Oxygen Saturation in Childhood at High Altitude: A Systematic Review

Abstract

Background:

It is well known that oxygen saturation as measured by pulse oximetry (SpO₂) decreases as altitude increases. However, how SpO₂ changes across childhood, and more specifically during sleep/wake states, at different high altitudes are less well understood. We aimed to perform a systematic review of all studies with direct SpO₂ measurement in healthy children living at high altitude (>2500 meters above sea level) to address these questions.

Methods:

MEDLINE, EMBASE, and SciELO databases were searched up to December 2018. Two independent reviewers screened the literature and extracted relevant data.

Results:

Of 194 references, 20 studies met the eligibility criteria. Meta-analysis was not possible due to the use of different oximeters and/or protocols for data acquisition and reporting of different SpO₂ central tendency and dispersion measures. The most relevant findings from the data were: (1) SpO₂ is lower as altitude increases; (2) at high altitude, SpO₂ improves with age through childhood; (3) SpO₂ is lower during sleep and feeding in comparison to when awake, this SpO₂ gap between wake and sleep states is more evident in the first months of life and narrows later in life; (4) SpO₂ dispersion (interindividual variation) is higher at younger ages, and more so during sleep; (5) In 6/20 studies, the SpO₂ values were nonnormally distributed with a consistent left skew.

Conclusions:

At high altitude, the mean/median SpO₂ increases in children with aging; a significant gap between wake and sleep states is seen in the first months of life, which narrows as the infant gets older; SpO₂ dispersion at high altitude is wider at younger ages; at high altitude, SpO₂ shows a nonnormal distribution skewed to the left; this bias becomes more evident as altitude increases, at younger ages and during sleep.

Introduction

Since its description in 1975, oxygen saturation as measured by pulse oximetry (SpO₂) (Severinghaus, 2007) has become a routine tool in clinical practice, to the point that it has been proposed as a new vital sign (Mower et al., 1997). SpO₂ is a cheap, accessible, and portable tool used extensively in acute care, being especially important for acute respiratory infection categorization in infants and children (Lazzerini et al., 2015).

Around 140 million people worldwide live permanently at high altitude (Moore, 2001), where settlements are characterized by hypobaric hypoxia. No prior systematic review has considered normal reference data for resting daytime SpO₂ levels across childhood at different high-altitude locations nor considered the impact of sleep and infant feeding on SpO₂. Only through understanding normal variation can treatment thresholds be determined for acute infections such as pneumonia and bronchiolitis and chronic illness such as bronchopulmonary dysplasia, cystic fibrosis, and postinfectious bronchiolitis obliterans. Expected SpO₂ characteristics during sleep at high altitude are also important to interpret polysomnographic findings (Hill et al., 2016b).

In the decade since Subhi et al. (2009) published a systematic review of normal SpO₂ values in high-altitude resident healthy infants and children up to 12 years of age, oximeter technology has progressed. We aimed to extend and update this review to include pulse oximetry studies at high altitude across childhood, including teenagers, and taking in account the differences in wake, sleep, and feeding states, especially in infants.

Methods

We identified published studies from MEDLINE, EMBASE, and SciELO (up to December 2018) databases using the search terms: “(oxygen saturation) AND (high altitude)” restricted to child (birth to 19 years old) without language restriction; high altitude was defined as altitude >2500 meters above sea level (masl) (Moore, 2001). Studies published solely in abstract form were excluded because the methods and results could not be fully analyzed. In addition, we searched other nonbibliographic data sources such as web searching. Only studies of healthy individuals habitually living at high altitude were included. Exclusion criteria included: (1) studies limited to SpO₂ measurement uniquely during the first 24 hours of life as SpO₂ has important changes during this period (Gonzáles and Salirrosas, 2005); (2) data collected across a range of altitudes greater than ±50 m (e.g., between 3800 and 4200 masl), where it was not possible to identify the number of individuals located at a specific altitude; and (3) studies that included children with acute or chronic cardiorespiratory disorders, chronic ill health, or history of prematurity.

Data extraction and assessment of risk of bias

Titles, abstracts, and citations were independently reviewed by two authors (S.U. and C.M.G.). Based on the full-text versions, all the studies were evaluated, and after obtaining full reports, eligibility was assessed. Disagreements were discussed and resolved by consensus, and, when necessary, advice was sought from the third review author (J.A.C.-R.). The risk of bias was evaluated according to the Newcastle–Ottawa Scale (Wells et al., 2011). For methodological and reporting quality, a prespecified data analysis included year, country, altitude, type of study, number of participants, age range, SpO₂ median/mean, and SpO₂ dispersion measurements. The pulse oximetry device used and the duration of SpO₂ recording were also searched. We also looked at whether hemoglobin (Hb) concentration measurements were reported.

Meta-analysis was not performed due to the variation between studies in oximeter devices, protocols, and the central tendency (e.g., mean/median) and dispersion measures (e.g., standard deviation, interquartile range, 5th/95th percentile) to report SpO₂ values.

Using original source data from Rojas-Camayo et al.'s (2018) study, the SpO₂ values from the first quartile were compared with those from the second to fourth quartiles grouped using RV 3.6.1 (July 2019 version) software. The comparison was made for three altitude levels (2500, 3600, and 5100 masl) and for two age groups (1–5 and 6–17 years old). Statistical differences were determined using the t-test when the Shapiro–Wilk test showed a p-value <0.05, otherwise the Mann–Whitney test was used.

Results

One hundred ninety-four studies were retrieved from the databases; of which, 20 were eligible for inclusion (Fig. 1). Most of the studies (n = 14) came from the South America Andean region (Peru = 4, Bolivia = 4, Colombia = 3, Ecuador = 2, Argentina = 1); three studies were from China, two from the United States, and one from Nepal. Seventeen studies reported SpO₂ data during wakefulness (Lozano et al., 1992; Nicholas et al., 1993; Niermeyer et al., 1993, 1995; Gamponia et al., 1998; Torres et al., 1999; Beall, 2000; Huicho et al., 2001; Ramírez-Cardich et al., 2004; Alduncin et al., 2005; Mattos et al., 2005; Weitz and Garruto, 2007; Schult and Canelo-Aybar, 2011; Shrestha et al., 2012; Duenas-Meza et al., 2015; Hill et al., 2016a, 2016b; Rojas-Camayo et al., 2018). Ten studies reported SpO₂ data during sleep (Niermeyer et al., 1993, 1995; Gamponia et al., 1998; Torres et al., 1999; Ramírez-Cardich et al., 2004; Duenas-Meza et al., 2015; Ucrós et al., 2015, 2017; Hill et al., 2016); four studies reported data in infants during feeding (Niermeyer et al., 1995; Gamponia et al., 1998; Torres et al., 1999; Ramírez-Cardich et al., 2004); and one study reported data when infants were crying (Gamponia et al., 1998). One study was published in a Colombian journal not identified by the search but was known by the authors (Torres et al., 1999). All studies were cross-sectional, except one, which had a longitudinal component (Duenas-Meza et al., 2015). The total number of SpO₂ records sampled in these 20 studies was 6877 during wakefulness, 923 during sleep, and 528 in infants during feeding. The pulse oximetry device used and the sampling protocols were diverse (Table 1). The quality assessment of most studies was high with a low risk of bias.

