Pulmonary Blood Flow Heterogeneity during Hypoxia and High-Altitude Pulmonary Edema
http://www.100md.com
美国呼吸和危急护理医学 2005年第1期
Department of Medicine, Division of Physiology
Department of Radiology, University of California, San Diego, La Jolla, California
ABSTRACT
Uneven hypoxic pulmonary vasoconstriction has been proposed to expose parts of the pulmonary capillary bed to high pressure and vascular injury in high-altitude pulmonary edema (HAPE). We hypothesized that subjects with a history of HAPE would demonstrate increased heterogeneity of pulmonary blood flow during hypoxia. A functional magnetic resonance imaging technique (arterial spin labeling) was used to quantify spatial pulmonary blood flow heterogeneity in three subject groups: (1) HAPE-susceptible (n = 5), individuals with a history of physician-documented HAPE; (2) HAPE-resistant (n = 6), individuals with repeated high-altitude exposure without illness; and (3) unselected (n = 6), individuals with a minimal history of altitude exposure. Data were collected in normoxia and after 5, 10, 20, and 30 minutes of normobaric hypoxia . Relative dispersion (SD/mean) of the signal intensity was used as an index of perfusion heterogeneity. Oxygen saturation was not different between groups during hypoxia. Relative dispersion was not different between groups (HAPE-susceptible 0.94 ± 0.05, HAPE-resistant 0.94 ± 0.05, unselected 0.87 ± 0.06; means ± SEM) during normoxia, but it was increased by hypoxia in HAPE-susceptible (to 1.10 ± 0.05 after 30 minutes, p < 0.0001) but not in HAPE-resistant (0.91 ± 0.05) or unselected subjects (0.87 ± 0.05). HAPE-susceptible individuals have increased pulmonary blood flow heterogeneity in acute hypoxia, consistent with uneven hypoxic pulmonary vasoconstriction.
Key Words: hypoxic pulmonary vasoconstriction magnetic resonance imaging pulmonary circulation
High-altitude pulmonary edema (HAPE) is a noncardiogenic high-permeability edema characterized by alveolar fluid with a high concentration of protein (1) that develops in otherwise healthy individuals after 24 to 72 hours of exposure to altitudes above 2,400 m ( 8,000 ft). Typically, radiographs in HAPE demonstrate a patchy distribution of edema, appearing first in the perihilar regions of the lung (2). The reported incidence is 2 to 15% of exposed individuals depending on rate of ascent and altitude reached (3eC5). Risk factors include rapid ascent, strenuous exercise, and a previous history of HAPE. Increased pulmonary artery pressure is a hallmark of HAPE (6, 7), and susceptible subjects have been shown to have increased pulmonary arterial pressures before the development of HAPE (8). Increased concentrations of red blood cells and protein in bronchoalveolar lavage fluid are observed in the early stages of HAPE before inflammatory cytokines are present (9), suggesting that mechanical injury to the pulmonary capillary bed (stress failure) may be important in the development of HAPE. Because hypoxic pulmonary vasoconstriction takes place in the precapillary arterioles, it is uncertain how capillaries might be exposed to high pressure. A mechanism initially proposed by Visscher (10) and later modified for HAPE by Hultgren (2) is that hypoxic pulmonary vasoconstriction is uneven in HAPE (2). If this were true, some parts of the capillary bed would be protected by upstream vasoconstriction, whereas the portions of the capillary that did not have upstream vasoconstriction would be exposed to high pressures and mechanical stress injury(11), thus inciting HAPE.
Recent animal work confirms that hypoxia results in increased spatial heterogeneity of pulmonary perfusion, suggesting that hypoxic pulmonary vasoconstriction is inherently uneven in the mammalian lung (12). However, the relationship between the development of uneven hypoxic pulmonary vasoconstriction and susceptibility to HAPE is unknown, and it is uncertain whether HAPE-susceptible subjects have increased heterogeneity of pulmonary perfusion in response to hypoxia. We hypothesized that HAPE-susceptible subjects would demonstrate increased pulmonary blood flow heterogeneity during hypoxia compared with subjects without a history of HAPE. To test this hypothesis, we measured pulmonary blood flow heterogeneity in response to acute hypoxia using a quantitative magnetic resonance (MR) imaging technique known as arterial spin labeling. Arterial spin labeling involves creating a magnetically tagged bolus by the use of specialized radiofrequency pulses to alter the magnetization of water in blood. By measuring the transit of the tagged bolus through organ vasculature, it is possible to measure the spatial distribution of blood flow. Arterial spin labeling has been widely applied to measure cerebral blood flow (13) and to a lesser extent has been used to evaluate perfusion in skeletal muscle (14), heart (15), kidney (16), and lung (17). We measured the distribution of pulmonary blood flow using arterial spin labeling in normoxia and acute hypoxia in subjects with a history of HAPE, and compared the results to those of individuals repeatedly exposed to altitude without HAPE and subjects unselected for HAPE-susceptibility by virtue of very limited high-altitude exposure. Some of the results of this study have been previously reported in abstract form (18, 19).
METHODS
The Human Subjects Research Protection Program of the University of California, San Diego approved this study, and the procedures followed conformed with this institution's guidelines. Healthy nonsmoking subjects were recruited by advertisement, informed of the risks of the study, and signed informed consent. A history was taken, with particular reference to the cardiopulmonary system and history of altitude exposure, and a physical examination was performed. The subjects were recruited from three groups: HAPE-susceptible individuals, a history of physician-documented HAPE on at least one occasion (n = 5); HAPE-resistant individuals, repeated high-altitude exposure greater than 5,400 m, without any history of altitude illness (n = 6), and unselected individuals, very limited altitude exposure history (see RESULTS, n = 6).
The subjects were screened for cardiac (ECG) and pulmonary abnormalities (chest radiograph and spirometry). Each subject underwent MR imaging with arterial spin labeling using a Vision 1.5 T whole-body MR Scanner (Siemens Medical Systems, Erlangen, Germany). MR imaging arterial spin labeling is a clinical technique used for the quantification of organ blood flow. The signal intensity of the image has been shown to be proportional to bulk flow in vitro using a tube-flow model (20), and it has been validated in heart (15) with microspheres, and skeletal muscle with venous occlusion plethysmography (21) with an excellent linear relationship between arterial spin labeling measurements and the validating technique. Arterial spin labeling inverts the proton magnetization of blood by applying a radiofrequency pulse, allowing these magnetically tagged protons in blood to act as an endogenous tracer for the evaluation of blood flow (see Figures E1 and E2 in the online data supplement). During each measurement two images are acquired. In the first image, an inversion pulse is applied to the entire volume of the lungs and the image of the selected lung slice is taken. In the second image, a second inversion pulse is used, but applied only to the image slice (a selective inversion). During the delay before the second image is acquired, any blood that enters the image slice from outside will have been missed by the selective inversion. The two images are then subtracted. Because stationary inverted protons from tissue will be present in both images, they will cancel one another out. Thus, the signal from each image voxel is then proportional to the amount of blood that entered the voxel in the delay between the selective inversion pulse and the image acquisition. The sequence is cardiac gated, and flow data are collected over one cardiac cycle encompassing one systole and portions of two diastoles.
