Abstract
The question addressed by the study Chronic exposure to hypoxia increases pulmonary artery pressure (PAP) in highlanders, but the criteria for diagnosis of high-altitude pulmonary hypertension (HAPH) are debated. We assessed cardiac function and PAP in highlanders at 3250 m and explored HAPH prevalence using different definitions.
Patients and methods Central Asian highlanders free of overt cardiorespiratory disease, permanently living at 2500–3500 m compared to age-matched lowlanders living <800 m. Participants underwent echocardiography close to their altitude of residence (at 3250 m versus 760 m).
Results 173 participants (97 highlanders, 76 lowlanders), mean±sd age 49±9 years (49% females) completed the study. Results in lowlanders versus highlanders were systolic PAP (23±5 versus 30±10 mmHg), right ventricular fractional area change (42±6% versus 39±8%), tricuspid annular plane systolic excursion (2.1±0.3 versus 2.0±0.3 cm), right atrial volume index (20±6 versus 23±8 mL·m−2), left ventricular ejection fraction (62±4% versus 57±5%) and stroke volume (64±10 versus 57±11 mL); all between-group comparisons p<0.05. Depending on criteria, HAPH prevalence varied between 6% and 35%.
The answer to the question Chronic exposure to hypoxia in highlanders is associated with higher PAP and slight alterations in right and left heart function compared to lowlanders. The prevalence of HAPH in this large highlander cohort varies between 6% according to expert consensus definition of chronic high-altitude disease to 35% according to the most recent definition of pulmonary hypertension proposed for lowlanders.
Abstract
This study of Central Asian highlanders living at altitudes >2500 m revealed slight alterations of right- and left-heart functions compared to lowlanders. Pulmonary hypertension prevalence depended on diagnostic criteria and was >6%. https://bit.ly/2zdpPb5
Introduction
Worldwide, >140 million people live at high altitude (>2500 m). In Kyrgyzstan, a country where mountains cover >90% of the territory, >200 000 people live above 2500 m [1]. Kyrgyz highlanders have only been exposed to high altitude for a relatively short period of time, as migration to the Tien-Shan and Pamir mountains can be dated to around the 9–10th centuries AD. Therefore, they seem to be less adapted to high altitude than Tibetan highlanders, for example [1]. Chronic altitude-related illnesses, such as high-altitude pulmonary hypertension (HAPH) and chronic mountain sickness (a condition associated with excessive erythrocytosis) have been estimated to affect 5–10% of the exposed population, but robust prevalence estimates are not available [2, 3]. HAPH and its deterioration to chronic right heart failure was first described in 1985 [1, 4–6]. Reported manifestations include dyspnoea, exercise intolerance, right heart failure and eventually premature death [5]. Responsible factors for variability in adaptive responses and the development of HAPH might be genetic, environmental or due to comorbidities. In Kyrgyz highlanders, we found an association of HAPH with sleep apnoea, reduced cognitive performance, exercise capacity and quality of life [7]. This emphasises the need to investigate HAPH in order to understand underlying aetiopathogenic mechanisms.
A meta-analysis assessing pulmonary artery pressure (PAP) in high altitude dwellers living between 3600 and 4350 m showed a significantly higher systolic PAP in comparison to lowlanders [8]. However, most studies included in the meta-analysis were conducted in native Andean highlanders living at an altitude >3500 m; compared echocardiographic parameters with mainly European lowlanders; and some of the studies included patients with chronic mountain sickness [8, 9]. Data from highlanders living between 2500 m and 3600 m of Asian origin are scarce.
Kojonazarov et al. [4] have shown that echocardiography can be used as a valid screening tool for pulmonary hypertension in highlanders when compared to right heart catheterisation. However, little is known about the differential presentation of left and right heart function and dimension in Kyrgyz highlanders (living 2500–3600 m) versus healthy Kyrgyz lowlanders.