FIG. 1.

Process of study selection.

Table 1.

Pulse Oximetry Device Used and Duration of SpO₂ Measurement

Study	Oximeter type	Duration of recording
Alduncin et al. (2005)	NONIN	No data during wakefulness. Overnight for nocturnal measurement (3:25–5:45 hours).
Beall (2000)	CRITICARE M 503	Six observations were recorded, and their average was used as the individual's SpO₂.
Duenas-Meza et al. (2015)	MASSIMO RAD 8	All the time during wakefulness and sleep during polysomnography study up to 10 hours. Minimum time and mean time: no data.
Gamponia et al. (1998)	NELLCOR N 10	At least 10 seconds.
Hill et al. (2016b)	NONIN, PLYMOUTH, (MN)	Sampling rate and data averaged over four successive pulse beats.
Hill et al. (2016a)	MASIMO RADICAL	At least 3 minutes for diurnal measurement. Overnight for nocturnal measurement (minimum 5 hours of artifact free data).
Huicho et al. (2001)	NELLCOR	Values were recorded after machine stabilization. When values oscillated, the average of three consecutive readings was used.
Lozano et al. (1992)	NELLCOR N 1O	At least 10 seconds. All readings were duplicated.
Mattos et al. (2005)	NO DATA	No data.
Nicholas et al. (1993)	NEONATAL FLEX IT	At least 2 minutes.
Niermeyer et al. (1993)	BIOX 3700BIO	Value was considered acceptable when the electrocardiographically measured heart rate was within 5 beats per minute of the oximetry-determined pulse rate, and the pulse waveform displayed on the oximeter was clear and full.
Niermeyer et al. (1995)	BIOX 3740	One-minute intervals for a total of 10 minutes.
Ramírez-Cardich et al. (2004)	NELLCOR N 20	One-minute intervals for a total of 10 minutes.
Rojas-Camayo et al. (2018)	NELLCOR 560	Every 10 seconds for a total of six measurements, and the average was the final value.
Schult and Canelo-Aybar (2011)	DEVON MEDICAL 300C-1	More than 10 seconds.
Shrestha et al. (2012)	CMS- 50DL	Several readings until a consistent value was displayed.
Torres et al. (1999)	MASIMO SET	Two registers, 1 minute each one, and the average was the final value.
Ucrós et al. (2015)	NONIN M 8008J	All the time during overnight polysomnography study. Minimum time 180 minutes − mean time 250 minutes.
Ucrós et al. (2017)	NONIN M 8008J	All the time during overnight polysomnography study. Minimum time 180 minutes − mean time 250 minutes.
Weitz and Garruto (2007)	NONIN 8500	Between 5 and 10 minutes.

SpO₂, oxygen saturation as measured by pulse oximetry.

Studies in the wake and sleep states

The results of studies for SpO₂ during the wake state are shown in Tables 2 –4 across three different geographic regions. Sleep data are shown in Table 5. The most relevant findings are as follows: diurnal awake SpO₂ is lower as altitude increases (for illustrative data, see Fig. 2) and is lower at the same altitude during sleep and infant feeding in comparison to the wake state; in younger children, the range of normal diurnal SpO₂ increases with increasing altitude, so, for example, varying by 5% in children aged 1–5 years at 2500 masl, in comparison with 12% in the same age group at 5100 masl, with this phenomenon being more marked during sleep. At high altitude, SpO₂ increases with age and a significant gap between wake and sleep states occurs in the first months of life and narrows as the age increases; for example, Hill et al. (2016a) report that at 3700 masl, the median difference in infants aged 6–10 months between the wake versus sleep stage is 6%, compared with 3% between the ages of 13 and 17 years (Fig. 3). SpO₂ dispersion at high altitude is wider at younger ages (for illustrative data, see Fig. 4), and SpO₂ peak values are attained at older ages as altitude increases.

FIG. 2.

SpO₂ 5th percentile at different altitudes in two age groups in Peru. Graphic built with data from Rojas-Camayo et al. (2018) 5th percentile data provided to offer potential thresholds below which children require further clinical evaluation. SpO₂, oxygen saturation as measured by pulse oximetry.

FIG. 3.

Gap between SpO₂ medians in wake and sleep states in children aged from 6 months to 17 years at 3700 masl in Bolivia. At low altitude, SpO₂ developmental differences were evident between infants and children in relation to 3% dips and were statistical significant (p = 0.019). These differences were more striking at high altitude (p < 0.001). Graphic built with data from Hill et al. (2016a). masl, meters above sea level.

FIG. 4.

SpO₂ at 2640 masl in awake and asleep infants aged from 3 to 15 months in Colombia. In 3.6 ± 0.5-month-old subjects (awake and asleep), SpO₂ nadir, time with SpO₂ <90%, and oxygen desaturation index were significantly higher in comparison with the two other age groups (p < 00.1). Graphic built with data from Duenas-Meza et al. (2015).

Table 2.

