Development of aptitude at altitude
Abstract
Millions of people currently live at altitudes in excess of 2500 metres, where oxygen supply is limited, but very little is known about the development of brain and behavioural function under such hypoxic conditions. We describe the physiological, cognitive and behavioural profile of a large cohort of infants (6–12 months), children (6–10 years) and adolescents (13–16 years) who were born and are living at three altitude locations in Bolivia (∼500 m, ∼2500 m and ∼3700 m). Level of haemoglobin oxygen saturation and end-tidal carbon dioxide were significantly lower in all age groups living above 2500 metres, confirming the presence of hypoxia and hypocapnia, but without any detectable detriment to health. Infant measures of neurodevelopment and behaviour yielded comparable results across altitude groups. Neuropsychological assessment in children and adolescent groups indicated a minor reduction in psychomotor speed with increasing altitude, with no effect of age. This may result from slowing of underlying brain activity in parallel with reduced cerebral metabolism and blood flow, evidenced here by reduced cerebral blood flow velocity, particularly in the basilar artery, in children and adolescents. The proportion of European, Native American and African genetic admixture was comparable across altitude groups, suggesting that adaptation to high altitude in these children occurred in response to chronic hypoxic exposure irrespective of ethnic origin. Thus, psychomotor slowing is proposed to be an adaptive rather than a deficient trait, perhaps enabling accuracy of mental activity in hypoxic conditions.
Introduction
It is estimated that more than 140 million people live at altitudes in excess of 2500 metres (Penaloza & Arias-Stella, 2007). At these altitudes oxygen supply is reduced, resulting in a state of chronic hypoxia. It is widely assumed that such populations are born fully adapted to their environment. There is a reasonable expectation of an inheritance of physical traits selected over hundreds of generations to confer protection against the low oxygen (hypoxia) experienced from, and possibly before, birth. Indeed, there may be more efficient uptake, transportation, and diffusion of oxygen in high-altitude natives than in lowland migrants (Frisancho, Borkan & Klayman, 1975). Increased red blood cell production is a well-documented mechanism for maximizing oxygen carrying capacity at high altitude (León-Velarde, Gamboa, Chuquiza, Esteba, Rivera-Chira & Monge, 2000), but very high levels may be maladaptive, increasing stroke risk (Gordeuk, Stockton & Prchal, 2005). Pulmonary hypertension allows adaptation to alveolar hypoxia by redistributing blood flow to well-ventilated lung segments, and may be present from early in life (Huicho, 2007), although heart function is generally normal (Huicho, Muro, Pacheco, Silva, Gloria, Marticorena, et al., 2005). The brain requires a finely balanced supply of oxygen, and may adapt by altering the responsiveness of cerebral circulation to changing oxygen and carbon dioxide levels (Norcliffe, Rivera-Ch, Claydon, Moore, León-Velarde, Appenzeller, et al., 2005). There is also evidence in South American Quechuan adults that the brain responds to chronically lowered haemoglobin oxygen saturation (SpO2) by reducing metabolic rate (Hochachka, Clark, Brown, Stanley, Stone, Nickles, et al., 1994). Although positron emission tomography (PET) studies are not practical at high altitude and are not justified in healthy young children, the possibility of a reduction in brain metabolic rate may be explored by non-invasive transcranial Doppler sonography. Specifically, cerebral blood flow velocity (CBFV) is a measure of cerebral blood flow (CBF), which is sensitive to brain metabolic rate (Brugneaux, Hodges, Hanly & Poulin, 2007).
Physiological adaptations to high-altitude hypoxia are thus well documented, but the reality may be variation in the expression of these traits between populations living at similar altitudes on different continents, for example in Tibetans, Andeans and Ethiopians (Beall, Brittenham, Strohl, Blangero, Williams-Blangero, Goldstein, et al., 1998; Kirkham & Datta, 2006), and also within populations.
Physical development in children living at high altitude has been described and indicates a combined influence of genetic inheritance, exposure to chronic hypoxia, and socio-economic status. Reduced growth, typically manifesting as shorter stature (Greksa, 2006; Bailey, Xu, Feng, Hu, Zhang & Qui, 2007), may be related to physiological changes induced by hypoxia (Bailey et al., 2007), but poor health and nutrition may also play a role (Leonard, 1989). By contrast, there are few data addressing the potential impact of high-altitude hypoxia on the development and adaptation of the most defining human feature, namely the capacity for higher cognitive function. Only a few studies on infants have been published (Leonard, 1989; Saco-Pollitt, 1981; Baker, 1960; Bender, Auer, Baran, Rodriguez & Simeonsson, 1994). Subtle behavioural/motor abnormalities were found in native high-altitude Peruvian neonates in one study (Saco-Pollitt, 1981), and in two further infant studies (Baker, 1960; Bender et al., 1994), although not in another (Hass, 1976). However, it is not clear if such data represent transient early adjustment to hypoxic exposure, or the start of an altered neurobehavioural trajectory.
In particular, hypoxia has been shown to alter brain white matter development in young rats bred at high altitude (Petropoulos, Dalal & Timiris, 1972), and in experimental paradigms at sea level (Kanaan, Farahani, Douglas, LaManna & Haddad, 2006). The importance of white matter growth for psychomotor speed has been documented in healthy children at sea level (Mabbott, Noseworthy, Bouffet, Laughlin & Rockel, 2006). Taken together, this evidence suggests that brain white matter is affected by chronic exposure to high-altitude hypoxia, and that this may result in a reduction in psychomotor processing speed.
In this report we describe the results of a cross-sectional study of physiological, behavioural and cognitive function in 168 infants, children and adolescents born and living at three altitude locations in Bolivia. Our aims were two-fold: first, to extend existing infant studies by sampling the cognitive and behavioural profile of children (6–10 years) and adolescents (13–16 years) growing up at high altitude; and second, to explore in children and adolescents the possibility that processing speed in particular may be influenced by high-altitude hypoxia. Such information is also of potential importance for our understanding of neurocognitive function in children living at sea level with pathological hypoxia.
Methods
Approval for the study was obtained from Institutional Review Committees at the University of Southampton, UK, and Univalle, Cochabamba, Bolivia.