Thus, the objective of this study was to assess the PAP and right and left heart function by echocardiography in Kyrgyz lowlanders and highlanders. In addition, we explored the prevalence of HAPH according to different proposed definitions [2, 10, 11]. We hypothesised that systolic PAP is higher and right ventricular systolic and diastolic function are lower in highlanders compared to lowlanders.
Methods
This study was conducted in the National Center for Cardiology and Internal Medicine, Bishkek (760 m) and in the Ak-Say region of the Tien Shan mountain range in Kyrgyzstan at an altitude of 3250 m from July 2017 to August 2018. Healthy lowlanders and highlanders of similar age were invited to participate in this study. The highlanders live as nomads in a huge altitude plane with an elevation of 2500–3600 m, most of them between 3000 and 3600 m. An exact constant altitude of residence can therefore not be determined. Patients were recruited from an existing cohort study investigating the progression of HAPH with age (registered at clinicaltrials.gov NCT03165656).
Participants
Lowlanders with Kyrgyz ethnicity (born, raised and living <800 m) and Kyrgyz highlanders (born, raised and currently living >2500 m) of both sexes, aged ≥18 years, were recruited among outpatients of the National Center for Cardiology and Internal Medicine in Bishkek and in the Ak-Say region, respectively. Highlanders were excluded if they had excessive erythrocytosis as an indicator of chronic mountain sickness (defined as haemoglobin >19 g·dL−1 in females and >21 g·dL−1 in males) or other relevant cardiopulmonary disease such as coronary heart disease, COPD or heavy smoking (>20 cigarettes·day−1). This study was conducted in accordance with the Declaration of Helsinki, approved by the ethics committee in Kyrgyzstan (01-8/433) and endorsed by the cantonal ethic review board Zurich (2017-00369). All participants gave written informed consent to participate in the study.
Assessments
Echocardiographic recordings were obtained with a real-time, phased array sector scanner (CX 50; Philips Respironics, Zofingen, Switzerland) with an integrated colour Doppler system and a transducer containing crystal sets for imaging (1–5 MHz) and for continuous-wave Doppler. Recording and analysis were performed according to guidelines of the European Association of Echocardiography [12]. Transtricuspid pressure gradient (TTPG) was calculated from maximal tricuspid regurgitation velocity (TR Vmax) obtained with continuous-wave Doppler using the modified Bernoulli equation: ΔP=4×Vmax2. Right atrial pressure (RAP) was estimated by the diameter of the inferior vena cava and its variation during inspiration. Systolic (s)PAP was calculated as TTPG+RAP. Areas of the right atrium and right ventricle (RV) were manually traced. Fractional area change of the RV was calculated (end-diastolic RV area ‒ end-systolic RV area/end-diastolic RV area). Right atrial volume index (RAVI) was calculated using the single plane area-length method from the apical four-chamber view at end systole divided by body surface area [13]. Tricuspid annular plane systolic excursion (TAPSE) was measured in M-mode. The tissue Doppler peak velocity of the RV free wall was assessed. Mean (m)PAP was calculated from sPAP as mPAP=0.61 sPAP+2 mmHg [14]. If sPAP was not available, mPAP was estimated from flow acceleration time measured with pulsed-wave Doppler in the RV outflow tract, as described by Kitabatake et al. [15] and validated against right-heart catheter measurements in Kyrgyz highlanders [4]. RV diastolic function was assessed by pulsed-wave Doppler measuring tricuspid inflow velocities, namely the E-wave (RVE) for early passive filling of the RV and A-wave (RVA) for late diastolic active filling of the RV as representation of the right atrial contraction. Tissue Doppler of the lateral tricuspid annulus was assessed in the four-chamber view for measuring peak velocities of systole (RVs′) and early diastole (RVe′) and late diastole during atrial contraction (RVa′). Cardiac output was estimated by the Doppler velocity time integral method from the left ventricular (LV) outflow tract [16]. The pulmonary artery wedge pressure was calculated using the Nagueh formula (pulmonary capillary wedge pressure (PCWP)=1.24×(E/e′)+1.9) [17] and pulmonary vascular resistance (PVR) was calculated as PVR=(mPAP−PCWP)/cardiac output. Values of PVR adjusted for haematocrit were determined according to the formula derived by the hindlimb vessel studies by Whittaker and Winton [18, 19]. To this end, PVR was computed for a standard haematocrit of 0.45 as (R0(45%)=R0(haematocrit) × ((1-φ1/3)/0.234))) [19]. RV-arterial coupling was estimated by TAPSE/sPAP [20]. LV parameters and function were measured according to guidelines [13].