Resting SpO₂ in High-Altitude Resident Andean Children During the Wake State by Age Group and Altitude (Ordered from Lower to Higher Altitude)

Study	Country	Altitude (masl)	Age range	n	% SpO₂ median/mean	SpO₂ dispersion measurements
Rojas-Camayo et al. (2018)	Perú	2500	1–5 years	87	96.0 (median)	p5th 93.0%, p95th 98.0%
Rojas-Camayo et al. (2018)	Perú	2500	5–17 years	234	97.0 (median)	p5th 94.0%, p95th 99.0%
Hill et al. (2016a)	Bolivia	2500	4–10 years	9	94.0 (median)	IQR 1%
Torres et al. (1999)	Colombia	2640	1–30 days	300	95.0 (median)	p5th 90.5%, p95th 98.8%
Lozano et al. (1992)	Colombia	2640	1–3 months	35	95.3 (median)	95% CI (90.5%–98.8%)
Duenas-Meza et al. (2015)	Colombia	2640	1 ± 0.3 months	106	92.5 (median)	p5th 88.0%, p95th 96.0%
Duenas-Meza et al. (2015)	Colombia	2640	3.6 ± 0.5 months	89	93.0 (median)	p5th 88.0%, p95th 96.0%
Duenas-Meza et al. (2015)	Colombia	2640	6.3 ± 0.8 months	89	93.0 (median)	p5th 90.0%, p95th 97.0%
Lozano et al. (1992)	Colombia	2640	7–12 months	23	93.4 (median)	95% CI (92.4%–94.4%)
Duenas-Meza et al. (2015)	Colombia	2640	13.2 ± 1.9 months	25	94.0 (median)	p5th 91.0%, p95th 96.0%
Rojas-Camayo et al. (2018)	Perú	2880	1–5 years	95	95.0 (median)	p5th 92.0%, p95th 98.0%
Rojas-Camayo et al. (2018)	Perú	2880	5–17 years	122	96.0 (median)	p5th 92.0%, p95th 98.0%
Rojas-Camayo et al. (2018)	Perú	3250	1–5 years	140	93.0 (median)	p5th 89.0%, p95th 95.0%
Rojas-Camayo et al. (2018)	Perú	3250	6–17 years	181	94.0 (median)	p5th 91.0%, p95th 97.0%
Mattos et al. (2005)	Bolivia	3600	1–2 weeks	60	85.3 (mean)	±1SD 10.5%
Rojas-Camayo et al. (2018)	Perú	3600	1–5 years	30	93.0 (median)	p5th 87.0%, p95th 95.0%
Rojas-Camayo et al. (2018)	Perú	3600	6–17 years	117	93.0 (median)	p5th 90%, p95th 95.0%
Hill et al. (2016a)	Bolivia	3700	6–12 months	7	94.0 (median)	IQR 2%
Hill et al. (2016a)	Bolivia	3700	4–10 years	18	94.0 (median)	IQR 1%
Hill et al. (2016a)	Bolivia	3700	13–17 years	7	94.0 (median)	IQR 2%
Ramírez-Cardich et al. (2004)	Perú	3750	2 weeks	30	91.0 (mean)	No data
Ramírez-Cardich et al. (2004)	Perú	3750	1 month	33	92.0 (mean)	No data
Ramírez-Cardich et al. (2004)	Perú	3750	2 months	32	92.0 (mean)	No data
Ramírez-Cardich et al. (2004)	Perú	3750	3 months	33	92.0 (mean)	No data
Ramírez-Cardich et al. (2004)	Perú	3750	4 months	30	92.0 (mean)	No data
Alduncin et al. (2005)	Argentina	3775	7–31 weeks	12	86.0 (mean)	±1SD 2.6%
Ramírez-Cardich et al. (2004)	Perú	3750	2 weeks	30	91.0 (mean)	No data
Ramírez-Cardich et al. (2004)	Perú	3750	1 month	33	92.0 (mean)	No data
Ramírez-Cardich et al. (2004)	Perú	3750	2 months	32	92.0 (mean)	No data
Ramírez-Cardich et al. (2004)	Perú	3750	3 months	33	92.0 (mean)	No data
Ramírez-Cardich et al. (2004)	Perú	3750	4 months	30	92.0 (mean)	No data
Gamponia et al. (1998)	Bolivia	4018	1–60 months	128	88.3(median)	95% CI (87.8%–88.1%)
Rojas-Camayo et al. (2018)	Perú	4338	1–5 years	47	86.0 (median)	p5th 82.0%, p95th 90.0%
Rojas-Camayo et al. (2018)	Perú	4338	6–17 years	126	88.0 (median)	p5th 82.0%, p95th 92.0%
Schult and Canelo-Aybar (2011)	Perú	4340	5–6 years	78	83.8 (mean)	±2SD 10%
Schult and Canelo-Aybar (2011)	Perú	4340	7–8 years	43	83.7 (mean)	±2SD 8.4%
Schult and Canelo-Aybar (2011)	Perú	4340	9–10 years	108	84.7 (mean)	±2SD 5.4%
Schult and Canelo-Aybar (2011)	Perú	4340	11–12 years	40	86.5 (mean)	±2SD 8.6%
Schult and Canelo-Aybar (2011)	Perú	4340	13–14 years	54	88.1 (mean)	±2SD 7.0%
Schult and Canelo-Aybar (2011)	Perú	4340	15–16 years	63	88.4 (mean)	±2SD 6.6%
Rojas-Camayo et al. (2018)	Perú	4500	1–5 years	36	85.0 (median)	p5th 78.0%, p95th 90.0%
Rojas-Camayo et al. (2018)	Perú	4500	6–17 years	121	85.0 (median)	p5th 79.0%, p95th 90.0%
Rojas-Camayo et al. (2018)	Perú	4715	1–5 years	28	83.0 (median)	p5th 78.0%, p95th 89.0%
Rojas-Camayo et al. (2018)	Perú	4715	6–17 years	133	86.0 (median)	p5th 79.0%, p95th 92.0%
Rojas-Camayo et al. (2018)	Perú	5100	1–5 years	74	80.0 (median)	p5th 73.0%, p95th 85.0%
Rojas-Camayo et al. (2018)	Perú	5100	6–17 years	168	81.0 (median)	p5th 75.0%, p95th 86.0%

CI, confidence interval; IQR, interquartile range; masl, meters above sea level; p5th, 5th percentile; p95th, 95th percentile; SD, standard deviation.

Table 3.

Resting SpO₂ in High-Altitude Resident Himalayan Children During the Wake State by Age Group and Altitude