Participants
A total of 254 participants were recruited from three locations: low altitude (Santa Cruz, ∼500m: ‘LA’), moderate altitude (Cochabamba, ∼2500m: ‘MA’) and high altitude (La Paz, ∼3700m: ‘HA’). At each location there were three age groups: infants (6–12 months), children (6–10 years) and adolescents (13–16 years). For inclusion in the present analysis, all children were born and currently living at the location where they were assessed; children and adolescents were permitted a maximum of three absences of up to 6 months duration from location, but not in the year prior to testing; all infants had remained at location since birth. This resulted in a total sample of 168 (Table 1 for group allocation). An additional inclusion criterion was that children spoke Spanish. Exclusion criteria included a history of birth injury (e.g. asphyxia), brain injury, epilepsy, diagnosis of a developmental disorder (e.g. autism; and/or currently receiving occupational and/or speech and language therapy) and/or learning disability (i.e. receiving special educational services). Most participants had at least one sibling (range 1–8), and 42.4% of the sample reported were first-born.
| Altitude group | |||
|---|---|---|---|
| LA | MA | HA | |
| n | |||
| Infants | 9 | 19 | 24 |
| Children | 18 | 22 | 19 |
| Adolescents | 19 | 16 | 22 |
| Mean age (SD) | |||
| Infants (mo) | 8.4 (2.4) | 8.4 (1.9) | 8.5 (1.8) |
| Children (yr) | 7.8 (1.5) | 7.4 (1.4) | 7.9 (1.0) |
| Adolescents (yr) | 14.2 (1.1) | 14.1 (1.7) | 14.3 (1.1) |
| Gender (M:F) | |||
| Infants | 5:4 | 7:12 | 14:10 |
| Children | 8:10 | 13:9 | 10:9 |
| Adolescents | 10:9 | 5:11 | 12:10 |
- LA, low altitude; MA, moderate altitude; HA, high altitude.
Genetic admixture
DNA was extracted from Buccal swabs (Web Scientific, Crewe, UK) and whole-genome-amplified using Genomiphi (GE Healthcare, Little Chalfont, UK). Individual Spanish and African admixture proportions as percentages were estimated using a panel of 16 ancestry-informative markers (AIMs) (rs1042602, rs1079598, rs146026, rs17203, rs1800404, rs1800498, rs326946, rs3287, rs518116, rs584059, rs723632, rs7349, rs930072, rs994174, rs2814778, rs285). These markers have previously been show to exhibit high-frequency differences in allele frequency between Native American, and European and African populations, which are a priori the main parental populations in this sample (Tsai, Kho, Shaikh, Choudhry, Naqvi, Navarro, et al., 2006; Brutsaert, Parra, Gamboa, Palacios, Rivera, Rodriguez, et al., 2004). Genotyping was undertaken using KASPar competitive allele-specific polymerase chain reaction (PCR; Kbiosciences, Hoddesdon, UK).
The admixture modelling program admixmap (Hoggart, Shriver, Kittles, Clayton & McKeigue, 2004) (http://homepages.ed.ac.uk/pmckeigu/admixmap/index.html) was used to model the distribution of admixture in the cohort and to generate individual ancestry estimates (IAEs). The program was run with 1,000 burn-in and 20,000 additional iterations under a dispersion model (dispersion parameter prior means 500, 50, 50; prior variances 5, 0.5, 0.5 for European, African and Native American populations), with input AIM ancestry-specific allele frequencies, genetic map locations and subject genotypes. The AIM ancestry-specific allele frequencies were estimated from their reported counts in modern European, African and Native American populations (Tsai et al., 2006; Brutsaert et al., 2004). Using admixmap’s convergence diagnostics, autocorrelation plots and stratification test, it was confirmed that the model converged and that there was no evidence of residual stratification (Bayesian P = .56).
Physiological measures
Anthropometry
Height/heel-to-crown length (cm), weight (kg) and head circumference (cm) were recorded for all participants. In addition, maternal height was measured when possible (133/168: 79.2% of the sample): it has been suggested that maternal height may be lower in any group with a greater percentage of high-altitude Amerindian compared to European ancestry (e.g. Mitchell Williams-Blangero, Chakraborty, Valdez, Hazuda, Haffner, et al., 1993). Height, body mass index (BMI) and head circumference values were standardized for analysis (Cole, Freeman & Preece, 1998).
Pulse oximetry (SpO2 and heart rate)
Pulse oximetry was used to non-invasively record haemoglobin oxygen saturation (SpO2) at rest, and to determine whether there was any evidence of right to left shunt, from all four limbs, using a MasimoTM oximeter. SpO2 is derived from the ratio of pulse-added signals of red and infrared light energy passed through tissue and is a good proxy for arterial saturation (SaO2) under most circumstances (Fanconi et al., 1985). It reflects the extent to which haemoglobin is saturated with oxygen at a given point in time, and is sensitive to the level of atmospheric oxygen, although it is also influenced by the oxygen-binding capacity of haemoglobin in red blood cells (which is an S-shaped curve, so rapid changes may occur at low mean saturation). After warming the hand/foot, a probe was sequentially attached to each ring finger, and to each big toe (infants) or second toe (children and adolescents). After determining that the signal had settled, SpO2 and heart rate were recorded every 10 seconds for 3 minutes. There were no significant differences between the recordings obtained from each limb, so the mean score for each participant was calculated across anatomical locations.
Capnography and respiratory rate
End-tidal CO2 (ETCO2) and respiratory rate were recorded at rest, using TIDALWAVE® (Respironics, Murrysville, PA) with nasal cannulae. A single result for each variable was recorded when the machine indicated a stable reading. (All pulse oximetry and ETCO2 recordings in infants were taken when the infants were settled and not crying or feeding.)
Blood pressure (children and adolescents only)
Blood pressure was recorded automatically, when the participant was seated.
Peak expiratory flow rate (children and adolescents only)
The mean of three attempts using a Wright’s Peak Flow MeterTM is reported.
Transcranial Doppler (TCD)
Based on an earlier study in adults (Norcliffe et al., 2005), we used TCD sonography to measure cerebral blood flow velocity (CBFV) non-invasively in the major intracranial arteries. CBFV was recorded by one of two experienced operators (F.J.K., A.M.H.) using a Nicolet EME Companion, with a 2-MHz probe. The anterior, middle and posterior cerebral arteries were insonated bilaterally through the temporal window immediately superior to the zygomatic arch. Multiple depths were insonated from 38 mm to 56 mm, and the highest CBFV (cm/sec) for each vessel across sides was recorded. The basilar artery was insonated at the back of the head, inferior to the inion, at depths ranging from 55 mm in infants to 60–70 mm in older children and adolescents. Again, the highest CBFV was recorded.
Medical and neurological examination
A brief medical examination was conducted by locally trained final-year medical students (all three locations) and by one junior doctor (La Paz), under the close supervision of two physician co-investigators (F.J.K, T.B.). Cardiorespiratory and musculo-skeletal systems were examined according to a checklist, and a number of other specific signs considered, for example evidence of: palpable thyroid, cyanosis, murmur compatible with patent ductus arteriosus. A more detailed neurological examination was performed, again under the supervision of F.J.K or T.B. This included examination of cranial nerves, muscle tone, reflexes, co-ordination and gait. Neurological examination of infants was conducted as part of the Bayley Infant Neurodevelopmental Screen, described below.
