CON: Most Climbers Do Not Develop Subclinical Interstitial Pulmonary Edema

The highly cited study by Cremona and colleagues (2002) reporting that 75% of climbers reaching 4559 m in a 2-day ascent has an increased closing volume (CV) is widely accepted as compelling evidence that interstitial pulmonary edema is very prevalent in otherwise healthy and successful mountaineers. This conclusion in a cohort of over 250 subjects corroborated the original hypothesis of Gray and colleagues (1973) that interstitial edema is common at high altitude, although somewhat ironically Gray and colleagues could not support the hypothesis with their own CV measurements. Since that time, numerous groups (over 30) have stalked this elusive entity using a variety of pulmonary function parameters, thoracic imaging, and other measures of lung water, with varying success. Important questions are raised by any data suggesting increased interstitial lung water: are these climbers at risk for the real threat of high altitude pulmonary edema (HAPE); does interstitial edema of this magnitude impair their ability to climb higher; and should anything be done to prevent the development of interstitial edema? However, before accepting the affirmative, as will be argued by my opponents, one must ask (1) whether the several indirect parameters used to infer that subclinical pulmonary edema is commonplace might be altered by other aspects of alpinism independent of increased interstitial lung water, and (b) how much pulmonary function measurements, gas exchange, and exercise capacity are altered when there is, in fact, unequivocal interstitial edema of the magnitude raised by the work of Cremona and colleagues (2002) and others?

Among pulmonary function changes taken as evidence for increased interstitial lung water are the original parameter of increased CV, reduction in vital capacity (VC), increase in residual volume (RV), increased peak expiratory flow rates, decreased lung compliance, and bronchial hyperreactivity (reviewed by Milic-Emili et al., 2001). As alluded to previously, not all studies (roughly only 60%) by one parameter or another, either at high altitude or in hypobaric chamber simulations, document these findings, in part because the changes are relatively small; different methodologies were employed; ascent rates, attained altitude, and duration at altitude varied; as did the time elapsed between any exercise and testing. Nonetheless, the changes, if and when they occur, are consistent with lung water accumulation, because increased interstitial fluid flux and lymphatic drainage along the bronchovascular bundles should increase lung elastic recoil and narrow small airways, leading to earlier closure of dependent airways, decreased lung compliance, lowered VC and elevated RV, and bronchial hyperreactivity (Hughes and Rosenzweig, 1970). Yet many of these same pulmonary function changes can also be evoked with cold air hyperpnea, hypocapnia, mild respiratory muscle weakness, and heavy exercise alone, all features of strenuous mountain climbing.

Breathing cold and dry alpine air for extended periods can increase basal bronchial tone and hyperreactivity to methacholine and reduce lung compliance, particularly in athletes engaged in extended endurance winter sports such as cross-country skiing. However, with lesser-duration exposure, effects are short-lived in nonasthmatics (O'Cain et al., 1980; Kaminsky et al., 2000) and are not likely to provide any contribution to the lung function changes that persist for hours to days later at altitude. This was recently reported by Pellegrino and colleagues (2009), who in fact showed no increase in methacholine sensitivity after a 2-day ascent to 4559 m, a surprising finding given the prevailing wisdom that increased interstitial lung water can provoke bronchial hyperreactivity, as is the case in congestive heart failure (Cabanes et al., 1989).

Hypocapnia occurs as a result of increased ventilation with hypoxic stimulation of the peripheral chemoreceptors. Lowered airway and alveolar CO₂ reduce parenchymal compliance and increase airway smooth muscle tone (reviewed by Swenson et al., 1998) in the ranges that may occur at high altitude. In humans, Cutillo and colleagues (1974) found that hypocapnic hyperventilation decreased dynamic lung compliance, but they did not measure static compliance to eliminate the influence of possible airway bronchoconstriction. The magnitude of hypocapnic pneumoconstriction separate from any changes in airway smooth muscle tension in rat lungs is about a 1.5% decrease per mmHg reduction in Pco₂ (Emery et al., 2007) over the Pco₂ ranges typical of high altitude. This could account for a significant fraction of the 5% to 10% reductions noted in VC, total lung capacity (TLC), and static lung compliance by many groups. This is supported by a correlation at Everest base camp of end-tidal Pco₂ with the fall in VC (Pollard et al., 1997) that accounted for about 15% of the variance in VC. Hypocapnic bronchoconstriction (Jamison et al., 1987) measureable at sea level may also be a partial explanation for those studies that find increased airways resistance, increased CV, and decreased dynamic compliance at high altitude.