Study	Country	Altitude (masl)	Age range	n	% SpO₂ mean	SpO₂ dispersion measurements
Shrestha et al. (2012)	Nepal	2700	2–14 years	56	95.0	±1SD 1.3%
Shrestha et al. (2012)	Nepal	2800	2–14 years	22	95.0	±1SD 1.2%
Weitz and Garruto (2007)	China	3200	5–9 years	157	90.8	±1SD 3.9%
Weitz and Garruto (2007)	China	3200	10–14 years	172	90.	±1SD 3.2%
Weitz and Garruto (2007)	China	3200	15–19 years	148	91.3	±1SD 3.0%
Shrestha et al. (2012)	Nepal	3550	2–14 years	19	91.0	±1SD 1.7%
Niermeyer et al. (1995)	China/Han	3658	1–16 weeks	15	87.0	±1SD 6.0%
Niermeyer et al. (1995)	China/Tibetan	3658	1–16 weeks	15	89.0	±1SD 3.0%
Beall (2000)	China	3800	0–9 years	735	89.3	No data
Beall (2000)	China	3800	10–19 years	735	90.9	No data
Shrestha et al. (2012)	Nepal	3800	2–14 years	9	91.0	±1SD 1.2%
Weitz and Garruto (2007)	China/Tibetan	3800	5–9 years	93	86.5	±1SD 3.9%
Weitz and Garruto (2007)	China/Han	3800	5–9 years	103	85.5	±1SD 4.1%
Weitz and Garruto (2007)	China	3800	10–14 years	217	86.6	±1SD 3.8%
Weitz and Garruto (2007)	China	3800	15–19 years	172	87.6	±1SD 2.8%
Beall (2000)	China	4200	0–9 years	294	85.5	No data
Beall (2000)	China	4200	10–19 years	294	88.7	No data
Weitz and Garruto (2007)	China	4300	5–9 years	49	84.0	±1SD 4.3%
Weitz and Garruto (2007)	China	4300	10–14 years	62	85.0	±1SD 3.0%
Weitz and Garruto (2007)	China	4300	15–19 years	47	85.8	±1SD 3.7%

Table 4.

Resting SpO₂ in High-Altitude Resident Children During the Wake State by Age Group and Altitude in Colorado, USA

Study	Country	Altitude (masl)	Age range	n	% SpO₂ mean	SpO₂ dispersion measurements
Nicholas et al. (1993)	USA	2774–2818	2 days–22 months	72	91.7	SD 2.0%
Niermeyer et al. (1993)	USA	3100	1 week	14	87.8	SD 4.8%
Niermeyer et al. (1993)	USA	3100	2 months	13	89.9	SD 2.4%
Niermeyer et al. (1993)	USA	3100	4 months	14	91.1	SD 1.7%

Table 5.

SpO₂ in Sleeping Children at Different Ages and High Altitudes Across the World

Study	Country	Altitude (masl)	Age range	n	% SpO₂ median/mean	SpO₂ dispersion measurement
Hill et al. (2016a)	Bolivia	2500	4–10 years	9	93.7 (median)	IQR 2.4%
Ucrós et al. (2015)	Ecuador	2560	1–4 months	36	92.0 (median)	p5th 86.0%, p95th 94.0%
Torres et al. (1999)	Colombia	2640	1–30 days	300	95.0 (median)	p5th 87.5%, p95th 96.0%
Duenas-Meza et al. (2015)	Colombia	2640	1 ± 0.3 months	106	92.5 (median)	p5th 87.5%, p95th 96.0%
Duenas-Meza et al. (2015)	Colombia	2640	3.6 ± 0.5 months	89	93.0 (median)	p5th 86.0%, p95th 94.0%
Duenas-Meza et al. (2015)	Colombia	2640	6.3 ± 0.8 months	89	93.0 (median)	p5th 90.0%, p95th 96.0%
Duenas-Meza et al. (2015)	Colombia	2640	13 ± 1.9 months	25	94.0 (median)	p5th 91.0%, p95th 96.0%
Niermeyer et al. (1993)	USA/active sleep	3100	1 week	14	83.0 (mean)	±1SD 5.6%
Niermeyer et al. (1993)	USA/quiet sleep	3100	1 week	14	80.6 (mean)	±1SD 5.3%
Niermeyer et al. (1993)	USA	3100	2 months	13	86.6 (mean)	±1SD 4.7%
Niermeyer et al. (1993)	USA/quiet sleep	3100	4 months	14	86.6 (mean)	±1SD 4.4%
Ucrós et al. (2017)	Ecuador	3200	1–4 months	18	92.0 (median)	p5th 66.0%, p95th 91.0%
Niermeyer et al. (1995)	China/Han	3658	1–16 weeks	15	84.0 (mean)	±1SD 9%
Niermeyer et al. (1995)	China/Tibetan	3658	1–16 weeks	15	87.0 (mean)	±1SD 5%
Hill et al. (2016b)	Bolivia	3650	7–10 years	26	89.0 (median)	IQR 3%
Hill et al. (2016b)	Bolivia	3650	13–16 years	26	89.0 (median)	IQR 3%
Hill et al. (2016a)	Bolivia	3700	6–12 months	7	87.5 (median)	IQR 7.7%
Hill et al. (2016a)	Bolivia	3700	4–10 years	18	90.3 (median)	IQR 6.1%
Hill et al. (2016a)	Bolivia	3700	13–17 years	7	91.2 (median)	IQR 3.1%
Ramírez-Cardich et al. (2004)	Peru	3750	2 weeks	14	86.2	No data
Ramírez-Cardich et al. (2004)	Peru	3750	1 month	15	86.0	No data
Ramírez-Cardich et al. (2004)	Peru	3750	2 months	20	87.5	No data
Ramírez-Cardich et al. (2004)	Peru	3750	3 months	17	88.0	No data
Ramírez-Cardich et al. (2004)	Peru	3750	4 months	12	89.2	No data
Gamponia et al. (1998)	Bolivia	4018	0–5 months	12	84.6 (median)	95% CI (81.3%–87.9%)

Studies during infant feeding and crying

Five studies analyzed SpO₂ during feeding in infants aged up to 5 months at high altitude (Table 6). The data show an SpO₂ decrease between 0.2% and 6% when wake and feeding states are compared, with a mean decrease of −2.2%. Only one study reported SpO₂ differences in infants when crying compared with quiet wakefulness (Gamponia et al., 1998); this study included 19 infants aged 0–5 months at 4018 masl and noted a 0.8% decrease in median SpO₂ in crying infants.

Table 6.

SpO₂ in Feeding Infants at Different Ages and High Altitudes Across the World

Study	Country	Altitude (masl)	Age range	n	% SpO₂ mean	Difference in SpO₂ % with the wake state
Torres et al. (1999	Colombia	2640	1–30 days	294	93.5	−1.5
Niermeyer et al. (1993)	USA	3100	1 week	14	85.8	−3.0
Niermeyer et al. (1993)	USA	3100	2 months	13	89.5	−0.4
Niermeyer et al. (1993)	USA	3100	4 months	14	90.9	−0.2
Niermeyer et al. (1995)	China/Han	3658	1 week	15	85	−2
Niermeyer et al. (1995)	China/Han	3658	1 month	15	81	−4
Niermeyer et al. (1995)	China/Han	3658	2 months	11	80	−2
Niermeyer et al. (1995)	China/Han	3658	4 months	9	79	−6
Niermeyer et al. (1995)	China/Tibetan	3658	1 week	15	87	−3
Niermeyer et al. (1995)	China/Tibetan	3658	1 month	15	86	−3
Niermeyer et al. (1995)	China/Tibetan	3658	2 months	13	86.5	−2
Niermeyer et al. (1995)	China/Tibetan	3658	4 months	13	86	−0.5
Ramírez-Cardich et al. (2004)	Peru	3750	2 weeks	13	88	−3.0
Ramírez-Cardich et al. (2004)	Peru	3750	1 month	16	89.5	−2.5
Ramírez-Cardich et al. (2004)	Peru	3750	2 months	21	90.5	−1.5
Ramírez-Cardich et al. (2004)	Peru	3750	3 months	13	91	−1.0
Ramírez-Cardich et al. (2004)	Peru	3750	4 months	20	91.5	−0.5
Gamponia et al. (1998)	Bolivia	4018	0–5 months	4	83.5	−3.5