Infant development
Bayley Infant Neurodevelopmental Screener – II (BINS) (Gonzalez et al., 2000)
The BINS was administered by an experienced assessor (A.M.H.) to all infants, except for one MA infant who was not interested in the items. The BINS assesses motor, cognitive and communication domains of function. Infants may be categorized by their raw score into one of two ‘risk’ domains for neurodevelopmental delay: ‘low risk’ or ‘moderate–high risk’; it was not possible to compare raw scores as the BINS has a different range for each age (6 months, 7–10 months, 11–12 months).
Infant Development Questionnaire – Revised (IDQ-R; Gartstein & Rothbart, 2003), translated into Spanish (Gonzalez, Hidalgo, Carranza & Ato, 2000)
The IDQ-R asks parents to rate infant temperament (7-point, Likert-type scale), as observed during the last week. 191 individual items yield multiple domain scores: ‘Approach’, ‘Vocal Reactivity’, ‘High Intensity Pleasure’, ‘Smile and Laughter’, ‘Activity Level’, ‘Perceptual Sensitivity’, ‘Sadness’, ‘Distress to Limitations’, ‘Fear’, ‘Falling Reactivity/Rate of Recovery from Distress’, ‘Low Intensity Pleasure’, ‘Cuddliness’, ‘Duration of Orienting’ and ‘Soothability’.
Search skill
This function was explored using the ‘Object Retrieval’ (OR) task, adapted for use from the description by Diamond (1990). Briefly, a desired toy was hidden under a clear Perspex box that, when on the table, effectively had only one side open. The infant was required to retrieve the toy without delay, that is, to inhibit the desire to reach straight ‘through’ the closed box top and to detour around the box to the open side. All trials were video-taped and graded off-line (A.M.H); furthermore, all administrations were observed by a trained psychology student to confirm consistency. Scores were averaged across box position (i.e. which direction the open side was facing). There were 13 trials in total, and the number of trials not attempted (i.e. as a result of infant disinterest) was comparable across locations, χ2(2) = 1.7, p = .421. The percentage of trials in which the toy was successfully retrieved (uncued) was recorded for each infant. Time scores were measured across trial type from the time the infant first looked down and moved their hand towards the box until the time at which their hand made contact with the toy under the box (a purposeful grasp). This task principally reflects the development of skills associated with inhibition and working memory, consistent with the development of frontal lobe function (Diamond, 1990).
Child and adolescent neuropsychology
All measures were administered by locally trained psychology students under the supervision of a clinical psychologist (J.V.). Psychomotor speed was of primary interest, but perceptual and learning/memory tasks have yielded deficits in altitude-naïve adults (e.g. Kramer, Coyne & Strayer, 1993). Considering also the infant data, documenting subtle behavioural changes in neonates (e.g. ‘less attentive’: Saco-Pollitt, 1981), it was considered important to assess behaviour and emotion in our children and adolescents.
Wechsler Intelligence Scale for Children IV (WISC-IV Sp: Escala de Inteligencia de Wechsler para Niños, 4th edn)
The following subtests were administered: ‘Block Design’ and ‘Digit Span’, each with a mean of 10 (±3, range 1–19). Two further subtests were administered: ‘Coding’ and ‘Symbol Search’, which together yield the Processing Speed Index (M = 100 ± 15).
Wide-Ranging Assessment of Visuo-Motor Abilities (WRAVMA) (Adams & Sheslow, 1995)
This test consists of three subtests: ‘Drawing’, ‘Pegboard’ and ‘Matching’ (M = 100 ± 15).
Finger Tapping, Hastead–Reitan Battery (Reitan & Wolfson, 2008)
This test measures the average number of taps over five trials per hand (dominant hand is reported in the present study); participants were asked to repeatedly tap a lever with their index finger as quickly as possible. Finger tapping is a measure of motor speed that is highly sensitive to mild brain damage (Lezak, Howieson & Loring, 2004).
Strengths & Difficulties Questionnaire (SDQ) (http://www.sdqinfo.com/b1.html)
The SDQ is a behavioural screening questionnaire for children aged 3–16 years. The Spanish version was administered to the parents of all children and adolescents. In addition, the self-report version was administered to adolescents. A total of 25 items yield five subscales: ‘Emotional Problems’, ‘Conduct Problems’, ‘Hyperactivity/Inattention’, ‘Peer Problems’ and ‘Prosocial Behaviour’, and a Total Difficulties score.
Procedure
Information letters were distributed to parents under the direction of the Bolivian Co-Investigator (A.B-B.). On receipt of consent for their child to participate, an appointment was offered by local co-ordinators. All assessments were conducted within one 1–3 hour session, depending on age, with a break for refreshment offered between anthropometric/medical and neuropsychological testing. Parents remained with their infant throughout assessment, and some chose to remain with their child, and a few with their adolescent. Feedback was provided to parents and children/adolescents on the day and in more detailed individualized letters a few weeks later. Assessments were conducted on University premises (Cochabamba, La Paz), or at the child’s school or nursery (all locations). No assessments reported in the present report were conducted in the participants’ homes. All participants received a small gift for their participation such as a toy or game (<$3 in value). Questionnaires were completed by parents on the day or within a month of their child’s assessment.
All assessments were conducted during the period September–October 2006, with the help of Bolivian psychologists, doctors, and final-year psychology and medical students. One week of training was provided for local professionals and medical and psychology student volunteers immediately prior to data collection, in order to standardize assessment. Two weeks of assessment were conducted at all three altitude locations, starting in Cochabamba, followed by Santa Cruz and ending in La Paz.
Test sheets were scored by students, checked by the relevant senior investigators (J.V., A.B-B, A.M.H., T.B., F.J.K. and R.S.B.), and entered into spss at location supervised by a data manager/statistician (R.S.B.). Buccal swabs for genetic admixture analysis were processed initially at Southampton University, UK (J.W.H., M.R-Z.), following which SNP genotype data were analysed in Perth, Australia (L.J.P., R.J.W.) for admixture.
Statistical analysis
Data were checked for multivariate normality. If they were not suitable for parametric descriptive and inferential statistics, non-parametric analyses were used. Transformation can often yield distributions suitable for parametric analysis but can make results complex to interpret. All data from children included in this study were entered in the analyses. Between-altitude and age-group differences were analysed using analysis of variance (ANOVA), of altitude (Low altitude, Moderate altitude, High altitude) by age group (infants, children, adolescents, or children and adolescents for variables not assessed in infants). Where non-parametric analysis was required, participants were split by age group (infants, children, adolescents), and between-location differences were explored using a Kruskal–Wallis one-way ANOVA. An alpha level of .05 was used throughout.
