Published ahead of print on August 28, 2002, doi:10.1164/rccm.200204-345OC
© 2002 American Thoracic Society
Characterization of High-Altitude Pulmonary Hypertension in the KyrgyzAssociation with Angiotensin-Converting Enzyme GenotypeNational Center for Cardiology and Internal Medicine, Bishkek, Kyrgyzstan; Section on Clinical Pharmacology, Hammersmith Hospital, Imperial College School of Medicine, London; and the Department of Medicine, University of Cambridge, Cambridge, United Kingdom Correspondence and requests for reprints should be addressed to Almaz A. Aldashev, M.D., Laboratory of Molecular and Cell Biology, National Center of Cardiology and Internal Medicine, Togolok Moldo St. 3, 720040 Bishkek, Kyrgyz Republic. E-mail: cardio{at}elcat.kg; or to Nicholas W. Morrell, M.D., Department of Medicine, University of Cambridge, Addenbrooke's Hospital, Hills Road, Cambridge CB2 2QQ, UK. E-mail: nwm23{at}cam.ac.uk
Previous studies have suggested a genetic component in susceptibility to hypoxia-induced pulmonary hypertension. We therefore estimated the prevalence of high-altitude pulmonary hypertension (HAPH) in a Kyrgyz population and whether the insertion/deletion (I/D) polymorphism of the angiotensin-converting enzyme (ACE) gene associates with HAPH. An electrocardiographic survey of 741 highlanders demonstrated electrocardiogram signs of cor pulmonale in 14% of subjects. Pulmonary artery hemodynamics measured in an independent group of 136 male highlanders with symptoms of dyspnea at altitude revealed established pulmonary hypertension (mean pulmonary artery pressure [MPAP]  25 mm Hg) in 20%. However, 26% of the normal subjects demonstrated an exaggerated response (twofold or greater increase in MPAP) to inhalation of 11% oxygen, and were classified as hyperresponsive. Ten-year follow-up of this group revealed increases in the MPAP, but not in normal subjects. Comparison of ACE I/D genotypes in the catheterized group revealed a threefold higher frequency of the I/I genotype in highlanders with HAPH, compared with normal highlanders ( 2 = 11.59, p = 0.003). In addition, MPAP was higher in highlanders with the I/I genotype (26.9 ± 4.0 mm Hg) compared with the I/D genotype (20.6 ± 1.2 mm Hg) or the D/D genotype (18.3 ± 0.9 mm Hg) (p < 0.05). We conclude that HAPH is associated with ACE I/D genotype among Kyrgyz highlanders and the development of HAPH in this population and may be predicted by hyperresponsiveness to acute hypoxia.
Key Words: altitude • angiotensin-converting enzyme • hypoxia • pulmonary hypertension
The normal adult pulmonary circulation is a low-pressure, high-flow vascular bed with a great capacity for vascular recruitment. A unique feature of the pulmonary, as opposed to the systemic, circulation is vasoconstriction observed at decreased levels of alveolar oxygen (1). When hypoxia is prolonged, for example, during residence at high altitude, structural changes occur in the walls of small pulmonary arteries, predominantly increased muscularization, leading to increased pulmonary vascular resistance and sustained elevation of pulmonary arterial pressure (2, 3). The resulting pulmonary hypertension leads to right ventricular hypertrophy and in some individuals to death from right ventricular failure. High-altitude pulmonary hypertension (HAPH) is a public health problem in the high-altitude areas of the world, including Kyrgyzstan in Central Asia, formerly part of the Soviet Union (4). In this region, the Tien-Shan and Pamir Mountains occupy about 90% of the territory and more than 200,000 people live at an altitude greater than 3,000 m above sea level. A number of factors are known to influence the development of HAPH, including the ventilatory response to hypoxia (4, 5), the presence of coexisting obstructive airway disease, and the severity of polycythemia (6). Genetic susceptibility to hypoxia-induced pulmonary hypertension in some individuals is suggested by a number of observations. First, it is known that susceptibility to high-altitude pulmonary hypertension in cattle ("brisket disease") is an inherited trait (7). In addition, strains of rat have been described with differing susceptibilities to hypoxia-induced pulmonary hypertension (8, 9). A quantitative trait locus was identified on rat chromosome 17 that influences the degree of right ventricular hypertrophy (10). In humans, large interindividual differences exist in the magnitude of the pulmonary pressure response to hypoxia (11), with some subjects demonstrating exaggerated increases in pulmonary arterial pressure (12). Furthermore, populations that have lived at high altitude for many generations, for example, the Tibetans, seem to have a blunted pulmonary pressor response to hypoxia, suggesting that this attenuated response confers an advantage at high altitude (13). The aim of the present study was to estimate the prevalence of HAPH among the high-altitude inhabitants of the Tien-Shan and Pamir Mountains of Kyrgyzstan, by (1) the frequency of electrocardiographic (ECG) signs of cor pulmonale and (2) pulmonary hemodynamics measured during right heart catheterization, as a basis for studies of genetic susceptibility. We previously suggested an association between HAPH and the insertion/deletion (I/D) polymorphism of intron 16 of the angiotensin-converting enzyme (ACE) gene (14), which is known to contribute to the level and activity of serum ACE (15, 16). Thus we studied ACE genotype, ACE levels, and ACE activity in a larger cohort of patients and control subjects to determine whether the ACE I/D polymorphism, and ACE levels and activity, influence susceptibility to high-altitude pulmonary hypertension in Kyrgyz highlanders.