Subjects rested in the supine position wearing a face mask (Hans Rudolph 8930; Hans Rudolph Inc, Kansas City, MO) equipped with a nonrebreathing valve (Hans Rudolph 2700; Hans Rudolph Inc). Pulmonary perfusion was quantified using a cardiac-gated pulmonary arterial spin labeling MR sequence as described (17). A detailed description of these methods is given in the online supplement. All sequence parameters were kept within U.S. Food and Drug Administration guidelines for clinical MR examinations. The spatial resolution of measurement with this technique is 1.5 x 3 x 15 mm ( 70 mm3).
Data were obtained during an 8- to 10-second breath hold at functional residual capacity in a 15-mm-thick image slice in the coronal plane in the posterior one-third of the lung using the posterior edge of the descending aorta as a reference point. Five measures were obtained at each time point and averaged. Baseline data were obtained after resting quietly in the magnet for 10 minutes breathing room air. Then, the inspired port of the Hans Rudolph valve was connected to a Douglas bag containing a ( 3,800 m equivalent altitude) and data collection was repeated after 5, 10, 20, and 30 minutes of hypoxia. Arterial oxygen saturation and heart rate were monitored (3150 MR Monitor; Invivo Research Inc, Orlando, FL). The signal intensity for each voxel within the lung parenchyma was determined for each image using MATLAB (The MathWorks, Natick, MA). Relative dispersion (SD/mean signal intensity), an index of blood flow heterogeneity where a greater number indicates a more heterogeneously distributed system, was calculated for all time points (22). Figure 1 presents reliability data for relative dispersion of perfusion heterogeneity obtained in 45 healthy nonsmoking volunteers using MR imaging arterial spin labeling.
Data were analyzed using analysis of variance (Statview 4.1, SAS Institute Inc., Cary, NC) with one randomized group (subject group, three levels: HAPE-susceptible, HAPE-resistant, and unselected subjects) and one repeated measure (duration of hypoxia, five levels: 0, 5, 10, 20, and 30 minutes) (23, 24). Where overall significance was determined, post hoc analysis of variance testing was applied to determine where the differences occurred. Linear regression was used to relate the change in blood flow heterogeneity to arterial oxygen saturation (24). Significance was accepted at p < 0.05, two tailed. All data are means ± SEM.
RESULTS
Subject descriptive data are given in Table 1. Four of the five HAPE-susceptible subjects were diagnosed on the basis of chest radiographs in addition to history and physical findings. The remaining subject had a classic history for HAPE in addition to typical physical findings, which were documented by his physician. Three of our HAPE-susceptible subjects were experienced climbers: two had experienced HAPE on more than one occasion and the third had experienced HAPE on one occasion after a rapid ascent. All three of these individuals continued to travel to high altitudes regularly. The other two HAPE-susceptible subjects had developed HAPE on their first time at high altitude and had not returned to high altitude because of concerns over a reoccurrence. All of the HAPE-resistant subjects were experienced alpinists and regularly traveled to high altitude, usually with very rapid ascent profiles. Four members of this group had Himalayan climbing experience; one had guided a successful climb of Mt. Everest (8,848 m). The remaining two had climbed in South America. One subject in the unselected group had a single history of brief exposure to an altitude of 4,300 m 2 years before the study. The remaining subjects had not traveled above 2,400 to 4,000 m, and all exposures at these altitudes were more than 1 year before the study, of brief (a few hours) duration, included no overnight stay, and occurred on a single occasion only. No subject had slept above 2,000 m.
There were no significant differences between groups for age, height, weight, or pulmonary function variables (Table 1). The HAPE-resistant subjects had reached significantly higher altitudes (6,736 ± 460 m) than either the HAPE-susceptible (5,907 ± 678 m) or the unselected subjects (3,322 ± 299 m). Arterial oxygen saturation (Figure 2, top panel) was 98.2 ± 0.2% at baseline and fell to 87.2 ± 1.3% after 30 minutes of hypoxia (p < 0.0001), but there were no significant differences between groups (p = 0.56), and no significant group-by-time interaction (p = 0.77), thus all subject groups experienced a similar level of hypoxia. Heart rate (Figure 2, bottom panel) measured by pulse oximetry was increased by hypoxia (p < 0.0001); however, this was not significantly different between groups (p = 0.57), and there was no significant group-by-time interaction (p = 0.63).
The overall quality of the MR imaging perfusion maps was excellent, and mean signal-to-noise ratio was 29:1. The mean signal did not differ between the three subject groups over the course of the study. The relative dispersion of signal intensity (Figure 3) was not different between groups at baseline (HAPE-susceptible, 0.94 ± 0.05; HAPE-resistant, 0.94 ± 0.05; and unselected subjects, 0.87 ± 0.06). However, relative dispersion was increased during hypoxia in HAPE-susceptible (to 1.10 ± 0.05 after 30 minutes, p < 0.0001) but not in HAPE-resistant (0.91 ± 0.05, p = 0.44) or unselected subjects (0.87 ± 0.05, p = 0.30). Strikingly, all of the HAPE-susceptible subjects increased pulmonary blood flow heterogeneity by at least 10% of their baseline values by 30 minutes of hypoxia (Figures 4 and 5).
Because there was no difference in the change in relative dispersion in response to hypoxia between the HAPE-resistant and unselected subjects, these data were pooled and linear regression analysis was performed. There was a significant relationship between arterial oxygen saturation at 30 minutes and the change in relative dispersion from baseline (Figure 4, R = eC0.60, p < 0.05). Overall, hypoxia was associated with small reductions in relative dispersion in these subjects, but three of the subjects who became the most hypoxic had small increases in heterogeneity resulting in a negative slope. When data from the five HAPE-susceptible subjects were plotted on the same figure, all of the values for the change in relative dispersion for HAPE-susceptible subjects fell above the +2 SD confidence limits for the relationship of the change in relative dispersion verses arterial oxygen saturation. In addition, there was a strong within-subject relationship between arterial oxygen saturation and relative dispersion for the HAPE-susceptible subjects, with an average correlation of eC0.82 within subjects across all time points (Figure 5). By contrast, the correlation between saturation and relative dispersion averaged +0.17 for the HAPE-resistant subjects and eC0.01 for the unselected subjects.
DISCUSSION
We describe a highly significant difference in the pulmonary vascular response to hypoxia between HAPE-susceptible individuals and those who have not suffered from HAPE. Clearly, we cannot determine whether this finding is a predisposing factor for or a consequence of having had HAPE. During hypoxia, all of the HAPE-susceptible subjects demonstrated an increase in pulmonary blood flow heterogeneity as measured by the relative dispersion of MR imaging signal intensity of at least 10% of baseline values. Although two of the other subjects, one unselected subject and one HAPE-resistant subject, had small increases in pulmonary blood flow heterogeneity during hypoxia, these were approximately one-half of the smallest response seen in any of the HAPE-susceptible subjects and were not significantly different from zero.