Numerous studies have reported that maximal inspiratory and expiratory muscle strength are reduced at high altitude with climbing or in hypobaric chamber simulations without exercise by as much as 10% to 20% (Deboeck et al., 2005; Fasano et al., 2007; Sharma and Brown, 2007). Reductions in maximal respiratory muscle strength of this magnitude will also contribute to reductions in TLC and VC and elevations in RV, independent of any other change in lung compliance or airways resistance.

Finally, heavy exercise itself, performed at sea level, may alter lung function and lead to mild respiratory muscle fatigue and reductions in maximal inspiratory lung volumes. Changes in CV, VC, and RV of the magnitude reported in the altitude studies cited previously occur with heavy exercise of the same intensity as mountaineering (Farrell et al., 1983; Miles et al., 1983; Miles and Schaffer, 1988) and are attributed to the high cardiac output of sustained heavy exercise increasing fluid filtration across a maximally recruited pulmonary vascular bed. It should be noted that, in climbing above 3500 m for many hours, mountaineers may be exercising at about 80% of their maximum heart rate, a level typical of marathon running (Bircher et al., 1994). Thus it is entirely plausible that most or all of the reported lung function changes at high altitude are essentially those of exercise. Although it is reasonable to assume that hypoxic exercise would be worse in this regard compared with equivalent normoxic exercise, this was not the case when studied in normobaric hypoxic conditions (Miles and Schaffer, 1988; Synder et al., 2006). Finally, because increased CV has received the greatest attention, could it be that increased blood flow and pulmonary vascular recruitment with exercise and soon thereafter, or with the 40% increase in resting cardiac output acutely at high altitude (Grollman, 1930), the lung will be heavier such that, analogous to a slinky, more dependent airways will be compressed and close off earlier in exhalation?

When more direct evidence of interstitial edema at high altitude has been sought, the picture is no clearer or robust, either by changes in thoracic electrical impedance or imaging. Lower thoracic electrical impedance as a noninvasive measure of increased lung water was found in several studies of climbing to high altitude in the first 1 to 2 days (Jaeger et al., 1979; Mason et al., 2003) and in those passively brought to high altitude by rail or road (Roy et al., 1974; Hoon et al., 1977); but others have reported no change with hypoxic exercise (Miles and Schaffer, 1988) or airlift to high altitude (Kanstrup et al., 1999). The heterogeneity of these few studies makes any definitive assessment difficult, because the technique is sensitive to changes in both thoracic extravascular water and intravascular blood volume. Most studies find that both pulmonary capillary and total thoracic blood volume rise with acute hypoxia (Jaeger et al., 1979; Synder et al., 2006), and no studies have attempted to make independent measurements of blood volume to deal with this confounding issue. Another problem not addressed is whether the high incidence of acute mountain sickness in some of these studies, with its attendant peripheral edema (Hackett et al., 1982) within the soft tissues of the thoracic wall, might also contribute.