Results
We describe results obtained from 168 infants, children and adolescents across the three altitude locations (see Table 1). All were considered to be growing normally, and most children and adolescents were classmates, indicating a similar educational level. The participant groups were well matched for age and gender (Table 1). Socio-economic status (SES), defined by maternal education and parental income, was comparable between altitude groups, confirming that the majority of participants were of middle-to-high social strata, many being the children of professionals (Table 2).
| Altitude group | Group difference | |||
|---|---|---|---|---|
| LA | MA | HA | ||
| Gestational age (weeks: mean (SD)) | 37.5 (2.5) | 37.7 (2.9) | 38.3 (2.3) | p = .375 |
| Birth weight (kg: mean (SD)) | 3.2 (0.4) | 3.2 (0.6) | 3.0 (0.6) | p = .112 |
| Maternal education (% completing high school equivalent) | 92 | 98 | 93 | p = .268 |
| Maternal height (metres: Mean (SD)) | 1.6 (0.1) | 1.6 (0.1) | 1.6 (0.1) | p = .338 |
| Family income (% earning ≥ $251 per mo) | 95 | 83 | 85 | p = .138 |
- LA, low altitude; MA, moderate altitude; HA, high altitude.
Ethnicity was examined using multiple means. All children spoke Spanish as their primary language, and maternal height was comparable across altitude groups (Table 2). Most parents described themselves and their child’s grandparents as being of Spanish-mixed origin (‘Mestizo’ or ‘Spanish’). However, cultural sensitivity may have biased these responses. Analysis of the ancestry-informative genetic markers suggested that the population was admixed, with mean population admixture proportions of 55% (95% CI = 49–60%) Native American, 41% (95% CI = 36–46%) European, and 4% (95% CI = 3–5%) African ancestry. Importantly, average levels of admixture did not significantly differ across altitude groups.
Anthropometry
Mean values and all statistical results are presented in Table 3. Standardized height significantly decreased with age, F(2,158) = 5.33, p = .006, and with altitude, F(2,158) = 5.72, p = .004. Shorter stature in children growing up at high altitude has already been documented (e.g. Greksa, 2006; Bailey et al., 2007), but the lack of altitude effect for maternal height in the present study (Table 2) suggests that the magnitude of height difference may decrease with development into adulthood. Conversely, body mass index increased with age, F(2,157) = 11.68, p < .001, but was comparable across altitude. Standardized head circumference also increased with age, F(2,157) = 6.30, p = .002, but did not change with altitude.
| Altitude group | Group difference | |||
|---|---|---|---|---|
| LA | MA | HA | ||
| Length / Height* | ||||
| Infants | .79 (2.1) | .56 (1.8) | −.41 (1.3) | Altitude: F(2,158) = 5.72, p = .004 |
| Children | 1.06 (1.3) | −.17 (.9) | −.27 (1.6) | Age: F(2,158) = 5.33, p = .006 |
| Adolescents | −.01 (.8) | −.58 (1.5) | −.89 (.9) | Altitude × Age: F(4,158) = 1.75, p = .141 |
| Body mass index* | ||||
| Infants | −.99 (2.2) | −.89 (2.1) | −.38 (1.5) | Altitude: F(2,157) < 1, p = .956 |
| Children | .85 (2.0) | .76 (1.0) | .45 (1.1) | Age: F(2,157) = 11.68, p < .001 |
| Adolescents | .41 (1.1) | .42 (1.2) | .63 (.7) | Altitude × Age: F(4,157) < 1, p = .700 |
| Head circumference* | ||||
| Infants | −.26 (1.1) | −.13 (1.0) | −.49 (1.1) | Altitude: F(2,157) = 2.80, p = .064 |
| Children | −.02 (1.0) | −.05 (1.2) | −.68 (.8) | Age: F(2,157) = 6.30, p = .002 |
| Adolescents | .22 (.9) | .67 (1.2) | .23 (1.1) | Altitude × Age: F(4,157) < 1, p = .895 |
| SpO2 | ||||
| Infants | 96.8 (1.2) | 93.4 (1.9) | 88.2 (3.3) | Altitude: F(2,158) = 176.91, p < .001 |
| Children | 98.4 (1.7) | 95.3 (1.5) | 92.1 (1.2) | Age: F(2,158) = 38.12, p < .001 |
| Adolescents | 98.7 (1.0) | 96.0 (1.4) | 93.1 (1.5) | Altitude × Age: F(4,158) = 3.86, p = .005 |
| ETCO2 | ||||
| Infants | 37.1 (4.3) | 33.2 (3.9) | 33.1 (3.9) | Altitude: F(2,151) = 5.10, p = .007 |
| Children | 38.1 (5.0) | 37.8 (2.1) | 38.1 (2.4) | Age: F(2,151) = 14.01, p < .001 |
| Adolescents | 40.3 (4.6) | 37.7 (5.6) | 37.5 (3.1) | Altitude × Age: F(4,151) = 1.22, p = .303 |
| Heart rate | ||||
| Infants | 137.6 (17.3) | 137.9 (12.6) | 136.2 (9.4) | Altitude: F(2,158) = 2.07, p = .129 |
| Children | 98.8 (11.8) | 90.8 (8.9) | 90.9 (11.9) | Age: F(2,158) = 348.99, p < .001 |
| Adolescents | 79.8 (11.1) | 85.5 (7.6) | 77.4 (9.6) | Altitude × Age: F(4,158) = 1.99, p = .099 |
| Respiratory rate | ||||
| Infants | 36.4 (16.4) | 41.3 (22.3) | 38.0 (15.1) | Altitude: F(2,151) < 1, p = .691 |
| Children | 21.9 (4.6) | 22.4 (4.9) | 21.8 (3.9) | Age: F(2,151) = 38.06, p < .001 |
| Adolescents | 19.7 (4.6) | 20.2 (3.6) | 21.2 (6.2) | Altitude × Age: F(4,151) < 1, p = .896 |
| Systolic BP** | ||||
| Infants | – | – | – | Altitude: F(2,109) = 4.59 p = .012 |
| Children | 106.9 (11.9) | 101.1 (13.7) | 97.9 (12.3) | Age: F(1,109) = 27.11, p < .001 |
| Adolescents | 117.7 (12.4) | 115.8 (13.7) | 109.4 (11.5) | Altitude × Age: F(4,151) < 1, p = .405 |
| Diastolic BP** | ||||
| Infants | – | – | – | Altitude: F(2,109) = 2.24, p = .111 |
| Children | 68.0 (13.8) | 60.1 (8.6) | 61.2 (8.7) | Age: F(1,109) = 8.78, p = .004 |
| Adolescents | 69.4 (6.8) | 70.0 (12.4) | 66.8 (9.8) | Altitude × Age: F(2,109) = 1.60, p = .206 |
| PEFR | ||||
| Infants | – | – | – | Altitude: F(2,108) = 5.32, p = .006 |
| Children | 205.0 (49.6) | 165.0 (40.9) | 183.9 (45.2) | Age: F(1,108) = 140.40, p < .001 |
| Adolescents | 366.8 (92.6) | 306.1 (94.7) | 333.1 (85.9) | Altitude × Age: F(2,108) < 1, p = .734 |
- LA, low altitude; MA, moderate altitude; HA, high altitude. BP, blood pressure; PEFR, Peak expiratory flow rate. *standardized. **unstandardized, so an age increase is inevitable. p-values given in bold are significant at < .05.