Electrocardiographic Survey Standard 12-lead electrocardiography (ECG) was performed on ethnic Kyrgyz subjects aged 16 to 75 years who were permanent residents of three villages in the Naryn area (altitude, 2,800 to 3,100 m above sea level). The total population of these villages is about 2,400. ECGs were performed on all available and consenting villagers by two investigators over a 2-month period. All subjects underwent health screening by history, physical examination, spirometry, blood pressure measurement, and ECG. Patients with obvious clinical or ECG signs of significant comorbidity (e.g., ischemic heart disease or obstructive airway disease) were excluded from subsequent analysis (n = 36). As a result, 741 ECGs were analyzed for signs of cor pulmonale, as evidenced by ECG criteria of right ventricular hypertrophy, defined by one of the following patterns (17): right axis deviation, defined as a frontal plane QRS axis of 90 degrees (Pattern A); R wave in lead V1 of 5 mm, an R/S ratio of > 1, and S > R in V5 or V6 (Pattern B); a leftward shift in the transition zone (the precordial lead with negative and positive deflections of equal amplitude) (Pattern C). To validate the use of ECG criteria as an index of pulmonary hypertension 11 subjects with all 3 ECG criteria traveled to Bishkek (600 m above sea level) to undergo right heart catheterization. All studies were performed within 7 days of leaving the high-altitude regions.
Right Heart Catheterization Studies To determine whether the above-described hemodynamic measurements could be used to predict the subsequent progression of pulmonary hypertension, we repeated cardiac catheterization in subjects who could be traced and who consented, at least 10 years after their initial catheterization.
Pathology
ACE Genotyping
ACE Levels and Activity Serum ACE activity was assayed by hydrolysis of the specific substrate Z-Phe-His-Leu as previously described (21).
Association Studies between HAPH and ACE Genotype and Activity
Statistic Analysis
Electrocardiography ECGs were performed on 741 permanent residents (347 males and 394 females; mean age, 34.3 ± 1.3 years) at high altitude and analyzed for signs of cor pulmonale. In total, 105 subjects (14%) had one or more ECG criteria of cor pulmonale as defined above. Of these, 80 were male and 25 were female. Thus the prevalence of ECG signs of cor pulmonale in the male highlanders was 23%, compared with 6% in females (p < 0.0001). Of the 80 males, 34 (43%) were smokers compared with none of the females. None had evidence of significant airflow obstruction on spirometry: mean FEV1, 3.58 ± 0.07 L; mean VC, 4.08 ± 0.06 L; and FEV1/VC ratio, greater than 70% predicted. Pattern A was observed in 81 subjects (10.9%), Pattern B was observed in 11 subjects (1.5%), and Pattern C was observed in 68 subjects (9.2%). Forty-six subjects had a combination of Patterns A and C (6.2%), and 18 subjects (2.4%) had a combination of Patterns A, B, and C. Eleven subjects (10 males and 1 female; mean age, 58 ± 4 years) with all 3 ECG criteria of cor pulmonale were studied further by right heart catheterization. The mean pulmonary arterial pressure in this group was 31.6 ± 3.9 mm Hg (range, 20–64 mm Hg). Eight of the 11 subjects had a resting pulmonary arterial pressure greater than 25 mm Hg. However, all subjects had a pulmonary vascular resistance greater than 200 dyn/second/m5 (mean pulmonary vascular resistance, 465 ± 62 dyn/second/m5; range, 283–925 dyn/second/m5).