Thoracic imaging by conventional chest X rays, computed tomography (CT), and magnetic resonance (MR) imaging has been investigated in order to capture the possible interstitial edema predicted by impedance and pulmonary function parameter changes occurring with hypoxia, exercise, and their combination. Anholm and colleagues (1999) studied athletic cyclists undertaking a variety of short, high-intensity sprints (up to 5 km) and longer endurance cycling (up to 131 km) at altitudes between 2200 and 3100 m. Using a composite score of numerous radiographic changes indicative of increased lung water, there was a small increase in the edema score that was mostly dominated by increased vascular markings (vessel recruitment and distension with high blood flow), but considerably less peribronchial cuffing, Kerley lines, enhanced fissures, pleural effusions, and hilar blurring, observations more indicative of transvascular fluid escape. In the case of these more telling findings of edema, the changes scored by four blinded radiologists were not statistically different. Nonetheless, VC and FEV₁ were reduced by about 5%, as many have shown with intense exercise. More recently, Hodges and colleagues (2007) used MR imaging and observed no increased lung density consistent with greater lung water in male cyclists in the first hour after exercising between 55% to 60% Vo₂max for 60 min, either in normoxia or breathing 15% oxygen. The same group also studied male (MacNutt et al., 2007) and female (Guenette et al., 2007) cyclists with CT imaging after subjects completed several 2.5- to 3.0-km intervals at near maximal heart rate while breathing 15% oxygen. Scans completed 45 to 70 min afterward showed no evidence of increased lung density. Finally, Synder and colleagues (2006) studied 25 subjects who, following a 17-h overnight normobaric hypoxic exposure, exercised to exhaustion in 15 min while breathing 12.5% O₂. With the overnight hypoxic exposure, there was a statistically significant reduction in lung density and a small but not statistically significant further reduction in lung water following exercise, rather than an increase. In part to establish the sensitivity of their measurements, a rapid infusion of 30 mL/kg of saline led to an increase in lung density equivalent to a gain of 60 to 70 mL in lung water, a value close to that predicted for proportional distribution of the saline into whole-body extracellular space. Thus the totality of the data by imaging to date is not supportive of increased lung water with hypoxia or hypoxic exercise. The very germane issue of whether pulmonary edema occurs in athletes performing heavy, sea-level exercise has been vigorously argued in a recent Point: Counterpoint series in the Journal of Applied Physiology (Hopkins et al., 2010) and reviewed earlier by Zavorsky (2007). Only under the circumstances of greatest cardiac output and lung lymphatic flux of high-intensity and long-duration exercise achievable at sea level can one detect MR or CT imaging evidence of slight increases in lung water, and even here the studies are not unanimous. Although CT and MR imaging come closest to a noninvasive gold-standard assessment of lung water, uncertainities remain around the problems of delay in getting subjects into a scanner after completing exercise; the recumbent position necessary to perform the scans; the problem of any mild exercise- induced air-trapping, which might counteract the increased density signal of increased lung water; and how to subtract away the increases in vascular volume to arrive at the interstitial fluid volume.

Beyond the technical problems of the methods employed and the alternative explanations for their changes enumerated previously that vitiate the argument for interstitial pulmonary edema occurring in most climbers, what are the consequences to climbers if interstitial edema of the purported magnitude does occur? Several studies of saline loading (Robertson et al., 2004; Prisk et al., 2010), as were performed by Synder and colleagues (2006), or water immersion (Derion et al., 1992), which increase interstitial lung water and equivalently alter some of the same pulmonary function parameters as at high altitude, find no changes in Va/Q heterogeneity, no deterioration in arterial oxygenation, no increase in the A–aPo₂ difference, and no increase in dead-space ventilation at rest. Saline loading reduced Vo₂max by only 5% and did not alter arterial blood gases or the A–aPo₂ difference during exercise (Robertson et al., 2004). So it is unlikely interstitial edema of this magnitude would have any significant effect on climbing ability. What of the risk of HAPE? If interstitial edema in most climbers is the form fruste of HAPE and is responsible for the changes in pulmonary function, one would predict that individuals with these changes might have higher pulmonary artery pressures. However, this was not found by Senn and colleagues (2006) at 4559 m in the case of increased CV, and none of their 26 subjects went on to develop HAPE. In a similar study of 34 subjects at 4559 m (Dehnert et al., 2010), we found no significant changes in pulmonary function in 30 asymptomatic climbers, and somewhat to our surprise the four subjects who developed mild HAPE by symptoms, exam, significant oxygen desaturation, and radiologically apparent interstitial and alveolar edema had only very modest changes in CV, VC, and static compliance, suggesting that these pulmonary function parameters may not be sensitive enough in all cases to detect interstitial fluid accumulation.

In conclusion it would appear that, like the fearsome Sasquatch of the Cascades of the Pacific Northwest, interstitial edema at high altitude is seldom seen but much discussed. The changes in pulmonary function taken to support interstitial edema, which are not reported in a number of studies, may be related to hypocapnia, very mild respiratory muscle weakness or fatigue, or simply the result of heavy exercise. Rather than occurring in a majority of climbers, the true incidence is probably much lower, as suggested by the observed 6% incidence in those without a history of HAPE susceptibility in the Alps (Vock et al., 1989), a value not greatly different from the 7% to 15% occurrence of either radiographic interstitial edema or crackles on chest auscultation in the subjects studied by Cremona and colleagues (2002). Therefore, it would be unnecessary and unwise to conclude that most climbers are at risk for HAPE should they have any indirect evidence of interstitial edema, particularly if this were to engender advice in the absence of symptoms to stop climbing or a widespread use of drugs for HAPE prophylaxis.