Physiology
Medical history (response to standard questions) and examination revealed no evidence of significant mountain sickness, and indicated that all participants were in good health at the time of assessment. Heart rate decreased significantly with age (see Table 3 for mean values and all statistical results), F(2,158) = 348.99, p < .001, but there were no altitude effects. There were also no differences in respiratory rate in these altitude-acclimatized and resting children. However, systolic blood pressure decreased slightly with increasing altitude, F(2,109) = 4.59 p = .012, suggesting subtle change in autonomic function. Peak expiratory flow rate (PEFR) was reduced in the MA children and adolescents, F(2,108) = 5.32, p = .006, but the prevalence of clinical asthma was not different between groups. Thus, there were few indications of physiological abnormality at higher altitudes.
The percentage of oxygen carried by blood (normally ≥97%: haemoglobin oxygen saturation ‘SpO2’; Figure 1a; Table 3) decreased linearly with increasing altitude, F(2,158) = 176.91, p < .001, confirming hypoxic exposure, and increased linearly with age, F(2,158) = 38.12, p < .001, with the greatest gain seen in the HA children, F(4,158) = 3.86, p = .005. By adolescence, SpO2 in the HA group had settled at 93.1%. ETCO2 also decreased with altitude, F(2,151) = 5.10, p = .007, and increased with age, F(2,151) = 14.01, p < .001, but not in association with an altered respiratory rate (see Table 3).
Mean (SE) haemoglobin oxygen saturation and cerebral blood flow velocity values in infants, children and adolescents, and Processing Speed Index scores in children and adolescents.
In summary, we were able to consider the degree to which CBFV and neurocognitive function was influenced by high-altitude hypoxia in a healthy population of native infants, children and adolescents, unconstrained by SES, and objectively accounting for ancestry.
Cerebral blood flow velocity (CBFV)
Altitude effects were absent in infants, but evident in older age groups, consistently for the basilar artery (Figure 1b; Table 4), the large unpaired artery supplying posterior brain regions and connected with the anterior brain circulation via the circle of Willis. In both children and adolescents, basilar CBFV decreased with increasing attitude (F(2, 56) = 8.50, p = .001, LA > MA p = .010, LA > HA p = .001; F(2, 49) = 16.09, p < .001, LA > MA p = .004, LA > HA p < .001, respectively).
| Altitude group | Group difference | |||
|---|---|---|---|---|
| LA | MA | HA | ||
| Anterior cerebral artery | ||||
| Infants | 52.4 (14.7) | 47.2 (10.1) | 51.7 (17.9) | F(2,40) < 1, p = .644 |
| Children | 51.6 (15.4) | 45.4 (14.2) | 42.3 (11.9) | F(2,56) = 2.11, p = .131 |
| Adolescents | 44.2 (8.6) | 36.0 (11.2) | 38.6 (11.5) | F(2,43) = 2.27, p = .115 |
| Middle cerebral artery | ||||
| Infants | 69.9 (10.5) | 68.3 (9.7) | 69.4 (10.1) | F(2,49) < 1, p = .909 |
| Children | 80.3 (12.9) | 70.9 (19.5) | 72.4 (17.9) | F(2,56) = 1.61, p = .207 |
| Adolescents | 69.8 (14.4) | 59.8 (11.1) | 54.9 (11.9) | F(2,51) = 6.8, p = .002 |
| Posterior cerebral artery | ||||
| Infants | 29.3 (9.6) | 26.5 (9.1) | 29.0 (7.5) | F(2,40) < 1, p = .646 |
| Children | 32.4 (10.8) | 30.9 (9.9) | 24.8 (6.6) | F(2,55) = 3.54, p = .036* |
| Adolescents | 31.5 (12.4) | 28.5 (11.4) | 25.9 (12.9) | F(2,46) < 1, p = .413 |
| Basilar artery | ||||
| Infants | 42.7 (11.0) | 40.0 (12.2) | 43.8 (9.4) | F(2,48) < 1, p = .521 |
| Children | 53.8 (8.4) | 44.4 (10.9) | 41.3 (8.9) | F(2,56) = 8.50, p = .001 |
| Adolescents | 48.3 (8.2) | 38.1 (10.3) | 32.9 (6.9) | F(2,49) = 16.09, p < .001 |
- LA, low altitude; MA, moderate altitude; HA, high altitude. *Owing to non-linear age effects in the majority of groups and vessels (see e.g. Figure 1b, for basilar artery), univariate analysis of variance was performed within each age group. Correcting the statistical significance threshold for the increased number of comparisons necessary to analyse age groups separately (p = .004) would leave all significant values intact except for the altitude effect on posterior cerebral artery velocity in children. p-values given in bold are significant at < .05.
Cognition and behaviour
The mean values and all statistical results related to cognitive and behavioural measures are provided in Table 5.
