Partial Pressure of Arterial Oxygen in Healthy Adults at High Altitudes: A Systematic Review and Meta-Analysis

Abstract

Importance: With increasing altitude, the partial pressure of inspired oxygen decreases and, consequently, the Pao2 decreases. Even though this phenomenon is well known, the extent of the reduction as a function of altitude remains unknown.

Objective: To calculate an effect size estimate for the decrease in Pao2 with each kilometer of vertical gain among healthy unacclimatized adults and to identify factors associated with Pao2 at high altitude (HA).

Data sources: A systematic search of PubMed and Embase was performed from database inception to April 11, 2023. Search terms included arterial blood gases and altitude.

Study selection: A total of 53 peer-reviewed prospective studies in healthy adults providing results of arterial blood gas analysis at low altitude (<1500 m) and within the first 3 days at the target altitude (≥1500 m) were analyzed.

Data extraction and synthesis: Primary and secondary outcomes as well as study characteristics were extracted from the included studies, and individual participant data (IPD) were requested. Estimates were pooled using a random-effects DerSimonian-Laird model for the meta-analysis.

Main outcomes and measures: Mean effect size estimates and 95% CIs for reduction in Pao2 at HA and factors associated with Pao2 at HA in healthy adults.

Results: All of the 53 studies involving 777 adults (mean [SD] age, 36.2 [10.5] years; 510 men [65.6%]) reporting 115 group ascents to altitudes between 1524 m and 8730 m were included in the aggregated data analysis; 13 of those studies involving 305 individuals (mean [SD] age, 39.8 [13.6] years; 185 men [60.7%]) reporting 29 ascents were included in the IPD analysis. The estimated effect size of Pao2 was -1.60 kPa (95% CI, -1.73 to -1.47 kPa) for each 1000 m of altitude gain (τ2 = 0.14; I2 = 86%). The Pao2 estimation model based on IPD data revealed that target altitude (-1.53 kPa per 1000 m; 95% CI, -1.63 to -1.42 kPa per 1000 m), age (-0.01 kPa per year; 95% CI, -0.02 to -0.003 kPa per year), and time spent at an altitude of 1500 m or higher (0.16 kPa per day; 95% CI, 0.11-0.21 kPa per day) were significantly associated with Pao2.

Conclusions and relevance: In this systematic review and meta-analysis, the mean decrease in Pao2 was 1.60 kPa per 1000 m of vertical ascent. This effect size estimate may improve the understanding of physiological mechanisms, assist in the clinical interpretation of acute altitude illness in healthy individuals, and serve as a reference for physicians counseling patients with cardiorespiratory disease who are traveling to HA regions.

Figure 1.. Study Selection Flowchart

ABG indicates arterial blood gas; AD, aggregated data; HAPE, high-altitude pulmonary edema; and IPD, individual participant data.

Figure 2.. Changes in Pao₂ by Altitude Based on Aggregated Data

A total of 51 studies^{,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,} are shown in the final forest plot; 2 studies^, included in the aggregated data analysis are not shown because of incomplete data. All values necessary to determine the mean effect size (calculated as the ratio of the difference in Pao₂ to the difference in altitude multiplied by 1000) for each study are provided in eTable 1 in Supplement 1. For example, Ainslie et al measured Pao₂ at 1400 m and 3840 m, and found that the corresponding means of Pao₂ were 9.92 kPa and 7.25 kPa, respectively. The mean effect size was therefore 1.09 kPa per 1000 m (9.92 kPa minus 7.25 kPa divided by 3840 m minus 1400 m then multiplied by 1000). Multilevel meta-analysis using a random-effects DerSimonian-Laird model was conducted to assess the pooled effect if a study had more than 1 measurement made at high altitude. The size of the squares corresponds to the weight of the effect size of the individual studies. The width of the diamond corresponds to the 95% CI of the point estimate of the pooled effect.

Figure 3.. Arterial Blood Gas Values by Altitude Based on Aggregated Data

The size of each bubble is proportional to the SE of each of the 51 studies^{,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,} shown in the figure; 2 studies^, included in the aggregated data analysis are not shown because of incomplete data. The 95% CI (gray area) is the SE of the estimate (ie, the SE of the point estimation for 1 observation).

Figure 4.. Lower and Upper Limits of Normal for Pao₂, Paco₂, and pH Based on Individual Participant Data

A total of 13 studies^,,,,,,,,,,,, were included in the analysis. Dots represent individual participant data, continuous lines represent means, and dashed lines represent 90% CIs. The lower dashed line represents the lower limit of normal and the upper dashed line represents the upper limit of normal at the respective altitude. The 90% CIs were not corrected for possible confounders, such as age, sex, and body mass index.