Change in pulmonary artery pressure upon acute exposure to high altitude: a systematic review and meta-analysis

Abstract

Background/research question

We investigated the increase in pulmonary artery pressure (PAP) with altitude gain in healthy low-altitude residents travelling to high altitude, analysing reported measures including systolic (sPAP) and mean pulmonary artery pressure (mPAP) and tricuspid regurgitation pressure gradient (TRPG).

Methods

A systematic literature search was performed in PubMed and Embase from database inception to 12 December 2024. Studies including healthy adults residing at <1000 m and reporting sPAP, mPAP or TRPG measured by echocardiography or right heart catheterisation at both low altitude and >1000 m within 30 days were included. Random effects meta-analyses and meta-regression assessed the increase in sPAP, mPAP and TRPG per 1000 m altitude gain.

Results

60 articles with 1220 participants (75% males, age 34±4 years) were eligible; 54 were included into the meta-analysis. sPAP was reported in 38 studies, mPAP in nine and TRPG in 11. At high altitude, absolute sPAP ranged from 24 to 50 mmHg and mPAP from 20 to 26 mmHg. Meta-regression showed a linear increase per 1000 m: sPAP 3.4±0.5 mmHg, mPAP 2.5±0.7 mmHg and TRPG 3.3±1.0 mmHg. The estimated upper limit of normal for the increase per 1000 m was 4.5 mmHg for all three measures.

Interpretation

This systematic review and meta-analysis of healthy low-altitude residents travelling to high altitudes demonstrates a linear increase in PAP with altitude gain, based predominantly on data from elevations >2500 m and younger male populations. It defines upper reference limits for the increase in sPAP, mPAP and TRPG per 1000 m. These population-derived reference ranges may assist in the interpretation of PAP measurements and in counselling individuals evaluated at high altitude, but should be applied with consideration of the underlying population and exposure characteristics.

Shareable abstract

In healthy low-altitude residents travelling to high altitudes, the pulmonary arterial pressure increases up to 4.5 mmHg per 1000 m altitude gained. These findings provide reference values to support clinician decision-making at high altitude. https://bit.ly/4cZcLW2

Introduction

With increasing global travel, exposure to high altitude has become more common, highlighting the need to better understand its effects on pulmonary haemodynamics [1]. At high altitude, reduced barometric pressure lowers inspired oxygen tension and rapidly induces hypoxic pulmonary vasoconstriction (HPV), which increases mean pulmonary artery pressure (mPAP) [2, 3]. During early exposure, sympathetic activity leads to an increase in heart rate and cardiac output (CO), further contributing to elevations in mPAP [1]. With ongoing exposure over hours to days, plasma volume decreases, due to fluid redistribution, causing stroke volume to fall and haemoglobin concentration to rise [4, 5]. Over longer periods, erythropoietin-mediated increases in red cell mass and sustained hyperventilation contribute to maintaining oxygen delivery, while also affecting blood viscosity and pulmonary haemodynamics [1]. These are physiological responses to improve ventilation–perfusion ratio and to maintain oxygen delivery to vital organs. However, extensive responses to high altitude might be associated with a greater risk for severe high-altitude illness such as acute mountain sickness or high-altitude pulmonary oedema (HAPE) [6, 7]. According to international guidelines, pulmonary hypertension is defined as mPAP >20 mmHg assessed by right heart catheterisation [8]. This definition is derived mainly from individuals living and being investigated <1000 m. However, whether this threshold is also valid for altitude newcomers who live at low altitude and travel to high altitude and whether a higher mPAP in these newcomers has clinical impact remains unknown [9].

Although increases in mPAP with altitude have been consistently demonstrated, reference values for healthy lowlanders during acute high-altitude exposure have not been established [3, 10]. Defining such values is essential to distinguish physiological adaptation from pathological responses.

The aim of this systematic review and meta-analysis was to quantify the altitude-dependent increase in mPAP (per 1000 m) in healthy lowlanders during acute exposure to altitudes >1000 m for up to 30 days.

Methods

This systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines [11]. The study was registered at PROSPERO (CRD42025638291).

