Point: Pulmonary edema does occur in human athletes performing heavy sea-level exercise

My colleagues (25) have a difficult task, because it is easier to prove a positive than a negative. A 1978 case report documents two athletes that presented, after running the Comrades marathon, with shortness of breath, cough productive of blood tinged and/or pink frothy sputum, crepitations on auscultation, and chest radiograph findings of pulmonary edema (18). There can be no doubt that these athletes had pulmonary edema, and although the Comrades marathon represents an extraordinary exercise task because of its length (90 km), the elevation of the route never exceeded 1,000 m. In addition, there are numerous reports in the literature that document the presence of pulmonary edema after exercise. These include cases of pulmonary edema occurring as a result of swimming (1), as distinct from pulmonary edema associated with diving, i.e., immersion pulmonary edema (26), as well as cases of overt edema occurring during sexual intercourse (29) and heavy or/prolonged exercise with very modest altitude exposure (2, 17) in otherwise healthy people. Consequently these case reports prove that pulmonary edema occurs during heavy sea-level exercise and this debate could stop here.

However, given the sporadic nature of these reports, perhaps a better question is why don't more athletes develop alveolar flooding during exercise? While the healthy human lung is overbuilt for a sedentary existence, during maximal exercise it is subjected to substantial physiological stressors. In particular, ventilation and cardiac output greatly increase and pulmonary blood volume and capillary blood volume rise (16). Even during submaximal exercise, pulmonary arterial and occlusion pressures (i.e., left atrial pressure) approximately double from rest, cardiac output increases approximately fourfold (13), and microvascular pressures rise. Starling's law is thought to apply to the lung microvasculature (see Ref. 28 for review), and thus fluid balance in the lung reflects the interplay of hydrostatic pressures and colloid osmotic pressures in the lung microvasculature and surrounding interstitium. Increases in microvascular pressure, combined with a greater surface area for fluid filtration are expected to increase fluid exchange between the lung vascular space and the interstitium (28). Given this, is not difficult to imagine that lung fluid balance might be altered, without completely overwhelming the lymphatic drainage, resulting in alveolar flooding. There are animal data supporting this idea. In dogs, when the capillary pressure is increased, a distinct pattern of edema formation emerges (Fig. 1). First, there is extravasation of fluid into the interstitial space surrounding the conducting vessels and airways (27). Following this, if the rise in pressure is prolonged, alveolar wall fluid is increased and then finally overt alveolar flooding, gas exchange defects, and possibly alveolar collapse occur (27). In keeping with this idea, lung lymphatic flow increases rapidly after the onset of exercise in sheep and further increases with increasing exercise duration (5, 20). On this basis it is to be expected that fluid balance in the human lung is likely to follow a similar pattern. However severe pulmonary edema with alveolar flooding has been documented during heavy sea-level exercise only under conditions where the development is exacerbated by factors such as very prolonged exercise (18), psychological stress (1), cold (29), moderate altitude exposure (2, 17), and water immersion (1, 26). This suggests that the normal lung has a large safety margin preventing the development of alveolar flooding except in the most extreme cases. However, interstitial edema is more likely.

Fig. 1.

The development of acute pulmonary edema as described by Staub et al. (27). A: the alveolar walls (left) are normal and there is no excess fluid in the interstitial space (right). In the first stages of pulmonary edema (B), the interstitial space is greatly increased (right) but the alveolar wall is minimally affected (left). Once the capacity of the interstitial space is overwhelmed (C), fluid progressively accumulates in the alveoli (D), leading to loss of stability and collapse (bottom inset). Figure borrowed with permission from Ref. 27.

In humans, the evidence for the development of interstitial edema is indirect, but very consistent. As seen in Fig. 1, interstitial edema would be expected to compress small airways and blood vessels, potentially affecting ventilation-perfusion (V̇_A/Q̇) matching. In many humans, V̇_A/Q̇ mismatch increases with exercise (6, 8, 9, 13, 14, 21, 22) and the pattern is consistent with the cause being interstitial edema. For example, those subjects who increase V̇_A/Q̇ mismatch during exercise, have greater V̇_A/Q̇ mismatch in recovery than those subjects that do not show an increase. This difference persists beyond the point at which ventilation and cardiac output normalize to pre-exercise levels (24). Exercise in normobaric hypoxia causes a greater increase in V̇_A/Q̇ mismatch than in normoxia, and this increase is relieved by breathing 100% oxygen (9). Since hypoxia and hyperoxia alter pulmonary arterial pressures and thus capillary filtration, this is also consistent. The extent of V̇_A/Q̇ mismatch increases with exercise duration even at relatively low exercise intensities (13). In subjects who have had the extent of V̇_A/Q̇ mismatch characterized by the multiple inert gas elimination technique (MIGET), prolonged exercise makes the distribution of pulmonary blood flow less spatially uniform (3), as would be expected from compression of small blood vessels, and the extent of this derangement is correlated with the degree of V̇_A/Q̇ mismatch previously measured by MIGET.

If, as these studies suggest, a great many healthy subjects develop interstitial edema with exercise, why don't more imaging studies show positive results? There are several potential explanations for negative imaging studies. The imaging has to be performed within a time frame that allows for detection of any pulmonary edema before it has resolved. In upright subjects, gas exchange alterations resolve ∼20 min after exercise (24), although in supine subjects who are not upright starting shortly after exercise, alterations in the spatial heterogeneity of blood flow persist for an hour (3). Since many studies have both long delays between the completion of exercise and the start of imaging (7, 11, 19) and long acquisition times (11, 19), this may partly explain the lack of findings. Mean lung density data are likely to be misleading. If interstitial edema affects small airways as well as blood vessels, this could lead to air trapping and no net change in density, although total lung water is increased, unless changes in lung air volume and inflation are specifically accounted for. To be sensitive to changes representing edema, the technique must account for fluid shifts between the vascular compartment and the interstitial space. Central blood volume and capillary blood volume is first increased by exercise and then decreased following exercise (10, 15), affecting mean lung density differently depending on the timing of the measure. Application of thresholds designed to eliminate the intravascular contribution to lung water (11, 19) have the unwanted effect of potentially removing lung regions of high density that represent extravascular fluid accumulation. Despite these challenges, a number of studies have shown positive results (4, 17, 19).

It is important to recognize that despite the evidence presented above, any V̇_A/Q̇ mismatch resulting from interstitial edema is unlikely to have a large effect on pulmonary gas exchange efficiency during exercise (reviewed in Ref. 12). This is because despite increasing mismatch, the overall increase in ventilation during exercise is greater than the increase in cardiac output, and thus relatively low V̇_A/Q̇ units of the lung are at a higher V̇_A/Q̇ ratio, minimizing the effect on oxygenation. In addition, as seen in Fig. 1, fluid accumulation is unlikely to affect the alveolar wall until the edema is advanced (27) and thus no diffusive limitation to O₂ transport is expected. This is consistent with recent work showing no detectable diffusion limitation when interstitial edema is created by volume overload in resting subjects (23).