| Altitude group | Group difference | ||||||
|---|---|---|---|---|---|---|---|
| LA | MA | HA | |||||
| Infants | BINSLow:Mod-High Risk | – | – | 7:2 | 11:7 | 21:3 | p = .13 |
| IDQ-RMedian [IQR] | Approach | – | 6.1 [1.5] | 5.8 [1.6] | 6.1 [1.3] | p = .631 | |
| Vocal reactivity | – | 5.1 [1.0] | 5.0 [1.1] | 4.5 [1.6] | p = .290 | ||
| High intensity pleasure | – | 5.6 [1.3] | 5.5 [1.7] | 6.0 [1.3] | p = .290 | ||
| Smile & laughter | – | 5.7 [0.9] | 5.7 [1.1] | 5.2 [1.3] | p = .054 | ||
| Activity level | – | 5.0 [0.7] | 4.6 [1.3] | 4.4 [1.5] | p = .302 | ||
| Perceptual sensitivity | – | 5.7 [1.4] | 4.7 [1.3] | 4.7 [1.9] | p = .128 | ||
| Sadness | – | 4.8 [1.8] | 3.7 [1.4] | 3.4 [1.2] | p = .111 | ||
| Distress to limitations | – | 4.7 [1.5] | 3.8 [0.8] | 4.0 [1.6] | χ2(2) = 10.1, p = .006 | ||
| Fear | – | 3.6 [1.3] | 3.0 [1.4] | 2.8 [0.7] | χ2(2) = 6.4, p = .040 | ||
| Falling reactivity/ Rate recovery from distress | – | 4.9 [1.1] | 4.8 [1.0] | 5.1 [1.2] | p = .587 | ||
| Low intensity pleasure | – | 4.6 [1.5] | 5.4 [0.8] | 5.2 [1.5] | p = .582 | ||
| Cuddliness | – | 5.8 [0.4] | 5.9 [0.8] | 5.7 [1.6] | p = .738 | ||
| Duration of orienting | – | 3.6 [2.6] | 4.5 [0.5] | 3.4 [1.7] | p = .067 | ||
| Soothability | – | 5.1 [1.9] | 5.0 [0.5] | 5.2 [0.6] | p = .919 | ||
| Object retrievalMedian [IQR] | % trials toy retrieved | – | 84.7 [35.2] | 50 [44.8] | 73.9 [38.9] | p = .215 | |
| Retrieval Time (sec) | – | 4.6 [4.2] | 5.1 [2.9] | 4.7 [4.7] | p = .960 | ||
| Children and adolescents | WISC-IV (Sp) Mean (SD) | Block design | Children | 8.5 (3.7) | 10.8 (3.3) | 11.6 (2.5) | Altitude: F(2,108) = 2.73, p = .070Age: F(1,108) < 1, p = .948Altitude × Age: F(2,108) = 2.3, p = .105 |
| Adolescents | 10.1 (3.4) | 10.6 (2.9) | 10.1 (3.0) | ||||
| Digit span – forwards | Children | 12.4 (3.0) | 11.6 (3.5) | 12.7 (3.6) | Altitude: F(2,108) = 1.23, p = .296Age: F(1,108) = 2.81, p = .096Altitude × Age: F(2,108) = 2.23, p = .070 | ||
| Adolescents | 9.6 (2.7) | 12.5 (3.7) | 11.4 (3.5) | ||||
| Digit span – backwards | Children | 10.3 (3.3) | 10.6 (3.0) | 11.3 (3.5) | Altitude: F(2,108) = 2.43, p = .093Age: F(1,108) < .01, p = .967Altitude × Age: F(2,108) = 1.73, p = .182 | ||
| Adolescents | 9.4 (2.5 | 12.1 (3.3) | 10.8 (2.4) | ||||
| Processing speed index | Children | 103.3 (12.3) | 100.3 (13.0) | 96.7 (12.5) | Altitude: F(2,108) = 3.29, p = .041Age: F(1,108) < 1, p = .540Altitude × Age: F(2,108) < 1, p = .868 | ||
| Adolescents | 103.3 (14.4) | 98.9 (15.8) | 93.1 (17.2) | ||||
| WRAVMAMean (SD) | Matching | Children | 89.9 (12.0) | 98.1 (17.8) | 98.7 (16.8) | Altitude: F(2,107) = 3.22, p = .044Age: F(1,107) < 1, p = .414Altitude × Age: F(2,107) < 1, p = .955 | |
| Adolescents | 92.8 (18.7) | 99.3 (15.6) | 102.1 (14.4) | ||||
| Drawing | Children | 108.0 (11.9) | 115.7 (14.4) | 108.3 (14.3) | Altitude: F(2,108) < 1, p = .713Age: F(1,108) < 1, p = .870Altitude × Age: F(2,108) = 2.35, p = .100 | ||
| Adolescents | 110.2 (12.3) | 107.6 (14.1) | 113.0 (13.1) | ||||
| Pegboard (dominant hand) | Children | 102.4 (14.3) | 103.8 (10.8) | 102.4 (15.3) | Altitude: F(2,109) < 1, p = .926Age: F(1,109) = 1.04, p = .308Altitude × Age: F(2,109) < 1, p = .736 | ||
| Adolescents | 105.8 (12.0) | 103.7 (10.9) | 107.8 (15.3) | ||||
| Finger tappingMean (SD) | Dominant hand | Children | 40.4 (11.8) | 34.5 (10.8) | 31.9 (7.9) | Altitude: F(2,108) = 3.42, p = .033Age: F(1,108) = 69.79, p < .001*Altitude × Age: F(2,108) = 1.73, p = .182 | |
| Adolescents | 52.1 (12.9) | 55.3 (8.3) | 48.7 (9.1) | ||||
| SDQ – Parent#Mean (SD) | Emotional symptoms | Children | 2.3 (1.9) | 2.4 (1.9) | 2.1 (1.4) | Altitude: F(2,103) < 1, p = .690Age: F(1,103) < 1, p = .439Altitude × Age: F(2,103) < 1, p = .823 | |
| Adolescents | 2.2 (2.1) | 2.9 (1.8) | 2.6 (2.2) | ||||
| Conduct problems | Children | 2.3 (2.2) | 1.9 (1.6) | 2.4 (1.7) | Altitude: F(2,103) < 1, p = .883Age: F(1,103) = 2.4, p = .123Altitude × Age: F(2,103) = 1.6, p = .214 | ||
| Adolescents | 2.4 (1.4) | 3.3 (2.2) | 2.5 (2.0) | ||||
| Hyperactivity-Inattention | Children | 5.2 (2.7) | 4.1 (2.2) | 4.4 (2.5) | Altitude: F(2,103) < 1, p = .686Age: F(1,103) = 1.2, p = .276Altitude × Age: F(2,103) = 1.2, p = .302 | ||
| Adolescents | 3.9 (2.4) | 4.6 (1.9) | 3.7 (2.8) | ||||
| Peer problems | Children | 1.6 (1.5) | 2.3 (1.7) | 2.1 (1.5) | Altitude: F(2,103) < 1, p = .556Age: F(1,103) < 1, p = .743Altitude × Age: F(2,103) < 1, p = .710 | ||
| Adolescents | 1.8 (1.8) | 1.8 (1.4) | 2.1 (1.5) | ||||
| Prosocial behavior | Children | 8.2 (1.8) | 8.0 (1.6) | 8.2 (2.2) | Altitude: F(2,103) < 1, p = .669Age: F(1,103) < 1, p = .370Altitude × Age: F(2,103) < 1, p = .910 | ||
| Adolescents | 7.9 (1.8) | 7.5 (1.8) | 8.0 (1.9) | ||||
| Total difficulties | Children | 11.4 (6.5) | 10.6 (4.3) | 10.9 (5.1) | Altitude: F(2,103) < 1, p = .813Age: F(1,103) < 1, p = .829Altitude × Age: F(2,103) < 1, p = .507 | ||
| Adolescents | 10.3 (6.1) | 12.6 (5.6) | 10.8 (6.2) | ||||
- LA, low altitude; MA, moderate altitude; HA, high altitude. * Not age-normative scores, so a large age effect is inevitable. # Adolescents (n = 48) also completed the self-report version of the Strengths and Difficulties Questionnaire (SDQ). The correlation coefficients between adolescent self-report and parental report scores are: Emotional Symptoms –R .571, Conduct Problems –R .632, Hyperactivity / Inattention –R .648, Peer Problems –R .603, Prosocial –R .546, Total Difficulties –R .655: all P < .001. In general, parent-reported scores were slightly higher than American averages (http://www.sdqinfo.com/b1.html), but there were no significant altitude effects. p-values given in bold are significant at < .05.