Literature search

PubMed and Embase were systematically screened for peer-reviewed articles from database inceptions to 12 December 2024. Search terms included “pulmonary artery pressure” and “altitude” (a full list of search terms is provided in the supplementary material). We screened the results in two stages using the software Covidence. Titles and abstracts were independently screened by two investigators (E.C. Wagner and D. Hertig) and full texts were obtained if they met the inclusion criteria. Full-text screening was conducted by two investigators independently based on the eligibility criteria.

Data extraction

Three investigators (E.C. Wagner, J. Müller and D. Hertig) extracted data from the included studies using a standardised form, one investigator for each study. If several articles reported data for the same cohort, the article with the most complete data was chosen.

Study selection criteria

Studies were included if pulmonary artery pressure (PAP) was measured at baseline at low altitude (<1000 m) and at high altitude (>1000 m). Exposure duration of <30 days was mandatory and studies had to include healthy adults (>18 years) residing at low altitude without relevant cardiopulmonary disease.

For inclusion, PAP had to be assessed by right heart catheterisation or by transthoracic echocardiography. Echocardiographic measurements were required to be based on peak tricuspid regurgitation velocity (TRV) to derive either the tricuspid regurgitation pressure gradient (TRPG) or systolic pulmonary artery pressure (sPAP), or mPAP using the Chemla equation [12, 13].

Studies assessing mPAP by pulmonary artery acceleration time were excluded, as were experiments conducted under normobaric hypoxia or drug trials without a placebo group. Only studies with to full-text availability in English, German, French or Spanish were considered. All study designs were eligible provided they represent original research and met the inclusion criteria. We extracted demographic data including age, sex, height, weight, body mass index (BMI) and duration of exposure.

Outcomes

The primary outcomes were the differences in sPAP, mPAP and TRPG between low altitude and at high altitude.

Risk of bias assessment

Three researchers (E.C. Wagner, A. Vella and D. Hertig) assessed independently the risk of bias (RoB) using the Quality Assessment Tool from National Heart, Lung and Blood Institute [14]. The tool is composed of 14 individual questions, each answered with low risk, high risk or unclear. The studies were classified as high RoB if <5/14 questions answered with low risk, intermediate RoB if 5–10/14 questions answered with low risk and low RoB if 11–14/14 questions answered with low risk. If the exposure time was unclear or the measurement of right atrial pressure (RAP) included in estimated sPAP was unclear, the study was rated as “high risk of bias” regardless of how many questions were answered with low risk. The detailed description and RoB assessment for each included study can be found in table S2.

Statistical analysis

We performed random-effects meta-analyses and meta-regression analyses to assess the increase in sPAP, mPAP and TRPG per 1000 m of altitude gain. Altitude differences were calculated by altitude reached minus baseline altitude. The upper limit of normal for the increase per 1000 m was defined using the upper bound of the 95% confidence interval of the meta-regression estimate. This approach was chosen as a conservative estimate of the population-average effect, while prediction intervals were not used for this purpose due to substantial residual between-study heterogeneity not explained by altitude, which would result in overly wide and clinically noninformative ranges. In studies reporting repeated measures within the same participants at the same altitude, values were pooled to obtain a single estimate per study and altitude level. For studies with repeated measurements at different altitudes, each altitude level was included as a separate data point. Studies were weighted according to sample size and the precision of the reported estimates. Standard errors were converted to standard deviation for the meta-analysis. Heterogeneity was assessed using the Cochran Q test, the Higgins I² statistic and between-study variance (τ²). Given the anticipated clinical and methodological heterogeneity across studies, including differences in altitude exposure, timing of assessment and measurement techniques, a random-effects model was applied to account for both within- and between-study variability. Residual heterogeneity was further explored using meta-regression analyses. To assess the functional form of the association between altitude difference and effect size, altitude difference was first modelled as a continuous moderator in a random-effects meta-regression. Linearity of the association was evaluated by comparing this model to nonlinear alternatives. Model performance was compared using the Akaike information criterion (AIC) and the proportion of between-study heterogeneity by each model was quantified using the R² statistic. Sensitivity analyses were performed according to measurement modality (right heart catheterisation versus echocardiography) and study quality (high RoB studies excluded) to assess the robustness of the findings. Reporting bias was assessed by visual inspection of funnel plots for each outcome (sPAP, mPAP, TRPG) and by the Egger regression test for funnel plot asymmetry, where sufficient studies were available. Analyses were done using R-Studio (version 2024.12.1+563).