There are several factors that might affect the absolute quantification of regional pulmonary blood flow in vivo. All flowing blood that enters the image plane from outside the selective tag region—from both arteries and veins—will contribute to the signal intensity of the perfusion-weighted image, thus changes in the heterogeneity of signal intensity could reflect alterations in local arterial blood flow, venous blood flow, or both. Also, the physical width of the tagging slice is greater than that of the imaging slice. This creates a gap through which protons in blood must travel before being imaged. Because some of the protons tagged within flowing blood may not reach the imaging slice, this may result in an underestimation of total blood flow, particularly in low velocity vessels. As data is collected for one systolic period, a small amount of flow during diastole will be excluded. However, it is unlikely that these factors that potentially affect the absolute quantification of flow and almost certainly the heterogeneity of flow would be altered significantly by hypoxia per se.
The study was designed to minimize the potential impact of these factors. Subjects remained in the same position in the magnet for the duration of the study. For each subject, the imaging parameters were held constant, the same anatomic level was used for all measurements, and all three subject groups followed the identical imaging protocol. Blood flow data were collected from a single anatomic location across multiple breath holds. Each subject was trained to achieve a consistent breath-hold volume, which was verified on the anatomic images obtained. In addition, this technique is highly reliable (Figure 1), thus changes are unlikely to be the result of a random variability. Importantly, this study did not seek to quantify absolute flow, but quantified flow heterogeneity within a subject. By using each subject as his or her own control, whereas there may be some error in estimation of absolute pulmonary blood flow, the magnitude of any error will be constant within a subject across time. Most importantly, the calculation of relative dispersion should significantly reduce the importance of any errors in quantification of absolute flow, as these potential errors should affect all vessels within the same subject in a relatively uniform fashion and should not change during the study. Thus, the observed changes in relative dispersion in the HAPE-susceptible subjects correspond directly to changes in the uniformity of regional pulmonary blood flow. It is unlikely that these differences in the pulmonary vascular response are related to ongoing altitude exposure in an particular subject group because increased perfusion heterogeneity was observed in all of the HAPE-susceptible subjects irrespective of recent altitude exposure and was not observed in either the HAPE-resistant subjects, who continued to travel to altitude regularly, or the unselected subjects, who had very limited lifetime exposure to altitude and had not traveled above 3,000 m in the last year.
The factors that determine an individual's susceptibly to HAPE are poorly understood (2, 5). HAPE-susceptible subjects have been shown to have exaggerated hypoxic pulmonary vasoconstriction compared with HAPE-resistant subjects (7, 25eC27), greater resting pulmonary vascular resistance (28), and higher pulmonary arterial and capillary wedge pressures (28) during exercise. In addition, it has been shown that capillary pressures are increased in HAPE-susceptible subjects during hypoxic exposure (29). These increased pulmonary vascular pressures are important because there is evidence that pulmonary capillaries suffer mechanical damage during heavy exercise (30, 31), and recent evidence indicates that damage to the capillary endothelium precedes the development of the leak in HAPE (9). How the pulmonary capillaries are exposed to increased microvascular pressure has been uncertain, because the site of vasoconstriction is proximal to the vulnerable capillary bed, leading to the hypothesis that hypoxic pulmonary vasoconstriction in uneven in HAPE.
Although uneven pulmonary blood flow (10) resulting from uneven hypoxic pulmonary vasoconstriction (2) has long been postulated as an important mechanism in HAPE, direct evidence for this idea has been difficult to obtain largely because of the lack of appropriate tools to measure regional pulmonary blood flow in humans. In pigs (12), recent work using fluorescent microspheres has demonstrated that hypoxic pulmonary vasoconstriction is inherently nonuniform. In these animals, a particular vasoconstrictive response is distributed among anatomic regions of the lung, and areas of relatively high flow during hypoxia tend to be adjacent to other areas of high flow, whereas low-flow areas are clustered near other low-flow areas. Also, there is an apparent influence of regional ventilation perfusion ratio, and areas of low ventilation-perfusion ratios tend to vasoconstrict at a higher inspired oxygen concentration than areas of high ventilation-perfusion ratios. In HAPE-susceptible humans, hypoxia is associated with a shift of pulmonary perfusion to the apical region of the lung, as measured by ventilation-perfusion scintigraphy (32). This regional shift in blood flow was not seen in control subjects. However, because HAPE-susceptible subjects are well known to experience higher pulmonary vascular pressures during hypoxia, this would be expected to recruit reserve vessels resulting in a more even overall distribution of pulmonary blood flow, accounting for the apical shift in perfusion. Thus, on the basis of these previous findings, it has been difficult to distinguish the effects of pressure on pulmonary blood flow distribution from uneven hypoxic pulmonary vasoconstriction.
We found an increase in pulmonary blood flow heterogeneity in the HAPE-susceptible subjects when they were exposed to hypoxia. This was not observed in either of the other subject groups. The extent that this heterogeneity contributes to the pathogenesis is uncertain, particularly in light of the short duration of the hypoxic exposure in this study. Although HAPE-susceptible subjects develop higher pulmonary vascular pressures when exposed to hypoxia, the changes in flow heterogeneity we observed in our HAPE-susceptible subjects are not likely to reflect higher pulmonary vascular pressures alone, because an increase in pressure would be expected to cause recruitment of blood vessels and more uniform blood flow rather than less uniform pulmonary blood flow. More likely, these changes in perfusion heterogeneity reflect a more heterogeneous pattern of hypoxic pulmonary vasoconstriction in the HAPE-susceptible subjects. Whether this difference reflects a greater heterogeneity in the stimulus, regional alveolar PO2 or the response to the stimulus, secondary to heterogeneity in the pulmonary vascular distribution of smooth muscle, is an intriguing but unresolved issue. Irrespective of mechanism, these data are consistent with the hypothesis that uneven hypoxic pulmonary vasoconstriction may play a role in the development of HAPE and could explain how high vascular pressures develop in the pulmonary capillaries in HAPE-susceptible individuals.
Acknowledgments
The authors thank the subjects for their enthusiastic participation and Rick Buxton, Tom Liu, Eric Wong, and Larry Frank for their assistance with the pulse sequence programming and discussions regarding the magnetic resonance imaging ASL-FAIRER measurements of regional pulmonary blood flow. They also thank an anonymous reviewer for pointing out the work of Maurice Visscher, who described the potential role of uneven pulmonary blood flow in the development of pulmonary edema (10, 33).
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
REFERENCES
Schoene RB, Hackett PH, Henderson WR, Sage EH, Chow M, Roach RC, Mills WJ, Martin TR. High-altitude pulmonary edema: characteristics of lung lavage fluid. JAMA 1986;256:63eC69.
Hultgren HN. High altitude medicine. Stanford, CA: Hultgren Publications; 1997.