Right Heart Catheterization Studies
Twenty-five subjects were successfully traced at an interval of 10 years and underwent repeat catheterization. The results are shown in Table 1 . In 8 subjects found to have a resting pulmonary arterial pressure of less than 20 mm Hg during the first study and a normal response to acute hypoxia there was no evidence of an increase in pulmonary arterial pressure at the second study. In contrast, 10 subjects classified as hyperresponders at the first study on the basis of an exaggerated pulmonary pressor response to hypoxia showed evidence of progression in that resting mean pulmonary arterial pressure was higher at the second study. The percentage change in MPAP between the two studies was higher in the hyperresponders (35.4 ± 10.9%) compared with the normal responders (6.0 ± 5.2%) (p = 0.04). Furthermore, in 7 subjects with a resting mean pulmonary arterial pressure greater than 20 mm Hg there was a further increase in pulmonary arterial pressure (19.6 ± 9.4%) after 10 years. Although there was a trend for this group of subjects to be older than those in the other two groups, age was not significantly different between the groups (p = 0.11).
Postmortem Studies Microscopy of elastic van Gieson-stained lung sections from two of five highlanders revealed essentially normal pulmonary vascular morphometry in that peripheral pulmonary arteries at the level of respiratory bronchioles were almost devoid of a smooth muscle layer (Figure 2) . These two subjects had normal RV/LV+S ratios, with a mean of 0.25 ± 0.02. In contrast, in three cases increased muscularization was seen in arteries accompanying respiratory bronchioles, with distal muscularization extending to alveolar ducts and alveolar wall arterioles (Figure 2). The mean RV/LV+S ratio was increased at 0.62 ± 0.03 in these cases. There were no signs of significant emphysema in the lungs examined.
ACE Genotyping The results of ACE genotyping of 212 adult male Kyrgyz lowlanders (residents of Bishkek) attending hospital for unrelated conditions and of 78 male highlanders who underwent cardiac catheterization, and from whom DNA samples were available, are shown in Table 2 . Comparison of allele frequencies revealed a significant difference in the ACE allele frequency between these two populations ( 2 = 4.23, p = 0.04), with an increased frequency of the I allele in the lowland population.
ACE I/D allele and genotype frequencies for the group of 78 male highlanders who underwent right heart catheterization are also shown in Table 3 . Highlanders were classified as normal or as showing evidence of HAPH on the basis of the results of right heart catheterization. The HAPH phenotype included subjects with established pulmonary hypertension and hyper-responders to acute hypoxia. Comparison of allele and genotype frequencies revealed a threefold higher frequency of the I/I genotype in highlanders with HAPH, compared with normal highlanders ( 2 = 11.59, p = 0.003). The frequency of the ACE I allele was also significantly higher (2 = 12.01, p = 0.0005) in highlanders with HAPH. The odds ratio for HAPH among highlanders carrying the I allele was 3.2 (95% confidence interval, 1.6–6.3).
To look for further evidence of an association between HAPH and ACE genotype we plotted MPAP according to ACE genotype (Figure 3) . This demonstrated significantly higher MPAP in highlanders with the I/I genotype (26.9 ± 4.0 mm Hg) compared with the I/D genotype (20.6 ± 1.2 mm Hg) or D/D genotype (18.3 ± 0.9 mm Hg) (p < 0.05).
ACE Activity and Protein Levels Although it is well recognized that ACE I/D genotype partly determines the variation in serum ACE activity (16), these data were obtained mainly during the studies of white populations in Europe and North America. To confirm these findings in the population under investigation we assayed ACE activity and protein levels in the Kyrgyz lowland population described above, from which samples were available, and determined the correlation with ACE genotyping. The mean ACE activities were as follows: I/I, 24.5 ± 1.1 mU/ml/minute; I/D, 31.9 ± 1.1 mU/ml/minute; and D/D, 36.6 ± 2.7 mU/ml/minute (Figure 4A) . The percent variance in serum ACE activity that could be attributed to the ACE polymorphism in this group was 16.3%. The corresponding ACE protein levels were as follows: I/I, 390 ± 20 ng/ml; I/D, 534 ± 21 ng/ml; and DD, 664 ± 54 ng/ml (Figure 4B). The percent variance in serum ACE levels that could be attributed to the ACE polymorphism was 22.3% (22).