Results

Study selection

The study flow chart is shown in figure 1. A total of 880 eligible studies were identified in Embase and 666 in PubMed. After removing the duplicates, the remaining 1112 studies were screened by title and abstract, which excluded 815 studies. Full-text screening of the remaining 297 studies excluded further 237 studies. In total, 60 original studies fulfilling the inclusion criteria were included.

FIGURE 1.

Study flow chart. mPAP: mean pulmonary arterial pressure; PAP: pulmonary arterial pressure; sPAP: systolic pulmonary arterial pressure; TRPG: tricuspid regurgitation pressure gradient.

43 studies reported sPAP, representing the largest group of the meta-analysis. The remaining five studies had to be excluded from the meta-analysis because they either did not report sd or se [15, 16] or reported median values of sPAP and quartiles [17–19].

12 studies [20–31] reported TRPG; among those, one also reported sPAP and mPAP and was included in the TRPG, mPAP and sPAP meta-analyses. One study reported median TRPG and could not be included in the meta-analysis [23].

Nine studies [21, 32–39] reported mPAP; among those, three studies [21, 33, 34] also reported sPAP and were included in both the mPAP and sPAP meta-analyses.

Overall, 54 of the 60 of the studies were included in the meta-analysis of either sPAP, mPAP or TRPG.

Systematic literature review

In total, 60 studies were identified, including 1220 participants. Six studies did not report sex. In 54 studies reporting sex, 75% were males. The mean±sd age was 33.8±4.4 years. The mean BMI of the participants was 23.3±1.3 kg·m⁻².

The range in baseline altitude was between 0 m and 880 m, with a mean baseline altitude of 305 m. The range in reported higher altitudes was between 1336 m and 7050 m in the mountains and up to 8000 m in hypobaric chamber studies. The average high altitude was 4170 m and the median high altitude was 4350 m.

12 studies were performed in Italy at 4559 m [16, 20, 26, 37, 40–47], eight in Nepal [17, 25, 30, 39, 48–51], five in Kyrgyzstan [21–23, 35, 52], five in the USA [36, 53–56], five in France [19, 24, 28, 31, 57], five in Switzerland [34, 58–61], three in Bolivia [29, 62, 63], three in Peru [27, 64, 65], two in China [18, 66], two in the UK [67, 68] and one each in Ecuador [69], Tanzania [33], Tibet [15], India [70], Germany [71], Spain [32] and Japan [72]. One study was conducted across multiple sites in Ecuador, Bolivia, Peru and Nepal [38]. Two studies did not report the country in which measurements were performed [73, 74]. Four studies (one in France, two in the UK and one with no reported country) were conducted in a hypobaric chamber [24, 67, 68, 74].

All studies showed an increase in sPAP, mPAP or TRPG with exposure to higher altitudes. mPAP was measured with a right heart catheter in three studies [34, 36, 37], while echocardiography was used in all others.

The mean time to measurement after arrival at target altitude was 3 days. Most studies (n=54) reported measurements within <5 days of arrival, including 25 studies within <24 h and eight studies reporting measurements at multiple time-points.

13 studies reported measurements at ≥5 days, of which nine included both early (<5 days) and later (≥5 days) time-points. Two studies did not report the exact timing of measurements; however, it was evident from the text that assessments were performed during acute exposure (<30 days).

The detailed characteristics for each study are described in the tables S3 and S4.

Risk of bias assessment

The RoB assessment was performed for all 60 studies. 17 studies were rated as “low risk RoB”, 24 as “intermediate RoB” and 19 as “high RoB”. The detailed RoB assessment for each included study can be found in table S2. Results of the Egger test indicated significant funnel plot asymmetry for sPAP (p<0.001), whereas no significant asymmetry was observed for mPAP (p=0.49) and TRPG (p=0.10) (see figure S6).