Hackett PH, Rennie D. The incidence, importance, and prophylaxis of acute mountain sickness. Lancet 1976;2:1149eC1155.
Cremona G, Asnaghi R, Baderna P, Brunetto A, Brutsaert T, Cavallaro C, Clark TM, Cogo A, Donis R, Lanfranchi P, et al. Pulmonary extravascular fluid accumulation in recreational climbers: a prospective study. Lancet 2002;359:303eC309.
Schoene R, Swenson E, Hultgren H. High altitude pulmonary edema. In: Hornbein T, Schoene R, editors. High altitude; an exploration of human adaptation. New York: Marcel Dekker Inc.; 2001. pp. 777eC814.
Hultgren HN, Lopez CE, Lundberg E, Miller H. Physiologic studies of pulmonary edema at high altitude. Circulation 1964;29:393eC408.
Hultgren HN, Grover RF, Hartley LH. Abnormal circulatory responses to high altitude in subjects with a previous history of high-altitude pulmonary edema. Circulation 1971;44:759eC770.
Bartsch P, Maggiorini M, Ritter M, Noti C, Vock P, Oelz O. Prevention of high-altitude pulmonary edema by nifedipine. N Engl J Med 1991;325:1284eC1289.
Swenson ER, Maggiorini M, Mongovin S, Gibbs JS, Greve I, Mairbaurl H, Bartsch P. Pathogenesis of high-altitude pulmonary edema: inflammation is not an etiologic factor. JAMA 2002;287:2228eC2235.
Visscher MB. Studies on embolization of lung vessels. Med Thorac 1962;19:334eC340.
Hultgren HN. Pulmonary hypertension and pulmonary edema. In: Loeppky JA, Riedesel ML, editors. Oxygen transport to human tissue. New York: Elsevier/North Holland; 1982. pp. 243eC254.
Hlastala MP, Lamm WJ, Karp A, Polissar NL, Starr IR, Glenny RW. Spatial distribution of hypoxic pulmonary vasoconstriction in the supine pig. J Appl Physiol 2004;96:1589eC1599. Epub 2003 Dec 29.
Wong EC, Buxton RB, Frank LR. Quantitative imaging of perfusion using a single subtraction (QUIPSS and QUIPSS II). Magn Reson Med 1998;39:702eC708.
Frank LR, Wong EC, Haseler LJ, Buxton RB. Dynamic imaging of perfusion in human skeletal muscle during exercise with arterial spin labeling. Magn Reson Med 1999;42:258eC267.
Poncelet BP, Koelling TM, Schmidt CJ, Kwong KK, Reese TG, Ledden P, Kantor HL, Brady TJ, Weisskoff RM. Measurement of human myocardial perfusion by double-gated flow alternating inversion recovery EPI. Magn Reson Med 1999;41:510eC519.
Roberts DA, Detre JA, Bolinger L, Insko EK, Lenkinski RE, Pentecost MJ, Leigh JS Jr. Renal perfusion in humans: MR imaging with spin tagging of arterial water. Radiology 1995;196:281eC286.
Mai VM, Berr SS. MR perfusion imaging of pulmonary parenchyma using pulsed arterial spin labeling techniques: FAIRER and FAIR. J Magn Reson Imaging 1999;9:483eC487.
Levin D, Garg J, Bolar D, Balouch J, Hopkins S. Pulmonary blood flow heterogeneity during hypoxia measured with ASL-FAIRER in subjects with prior high altitude pulmonary edema (HAPE) . Proc Intl Soc Mag Reson Med 12th Meeting 2004:644.
Garg J, Levin D, Bolar D, Hopkins S. Pulmonary blood flow heterogeneity during hypoxia in subjects with a history of high altitude pulmonary edema (HAPE) . High Alt Med Biol 2003;4:429.
Andersen IK, Sidaros K, Gesmara H, Rostrup E, Larsson HB. A model system for perfusion quantification using FAIR. Magn Reson Imaging 2000;18:565eC574.
Raynaud JS, Duteil S, Vaughan JT, Hennel F, Wary C, Leroy-Willig A, Carlier PG. Determination of skeletal muscle perfusion using arterial spin labeling NMRI: validation by comparison with venous occlusion plethysmography. Magn Reson Med 2001;46:305eC311.
Glenny RW. Heterogeneity in the lung:concepts and measures. In: Hlastala MP, Robertson HT, editors. Complexity in structure and function in the lung. New York: Marcel Dekker Inc.; 1998. pp. 571eC609.
Kusuoka H, Hoffman JI. Advice on statistical analysis for Circulation Research. Circ Res 2002;91:662eC671.
Glantz SA, Slinker BK. Primer of applied regression and analysis of variance, 2nd ed. New York: McGraw-Hill; 2001.
Kawashima A, Kubo K, Kobayashi T, Sekiguchi M. Hemodynamic responses to acute hypoxia, hypobaria, and exercise in subjects susceptible to high-altitude pulmonary edema. J Appl Physiol 1989;67:1982eC1989.
Viswanathan R, Jain SK, Subramanian S, Subramanian TA, Dua GL, Giri J. Pulmonary edema of high altitude: II. Clinical, aerohemodynamic, and biochemical studies in a group with history of pulmonary edema of high altitude. Am Rev Respir Dis 1969;100:334eC341.
Yagi H, Yamada H, Kobayashi T, Sekiguchi M. Doppler assessment of pulmonary hypertension induced by hypoxic breathing in subjects susceptible to high altitude pulmonary edema. Am Rev Respir Dis 1990;142:796eC801.
Eldridge MW, Podolsky A, Richardson RS, Johnson DH, Knight DR, Johnson EC, Hopkins SR, Michimata H, Grassi B, Feiner J, et al. Pulmonary hemodynamic response to exercise in subjects with prior high-altitude pulmonary edema. J Appl Physiol 1996;81:911eC921.
Maggiorini M, Melot C, Pierre S, Pfeiffer F, Greve I, Sartori C, Lepori M, Hauser M, Scherrer U, Naeije R. High-altitude pulmonary edema is initially caused by an increase in capillary pressure. Circulation 2001;103:2078eC2083.
Hopkins SR, Schoene RB, Henderson WR, Spragg RG, Martin TR, West JB. Intense exercise impairs the integrity of the pulmonary blood-gas barrier in elite athletes. Am J Respir Crit Care Med 1997;155:1090eC1094.
West JB, Mathieu-Costello O, Jones JH, Birks EK, Logemann RB, Pascoe JR, Tyler WS. Stress failure of pulmonary capillaries in racehorses with exercise-induced pulmonary hemorrhage. J Appl Physiol 1993;75:1097eC1109.
Hanaoka M, Tanaka M, Ge R-L, Droma Y, Ito A, Miyahara T, Koizumi T, Fujimoto K, Fujii T, Kobayashi T, et al. Hypoxia-induced pulmonary blood redistribution in subjects with a history of high-altitude pulmonary edema. Circulation 2000;101:1418eC1422.