Arterial blood gas analyses (RapidPoint 405; Siemens, Zurich, Switzerland) were assessed from the radial artery, after waking-up in the morning. Arterial oxygen content (CaO2) was calculated as CaO2 (mL O2·dL−1)=(1.34×haemoglobin concentration×SaO2)+(0.0031×PaO2) (SaO2: arterial oxygen saturation; PaO2: arterial oxygen tension). Oxygen delivery (mL·min−1) was calculated as CaO2×cardiac output or CaO2×cardiac index.
Respiratory sleep studies (AlicePDx; Philips AG Respironics) were performed including nightly oxygen saturation, indices of oxygenation and apnoea/hypopnoea index (mean number of apnoeas/hypopnoeas per hour of time in bed), as described previously [21].
Outcomes
Outcomes of the study were PAP estimated by transthoracic echocardiography and further parameters of right and left heart function and dimensions obtained during standard echocardiography. The prevalence of HAPH was explored according to various proposed PAP thresholds: the consensus statement on chronic and subacute high altitude disease [2] using a cut-off for sPAP >50 mmHg or mPAP >30 mmHg, the 2015 European Society of Cardiology/European Respiratory Society guidelines for the diagnosis and treatment of pulmonary hypertension with a cut-off for mPAP ≥25 mmHg [10] and the latest proposal by the 6th World Symposium on Pulmonary Hypertension with a cut-off for mPAP >20 mmHg and PVR ≥3 Wood Units [11].
Statistical analysis
Outcomes were analysed per protocol and values are presented as mean±sd and mean differences (95% CI) to account for normal distribution in the vast majority of the data. Missing values were not replaced. Differences between lowlanders and highlanders were compared by t-tests or Fisher's exact test for categorical data. Predictors for sPAP were assessed by univariate regression analyses. With a p-value <0.2 in univariate regression analysis, the predictor was included in a multivariable regression analysis using backward elimination. Analysis was performed using STATA 15 (StataCorp, College Station, TX, USA) and a p-value <0.05 or 95% confidence interval of differences not including zero were considered statistically significant.
Results
173 participants were included into the study: 76 lowlanders and 97 highlanders (figure 1). Baseline characteristics of the groups are displayed in table 1. Of interest, highlanders were smaller than lowlanders. As expected, PaO2 and SaO2 were significantly lower in highlanders compared to lowlanders. Highlanders had higher haemoglobin concentration, haematocrit, heart rate, diastolic and systolic blood pressure compared to lowlanders.
Study flow chart. GOLD: Global Initiative for Chronic Obstructive Lung Disease.
Baseline characteristics, including blood gases and lung function
The sPAP could be obtained in 64 out of 76 lowlanders and 83 out of 97 highlanders. Measurements were 23±5 mmHg in lowlanders and 30±10 mmHg in highlanders (p<0.001). mPAP was 17±4 mmHg in lowlanders and 22±6 mmHg in highlanders (p<0.001). The effect size (95% CI) of high-altitude residence for sPAP and mPAP were 0.85 (0.51–1.19) and 0.95 (0.64–1.27) compared to low-altitude residence. Parameters of RV systolic function (fractional area change, TAPSE and tricuspid annular systolic velocity) were lower in highlanders compared to lowlanders. RVE/A, as a surrogate of the diastolic RV function, was significantly lower, and RAVI, RAP, RV wall thickness and PVR were found to be higher in highlanders compared to and lowlanders (table 2, figure 2).