This study has determined that HAPH, as defined by the prevalence of ECG signs of right ventricular hypertrophy, in the absence of chronic airflow obstruction, is common among the highland population of Kyrgyzstan. Indeed, the frequency of ECG signs is likely to underestimate the true prevalence of pulmonary hypertension and right ventricular hypertrophy because the sensitivity of ECG in the diagnosis of right ventricular hypertrophy is as low as 20%, depending on the criterion used (17). The prevalence of ECG signs of right ventricular hypertrophy was higher in men than women. This observation may be accounted for by increased cigarette smoking among males, but this seems unlikely because none of the male subjects had evidence of airflow obstruction. There are probably other important factors such as sex hormone- and sex gene-determined effects, as female rats also develop less severe hypoxia-induced pulmonary hypertension than males (23). Our measurements of pulmonary hemodynamics in highlanders are likely to overestimate the true prevalence of HAPH in highlanders because we selected subjects with exertional dyspnea. We consider it difficult to justify performing right heart studies in asymptomatic individuals. We found that highlanders could be divided into three groups on the basis of the results of hemodynamic studies: normal, hyperresponsive to acute hypoxia, and those with established pulmonary hypertension. Repeat right heart catheterization studies performed after an interval of 10 years demonstrated increases in pulmonary arterial pressure in those with pre-existing elevation of pulmonary vascular resistance. The rate of increase (0.46 mm Hg/year) was similar to that observed previously in patients with pulmonary hypertension secondary to chronic obstructive pulmonary disease (24). Interestingly, subjects who were hyperresponsive to acute hypoxia at the initial study were also more likely to demonstrate progression of pulmonary arterial pressure, suggesting excessive acute hypoxic pulmonary vasoconstriction as a risk factor for the development of pulmonary hypertension. This potentially important observation requires confirmation in a larger cohort. Differences in the cardiovascular response to hypoxia have been observed previously in humans. Large interindividual differences have been reported in the magnitude of the pulmonary artery pressor response to acute hypoxia (25, 26). In addition, the degree of pulmonary hypertension measured at right heart catheterization varied widely in one community living at 3,100 m above sea level in the Colorado Rockies (12). Because a hyper-responsive pulmonary circulation would appear to be a risk factor for the development of established pulmonary hypertension, we included this group in our definition of the HAPH phenotype for subsequent genetic association studies. Our limited postmortem studies support the possibility that interindividual differences in the response to chronic hypoxia may include the extent of pulmonary vascular remodeling, in addition to hemodynamic parameters. Interestingly, Kyrgyz highlanders have a higher mean pulmonary arterial pressure and higher prevalence of HAPH compared with Tibetans residing at 3,658 m above sea level, whose pulmonary arterial pressure is comparable to that of lowlanders (13). This variation in the magnitude of HAPH between different populations has been noted previously. For example, the Tibetans exhibit less pulmonary hypertension than the inhabitants of the Peruvian Andes (13), who in turn are less susceptible than recent immigrants to high altitude, such as white Americans in Peru (25), and the Han Chinese in Tibet (27, 28). Interestingly, the least susceptible population, the Tibetans, have lived at altitude for many thousands of years, making it genetically advantageous to exhibit a blunted cardiovascular response to chronic hypoxia (13). Conversely, the Kyrgyz migrated to the Tien-Shan and Pamir Mountains from lower altitudes relatively recently, in about the 9th and 10th centuries AD, with a shorter period of time for selection against HAPH. The renin–angiotensin system appears to play an important role in the pathogenesis of hypoxia-induced pulmonary hypertension in the rat model. For example, ACE expression is increased in small peripheral pulmonary arteries in the chronically hypoxic rat (21), and either ACE inhibitors or angiotensin II receptor antagonists have been shown to inhibit the development of hypoxia-induced pulmonary hypertension and pulmonary vascular remodeling (29–32). In addition, mice deficient in the ACE gene demonstrate less hypoxia-induced pulmonary vascular remodeling compared with wild-type littermates (33). Because the expression of ACE is partly genetically determined, with the D allele associated with higher blood ACE levels and activity (16), we hypothesized that the D allele and DD genotype would be associated with increased susceptibility to HAPH in a highland population. In fact, our results show that it is the ACE I/I genotype, associated with reduced serum and tissue ACE activity, that associates with HAPH. We also found that the I allele frequencies differed significantly between Kyrgyz highlanders and lowlanders, the I allele being more frequent in lowlanders. Taken together with the association of the