Visscher MB. Normal and abnormal pulmonary circulation. Paper presented at 5th Conference on Research in Emphysema. 1962, Aspen, CO.(Susan R. Hopkins, Joy Gar)
Department of Radiology, University of California, San Diego, La Jolla, California
ABSTRACT
Uneven hypoxic pulmonary vasoconstriction has been proposed to expose parts of the pulmonary capillary bed to high pressure and vascular injury in high-altitude pulmonary edema (HAPE). We hypothesized that subjects with a history of HAPE would demonstrate increased heterogeneity of pulmonary blood flow during hypoxia. A functional magnetic resonance imaging technique (arterial spin labeling) was used to quantify spatial pulmonary blood flow heterogeneity in three subject groups: (1) HAPE-susceptible (n = 5), individuals with a history of physician-documented HAPE; (2) HAPE-resistant (n = 6), individuals with repeated high-altitude exposure without illness; and (3) unselected (n = 6), individuals with a minimal history of altitude exposure. Data were collected in normoxia and after 5, 10, 20, and 30 minutes of normobaric hypoxia . Relative dispersion (SD/mean) of the signal intensity was used as an index of perfusion heterogeneity. Oxygen saturation was not different between groups during hypoxia. Relative dispersion was not different between groups (HAPE-susceptible 0.94 ± 0.05, HAPE-resistant 0.94 ± 0.05, unselected 0.87 ± 0.06; means ± SEM) during normoxia, but it was increased by hypoxia in HAPE-susceptible (to 1.10 ± 0.05 after 30 minutes, p < 0.0001) but not in HAPE-resistant (0.91 ± 0.05) or unselected subjects (0.87 ± 0.05). HAPE-susceptible individuals have increased pulmonary blood flow heterogeneity in acute hypoxia, consistent with uneven hypoxic pulmonary vasoconstriction.
Key Words: hypoxic pulmonary vasoconstriction magnetic resonance imaging pulmonary circulation
High-altitude pulmonary edema (HAPE) is a noncardiogenic high-permeability edema characterized by alveolar fluid with a high concentration of protein (1) that develops in otherwise healthy individuals after 24 to 72 hours of exposure to altitudes above 2,400 m ( 8,000 ft). Typically, radiographs in HAPE demonstrate a patchy distribution of edema, appearing first in the perihilar regions of the lung (2). The reported incidence is 2 to 15% of exposed individuals depending on rate of ascent and altitude reached (3eC5). Risk factors include rapid ascent, strenuous exercise, and a previous history of HAPE. Increased pulmonary artery pressure is a hallmark of HAPE (6, 7), and susceptible subjects have been shown to have increased pulmonary arterial pressures before the development of HAPE (8). Increased concentrations of red blood cells and protein in bronchoalveolar lavage fluid are observed in the early stages of HAPE before inflammatory cytokines are present (9), suggesting that mechanical injury to the pulmonary capillary bed (stress failure) may be important in the development of HAPE. Because hypoxic pulmonary vasoconstriction takes place in the precapillary arterioles, it is uncertain how capillaries might be exposed to high pressure. A mechanism initially proposed by Visscher (10) and later modified for HAPE by Hultgren (2) is that hypoxic pulmonary vasoconstriction is uneven in HAPE (2). If this were true, some parts of the capillary bed would be protected by upstream vasoconstriction, whereas the portions of the capillary that did not have upstream vasoconstriction would be exposed to high pressures and mechanical stress injury(11), thus inciting HAPE.
Recent animal work confirms that hypoxia results in increased spatial heterogeneity of pulmonary perfusion, suggesting that hypoxic pulmonary vasoconstriction is inherently uneven in the mammalian lung (12). However, the relationship between the development of uneven hypoxic pulmonary vasoconstriction and susceptibility to HAPE is unknown, and it is uncertain whether HAPE-susceptible subjects have increased heterogeneity of pulmonary perfusion in response to hypoxia. We hypothesized that HAPE-susceptible subjects would demonstrate increased pulmonary blood flow heterogeneity during hypoxia compared with subjects without a history of HAPE. To test this hypothesis, we measured pulmonary blood flow heterogeneity in response to acute hypoxia using a quantitative magnetic resonance (MR) imaging technique known as arterial spin labeling. Arterial spin labeling involves creating a magnetically tagged bolus by the use of specialized radiofrequency pulses to alter the magnetization of water in blood. By measuring the transit of the tagged bolus through organ vasculature, it is possible to measure the spatial distribution of blood flow. Arterial spin labeling has been widely applied to measure cerebral blood flow (13) and to a lesser extent has been used to evaluate perfusion in skeletal muscle (14), heart (15), kidney (16), and lung (17). We measured the distribution of pulmonary blood flow using arterial spin labeling in normoxia and acute hypoxia in subjects with a history of HAPE, and compared the results to those of individuals repeatedly exposed to altitude without HAPE and subjects unselected for HAPE-susceptibility by virtue of very limited high-altitude exposure. Some of the results of this study have been previously reported in abstract form (18, 19).
METHODS
The Human Subjects Research Protection Program of the University of California, San Diego approved this study, and the procedures followed conformed with this institution's guidelines. Healthy nonsmoking subjects were recruited by advertisement, informed of the risks of the study, and signed informed consent. A history was taken, with particular reference to the cardiopulmonary system and history of altitude exposure, and a physical examination was performed. The subjects were recruited from three groups: HAPE-susceptible individuals, a history of physician-documented HAPE on at least one occasion (n = 5); HAPE-resistant individuals, repeated high-altitude exposure greater than 5,400 m, without any history of altitude illness (n = 6), and unselected individuals, very limited altitude exposure history (see RESULTS, n = 6).
The subjects were screened for cardiac (ECG) and pulmonary abnormalities (chest radiograph and spirometry). Each subject underwent MR imaging with arterial spin labeling using a Vision 1.5 T whole-body MR Scanner (Siemens Medical Systems, Erlangen, Germany). MR imaging arterial spin labeling is a clinical technique used for the quantification of organ blood flow. The signal intensity of the image has been shown to be proportional to bulk flow in vitro using a tube-flow model (20), and it has been validated in heart (15) with microspheres, and skeletal muscle with venous occlusion plethysmography (21) with an excellent linear relationship between arterial spin labeling measurements and the validating technique. Arterial spin labeling inverts the proton magnetization of blood by applying a radiofrequency pulse, allowing these magnetically tagged protons in blood to act as an endogenous tracer for the evaluation of blood flow (see Figures E1 and E2 in the online data supplement). During each measurement two images are acquired. In the first image, an inversion pulse is applied to the entire volume of the lungs and the image of the selected lung slice is taken. In the second image, a second inversion pulse is used, but applied only to the image slice (a selective inversion). During the delay before the second image is acquired, any blood that enters the image slice from outside will have been missed by the selective inversion. The two images are then subtracted. Because stationary inverted protons from tissue will be present in both images, they will cancel one another out. Thus, the signal from each image voxel is then proportional to the amount of blood that entered the voxel in the delay between the selective inversion pulse and the image acquisition. The sequence is cardiac gated, and flow data are collected over one cardiac cycle encompassing one systole and portions of two diastoles.