Right heart function measured with echocardiography
Mean differences (95% CI) of various echocardiographic and additional variables between highlanders and lowlanders. sPAP: systolic pulmonary artery pressure; mPAP: mean pulmonary artery pressure; CO: cardiac output; RAVI: right atrial volume index; TAPSE: tricuspid annular plane systolic excursion; FAC: fractional area change; RVE/A: right ventricular ratio of the early (E) to late (A) ventricular filling velocities; LVEF: left ventricular ejection fraction; LAVI: left atrial volume index; SpO2: oxygen saturation by pulse oximetry.
LV ejection fraction and stroke volume were significantly lower in highlanders compared to lowlanders, but cardiac index was preserved in highlanders compared to lowlanders, as shown in table 3. CaO2 was higher in highlanders compared to lowlanders, due to the increased haemoglobin, therefore oxygen delivery was higher in highlanders compared to lowlanders (table 3 and figure 2).
Left heart function measured with echocardiography
Regression analysis revealed that sPAP was associated with age and nocturnal oxygen saturation in a multivariable model (table 4).
Univariate and multivariable regression analysis with systolic pulmonary artery pressure as dependent variable
The prevalence of HAPH according to the different sPAP/mPAP criteria for pulmonary hypertension varied between 6% and 35%, as shown in table 5.
Prevalence of high-altitude pulmonary hypertension (HAPH) according to different criteria
Discussion
This is the largest investigation of PAP and heart function evaluated in detail using echocardiography in highlanders of both sexes in comparison to lowlanders of the same ethnicity and of similar age, sex and body mass index. Consistent with our hypothesis and previous observations [7, 8], sPAP was higher in highlanders compared to lowlanders and nocturnal oxygen saturation was a significant predictor of sPAP when controlled for age.
A meta-analysis showed that high-altitude residents, 70% of whom lived in the Andes, but also in Tibet, Nepal and in Ethiopia, at altitudes between 3600 and 4350 m, had a significantly higher sPAP compared to healthy, mainly European low-altitude control cohorts. Importantly, sPAP was taken as equivalent to TTPG by these authors not accounting for RAP [8]. Data on Central Asian high-altitude residents and from a comparably lower altitude range of 2500–3600 m are scarce. Thus, the current study extends the existing data by assessing a large Central Asian cohort of highlanders living between 2500 m and 3600 m with a mean age of 49 years, oxygen saturation of 91% and a TTPG of 26 mmHg measured at 3250 m. The comparison to (mostly Andean) highlanders with a mean age of 38 years assembled in the meta-analysis reveals a similar oxygen saturation (i.e. 92%) and mean TTPG of 25 mmHg at higher altitudes of 3650–4350 m [8]. Thus, the younger age but higher altitude of residence of participants in the meta-analysis compared to those of the current study may have had opposing influences on the TTPG, resulting in similar values in the two cohorts. Regression analysis revealed an independent association of sPAP with both age and nocturnal oxygen saturation by pulse oximetry (SpO2), confirming previous findings of an age-dependent increase in sPAP [22–24]. Thus, our regression analysis revealed that every decade of life would increase the sPAP by ∼1.5 mmHg.
A correlation of sPAP with nocturnal, but not daytime SpO2, as observed in the current study (table 4), has been observed in previous studies at altitude and might be explained by the increased PVR, related to a sleep-related reduction in ventilatory drive and associated reduction in ventilation and nocturnal SpO2, in particular in highlanders with sleep apnoea [7, 25]. Parameters of systolic right heart function, such as TAPSE and fractional area change, were only slightly lower in highlanders compared to lowlanders and may indicate a higher load to the RV due to the chronic elevation of the PAP. While the effect size of altitude residence was high for the PAP, the differences in indices of RV function were not pronounced. Similar to the present cohort, Peruvian highlanders with chronic mountain sickness, a condition different from HAPH and associated with excessive erythrocytosis, also revealed an enlarged RV end-diastolic area and a higher TTPG compared to nonaffected highlanders and lowlanders, but in contrast to our cohort these patients did not reveal an impaired RV systolic function [26]. In Bolivian patients with chronic mountain sickness who had lower resting RV function, RV contractile reserve during stress echocardiography was found to be similar to healthy highlanders and the authors interpreted these findings as a physiological adaption [27]. However, Bolivian lowlanders were not included in this study. In our cohort, right ventricular-arterial coupling (TAPSE/sPAP) was lower in highlanders than lowlanders, but still above the threshold of 0.31 mm·mmHg−1 that would indicate uncoupling. Whether those findings persist during exercise remains to be studied.