Linearity of the association

Meta-regression demonstrated a strong positive association between altitude difference and effect size when modelled as a continuous variable (β=3.38 per km, se=0.47, p<0.001). When altitude difference was instead modelled categorically, the resulting pattern of estimates did not indicate a consistent nonlinear relationship across altitude bands. Model comparison favoured the linear specification, with a lower AIC compared to the categorical model (AIC 296.5 versus 303.6; ΔAIC=7.1), indicating substantially better fit. Furthermore, the linear model explained a greater proportion of between-study heterogeneity (R²=60.6%) than the categorical model (R²=43.0%). Taken together, these findings support the use of altitude difference as a continuous linear moderator, with the categorical approach resulting in reduced explanatory power and poorer model fit.

Random-effects model meta-analysis for sPAP

Among 54 studies included in the meta-analysis in total, 38 reported the sPAP. The analysis included 691 participants. Considering high heterogeneity, a random-effects model was chosen. The first preliminary analysis of mean difference in sPAP at low versus high altitude is shown in the forest plot in figure 2 and shows that sPAP significantly increases with altitude. In figure 2 the forest plot including the 38 studies per 1000 m altitude gain shows an overall increase in sPAP of 2.66 mmHg per 1000 m (95% CI 2.34–2.98 mmHg). Absolute sPAP ranged from 24 to 50 mmHg. Figure S3 shows the forest plot of a meta-analysis for sPAP per 1000 m of altitude, stratified by exposure days. The χ² test for subgroup differences indicated no statistically significant difference between exposures ≥5 days and <5 days. The studies reporting sPAP <5 days of exposure to high altitude showed an sPAP increase of 2.77 mmHg (95% CI 2.40–3.13 mmHg) per 1000 m, while ≥5 days of exposure time sPAP increased by 2.66 mmHg (95% CI 2.34–2.98 mmHg) per 1000 m.

FIGURE 2.

Forest plot of the random-effects meta-analyses for the increase in systolic pulmonary arterial pressure per 1000 m of altitude gained.

Meta-regression demonstrated a linear increase in sPAP with altitude, with an estimated increase of 3.4 mmHg per 1000 m (95% CI 2.4–4.3 mmHg) (see figure 3a). The upper bound of the confidence interval (4.3 mmHg) was therefore used as a conservative estimate of the upper limit of normal. The corresponding prediction interval was wide, reflecting substantial residual heterogeneity across studies. Figure 3a also shows whether participants were exposed to hypoxia in a hypobaric chamber or in the mountains. Plotting the absolute values confirmed assumable linearity of sPAP with increasing altitude (see figure 3b).

FIGURE 3.

a–e) Results of meta-regressions of the increase in pulmonary arterial pressure per change in altitude. The size of the dots represents the sample size and precision of each study. a) Increase in systolic pulmonary arterial pressure (sPAP), stratified by whether hypobaric hypoxia was applied in a hypobaric chamber or in the mountains. c) Increase in mean pulmonary arterial pressure (mPAP), stratified by measurement method. e) Increase in tricuspid regurgitation pressure gradient (TRPG), stratified by whether hypobaric hypoxia was applied in a hypobaric chamber or in the mountains. b), d) and f) represent absolute values for the pulmonary arterial pressure (PAP) at absolute altitudes, with baseline measurements shown in purple and altitude measurements shown in yellow. Echo: echocardiogram; RHC: right heart catheterisation. ^#: Increase of PAP per 1000 m.

Random-effects model meta-analysis for mPAP

The meta-analysis for mPAP included data from nine studies with 215 participants. The forest plot in figure S1 shows an increase in mPAP with high altitude. Per 1000 m altitude gain, mPAP increases by 1.98 mmHg (95% CI 1.60–2.35 mmHg) per 1000 m. Absolute mPAP ranged from 20 to 26 mmHg. The meta-regression indicated an increase in mPAP of 2.5±0.7 mmHg per 1000 m (figure 3c). Figure 3c also shows whether mPAP was assessed by echocardiography or by right heart catheterisation. The upper limit of normal was 3.8 mmHg.