(i) Infants
Neurodevelopment in infants was assessed using the Bayley Infant Neurodevelopmental Screen and found to be normal. Accuracy and speed of search on the Object Retrieval task was also comparable between altitude groups. Consistent with this, significant group differences were found in only 2/14 domains assessed by the Infant Development Questionnaire, namely ‘Distress to Limitations’ (excessive fussing/crying when restrained; χ2(2) = 10.1, p = .006) and ‘Fear’ (χ2(2) = 6.4, p = .040), but LA infants had significantly more distress to limitation than both MA (p = .007) and HA (p = .039) infants, and more fear than MA and HA infants (both p = .009). It is not certain why our LA infants differed only in these two domains. There were no altitude-group differences in terms of a number of variables that might influence parent–child interaction and thus infant emotional development, including experience of breast-feeding (p = .140) and number of older siblings (p = .708), and none of the infants in this study had been carried (swaddled) on their mother’s back as may be observed in Quechua/Aymaran culture. However, significantly more LA infants were sleeping in a cot in a room on their own (4/9, 44.4%), compared to our MA and HA infants, who tended to sleep in the same room as their parents (2/19, 10.5%; 2/24, 8.3%, respectively, slept on their own) (p = .029). In addition, whereas approximately one-third of infants in each of the LA and MA groups were cared for in a nursery for one or more days a week, this was not the case for any of the HA infants, none of whom attended a nursery (p = .021). It is possible, therefore, that subtle differences in child-rearing practice may account for the group differences on the IDQ-R.
(ii) Children and adolescents
Behaviour (Strengths & Difficulties Questionnaire) was within normal limits in all altitude and age groups (Table 5). The altitude groups were equally proficient at copying 2-D designs using pencil and paper (‘Drawing’; WRAVMA), and at recalling digits forwards and backwards (working memory; WISC-IV). HA children and adolescents were particularly adept at ‘Matching’ (WRAVMA) rotated and flipped shapes to a target, a task requiring visuo-spatial and working-memory skill, F(2,107) = 3.22, p = .044 (LA < HA p = .048). By contrast, finger-tapping rate, a measure of motor speed, decreased with increasing altitude, F(2,108) = 3.42, p = .033. The Processing Speed Index (PSI: WISC) is derived from tests of fast matching and copying figures, and thus reflects both cognitive and motor (psychomotor) speed. PSI scores also decreased with increasing altitude (Figure 1c), F(2,108) = 3.29, p = .041; LA > HA p = .030; although both components of the PSI score (Symbol Search and Coding subtests) showed a stepwise decline in scores with increasing altitude, this was only significant for Symbol Search, F(2,107) = 3.81, p = .025.
Discussion
The results of this study suggest a pattern of successful neurocognitive adaptation in children growing up at high altitude in Bolivia. Infant neurodevelopment, search skill and behaviour were explored, yielding comparable results across altitude groups. Earlier infant studies described no (Hass, 1976) or only subtle (Saco-Pollitt, 1981; Baker, 1960; Bender et al., 1994)1 effects of altitude on infant development. With increasing altitude, only a minor trade-off in subtly reduced psychomotor speed was found in children and adolescents, consistent with the reduction in CBFV. We suggest that this may reflect a gradual slowing of underlying brain activity, an adaptation that appears to be more responsive to chronic hypoxic exposure than to genetic ancestry in these children of mixed-ethnic background.
Reduced psychomotor speed in children and adolescents suggests two possibilities. It could be driven by physiological constraints on brain function, or by a more conscious cognitive strategy of the kind that leads to a speed–accuracy trade-off. We conclude that the former is more likely, owing to indirect evidence of altered brain physiology (reduction in CBFV). Slowing of nervous system function has previously been observed in event-related potential studies under hypobaric hypoxia (Wesensten, Crowley, Balkin, Kamimori, Iwanyk & Devine, 1993) and with cognitive testing of adult mountaineers (Kennedy et al., 1989; Kramer et al., 1993), but may actually help to prevent errors of judgement (Kennedy, Dunlap, Banderet, Smith & Houston, 1989). To our knowledge, there are no studies of psychomotor speed in the native high-altitude living adult. However, this study demonstrates that psychomotor slowing is present by mid-childhood, and remains into adolescence, indicating that it may be a lasting feature of high-altitude neurocognitive function. It should nevertheless be emphasized that reduction in psychomotor speed was subtle, placing HA adolescents only two-thirds of a standard deviation (for PSI, the normative SD = 15) below the score obtained by LA adolescents, and within the clinically normal range (>80).
It is interesting that psychomotor speed was responsive to only a small reduction in haemoglobin oxygen saturation. Only a few studies have published SpO2 reference values by age, and typically focus on very high-altitude groups (∼4000m) (e.g. Gamponia, Babaali, Yugar & Gilman, 1998; Beall, Almasy, Blangero, Williams-Blangero, Brittenham & Strohl, 1999). These studies, nevertheless, indicate a slight increase in SpO2 with increasing age, which is consistent with the findings in the present cohort. Specifically, we show that significant age-related increases in SpO2 occur between infancy and mid-childhood, and only at the highest altitude tested (3700 m). Similar reference values for carbon dioxide, the chief determinant of respiratory drive in healthy individuals, are lacking, but we demonstrate here that ETCO2 follows a similar trajectory to SpO2, in that it decreases with altitude and increases with age. Interestingly, however, a significant increase in ETCO2 between infancy and mid-childhood occurred at moderate- (2500 m: p = .005) and at high- (3700 m: p < .001) altitude locations. Together, these data suggest that children gradually adapt to hypoxic conditions, and that the threshold for such adaptation may be a relatively small reduction in atmospheric oxygen. An important limitation of the present study is the lack of haematological measures of oxygen-carrying capacity, such as haemoglobin or haematocrit. It was felt that anxiety about blood tests may have adversely influenced the children’s willingness to participate in the study, and/or their performance on cognitive tasks (our focus of interest). It is anticipated that the development of non-invasive methods for the determination of haematocrit will make this feasible in future studies.