In the present study, hypoxaemia along with a higher heart rate, haematocrit and haemoglobin concentration in highlanders compared to lowlanders was associated with a slightly reduced stroke volume but preserved cardiac output in highlanders. Lower stroke volumes have been reported in Sherpas before, despite having higher plasma volumes [28]. Whether plasma volume plays a role in high-altitude acclimatisation in the Kyrgyz highlanders has not been studied yet and remains of interest. The oxygen delivery in highlanders was higher, possibly due to higher metabolic rates related to a higher sympathetic tone and higher diaphragmatic and cardiac muscular activity with higher breath and heart rates, although this has not been confirmed by metabolic measurements.
In our highlander cohort, we found mild significant, albeit most probably not clinically relevant impairments of systolic and diastolic left heart function. These differences might help us understand further acclimatisation mechanisms in the studied highland population. While hypoxia is a strong activator of the sympathetic nervous system in altitude-naïve lowlanders, baroreflex muscle sympathetic nerve activity and resting sympathetic vasomotor outflow are lower in acclimatised Sherpas [29]. In the Kyrgyz highlanders we found a higher systemic blood pressure and heart rate compared to lowlanders that may reflect chronic hypoxia-related stimulation of the sympathetic nervous system, which might be attributed to the fact that Kyrgyz highlanders have only migrated to high altitude relatively recently [28]. It is known from general population studies that the PAP and the proportion of heart dysfunction increases with age [22]. In altitude natives, a higher age not only represents a longer exposure time to hypoxia, but also longer acclimatisation. Whether the changes discussed above are signs of good or poor adaption remains to be clarified.
Depending on the criteria used, HAPH could be detected in 6–35% of highlanders (table 5). In our cohort, 12% fulfilled the criteria proposed by a consensus statement on high-altitude diseases, i.e. mPAP >30 mmHg. The prevalence increased to 27% and 35% if lower thresholds for mPAP of ≥25 mmHg or >20 mmHg in association with PVR ≥3 mmHg, respectively, were selected as recommended recently for assessment of pulmonary hypertension near sea level. In our opinion, the criteria proposed in the consensus paper 2005 by Léon-Velarde et al. [2] (mPAP >30 mmHg and sPAP >50 mmHg) are pragmatic in order to avoid overdiagnosing the disease in asymptomatic patients. However, as early diagnosis of pulmonary hypertension is of importance in symptomatic patients, we suggest the application of more-recent recommendations and take into account several echocardiographic markers such as RV/LV diameter, LV eccentricity index >1.1, right atrial area or vena cava diameter, among others, especially if mPAP is >20 mmHg [11].
To compare our cohort to data from the previous meta-analysis and another, more recent meta-analysis in Andean patients with chronic mountain sickness, we derived estimates of mPAP from the reported mean TTPG of 25 [8] and 28 mmHg [9], respectively, by adding an assumed RAP of 5 mmHg [30] and converting to mPAP according to (mPAP=(sPAP×0.61)+2) [14]. This revealed estimated mean mPAP of 20 and 22 mmHg for the meta-analysed healthy highlanders [8] and patients with chronic mountain sickness [9], respectively. The reason for the discrepancy between studies in Andean, Tibetan, Nepalese and Ethiopian highlanders reported previously and the current study remains elusive. As in the studies included in the meta-analysis, we cannot exclude a potential selection bias in our study, as highlanders were included upon self-presentation, which may have encouraged individuals with concerns about their health to undergo examinations. Differences in susceptibility to HAPH between ethnicities is another explanation for the high prevalence of HAPH in Kyrgyz highlanders since genetic factors have been shown to be relevant in the pulmonary vascular response to hypoxia [31] and specific genetic variants associated with HAPH have been described in Kyrgyz highlanders previously [32].
