small reduction in overall metabolic rate.

The available evidence therefore shows that ‘mean’ heart rate does not fall sufficiently to produce measurable decreases in metabolic rate. And anyway during such breath-holds metabolic rate does not measurably decrease!

Conclusions

None of the scientific literature I can find demonstrates that ‘elite’ breath-hold divers have any unique abilities to reduce both mean heart and metabolic rate during breath-holding at rest and without immersion. So the current scientific literature exposes some of the mythology around breath-holding.

What might happen during immersion, especially in freezing water is, however, another story. The additional variables with immersion (how reliable are ECG measurements underwater? what duration? what depth? what water temperature? how hard were they swimming?) make this an even more complex question to unravel.

References

Cooper HE, Parkes MJ & Clutton-Brock TH (2003). CO2-dependent components of sinus arrhythmia from the start of breath-holding in Man. Am J Physiol Heart Circ Physiol 285, H841–H848.

Hong SK, Lin YC, Lally DA, Yim BJB, Kominami N, Hong PW & Moore TO (1971). Alveolar gas exchanges and cardiovascular functions during breath holding with air. J Appl Physiol 30, 540–547.

Klocke FJ & Rahn H (1959). Breath holding after breathing oxygen. J Appl Physiol 14, 689–693.

Lin YC (1982). Breath-hold diving in terrestrial mammals. Exerc Sport Sci Rev 10, 270–307.

Lin YC & Hong SK (1996). Hyperbaria: breath-hold diving. In Handbook of Physiology, vol. II, Environmental Physiology, section 4, pp. 979–995. American Physiological Society, Washington DC.

Parkes MJ (2006). Breath-holding and its breakpoint. Exp Physiol 91, 1–15.

Stevens CD, Ferris EB, Webb JP, Engel GL & Logan M (1946). Voluntary breath holding 1. Pulmonary gas exchange during breath holding. J Clin Invest 25, 723–728.

Takaoka H, Takeuchi M, Odake M & Yokoyama M (1992). Assessment of myocardial oxygen consumption (VO2) and systolic pressure-volume area (PVA) in human hearts. Eur Heart J 13, 85–90.