Studies reporting ethnicity and sex influences

Three studies reported the influence of ethnicity and sex on waking SpO₂ variables. One study in Tibet at 3200 masl (Weitz and Garruto, 2007) reported a significant sex-by-ethnicity interaction in 15–19 year-olds, such that Tibetan females had higher SpO₂ values than Tibetan males, but among Chinese Han born at this high altitude, males had higher SpO₂ values than females (the Han population originate from lowland plains in Eastern China). A study from Peru at 4100 masl found no differences in SpO₂ values between males and females, but children from the Nuñoa ethnicity had higher SpO₂ levels in comparison with children from Tintayá and Marquirí ethnicities (Huicho et al., 2001), all ethnicities being highlanders. Finally, a study at 3800 masl reported that Tibetan 5–9 year-olds had significantly higher SpO₂ values than Han children did at the same age (Niermeyer et al., 1995).

Statistical distribution

In 12/20 studies, a normal distribution was assumed by the authors who reported mean values, but without specification of the statistical test used to reach this conclusion. In 2/20 studies, a nonnormal distribution can be deducted from the data, but the statistical bias is not known; in 6/20 studies, a left-skewed nonnormal distribution was found (Torres et al., 1999; Ucrós et al., 2015, 2017; Hill et al., 2016a, 2016b; Rojas-Camayo et al., 2018). A review of the source data from studies by Hill and Rojas-Camayo confirmed that the data distribution has a negative bias. In Figure 5, the SpO₂ left-skewed distribution at different altitudes and ages is shown. In Figure 6, the increase in the difference between 5th percentile and 95th percentile as altitude increases can be seen reflecting how the left bias is more evident as altitude increases.

FIG. 5.

Consistent SpO₂ left skew at different altitudes and age groups. (A–C) Built with data from Rojas-Camayo et al. (2018). (D) Built with data from Sime et al. (1963). The original article from Sime et al. shows results in means and SD. As the article delivers data of each of the children, we looked for the statistical distribution for SpO₂ and built the graphic. SD, standard deviation.

FIG. 6.

SpO₂ difference between 5th percentile and 95th percentile in two age groups at different altitudes. Graphic built with data from Rojas-Camayo et al. (2018).

The comparison of SpO₂ values from the first quartile with those from the second to fourth quartiles grouped, using data from Rojas-Camayo et al.'s study, indicated statistically significant differences in all altitude/age groups analyzed (Table 7).

Table 7.

Differences in SpO₂ Between Quartiles 1 and Quartiles 2–4 at Several Altitudes and Age Groups

Age (years)	Altitude (masl)	SpO₂ median quartile 1 (n)	SpO₂ median quartiles 2–4 (n)	SpO₂ difference (%)	p
1–5	2500	94% (21)	96% (66)	2	<0.001
6–17	2500	95% (58)	97% (176)	2	<0.001
1–5	3600	88% (8)	93% (30)	5	<0.001
6–17	3600	91% (29)	93% (88)	2	<0.001
1–5	5100	75% (19)	81% (168)	6	<0.001
6–17	5100	77% (41)	82% (131)	5	<0.001

Table built with data from Rojas-Camayo et al. (2018).

Hb concentration measurement

Hb concentration was only measured in two studies (Ramírez-Cardich et al., 2004; Mattos et al., 2005). Oxygen-carrying capacity was not calculated in any of the studies.

Discussion

To the best of our knowledge, this is the first systematic review to examine the influence of high altitude on SpO₂ across childhood from 0 to 19 years of age. It builds on data published by Subhi et al. (2009), which was limited to six studies above 2500 masl (rather than the 20 included here), and to pre-adolescent children. Furthermore, we included data on infants when feeding and report key differences between wake and sleep states.

Data provided by this review show that SpO₂ at high altitudes increases through childhood; this phenomenon was highlighted by Beall (2000) who reported that at altitudes between 3800 and 4200 masl, SpO₂ peak values were attained by 11 years of age and were 7% higher than values in young infants. The SpO₂ improvement with age can be viewed as a marker of physiological adaptation or “SpO₂ maturation” (Hill et al., 2016a); interestingly, SpO₂ maturation is not a phenomenon exclusive to high altitude and is also seen at sea level (Schlüter et al., 2001). In addition, the present review shows that interindividual variability in SpO₂ is higher at younger ages (Figs. 3, 4, and 6). This point was also noted by Beall (2000) and has been observed in subsequent studies (Duenas-Meza et al., 2015). This interindividual variation increases with altitude and is more marked during sleep (Figs. 3, 4, and 6) (Duenas-Meza et al., 2015; Ucrós et al., 2015, 2017; Hill et al., 2016a) and during feeding in young infants (Niermeyer et al., 1993, 1995; Torres et al., 1999). In consequence at altitudes >2500 masl, SpO₂ treatment thresholds may change depending on sleep–wake or feeding status, especially in young infants.

The lower SpO₂ values during sleep are normal in children, even at sea level (MacLean et al., 2015), and could be related with drops in respiratory rate and functional residual capacity, as well as increase in upper airway resistance (Marcus, 2001). What emerges from the present review is that the SpO₂ gap between wake and sleep stages is notable at high altitude. The explanation for lower SpO₂ values during feeding in infants is less certain; a recent study in infants, aged 2 weeks–3 months with cardiac disease, showed significant desaturation (mean −2.8%) during feeding, but no drop in a healthy age-similar control group (Miranda et al., 2019). The lower baseline SpO₂ values in the infants with cardiac disease are similar to those seen in Andean infants at the threshold of high altitude; this suggests that minor perturbations in ventilation associated with infant feeding may become critical when baseline SpO₂ is lowered. As it is known, from the O₂ dissociation curve, at high altitude, mild changes in P_aO₂ induce significant SpO₂ drops (Chernick and West, 1990).