Random-effects model meta-analysis for TRPG

The meta-analysis for TRPG included the data from 11 studies with 183 participants. The forest plot in figure S2 shows an increase in TRPG with high altitude. Per 1000 m altitude gain, TRPG increases by 2.86 mmHg (95% CI 2.23–3.49 mmHg) per 1000 m. Absolute TRPG ranged from 24 to 40 mmHg. The meta-regression indicated an increase in TRPG of 3.3±1.0 mmHg per 1000 m, in figure 3e. Figure 3e also shows whether participants were exposed to hypoxia in a hypobaric chamber or in the mountains. The upper limit of normal was 5.3 mmHg.

Heterogeneity

Meta-regression demonstrated a significant linear increase in sPAP, mPAP and TRPG with altitude (QM p<0.0001), explaining 61% of the between-study heterogeneity. Despite this, substantial residual heterogeneity remained (τ²=9.33; I²=86.1%; QE p<0.0001), indicating that variability beyond altitude, such as differences in measurement methods, timing of assessment and study populations, contributed to the observed effects.

Sensitivity analyses

Sensitivity analysis by measurement modality showed that the increase in mPAP per 1000 m was consistent across methods, with no statistically significant difference between right heart catheterisation and echocardiography (test for subgroup differences, p=0.073), supporting the robustness of the findings (figure 4).

FIGURE 4.

Results of the sensitivity analysis for the increase in mean pulmonary arterial pressure (mPAP) per 1000 m altitude gained, stratified by measurement method. Echo: echocardiography; RHC: right heart catheterisation.

Sensitivity analyses excluding studies rated as high RoB confirmed the robustness of the findings for both sPAP and mPAP (figures 2 and S4; figures S1 and S5). The estimated increase per 1000 m remained consistent, with no meaningful change in effect size. Heterogeneity estimates were comparable to the primary analyses, suggesting that the results were not driven by lower-quality studies.

Discussion

This systematic review with meta-analysis of acute changes in PAP in healthy low-altitude residents exposed to hypobaric hypoxia (1336–5200 m) demonstrates an approximately linear increase in pressure with altitude. The increase per 1000 m was 3.4±0.5 mmHg for sPAP, 2.5±0.7 mmHg for mPAP and 3.3±1.0 mmHg for TRPG, based predominantly on data from young male populations and elevations >2500 m. These results provide a clinically relevant estimates of expected increase PAP at high altitudes.

The approximately linear increase in mPAP with altitude, despite the nonlinear decline in inspired oxygen tension, reflects the progressive activation of HPV, also known as the von Euler–Liljestrand mechanism, which optimises the ventilation–perfusion ratio [75]. Under global alveolar hypoxia at high altitude, this response becomes generalised, resulting in diffuse pulmonary vasoconstriction and increased mPAP [1, 75]. HPV is mediated by membrane depolarisation and calcium influx and is associated with reduced nitric oxide (NO) bioavailability [16, 76]. Its magnitude varies between individuals and has direct clinical consequences; exaggerated HPV leads to disproportionate increases in mPAP and a higher risk of high-altitude pulmonary oedema (HAPE) [1, 77, 78]. While moderate mPAP elevation is part of physiological adaptation, excessive vasoconstriction promotes capillary stress and alveolar oedema [75, 77]. HAPE-susceptible individuals also show reduced pulmonary NO bioavailability during hypoxia, which correlates inversely with mPAP [16, 79]. These mechanisms support the need for reference ranges to distinguish adaptive from maladaptive mPAP responses. The 30-day exposure captures both acute HPV (occurring within seconds to hours) and the early phases of pulmonary vascular remodelling (developing over days to weeks). While acute HPV is reversible, sustained hypoxia activates hypoxia-inducible factor-1α, leading to smooth muscle cell proliferation and structural vascular remodelling, contributing to persistent mPAP elevation and, potentially, to inter-study variability [75, 80].

The forest plot summarises the average across included studies. In contrast, the meta-regression quantifies the continuous increases in pressure with altitude, enabling estimations of sPAP, mPAP and TRPG at specific altitudes. In clinical practice, this approach allows estimation of expected changes in mPAP during acute altitude exposure. The derived increase in mPAP per 1000 m may assist in distinguishing physiological from exaggerated haemodynamic responses in individuals newly exposed to altitude.

