Is the healthy respiratory system (always) built for exercise?

I am honoured that the editors at The Journal of Physiology have asked me to place into perspective the manuscript published 22 years ago by Peter Hanson, Kathleen Henderson and myself.

The study (Dempsey et al. 1984) was primarily a descriptive one in humans which showed that about half the members of a group of endurance-trained runners experienced significant arterial hypoxaemia (EIAH) during exercise due primarily to an excessively widened difference between alveolar and arterial P_O2 (indicating a highly inefficient gas exchange) and to a less consistent extent to an inadequate compensatory hyperventilation. The study extended previous anecdotal reports of EIAH and challenged the traditional concept that the healthy respiratory system was ‘overbuilt’ for purposes of providing near-perfect gas exchange during exercise. These findings were purely serendipitous as the original study had an entirely different purpose, until we observed significant EIAH in the initial few athletes studied and confirmed these observations with repeat testing and by verifying our blood gas measurements with traditional (VanSlyke) manometric measurements of HbO₂ content and saturation.

In the ensuing years EIAH has also been shown to occur in young adult females, elderly endurance-trained athletes, in Thoroughbred horses (but not in most other mammals with high V˙_O2max) and also during high intensity endurance exercise, including time trials (Dempsey & Wagner, 1999). During endurance exercise, time-dependent increases in temperature and arterial acidity also become major effectors of arterial O₂ desaturation in all subjects, with a minority showing a reduced P_aO2. Use of supplemental inspired O₂ to prevent EIAH showed that it negatively impacted V˙_O2max and contributed significantly to the development of exercise-induced locomotor muscle peripheral fatigue (Romer et al. 2006).

The implication from these studies was that lung structure in humans does not adapt to chronic physical training, as it does to chronic hypoxia or lung resection. Paradoxically, the more fit someone is the more likely they are to experience these gas exchange inefficiencies. Accordingly, EIAH contributes significantly to exercise performance limitation in many trained athletes at sea-level and especially with even mild elevations in altitude. Further studies have also revealed that the work output required by the respiratory muscles during sustained heavy exercise will also contribute to performance limitation through its effects on diaphragm fatigue, limb blood flow and limb muscle fatigue (Dempsey et al. 2006). Accordingly, under special circumstances, functions of the lung (in the highly trained) and/or the respiratory muscles (in trained and untrained) will impede performance. However, the major, universal contributors to exercise performance limitation in health reside primarily in the relatively ‘weak links’ to O₂ transport and utilization provided by limitations to skeletal muscle blood flow and O₂ utilization by skeletal muscle and their sequellae (Saltin et al. 2006).

Many questions remain, including the fundamental causes of the reduced arterial P_O2. The widening of the alveolar-to-arterial O₂ difference (A–aD_O2) during exercise has been attributed to a combination of ventilation to perfusion maldistribution plus a small anatomic shunt or to a diffusion limitation or to exercise-induced intrapulmonary shunts (Dempsey & Wagner, 1999). However, these suggestions are based primarily on indirect evidence. Broncho-alveolar lavage analysis suggests that alveolar–capillary ‘stress failure’ may occur following exhaustive exercise in the athlete (Hopkins et al. 1997), but this occurrence has not yet been causally linked to inefficiencies in gas exchange. It is important to note that the excessive A–aD_O2 values in the athlete often begin to appear during submaximal exercise intensities, suggesting that EIAH is in part due to the inherent characteristics of the athletic lung as well as to the extraordinary demands for gas exchange demanded by their very high maximum work loads. This implies that chronic physical training per se might precipitate maladaptation of lung structures at the alveolar : capillary interface, just as it does in the airways when heavy exercise training is carried out in cold environments (Karjalainen et al. 2000). EIAH involving a reduction in P_aO2 also appears to be much more prevalent during running exercise, as opposed to cycling.

Inadequate compensatory hyperventilation has subsequently been explained, in part, by mechanical constraint of airflow, as the extraordinarily high levels of ventilation required by the athlete exceed the limits of the maximum flow : volume envelope. However, there is also a deficiency in ‘sensitivity’ to the ventilatory stimuli associated with heavy exercise in some of these athletes, and this apparent ‘unwillingness’ to increase ventilation in proportion to demand remains unexplained.

Particularly perplexing is why some athletes do and some do not manifest highly inefficient gas exchange leading to EIAH. Given the dramatic reductions in mixed venous O₂ content and the high velocities of pulmonary blood flow and airway flow rates in heavy intensity exercise, very small inter-individual structural differences in airway and vessel diameters and their spatial distribution or in pulmonary capillary blood volume or in shunt pathways would have major effects on gas exchange efficiency.

We are especially pleased with the questions and research initiatives on the determinants of respiratory system inefficiencies and their consequences that may have been stimulated in part by our initial descriptive study. Since the phenomenon of gas exchange inefficiency and EIAH in health is exclusive to the human, further progress requires the application of relatively non-invasive methodology to the exercising state. Hopefully, recent introductions of powerful imaging techniques for the lung will eventually be applicable to the exercising human, so that critical determinants of O₂ and CO₂ exchange such as red cell transit time distribution, extra-vascular lung water accumulation, pre-capillary gas exchange and exercise-induced intra- and extra-pulmonary shunts may be quantified.