Subjects rested in the supine position wearing a face mask (Hans Rudolph 8930; Hans Rudolph Inc, Kansas City, MO) equipped with a nonrebreathing valve (Hans Rudolph 2700; Hans Rudolph Inc). Pulmonary perfusion was quantified using a cardiac-gated pulmonary arterial spin labeling MR sequence as described (17). A detailed description of these methods is given in the online supplement. All sequence parameters were kept within U.S. Food and Drug Administration guidelines for clinical MR examinations. The spatial resolution of measurement with this technique is 1.5 x 3 x 15 mm ( 70 mm3).
Data were obtained during an 8- to 10-second breath hold at functional residual capacity in a 15-mm-thick image slice in the coronal plane in the posterior one-third of the lung using the posterior edge of the descending aorta as a reference point. Five measures were obtained at each time point and averaged. Baseline data were obtained after resting quietly in the magnet for 10 minutes breathing room air. Then, the inspired port of the Hans Rudolph valve was connected to a Douglas bag containing a ( 3,800 m equivalent altitude) and data collection was repeated after 5, 10, 20, and 30 minutes of hypoxia. Arterial oxygen saturation and heart rate were monitored (3150 MR Monitor; Invivo Research Inc, Orlando, FL). The signal intensity for each voxel within the lung parenchyma was determined for each image using MATLAB (The MathWorks, Natick, MA). Relative dispersion (SD/mean signal intensity), an index of blood flow heterogeneity where a greater number indicates a more heterogeneously distributed system, was calculated for all time points (22). Figure 1 presents reliability data for relative dispersion of perfusion heterogeneity obtained in 45 healthy nonsmoking volunteers using MR imaging arterial spin labeling.
Data were analyzed using analysis of variance (Statview 4.1, SAS Institute Inc., Cary, NC) with one randomized group (subject group, three levels: HAPE-susceptible, HAPE-resistant, and unselected subjects) and one repeated measure (duration of hypoxia, five levels: 0, 5, 10, 20, and 30 minutes) (23, 24). Where overall significance was determined, post hoc analysis of variance testing was applied to determine where the differences occurred. Linear regression was used to relate the change in blood flow heterogeneity to arterial oxygen saturation (24). Significance was accepted at p < 0.05, two tailed. All data are means ± SEM.
RESULTS
Subject descriptive data are given in Table 1. Four of the five HAPE-susceptible subjects were diagnosed on the basis of chest radiographs in addition to history and physical findings. The remaining subject had a classic history for HAPE in addition to typical physical findings, which were documented by his physician. Three of our HAPE-susceptible subjects were experienced climbers: two had experienced HAPE on more than one occasion and the third had experienced HAPE on one occasion after a rapid ascent. All three of these individuals continued to travel to high altitudes regularly. The other two HAPE-susceptible subjects had developed HAPE on their first time at high altitude and had not returned to high altitude because of concerns over a reoccurrence. All of the HAPE-resistant subjects were experienced alpinists and regularly traveled to high altitude, usually with very rapid ascent profiles. Four members of this group had Himalayan climbing experience; one had guided a successful climb of Mt. Everest (8,848 m). The remaining two had climbed in South America. One subject in the unselected group had a single history of brief exposure to an altitude of 4,300 m 2 years before the study. The remaining subjects had not traveled above 2,400 to 4,000 m, and all exposures at these altitudes were more than 1 year before the study, of brief (a few hours) duration, included no overnight stay, and occurred on a single occasion only. No subject had slept above 2,000 m.
There were no significant differences between groups for age, height, weight, or pulmonary function variables (Table 1). The HAPE-resistant subjects had reached significantly higher altitudes (6,736 ± 460 m) than either the HAPE-susceptible (5,907 ± 678 m) or the unselected subjects (3,322 ± 299 m). Arterial oxygen saturation (Figure 2, top panel) was 98.2 ± 0.2% at baseline and fell to 87.2 ± 1.3% after 30 minutes of hypoxia (p < 0.0001), but there were no significant differences between groups (p = 0.56), and no significant group-by-time interaction (p = 0.77), thus all subject groups experienced a similar level of hypoxia. Heart rate (Figure 2, bottom panel) measured by pulse oximetry was increased by hypoxia (p < 0.0001); however, this was not significantly different between groups (p = 0.57), and there was no significant group-by-time interaction (p = 0.63).
The overall quality of the MR imaging perfusion maps was excellent, and mean signal-to-noise ratio was 29:1. The mean signal did not differ between the three subject groups over the course of the study. The relative dispersion of signal intensity (Figure 3) was not different between groups at baseline (HAPE-susceptible, 0.94 ± 0.05; HAPE-resistant, 0.94 ± 0.05; and unselected subjects, 0.87 ± 0.06). However, relative dispersion was increased during hypoxia in HAPE-susceptible (to 1.10 ± 0.05 after 30 minutes, p < 0.0001) but not in HAPE-resistant (0.91 ± 0.05, p = 0.44) or unselected subjects (0.87 ± 0.05, p = 0.30). Strikingly, all of the HAPE-susceptible subjects increased pulmonary blood flow heterogeneity by at least 10% of their baseline values by 30 minutes of hypoxia (Figures 4 and 5).
Because there was no difference in the change in relative dispersion in response to hypoxia between the HAPE-resistant and unselected subjects, these data were pooled and linear regression analysis was performed. There was a significant relationship between arterial oxygen saturation at 30 minutes and the change in relative dispersion from baseline (Figure 4, R = eC0.60, p < 0.05). Overall, hypoxia was associated with small reductions in relative dispersion in these subjects, but three of the subjects who became the most hypoxic had small increases in heterogeneity resulting in a negative slope. When data from the five HAPE-susceptible subjects were plotted on the same figure, all of the values for the change in relative dispersion for HAPE-susceptible subjects fell above the +2 SD confidence limits for the relationship of the change in relative dispersion verses arterial oxygen saturation. In addition, there was a strong within-subject relationship between arterial oxygen saturation and relative dispersion for the HAPE-susceptible subjects, with an average correlation of eC0.82 within subjects across all time points (Figure 5). By contrast, the correlation between saturation and relative dispersion averaged +0.17 for the HAPE-resistant subjects and eC0.01 for the unselected subjects.
DISCUSSION
We describe a highly significant difference in the pulmonary vascular response to hypoxia between HAPE-susceptible individuals and those who have not suffered from HAPE. Clearly, we cannot determine whether this finding is a predisposing factor for or a consequence of having had HAPE. During hypoxia, all of the HAPE-susceptible subjects demonstrated an increase in pulmonary blood flow heterogeneity as measured by the relative dispersion of MR imaging signal intensity of at least 10% of baseline values. Although two of the other subjects, one unselected subject and one HAPE-resistant subject, had small increases in pulmonary blood flow heterogeneity during hypoxia, these were approximately one-half of the smallest response seen in any of the HAPE-susceptible subjects and were not significantly different from zero.