Most studies assessed PAP using echocardiography, reflecting its feasibility in remote high-altitude settings and its suitability for studies in healthy subjects. Most studies derived sPAP from TRPG with the addition of estimated RAP. However, RAP estimation varied considerably across studies, contributing to variability in sPAP. Given the limited agreement between estimated and invasively measured RAP, this represents a relevant source of methodological variability in sPAP. In this review, three studies reported comparably high sPAP values at 4559 m (45.4 mmHg [44], 49.7 mmHg [42] and 44.3 mmHg [43]), which are consistent with differences in RAP estimation, particularly when based on clinically assessed jugular venous pressure. Across the included studies, multiple approaches of RAP estimation were used, fixed values of 5 mmHg [15, 17, 18, 32, 35, 38, 41, 55, 59, 61–68, 72, 73] or 10 mmHg [74], inferior vena cava collapsibility [19, 21, 47, 49, 51, 54, 56, 60, 70], clinically assessed jugular venous pressure [16, 40, 42–46] or alternative anatomic references [53]. In several studies, the method was not stated [33, 34, 39, 48, 50, 52, 57, 58, 69, 71]. This heterogeneity likely contributes to between-study variability in sPAP. Notably, funnel plot asymmetry was observed for sPAP but not for TRPG or mPAP. As sPAP incorporates estimated RAP, this finding suggests that variability in RAP estimation contributes to small-study effects, whereas TRPG and derived mPAP appear less affected by this source of bias. This supports the interpretation that the observed asymmetry is more likely driven by measurement heterogeneity than publication bias alone. Although recent guidelines recommend the use of inferior vena cava collapsibility for estimating RAP, its application and interpretation may still vary across studies [13]. These factors likely explain the discrepancies between sPAP and TRPG observed in the present meta-analyses. To reduce this bias, the European Society of Cardiology and European Respiratory Society recommend reporting peak TRV without mathematical transformation or the addition of imprecise RAP estimations [8]. In this review, only two studies reported TRV, underscoring the limited implementation of this recommendation [53, 66]. To address variability related to RAP estimation, a pooled upper limit of normal for pressure increases with altitude of approximately 4.5 mmHg per 1000 m was derived for sPAP, mPAP and TRPG.

Three studies reported comparable high baseline sPAP values of 30 mmHg [67], 30.4 mmHg [53] and 31.9 mmHg [74]. In the latter study, the higher value corresponds to the use of a fixed RAP of 10 mmHg, whereas the other two studies applied a fixed RAP of 5 mmHg without further specification.

One study reported a baseline mPAP of 20 mmHg [35] in a cohort of healthy, young, male residents of Kyrgyzstan living ≤800 m. In this study, RAP was estimated to be 5 mmHg, which should be considered.

Regarding clinical application, approximately 10% of healthy Caucasian individuals demonstrated an exaggerated HPV response during a 2-h exposure to normobaric hypoxia (inspiratory oxygen fraction 12%) [81]. In a larger cohort of 421 healthy individuals exposed to the same conditions, those with abnormal HPV responses showed sPAP values at 4559 m comparable to those measured during normobaric hypoxia at matched arterial oxygen tension [82]. Individuals with normal HPV responses had sPAP values of 24 mmHg at low altitude and 35 mmHg at 4559 m, consistent with the estimates derived in the present analysis (figure 3b) [82]. These findings indicate that hypoxic challenge testing can identify individuals at risk for exaggerated pressure response and provide a reference for interpreting mPAP changes during altitude exposure. However, other studies comparing hypobaric and normobaric hypoxia report more pronounced haemodynamic changes under hypobaric conditions, including higher mPAP and greater hypoxaemia at equivalent inspired oxygen tensions [83, 84].