I allele with HAPH found in the highland population, it is possible that there may be a selection pressure against the I allele in highlanders because of the increased susceptibility to HAPH. However, this requires confirmation, because our "lowland control" group may not be entirely representative of the Bishkek male population, as they were recruited from patients attending hospital, albeit for unrelated conditions. We went on to confirm that ACE genotype is related to ACE levels and activity in the Kyrgyz population, because most of the evidence linking ACE activity to ACE I/D genotype has been obtained in the studies of white populations. In addition, we studied the distribution of ACE genotypes and allele frequencies in a lowland Kyrgyz population. Interestingly, the frequency of the II genotype was 3.5-fold higher than that of the DD genotype in this population. It is known that allele frequencies of the ACE I/D polymorphism vary between different ethnic groups. The allele frequencies in our lowland Kyrgyz population are similar to those found in the Japanese (I, 0.63; D, 0.37) (34) and Asians in the UK (I, 0.61; D, 0.39) (35). Frequencies in the white population in the UK are 0.43 (I) and 0.57 (D) (35). The frequency of the I/I genotype is therefore much higher in the Kyrgyz population (42.4%) compared with the white population (15–21%), but similar to that found in other Asian populations. We found that the variance in plasma ACE activity that could be attributed to the ACE I/D polymorphism was 16.3% in this study, compared with 28% in a white population (22). One explanation for this difference is that plasma ACE activity in the Kyrgyz population may be under greater control from other polymorphisms in the ACE gene, or other regulatory genes. Because our data suggest that the I/I genotype is associated with HAPH in the Kyrgyz population and that the ACE I/D genotype, at least partly, influences ACE activity and levels in the same way that was described previously, is there a plausible biological explanation for the association of lower ACE levels with HAPH? Most of the available evidence supports a role for angiotensin II in the development of hypoxia-induced pulmonary hypertension (21, 29, 33, 36). In addition, decreased availability of bradykinin as a consequence of increased ACE activity should also favor the development of pulmonary hypertension by decreasing production of nitric oxide from endothelial cells (37). However, we have previously demonstrated that changes in serum ACE activity do not necessarily reflect changes in local tissue. For example, in rats exposed to chronic hypoxia serum ACE is reduced by 50%, whereas ACE expression is increased in small pulmonary arteries undergoing remodeling (21). Therefore the relationship between ACE I/D genotype and lung, rather than serum, ACE activity, and changes in local expression of lung ACE in hypoxia, deserve further study (38). A further potential explanation for the association of I/I genotypes with HAPH is that subjects with the I/I genotype may have enhanced endurance performance, as described in elite athletes (39, 40). Thus it is conceivable that highlanders with the I/I genotype are capable of a more sustained level of exercise at altitude. An increased level of exercise would lead to an elevation of cardiac output and pulmonary artery pressure, exacerbating any tendency to HAPH. It is well known that exercise at high altitude does further increase the pulmonary arterial pressure (41). This hypothesis could be addressed by studies of the exercise capacity of highlanders with different ACE genotypes at altitude. Finally, it remains possible that the ACE I/I genotype is in linkage disequilibrium with another important disease susceptibility locus, and represents a genetic marker for HAPH susceptibility, independent of its effects on ACE activity. Our data are broadly consistent with a study demonstrating that the ACE D/D genotype is infrequent among male patients with ECG evidence of right ventricular hypertrophy secondary to severe chronic obstructive pulmonary disease (42). In that study the prevalence of right ventricular hypertrophy determined by ECG criteria in I/I and I/D genotypes was 46.5 versus 21.1% in D/D homozygotes. In contrast, a study of patients with primary pulmonary hypertension found that the ACE D/D genotype occurred more frequently in these patients compared with control subjects (43). In this group of patients with primary pulmonary hypertension, the DD genotype seemed to be protective against right ventricular failure, perhaps allowing a greater degree of right ventricular hypertrophy in the face of elevated pulmonary artery pressure. Sex differences in disease predisposition may also partly explain the differences between the two forms of pulmonary hypertension, because in our study HAPH was more common in men, whereas primary pulmonary hypertension was more common in females. Clearly, different mechanisms may be operating in hypoxia-induced versus primary pulmonary hypertension.
Supported by an International Grant from the Royal Society and by an International Collaborative Research Grant from the Wellcome Trust. Received in original form April 18, 2002; accepted in final form August 23, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