There are several factors that might affect the absolute quantification of regional pulmonary blood flow in vivo. All flowing blood that enters the image plane from outside the selective tag region—from both arteries and veins—will contribute to the signal intensity of the perfusion-weighted image, thus changes in the heterogeneity of signal intensity could reflect alterations in local arterial blood flow, venous blood flow, or both. Also, the physical width of the tagging slice is greater than that of the imaging slice. This creates a gap through which protons in blood must travel before being imaged. Because some of the protons tagged within flowing blood may not reach the imaging slice, this may result in an underestimation of total blood flow, particularly in low velocity vessels. As data is collected for one systolic period, a small amount of flow during diastole will be excluded. However, it is unlikely that these factors that potentially affect the absolute quantification of flow and almost certainly the heterogeneity of flow would be altered significantly by hypoxia per se.
The study was designed to minimize the potential impact of these factors. Subjects remained in the same position in the magnet for the duration of the study. For each subject, the imaging parameters were held constant, the same anatomic level was used for all measurements, and all three subject groups followed the identical imaging protocol. Blood flow data were collected from a single anatomic location across multiple breath holds. Each subject was trained to achieve a consistent breath-hold volume, which was verified on the anatomic images obtained. In addition, this technique is highly reliable (Figure 1), thus changes are unlikely to be the result of a random variability. Importantly, this study did not seek to quantify absolute flow, but quantified flow heterogeneity within a subject. By using each subject as his or her own control, whereas there may be some error in estimation of absolute pulmonary blood flow, the magnitude of any error will be constant within a subject across time. Most importantly, the calculation of relative dispersion should significantly reduce the importance of any errors in quantification of absolute flow, as these potential errors should affect all vessels within the same subject in a relatively uniform fashion and should not change during the study. Thus, the observed changes in relative dispersion in the HAPE-susceptible subjects correspond directly to changes in the uniformity of regional pulmonary blood flow. It is unlikely that these differences in the pulmonary vascular response are related to ongoing altitude exposure in an particular subject group because increased perfusion heterogeneity was observed in all of the HAPE-susceptible subjects irrespective of recent altitude exposure and was not observed in either the HAPE-resistant subjects, who continued to travel to altitude regularly, or the unselected subjects, who had very limited lifetime exposure to altitude and had not traveled above 3,000 m in the last year.
The factors that determine an individual's susceptibly to HAPE are poorly understood (2, 5). HAPE-susceptible subjects have been shown to have exaggerated hypoxic pulmonary vasoconstriction compared with HAPE-resistant subjects (7, 25eC27), greater resting pulmonary vascular resistance (28), and higher pulmonary arterial and capillary wedge pressures (28) during exercise. In addition, it has been shown that capillary pressures are increased in HAPE-susceptible subjects during hypoxic exposure (29). These increased pulmonary vascular pressures are important because there is evidence that pulmonary capillaries suffer mechanical damage during heavy exercise (30, 31), and recent evidence indicates that damage to the capillary endothelium precedes the development of the leak in HAPE (9). How the pulmonary capillaries are exposed to increased microvascular pressure has been uncertain, because the site of vasoconstriction is proximal to the vulnerable capillary bed, leading to the hypothesis that hypoxic pulmonary vasoconstriction in uneven in HAPE.
Although uneven pulmonary blood flow (10) resulting from uneven hypoxic pulmonary vasoconstriction (2) has long been postulated as an important mechanism in HAPE, direct evidence for this idea has been difficult to obtain largely because of the lack of appropriate tools to measure regional pulmonary blood flow in humans. In pigs (12), recent work using fluorescent microspheres has demonstrated that hypoxic pulmonary vasoconstriction is inherently nonuniform. In these animals, a particular vasoconstrictive response is distributed among anatomic regions of the lung, and areas of relatively high flow during hypoxia tend to be adjacent to other areas of high flow, whereas low-flow areas are clustered near other low-flow areas. Also, there is an apparent influence of regional ventilation perfusion ratio, and areas of low ventilation-perfusion ratios tend to vasoconstrict at a higher inspired oxygen concentration than areas of high ventilation-perfusion ratios. In HAPE-susceptible humans, hypoxia is associated with a shift of pulmonary perfusion to the apical region of the lung, as measured by ventilation-perfusion scintigraphy (32). This regional shift in blood flow was not seen in control subjects. However, because HAPE-susceptible subjects are well known to experience higher pulmonary vascular pressures during hypoxia, this would be expected to recruit reserve vessels resulting in a more even overall distribution of pulmonary blood flow, accounting for the apical shift in perfusion. Thus, on the basis of these previous findings, it has been difficult to distinguish the effects of pressure on pulmonary blood flow distribution from uneven hypoxic pulmonary vasoconstriction.
We found an increase in pulmonary blood flow heterogeneity in the HAPE-susceptible subjects when they were exposed to hypoxia. This was not observed in either of the other subject groups. The extent that this heterogeneity contributes to the pathogenesis is uncertain, particularly in light of the short duration of the hypoxic exposure in this study. Although HAPE-susceptible subjects develop higher pulmonary vascular pressures when exposed to hypoxia, the changes in flow heterogeneity we observed in our HAPE-susceptible subjects are not likely to reflect higher pulmonary vascular pressures alone, because an increase in pressure would be expected to cause recruitment of blood vessels and more uniform blood flow rather than less uniform pulmonary blood flow. More likely, these changes in perfusion heterogeneity reflect a more heterogeneous pattern of hypoxic pulmonary vasoconstriction in the HAPE-susceptible subjects. Whether this difference reflects a greater heterogeneity in the stimulus, regional alveolar PO2 or the response to the stimulus, secondary to heterogeneity in the pulmonary vascular distribution of smooth muscle, is an intriguing but unresolved issue. Irrespective of mechanism, these data are consistent with the hypothesis that uneven hypoxic pulmonary vasoconstriction may play a role in the development of HAPE and could explain how high vascular pressures develop in the pulmonary capillaries in HAPE-susceptible individuals.
Acknowledgments
The authors thank the subjects for their enthusiastic participation and Rick Buxton, Tom Liu, Eric Wong, and Larry Frank for their assistance with the pulse sequence programming and discussions regarding the magnetic resonance imaging ASL-FAIRER measurements of regional pulmonary blood flow. They also thank an anonymous reviewer for pointing out the work of Maurice Visscher, who described the potential role of uneven pulmonary blood flow in the development of pulmonary edema (10, 33).
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
REFERENCES
Schoene RB, Hackett PH, Henderson WR, Sage EH, Chow M, Roach RC, Mills WJ, Martin TR. High-altitude pulmonary edema: characteristics of lung lavage fluid. JAMA 1986;256:63eC69.
Hultgren HN. High altitude medicine. Stanford, CA: Hultgren Publications; 1997.
Hackett PH, Rennie D. The incidence, importance, and prophylaxis of acute mountain sickness. Lancet 1976;2:1149eC1155.