Most studies included in this review were conducted at altitudes between 2500 and 4500 m; accordingly, estimates of PAP are most robust within this range. At altitude differences below 2000 m, six studies reported sPAP, of which four demonstrated only minimal deviations from baseline [50, 51, 55, 67]. The findings suggest a relevant increase in PAP in healthy individuals at altitudes above approximately 2000 m (barometric pressure >79 kPa). This range corresponds to physiological thresholds for HPV, which is typically initiated when alveolar oxygen partial pressure falls below ∼10 kPa (76 mmHg). Such levels are generally reached above 2000 m in healthy individuals, although inter-individual variability in ventilatory and pulmonary vascular responses may result in earlier onset in some cases. This threshold has practical clinical implications, as commercial aircraft cabins are pressurised to a maximum equivalent altitude of 2438 m to maintain arterial oxygen levels within safe limits [85, 86]. The present findings are consistent with a modest increase in mPAP under such conditions in healthy individuals, which may be relevant when counselling patients with pre-existing cardiopulmonary disease prior to air travel. In patients with hypoxaemic lung disease, such as COPD or pulmonary hypertension, mPAP may increase at moderate altitude, with substantial inter-individual variability. Short-term exposure to 2000–2500 m is not associated with a clinically relevant increase in mPAP in pulmonary vascular disease, consistent with healthy individuals, whereas COPD patients show greater increases with altitude [87–89]. Studies including individuals above 5000 m are subject to selection bias, as participants are typically highly fit and less susceptible to altitude-related complications [90].

The reference values established in this meta-analysis should be interpreted within the broader context of altitude physiology. The landmark Operation Everest II studies, which remain the only invasive haemodynamic measurements at simulated extreme altitude, demonstrated that mPAP increases progressively from 15 mmHg at sea level to 34 mmHg at 7620 m, with pulmonary vascular resistance increasing nearly fourfold, changes that were not immediately reversible with oxygen supplementation, suggesting early structural remodelling even during acute exposure [91, 92].

Stratified analysis by exposure duration demonstrated slightly lower increase in sPAP for exposures ≥5 days compared to shorter exposures; however, this difference did not reach statistical significance. This pattern is consistent with partial attenuation of PAP during acclimatisation. Proposed mechanisms include augmented NO bioavailability as hypoxic stress stabilises, enhanced endothelial vasodilatory responses and a reduction in CO due to plasma volume contraction and decreased stroke volume during acclimatisation [1, 4, 93]. However, this binary categorisation may not fully capture the time-dependent nature of physiological adaptation to hypoxia. Previous studies similarly report modest reductions in sPAP over several days in both healthy individuals and patients with COPD [94, 95], consistent with observations in some studies included in this review showing lower values on repeated measurements within the first days at altitude [15, 28, 45, 47, 58, 59]. In contrast, individuals developing HAPE exhibit a marked and sustained increase in mPAP and were therefore excluded from this analysis. HAPE-susceptible individuals show exaggerated HPV, with mPAP values of 28 mmHg compared to 19 mmHg in resistant individuals during acute hypoxia exposure [96].

Participants exposed to hypoxia in hypobaric chambers showed lower sPAP and TRPG compared to those studied in mountainous settings [24, 67, 68, 74]. This difference may relate to variations in physical activity, as increased CO during exertion augments PAP. Exercise is known to substantially increase mPAP at altitude; invasive studies demonstrate that moderate exercise approximately doubles mPAP at high altitude compared to a ∼50% increase at sea level for equivalent workloads [97]. Similarly, in the Operation Everest II studies, mPAP increased from 33 mmHg at sea level to 54 mmHg at 7620 m during near-maximal exercise [91]. While this analysis focused on acute newcomers, chronic high altitude maladaptation syndromes such as chronic mountain sickness demonstrate that sustained hypoxic exposure leads to further pressure elevation, particularly during exercise, where sPAP reaches 48 mmHg compared to 36 mmHg in healthy high-altitude dwellers [98]. In one study, the participants ascended to 4350 m for pre-acclimatisation for 7 days before returning to sea level and going in the hypobaric chamber one day later, which may account for lower sPAP compared to nonacclimatised exposure [24]. In the remaining studies, no prior acclimatisation was reported. Overall, these findings indicate that exposure conditions may influence PAP responses. Notably, previous data show no significant acclimatisation effect after re-exposure following a 1-week return to low altitude [94]. The consistency of results across measurement modalities (right heart catheterisation and echocardiography) supports the robustness of the observed relationship between altitude and mPAP (see figure 4), a conclusion further supported by sensitivity analyses excluding studies at high risk of bias, which yielded comparable results (figures S4 and S5).