In relation with lower SpO₂ values during sleep in children, it is known at sea level that normal sleep onset prompts a fall in respiratory rate, functional residual capacity, and an increase in upper airway resistance (Marcus, 2001) and that small decrements in oxygen saturation (e.g., 1%–2%) are acceptable (MacLean et al., 2015). Regarding differences by sex, it has been found in adults aged 30–40 years at high altitude, higher SpO₂ levels in females than in males (Beall, 2000), but similar findings are not evidenced in the pediatric literature (Huicho et al., 2001).

An interesting question is why SpO₂ maturation occurs. It could be hypothesized that since periodic breathing (PB) is much more important at high altitude during the first months of life (Duenas-Meza et al., 2015; Ucrós et al., 2015, 2017), there could be a link between SpO₂ increase and decrease in PB with age. We found two studies in infants, which explored this possible association. The first was carried out by Parkins et al. (1998), who exposed 34 healthy infants to a 15% oxygen environment during a mean of 6.3 hours of sleep and measured oxygen saturation, frequency of isolated and periodic apnea, and frequency of desaturation. In the later published by Ucrós et al. (2015), oxygen saturation, PB, and apnea indexes were measured during a mean of 4.1 hours of sleep at 2560 masl in infants aged between 1 and 4 months. Neither study found an association between the mean SpO₂ and PB. Similarly, studies carried out in adults have not found this correlation (Salvaggio et al., 1998; Insalaco et al., 2012; Ainslie et al., 2013). A second hypothesis could be related to possible changes in ventilation across childhood at high altitude, nevertheless studies undertaken by Beall on different populations do not support this theoretical mechanism; this research showed that Tibetans were more hypoxic than Aymara, despite a higher resting ventilation and hypoxic ventilatory responses (Beall et al., 1997). A third mechanism, and probably the most plausible hypothesis, is that as pulmonary pressure falls through childhood, the SpO₂ gradually increases (Penaloza and Arias-Stella, 2007). The question that would then remain is why pulmonary pressure shows such behavior in children living at high altitude.

In six studies, data explicitly demonstrated a nonnormal SpO₂ left-skewed distribution, meaning significant interindividual differences in SpO₂ at a same age and altitude. This phenomena was also noted when data from an early study published in 1963 were reanalyzed; in this research, oxygen saturation was measured in blood of children living around 4400 masl (Sime et al., 1963) (Fig. 5). The striking feature of a left-skewed SpO₂ distribution has been found consistently across different altitudes, ages, and sleep versus wake states. At high altitude, the SpO₂ nonnormal statistical distribution delivers a parabolic curve, which has to be differentiated from the sinusoidal oxygen dissociation curve related to the capacity of Hb to carry oxygen. This concept is illustrated in Figure 7 (Chernick and West, 1990; Penaloza and Arias-Stella, 2007).

FIG. 7.

Comparison between the oxygen–hemoglobin dissociation curve and SpO₂ curve as oxygen partial pressure increases. Graphic built with data from Chernick and West (1990) and Penaloza and Arias-Stella (2007).

We speculate that children with low steady-state oxygen saturation measures during health may be at increased risk of later life chronic mountain sickness (CMS) in its different forms. These include CMS with altitude polycythemia or Monge's disease (Villafuerte and Corante, 2016); high-altitude heart disease (HAHD) or high-altitude pulmonary hypertension, which occurs without polycythemia (Ge and Helun, 2001; Aldashev et al., 2002; Reeves and Grover, 2005; Kojonazarov et al., 2007; Xu and Jing, 2009); and mixed CMS, which includes both excessive polycythemia and pulmonary hypertension-related HAHD (Ge and Helun, 2001). These different kinds of progressive life-limiting conditions affect 5%–10% of high-altitude residents in the first category (Leon-Velarde et al., 2005), and 5%–18% in the second one (Mirrakhimov and Strohl, 2016). To the best of our knowledge, childhood predictors of CMS have not been reported. The etiology of this condition is likely to be multifactorial, including genetically determined differences in erythropoiesis regulation, as demonstrated by Stobdan et al. (2017) who identified novel genes involved in high-altitude adaptation in the Andes; alterations in ventilatory responses to hypoxia (Villafuerte and Corante, 2016); hypoxia gene expressions related to CMS (Xing et al., 2008; Stobdan et al., 2017); and levels of mediators related to lung vascular reactivity (Aldashev et al., 2002; Kojonazarov et al., 2012). Longitudinal studies of SpO₂ across childhood would be of value in this context.

The present study has several limitations. First, different oximeter devices and/or protocols for measuring SpO₂ were used across the different studies, and in most studies, we do not have information relating to important equipment settings, such as sampling rate and averaging time (Hill and Evans, 2016). Second, the data available across the publications used different approaches for central tendency measures, with some studies reporting means while others reported medians. Similarly, in relation to dispersion, different measures were used including percentiles, standard deviation, confidence intervals, and interquartile ranges. Third, ethnic background is a key influence on high-altitude adaptation, and data presented here represent populations with different adaptive mechanisms. Fourth, only two studies measured Hb concentration levels. Importantly, SpO₂ alone is a limited indicator of oxygen delivery to the tissues, especially in the setting of polycythemia.

However, this review also has notable strengths. In total, 20 studies were analyzed, which included 8347 SpO₂ data; furthermore, the quality assessment of most studies was high with a low risk of bias. Importantly, diurnal–nocturnal variation highlighted from sleep data cannot be accounted for by Hb concentration fluctuations, suggesting genuinely compromised adaptation to high altitude in some children that may have implications for later life CMS vulnerability. Finally, the data synthesized in this review offer interesting insights into the concept of SpO₂ maturation at high altitude.

The need for international reference data on ranges of SpO₂ across childhood at high altitude remains pressing. Future studies should use standardized oximetry protocols (uniform devices and settings) and standardized reporting of data to confidently establish such reference data in healthy children to support clinical decision-making. In the meantime, the data synthesized in this review provide a preliminary guide for clinicians managing children with cardiorespiratory disorders at high altitude.

Conclusions

At high altitude, the mean/median SpO₂ improves in infants and children with aging; a significant gap between wake and sleep states is seen in the first months of life, which narrows as the infant gets older; SpO₂ dispersion at high altitude is wider at younger ages; at high altitude, SpO₂ shows a nonnormal distribution skewed to the left; this bias becomes more evident as altitude increases, at younger ages and during sleep. SpO₂ has interindividual differences at a same given age and altitude, being significantly lower in one of four children. Collaborative studies using similar oximeters/protocols for measuring SpO₂ at different altitudes in healthy children across high-altitude geographical regions are urgently needed.