Cremona G, Asnaghi R, Baderna P, Brunetto A, Brutsaert T, Cavallaro C, Clark TM, Cogo A, Donis R, Lanfranchi P, et al. Pulmonary extravascular fluid accumulation in recreational climbers: a prospective study. Lancet 2002;359:303eC309.
Schoene R, Swenson E, Hultgren H. High altitude pulmonary edema. In: Hornbein T, Schoene R, editors. High altitude; an exploration of human adaptation. New York: Marcel Dekker Inc.; 2001. pp. 777eC814.
Hultgren HN, Lopez CE, Lundberg E, Miller H. Physiologic studies of pulmonary edema at high altitude. Circulation 1964;29:393eC408.
Hultgren HN, Grover RF, Hartley LH. Abnormal circulatory responses to high altitude in subjects with a previous history of high-altitude pulmonary edema. Circulation 1971;44:759eC770.
Bartsch P, Maggiorini M, Ritter M, Noti C, Vock P, Oelz O. Prevention of high-altitude pulmonary edema by nifedipine. N Engl J Med 1991;325:1284eC1289.
Swenson ER, Maggiorini M, Mongovin S, Gibbs JS, Greve I, Mairbaurl H, Bartsch P. Pathogenesis of high-altitude pulmonary edema: inflammation is not an etiologic factor. JAMA 2002;287:2228eC2235.
Visscher MB. Studies on embolization of lung vessels. Med Thorac 1962;19:334eC340.
Hultgren HN. Pulmonary hypertension and pulmonary edema. In: Loeppky JA, Riedesel ML, editors. Oxygen transport to human tissue. New York: Elsevier/North Holland; 1982. pp. 243eC254.
Hlastala MP, Lamm WJ, Karp A, Polissar NL, Starr IR, Glenny RW. Spatial distribution of hypoxic pulmonary vasoconstriction in the supine pig. J Appl Physiol 2004;96:1589eC1599. Epub 2003 Dec 29.
Wong EC, Buxton RB, Frank LR. Quantitative imaging of perfusion using a single subtraction (QUIPSS and QUIPSS II). Magn Reson Med 1998;39:702eC708.
Frank LR, Wong EC, Haseler LJ, Buxton RB. Dynamic imaging of perfusion in human skeletal muscle during exercise with arterial spin labeling. Magn Reson Med 1999;42:258eC267.
Poncelet BP, Koelling TM, Schmidt CJ, Kwong KK, Reese TG, Ledden P, Kantor HL, Brady TJ, Weisskoff RM. Measurement of human myocardial perfusion by double-gated flow alternating inversion recovery EPI. Magn Reson Med 1999;41:510eC519.
Roberts DA, Detre JA, Bolinger L, Insko EK, Lenkinski RE, Pentecost MJ, Leigh JS Jr. Renal perfusion in humans: MR imaging with spin tagging of arterial water. Radiology 1995;196:281eC286.
Mai VM, Berr SS. MR perfusion imaging of pulmonary parenchyma using pulsed arterial spin labeling techniques: FAIRER and FAIR. J Magn Reson Imaging 1999;9:483eC487.
Levin D, Garg J, Bolar D, Balouch J, Hopkins S. Pulmonary blood flow heterogeneity during hypoxia measured with ASL-FAIRER in subjects with prior high altitude pulmonary edema (HAPE) . Proc Intl Soc Mag Reson Med 12th Meeting 2004:644.
Garg J, Levin D, Bolar D, Hopkins S. Pulmonary blood flow heterogeneity during hypoxia in subjects with a history of high altitude pulmonary edema (HAPE) . High Alt Med Biol 2003;4:429.
Andersen IK, Sidaros K, Gesmara H, Rostrup E, Larsson HB. A model system for perfusion quantification using FAIR. Magn Reson Imaging 2000;18:565eC574.
Raynaud JS, Duteil S, Vaughan JT, Hennel F, Wary C, Leroy-Willig A, Carlier PG. Determination of skeletal muscle perfusion using arterial spin labeling NMRI: validation by comparison with venous occlusion plethysmography. Magn Reson Med 2001;46:305eC311.
Glenny RW. Heterogeneity in the lung:concepts and measures. In: Hlastala MP, Robertson HT, editors. Complexity in structure and function in the lung. New York: Marcel Dekker Inc.; 1998. pp. 571eC609.
Kusuoka H, Hoffman JI. Advice on statistical analysis for Circulation Research. Circ Res 2002;91:662eC671.
Glantz SA, Slinker BK. Primer of applied regression and analysis of variance, 2nd ed. New York: McGraw-Hill; 2001.
Kawashima A, Kubo K, Kobayashi T, Sekiguchi M. Hemodynamic responses to acute hypoxia, hypobaria, and exercise in subjects susceptible to high-altitude pulmonary edema. J Appl Physiol 1989;67:1982eC1989.
Viswanathan R, Jain SK, Subramanian S, Subramanian TA, Dua GL, Giri J. Pulmonary edema of high altitude: II. Clinical, aerohemodynamic, and biochemical studies in a group with history of pulmonary edema of high altitude. Am Rev Respir Dis 1969;100:334eC341.
Yagi H, Yamada H, Kobayashi T, Sekiguchi M. Doppler assessment of pulmonary hypertension induced by hypoxic breathing in subjects susceptible to high altitude pulmonary edema. Am Rev Respir Dis 1990;142:796eC801.
Eldridge MW, Podolsky A, Richardson RS, Johnson DH, Knight DR, Johnson EC, Hopkins SR, Michimata H, Grassi B, Feiner J, et al. Pulmonary hemodynamic response to exercise in subjects with prior high-altitude pulmonary edema. J Appl Physiol 1996;81:911eC921.
Maggiorini M, Melot C, Pierre S, Pfeiffer F, Greve I, Sartori C, Lepori M, Hauser M, Scherrer U, Naeije R. High-altitude pulmonary edema is initially caused by an increase in capillary pressure. Circulation 2001;103:2078eC2083.
Hopkins SR, Schoene RB, Henderson WR, Spragg RG, Martin TR, West JB. Intense exercise impairs the integrity of the pulmonary blood-gas barrier in elite athletes. Am J Respir Crit Care Med 1997;155:1090eC1094.
West JB, Mathieu-Costello O, Jones JH, Birks EK, Logemann RB, Pascoe JR, Tyler WS. Stress failure of pulmonary capillaries in racehorses with exercise-induced pulmonary hemorrhage. J Appl Physiol 1993;75:1097eC1109.
Hanaoka M, Tanaka M, Ge R-L, Droma Y, Ito A, Miyahara T, Koizumi T, Fujimoto K, Fujii T, Kobayashi T, et al. Hypoxia-induced pulmonary blood redistribution in subjects with a history of high-altitude pulmonary edema. Circulation 2000;101:1418eC1422.
Visscher MB. Normal and abnormal pulmonary circulation. Paper presented at 5th Conference on Research in Emphysema. 1962, Aspen, CO.(Susan R. Hopkins, Joy Gar)