Since age and sex were rarely reported among the included studies, no stratification between age groups was possible. Most studies included young males and the results may not be transferred to an older population. At low altitude, mPAP in known to rise with age [99]. Only two studies included in this systematic review compared younger and older participants. One study found a slightly, but not significantly higher mPAP at high altitude in older people [33]. The other study, however, found a statistically significantly higher sPAP at high altitude in older compared to younger participants [67]. More studies are needed to explore PAP at high altitude for different ages and sexes. Although altitude explained a substantial proportion of the variability (R²=61%), considerable residual heterogeneity remained, likely reflecting differences in measurement techniques, timing of assessment and study populations. As a result, prediction intervals were wide and primarily reflect between-study variability rather than physiological limits in individuals. Therefore, the upper bound of the confidence interval was used as a conservative and clinically applicable estimate of the upper limit of normal. This approach provides a pragmatic reference value for clinical interpretation while acknowledging underlying variability across studies. While these reference ranges are derived from population-level data and may aid in clinical interpretation, they should be applied cautiously at the individual level given the substantial interindividual variability in pulmonary vascular responses to hypoxia.

Limitations

In the current meta-analysis, it was not feasible to correct for ascent time and ascent protocol. Furthermore, the methodology regarding the measurement or estimation of RAP was not always accurately described. These studies with unclear measurements of RAP were considered as high RoB. Barometric pressure, which can vary at the same altitude at different locations and times, could not be considered in this analysis. The heterogeneity observed in the forest plots is explained by the nature of the data, as per definition different altitudes were analysed; however, this heterogeneity is accounted for through meta-regression. Most data were published from altitudes >2500 m, limiting the interpretability of findings at moderate altitudes.

Conclusion

This systematic review and meta-analysis quantified the normal ranges of sPAP, mPAP and TRPG, as well as their increases, in healthy low-altitude residents travelling to higher altitudes. The physiological increase in PAP was approximately linear with altitude. The increase per 1000 m was 3.4±0.5 mmHg in sPAP, 2.5±0.7 mmHg in mPAP and 3.3±1.0 mmHg in TRPG, based predominantly on data from elevations >2500 m. The pooled upper reference limit for the increase was 4.5 mmHg per 1000 m for all three measures. At high altitude, absolute sPAP ranged from 24 to 50 mmHg and mPAP from 20 to 26 mmHg. These findings are primarily applicable to young, predominantly male, healthy low-altitude residents and to exposures above 2500 m. They underscore the need for a unified methodology for PAP assessment by echocardiography. In line with current guidelines, we recommend reporting peak TRV in future studies.

The resulting population-derived reference ranges may aid in contextualising sPAP, mPAP and TRPG measurements in low-altitude residents at higher altitudes, including during intercurrent illness. However, given the substantial interindividual variability in pulmonary vascular responses to hypoxia, these estimates should be applied cautiously at the individual level and not considered definitive diagnostic thresholds.

Points for clinical practice

• In healthy low-altitude residents travelling to high altitudes, PAP increases by up to 4.5 mmHg per 1000 m altitude gained, based predominantly on data from elevations above 2500 m and in young, mostly male populations.

• These population-derived reference values may assist in contextualising PAP measurements at high altitude, but should be applied cautiously at the individual level and outside these settings.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material and tables S1-S4

ERR-0069-2026.SUPPLEMENT

Supplementary figure S1

ERR-0069-2026.SUPPLEMENT2

Supplementary figure S2

ERR-0069-2026.SUPPLEMENT3

Supplementary figure S3

ERR-0069-2026.SUPPLEMENT4

Supplementary figure S4

ERR-0069-2026.SUPPLEMENT5

Supplementary figure S5

ERR-0069-2026.SUPPLEMENT6

Supplementary figure S6: Funnel plots assessing the risk of publication bias

ERR-0069-2026.SUPPLEMENT7