Footnotes

Acknowledgments

The authors thank Mr. Gerardo Ardila, Statistician, MSc, of the Research Department of the Fundación Santa Fe de Bogotá, for his valuable help in the construction of graphics and statistical analysis; Dr. Jose Rojas-Camayo of the Instituto de Investigaciones de la Altura, Universidad Peruana Cayetano Heredia, Lima, Peru, for providing raw data from the publication “Reference values for oxygen saturation from sea level to the highest human habitation in the Andes in acclimatised persons” (Rojas-Camayo et al., 2018); and Prof. Romola Bucks of the University of Western Australia for providing raw data from the DeSat study (Hill et al., 2016a, ).

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

Ainslie

, Lucas

, and Burgess

. (2013). Breathing and sleep at high altitude. Respir Physiol Neurobiol, 188:233–256.

Aldashev

, Sarybaev

, Sydykov

, Kalmyrzaev

, Kim

, Mamanova

, Maripov

, Kojonazarov

, Mirrakhimov

, Wilkins

, and Morrell

. (2002). Characterization of high-altitude pulmonary hypertension in the Kyrgyz: Association with angiotensin-converting enzyme genotype. Am J Respir Crit Care Med, 166:1396–1402.

Alduncin

, Grañana

, Follett

, Musante

, Milberg

, Vogler

, and Rocca Rivarola

. (2005). Breathing problems during sleep in infants native from the Argentine highlands. Arch Argent Pediatr, 103:14–22.

Beall

. (2000). Oxygen saturation increases during childhood and decreases during adulthood among high altitude native Tibetians residing at 3,800–4,200 m. High Alt Med Biol, 1:25–32.

Beall

, Strohl

, Blangero

, Williams-Blangero

, Almasy

, Decker

, Worthman

, Goldstein

, Vargas

, Villena

, Soria

, Alarcon

, and Gonzales

. (1997). Ventilation and hypoxic ventilatory response of Tibetan and Aymara high altitude natives. Am J Phys Anthropol, 104:427–447.

Chernick

, and West

. (1990). The functional basis of respiratory disease. In: Kendig's Disorders of the Respiratory Tract in Children. Saunders-Elsevier, Philadelphia; pp. 52.

Duenas-Meza

, Bazurto

, Gozal

, González-García

, Durán-Cantolla

, and Torres-Duque

. (2015). Overnight polysomnographic characteristics and oxygen saturation of healthy infants, 1 to 18 months of age, born and residing at high altitude (2,640 meters). Chest, 148:120–127.

Gamponia

, Babaali

, Yugar

, and Gilman

. (1998). Reference values for pulse oximetry at high altitude. Arch Dis Child, 78:461–465.

, and Helun

. (2001). Current concept of chronic mountain sickness: Pulmonary hypertension-related high-altitude heart disease. Wilderness Environ Med, 12:190–194.

10.

Gonzáles

, and Salirrosas

. (2005). Arterial oxygen saturation in healthy newborns delivered at term in Cerro de Pasco (4340 m) and Lima (150 m). Reprod Biol Endocrinol, 3:46.

11.

Hill

, Baya

, Gavlak

, Carroll

, Heathcote

, Dimitriou

, L'Esperance

, Webster

, Holloway

, Virues-Ortega

, Kirkham

, Bucks

, and Hogan

. (2016a). Adaptation to life in the high Andes: Nocturnal oxyhemoglobin saturation in early development. Sleep, 39:1001–1008.

12.

Hill

, Carroll

, Dimitriou

, Gavlak

, Heathcote

, L'Esperance

, Baya

, Webster

, Pushpanathan

, and Bucks

. (2016b). Polysomnography in Bolivian children native to high altitude compared to children native to low altitude. Sleep, 39:2149–2155.

13.

Hill

, and Evans

. (2016). The investigation of sleep disordered breathing: Seeing through a glass darkly. Arch Dis Child, 101:1082–1083.

14.

Huicho

, Pawson

, León-Velarde

, Rivera-Chira

, Pacheco

, Muro

, and Silva

. (2001). Oxygen saturation and heart rate in healthy school children and adolescents living at high altitude. Am J Hum Biol, 13:761–770.

15.

Insalaco

, Romano

, Salvaggio

, Pomidori

, Mandolesi

, and Cogo

. (2012). Periodic breathing, arterial oxyhemoglobin saturation, and heart rate during sleep at high altitude. High Alt Med Biol, 13:258–262.

16.

Kojonazarov

, Isakova

, Imanov

, Sovkhozova

, Sooronbaev

, Ishizaki

, and Aldashev

. (2012). Bosentan reduces pulmonary artery pressure in high altitude residents. High Alt Med Biol, 13:217–223.

17.

Kojonazarov

, Imanov

, Amatov

, Mirrakhimov

, Naeije

, Wilkins

, and Aldashev

. (2007). Noninvasive and invasive evaluation of pulmonary arterial pressure in highlanders. Eur Respir J, 29:352–356.

18.

Lazzerini

, Sonego

, and Pellegrin

. (2015). Hypoxaemia as a mortality risk factor in acute lower respiratory infections in children in low and middle-income countries: Systematic review and meta-analysis. PLoS One, 10:e0136166.

19.

Leon-Velarde

, Maggiorini

, Reeves

, Aldashev

, Asmus

, Bernardi

, Ge

, Hackett

, Kobayashi

, Moore

, Penaloza

, Richalet

, Roach

, Wu

, Vargas

, Zubieta-Castillo

, and Zubieta-Calleja

. (2005). Consensus statement on chronic and subacute high altitude diseases. High Alt Med Biol, 6:147–157.

20.

Lozano

, Duque

, Buitrago

, and Behaine

. (1992). Pulse oximetry reference values at high altitude. Arch Dis Child, 67:299–301.

21.

MacLean

, Fitzgerald

, and Waters

. (2015). Developmental changes in sleep and breathing across infancy and childhood. Paediatr Respir Rev, 6:276–284.

22.

Marcus

. (2001). Sleep-disordered breathing in children. Am J Respir Crit Care Med, 164:16–30.

23.

Mattos

, Caballero

, and Baros

. (2005). Gasometry, hematocrit and pulse oximetry in newborns at 3,600 meters above sea level. Rev Soc Boliv Pediatr, 44:158–160.

24.

Miranda

VSG

, Souza

, Etges

, and Barbosa

. (2019). Cardiorespiratory parameters in infants cardiopathy: Variations during feeding. Codas, 31:e20180153.

25.