The brain is particularly sensitive to oxygen supply and may adapt to hypoxia and hypocapnia (lowered ETCO2) by altering the responsiveness of cerebral circulation (Norcliffe et al., 2005), with implications for neurocognitive function. Cerebral blood flow (CBF) increases after acute exposure to isocapnic hypoxia (Cohen, Alexander, Smith, Reivich & Wollman 1967). In sea-level residents acclimatized to altitude, CBF and cerebral metabolic rate for oxygen and glucose do not change on return to sea level (Møller, Paulson, Hornbein, Colier, Paulson, Roach, Holm & Knudsen 2002) but the available data suggest that the cerebral metabolic rate for glucose is reduced in Quecha natives living in the Andes (Hochachka, Clark, Brown, Stanley, Stone & Nickles, 1994) but not in Tibetan Sherpas (Hochachka, Clark, Monge, Stanley, Brown, Stone, Nickles & Holden 1996). As the techniques for measuring CBF and cerebral metabolic rate are invasive there are no data for children.
Assuming that blood vessel diameter remains constant, CBFV, assessed by non-invasive transcranial Doppler sonography, is a measure of blood flow to the brain, which is typically tightly coupled to metabolic rate at normal levels of SpO2 (Brugneaux et al., 2007). By adolescence, basilar artery CBFV in the MA and HA groups was lower compared to the sea-level standard (Bode & Wais, 1988) by approximately 1.0 and 1.5 standard deviations, respectively. Lower CBFV in children and adolescents at higher altitudes is consistent with the finding of reduced brain metabolism in native high-altitude adults (Hochachka et al., 1994). There is evidence for heterogeneity of blood flow distribution in response to acute isocapnic hypoxia, with the least increase in the right occipital lobe (Binks, Cunningham, Adams & Banzett, 2008); the distribution of our CBFV data with a particular reduction in basilar CBFV suggests that this may also be the case in chronic hypoxia, as has been demonstrated in the foetus (Dubiel, Gunnarsson & Gudmundsson, 2002). Other mechanisms, such as dilatation of the large vessels without metabolic change (Steen, Langston, Ogg, Manci, Mulher & Wang, 1998), increased haemoglobin and/or increased capillary density induced by chronic hypoxia (Xu & Lamanna, 2006), may improve brain perfusion and also explain the presence of reduced rather than increased CBFV in our children. Nevertheless, with no increased incidence of neurological signs in the high-altitude groups, we interpret our CBFV data as reflecting adaptation rather than pathology. Interestingly, MA and HA groups demonstrated equivalent reduction in CBFV, suggesting that, as for ETCO2, such changes may occur at comparatively low altitude (MA = ∼2500m); it is noteworthy that many countries, including the United States, have populations living above this altitude. In addition, reduction in CBFV in these groups was only found from childhood, consistent with the timing of physiological change in related body systems. For example, a recent review of cardio-pulmonary function at high altitude documented delayed post-natal change from foetal to infant morphology and function, with the greatest change occurring after age 5 years (Penaloza & Arias-Stella, 2007). This is also compatible with the presence of psychomotor slowing in childhood (defined as 6–10 years in the present study). Although the lack of younger age groups, for example preschool, precludes the conclusion that psychomotor slowing emerges in mid-childhood, the lack of group difference for time-scores in the Object Retrieval task in our infants provides some support for this hypothesis. In general, we interpret our data as supporting the hypothesis that physiological adaptation constrains only certain cognitive functions, but that this cost is apparently minor and perhaps adaptive in and of itself: an interpretation of slower facilitating surer at high altitude (cf. Kennedy et al., 1989) may even be reflected in the higher ‘Matching’ scores in the HA groups.
Any mechanism relating hypoxia and cerebrovascular change to reduction in psychomotor speed is likely to be complex. A lack of correlation with SpO2 may be expected, as oxygenation of the brain may be different from that of the periphery and is determined by uptake as well as by delivery of oxygen. Such a relationship may only appear with greater degree of hypoxia: even at 78% SpO2 only subtle cognitive deficit is found (Noble, Jones & Davies, 1993). Faster performance on cognitive tests is associated with higher mean CBFV in anterior cerebral vessels in sea-level adults (Schuepbach, Bader, Hell & Baumgartner, 2004; Duschek, Schuepbck & Schandry, 2008). However, in the present study no correlation between psychomotor speed and CBFV was found.
It is often assumed that populations living at high altitude are protected by their greater proportion of hypoxia-adapted ancestry, but this may not be the case for many of today’s high-altitude dwellers, at least not for those of mixed-ethnic origin who were part of our cohort. Indeed, in South America, substantial migration between lowlands and Andean highlands has occurred: during the pre-contact period (Weinstein, 2007), followed by European admixture resulting from Spanish colonization and more recent immigration resulting from macroeconomic expansion (Ehrlich & Canavire Bacarreza, 2006). It is reassuring, therefore, that chronic hypoxic exposure rather than an Andean genetic inheritance appears to protect neurocognitive development at high altitude. The result is subtle psychomotor slowing, which we interpret as an adaptive rather than a deficient trait, perhaps enabling accuracy of mental activity in hypoxic conditions.
Footnotes
Acknowledgements
BoCLA 2006 (Bolivian Children Living at Altitude) was funded by grants from The British Academy and The Gerald Kerkut Trust to Alexandra M. Hogan, who initiated BoCLA whilst based at the University of Southampton. We are indebted to Univalle (Cochabamba, La Paz) and USPA (Santa Cruz) universities, and to their student volunteers. In particular, we thank Patty Mendieta (Univalle, La Paz), and Mirta W. Handal, Claudia Arteaga and Claudia Ortiz (UPSA, Santa Cruz) for their contribution to recruitment and local co-ordination, and participating schools (Colegio Británico (Santa Cruz), Colegio Anglo-Americano (Cochabamba & La Paz)) for allowing us access to their children and for providing rooms for testing. Our gratitude extends to Melanie Marshall, Cranfield University, UK, for training on the capnography and oximetry equipment, and to Maria Gartstein, Washington State University, USA, for her help with the IDQ-R. We also acknowledge the assistance of the Australian Medical Bioinformatics Resource (a National Health and Medical Research Council of Australia Medical Bioinformatics Genomics Proteomics Program).