Right heart catheterisation is the gold standard to diagnose pulmonary hypertension, but such invasive measurements were not feasible and not considered ethical in the remote high-altitude setting of the current study. Therefore, we employed echocardiography, which is well established as noninvasive measure to estimate PAP in groups of individuals. Comparisons to right heart catheter-derived mPAP in Kyrgyz highlanders have demonstrated satisfactory accuracy of mPAP derived by echocardiography [4] and other studies have validated estimates of sPAP based on trans-tricuspid flow velocity [30]. The investigators performing echocardiography were highly experienced and individual measurements were reviewed by two independent investigators to minimise bias.
Conclusion
We studied a large cohort of Central Asian highlanders at an altitude of 3250 m in comparison to lowlanders. Our data confirm that chronic exposure to the hypoxic environment at altitude caused a significantly lower blood oxygenation along with an adaptive higher in heart rate, haematocrit and haemoglobin in highlanders compared to lowlanders. We extend previous findings on an increased PAP in highlanders living between 3600 and 5050 m [8] by showing for the first time that an altitude of residence between 2500 and 3500 m is associated with a higher PAP and minor impairments of right and left heart function compared to lowlanders. HAPH was suspected in 6–12% of highlanders when applying the recommended criteria of high-altitude experts [2], with the prevalence rising to 35% when applying recent criteria proposed for lowlanders [11]. Longitudinal observations are warranted to determine the value of echocardiographic parameters as risk indicators for high-altitude diseases, and to learn about the natural course, further adaptive mechanisms, clinical manifestations and possible treatments of this disease.
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-02474-2019.Shareable
Footnotes
This study was registered at ClinicalTrials.gov as NCT03165656.
Author contributions: M. Lichtblau, S. Saxer, S. Ulrich: data acquisition, analysing and interpretation, drafting of the manuscript, final approval of the version to be published. M. Furian, L. Mayer, P.R. Bader, P.M. Scheiwiller, M. Mademilov, U. Sheraliev, F.C. Tanner, T.M. Sooronbaev: data acquisition, critical revision for important intellectual content, final approval of the version to be published. M. Lichtblau, S. Saxer, K.E. Bloch, S. Ulrich: substantial contributions to the conception and design of the study, interpretation of the data, critical revision for important intellectual content, final approval of the version to be published. S. Ulrich is the guarantor of the paper.
Conflict of interest: M. Lichtblau has nothing to disclose. Conflict of interest: S. Saxer has nothing to disclose.
Conflict of interest: M. Furian has nothing to disclose.
Conflict of interest: L. Mayer has nothing to disclose.
Conflict of interest: P.R. Bader has nothing to disclose.
Conflict of interest: P.M. Scheiwiller has nothing to disclose.
Conflict of interest: M. Mademilov has nothing to disclose.
Conflict of interest: U. Sheraliev has nothing to disclose.
Conflict of interest: F.C. Tanner has nothing to disclose.
Conflict of interest: T.M. Sooronbaev has nothing to disclose.
Conflict of interest: K.E. Bloch reports grants from Zurich Lung League and Swiss National Science Foundation, during the conduct of the study.
Conflict of interest: S. Ulrich reports grants from Zurich Lung League and Swiss National Science Foundation, during the conduct of the study; grants and personal fees from Actelion SA and Orpha Swiss, personal fees from Bayer SA and MSD, outside the submitted work.
Support statement: The study was funded by grants from the Swiss National Science Foundation and the Zurich Lung. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received January 7, 2020.
- Accepted April 17, 2020.
- Copyright ©ERS 2020
References