Mirrakhimov

, and Strohl

. (2016). High-altitude pulmonary hypertension: An update on disease pathogenesis and management. Open Cardiovasc Med J, 10:19–27.

26.

Moore

. (2001). Human genetic adaptation to high altitude. High Alt Med Biol, 2:257–279.

27.

Mower

, Sachs

, Nicklin

, and Baraff

. (1997). Pulse oximetry as a fifth pediatric vital sign. Pediatrics, 99:681–686.

28.

Nicholas

, Yaron

, and Reeves

. (1993). Oxygen saturation in children living at moderate altitude. J Am Board Fam Pract, 6:452–456.

29.

Niermeyer

, Shaffer

, Thilo

, Corbin

, and Moore

. (1993). Arterial oxygenation and pulmonary arterial pressure in healthy neonates and infants at high altitude. J Pediatr, 123:767–772.

30.

Niermeyer

, Yang

, Shanmina

Drolkar

, Zhuang

, and Moore

. (1995). Arterial oxygen saturation in Tibetan and Han infants born in Lhasa, Tibet. N Engl J Med, 333:1248–1252.

31.

Parkins

, Poets

, O'Brien

, Stebbens

, and Southall

. (1998). Effect of exposure to 15% oxygen on breathing patterns and oxygen saturation in infants: Interventional study. BMJ, 316:887–891.

32.

Penaloza

, and Arias-Stella

. (2007). The heart and pulmonary circulation at high altitudes: Healthy highlanders and chronic mountain sickness. Circulation, 115:1132–1146.

33.

Ramírez-Cardich

, Saito

, Gilman

, Escate

, Strouse

, Kabrhel

, Johnson

, Galchen

, and Bautista

. (2004). Effect of maternal anemia at high altitude on infant hematocrit and oxygenation. Am J Trop Med Hyg, 70:420–424.

34.

Reeves

, and Grover

. (2005). Insights by Peruvian scientists into the pathogenesis of human chronic hypoxic pulmonary hypertension. J Appl Physiol, 98:384–389.

35.

Rojas-Camayo

, Mejia

, Callacondo

, Dawson

, Posso

, Galvan

, Davila-Arango

, Bravo

, Loescher

, Padilla-Deza

, Rojas-Valero

, Velasquez-Chavez

, Clemente

, Alva-Lozada

, Quispe-Mauricio

, Bardalez

, and Subhi

. (2018). Reference values for oxygen saturation from sea level to the highest human habitation in the Andes in acclimatised persons. Thorax, 73:776–778.

36.

Salvaggio

, Insalaco

, Marrone

, Romano

, Braghiroli

, Lanfranchi

, Patruno

, Donner

, and Bonsignore

. (1998). Effects of high-altitude periodic breathing on sleep and arterial oxyhaemoglobin saturation. Eur Respir J, 12:408–413.

37.

Schlüter

, Buschatz

, and Trowitzsch

. (2001). Polysomnographic reference curves for the first and second year of life. Somnologie, 5:3–16.

38.

Schult

, and Canelo-Aybar

. (2011). Oxygen saturation in healthy children aged 5 to 16 years residing in Huayllay, Peru at 4340 m. High Alt Med Biol, 12:89–92.

39.

Severinghaus

. (2007). Takuo Aoyagi: Discovery of pulse oximetry. Anesth Analg, 105:S1–S4.

40.

Shrestha

, Shrestha

, and Bhandary

. (2012). Oxygen saturation of hemoglobin in healthy children of 2–14 years at high altitude in Nepal. Kathmandu Univ Med J (KUMJ), 10:40–43.

41.

Sime

, Banchero

, Penaloza

, Gamboa

, Cruz

, and Marticorena

. (1963). Pulmonary hypertension in children born and living at high altitudes. Am J Cardiol, 11:143–149.

42.

Sime

, Monge

, and Whittembury

. (1975). Age as a cause of chronic mountain sickness (Monge's disease). Int J Biometeorol, 19:93–98.

43.

Stobdan

, Akbari

, Azad

, Zhou

, Poulsen

, Appenzeller

, Gonzales

, Telenti

, Wong

EHM

, Saini

, Kirkness

, Venter

, Bafna

, and Haddad

. (2017). New insights into the genetic basis of Monge's disease and adaptation to high-altitude. Mol Biol Evol, 34:3154–3168.

44.

Subhi

, Smith

, and Duke

. (2009). When should oxygen be given to children at high altitude? A systematic review to define altitude-specific hypoxaemia. Arch Dis Child, 94:6–10.

45.

Torres

, Osorio

, and Ramos

. (1999). Measurement of pulse oximetry values in healthy newborns in Bogotá-city during sleep, wake-up and feeding. Avances Pediátr, 1:2–8.

46.

Ucrós

, Granados

, Parejo

, Guillén

, Ortega

, Restrepo

, Gil

, and Guillén

. (2015). Oxygen saturation, periodic breathing and apnea during sleep in infants 1 to 4 month old living at 2,560 meters above sea level. Arch Argent Pediatr, 113:341–344.

47.

Ucrós

, Granados

, Parejo

, Ortega

, Guillén

, Restrepo

, Gil

, and Guillén

. (2017). Oxygen saturation, periodic breathing, and sleep apnea in infants aged 1–4 months old living at 3200 meters above sea level. Arch Argent Pediatr, 115:54–57.

48.

Villafuerte

, and Corante

. (2016). Chronic mountain sickness: Clinical aspects, etiology, management, and treatment. High Alt Med Biol, 17:61–69.

49.

Weitz

, and Garruto

. (2007). A comparative analysis of arterial oxygen saturation among Tibetans and Han born and raised at high altitude. High Alt Med Biol, 8:13–26.

50.

Wells

, Shea

, O'Connell

, Peterson

, Welch

, Losos

, and Tugwell

. (2011). The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomized studies in meta-analysis. www.ohri.ca/programs/clinicalepidemiology/oxford.asp. 2011. Last assessed April 1, 2019.

51.

Xing

, Qualls

, Huicho

, Rivera-Ch

, Stobdan

, Slessarev

, Prisman

, Ito

, Wu

, Norboo

, Dolma

, Kunzang

, Norboo

, Gamboa

, Claydon

, Fisher

, Zenebe

, Gebremedhin

, Hainsworth

, Verma

, and Appenzeller

. (2008). Adaptation and mal adaptation to ambient hypoxia; Andean, Ethiopian and Himalayan patterns. PLoS One, 3:e2342.

52.

, and Jing

. (2009). High-altitude pulmonary hypertension. Eur Respir Rev, 18:13–17.