Protected fetal growth in highland Andeans

Material Information

Protected fetal growth in highland Andeans the role of maternal antioxidant status and oxidate stress
Julian, Colleen Glyde
Publication Date:
Physical Description:
xiv, 203 leaves : ; 28 cm


Subjects / Keywords:
Fetus -- Growth ( lcsh )
Reproduction -- Effect of altitude on -- Andes Region ( lcsh )
Preeclampsia -- Andes Region ( lcsh )
Oxidative stress -- Andes Region ( lcsh )
Antioxidants ( lcsh )
Antioxidants ( fast )
Fetus -- Growth ( fast )
Oxidative stress ( fast )
Preeclampsia ( fast )
Reproduction -- Effect of altitude on ( fast )
Andes Region ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 163-203).
Statement of Responsibility:
by Colleen Glyde Julian.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
227207134 ( OCLC )
LD1193.L566 2007d J84 ( lcc )


This item has the following downloads:

Full Text
Colleen Glyde Julian
B.S., University of Colorado, Boulder, 1997 M.S., University of Colorado, Boulder, 2001
A thesis submitted to the
University of Colorado at Denver and Health Sciences Center in partial fulfillment of the requirements for the degree of Doctor of Philosophy Health and Behavioral Science 2007

by Colleen Glyde Julian All rights reserved.

This thesis for the Doctor of Philosophy degree by
Colleen Glyde Julian has been approved by

Lorna G. Moore, PhD f\

David P. Tracer, PhD
n-2-S- zd

Julian, Colleen Glyde (Ph.D., Health and Behavioral Science)
Protected fetal growth in highland Andeans: the role of maternal antioxidant status and oxidative stress
Thesis directed by Professor Lorna G. Moore
Permanent high-altitude (HA) residence compromises fetal growth and increases the incidence of preeclampsia (PE). Highland ancestry groups (i.e. Andeans) experience a lesser degree of birth weight reduction than lowland ancestry groups (i.e. Europeans). Oxidative stress and reduced antioxidant capacity have been implicated in the maternal vascular dysfunction hallmark to PE and reduced fetal growth at low altitude (LA). Since hypoxia increases oxidative stress, we hypothesized 1) that altered redox status contributes to reduced fetal growth and PE at HA by causing endothelial damage and reducing uterine artery (UA) blood flow and 2) that enhanced antioxidant activity contributes to the protective effect of Andean ancestry. Studies (n=310) were conducted in non-pregnant (NP) or pregnant (20w, 36w) women of Andean or European ancestry residing at HA (3600- 4100m, La Paz) or LA (400m, Santa Cruz) in Bolivia. Maternal oxidative stress (8-iso-PGF2a), antioxidant status (superoxide dismutase [SOD]; catalase) and UA blood flow were assessed at each time point. Fetal biometry was assessed by ultrasound at 20w and 36w. Erythrocyte SOD and catalase activity were quantified by spectrophotometry. Plasma 8-iso-PGF2a was measured by mass-spectrometry. Medical records were reviewed to obtain newborn information. At HA Andeans had greater UA blood flows at all time points, tended to be heavier at birth and to be SGA less often than Europeans. Catalase and SOD (36w) were lower during pregnancy in European than Andeans at HA. Catalase tended to be lower in Europeans who delivered SGA versus non-SGA infants (20w). SOD was lower in mothers of SGA infants in all women and at HA alone (20w). 8-iso-PGF2a reduced UA blood flow at HA in all women or in Europeans considered alone. Our data support the hypothesis that maternal antioxidant status has a protective effect on UA blood flow, blood pressure and fetal growth at HA. Further, our results strongly suggest that augmented antioxidant activity contributes to the preservation of fetal growth at altitude in Andean populations.
This abstract accurately represents the content of the candidates thesis. I recommend its publication.

To my husband, Pete, for always encouraging me to carve my own trails; to my parents, for giving me legs both literally and metaphorically to take me down the routes I have chosen; and to my son, Wade, for reminding me of the importance of living with purpose.

I would first like to thank all of the women who so graciously participated in these studies. Likewise, I am indebted to everyone at the Bolivian Institute of Altitude Biology (IBBA) in La Paz, Bolivia, but particularly to Dr. Enrique Vargas, Dra. Wilma Tellez, Dr. Armando Rodriguez, Cristina Gonzalez and Ana Maria Alarcon. I would like to thank Dr. Carmelo Rodriguez, Dr. Henry Yamashiro, Dra. Pilar Garcia, Dra. Jessica Pardo, Dra. Jessika Schayman, Marta Cardenas, Julio Roca, Ruliana Arce, Maria Elena Gira and Ida Gonzales of Clinica Sirani in Santa Cruz de la Sierra, Bolivia and Dra. Miriam Lopez at Clinica del Sur in La Paz for all of your contributions. To UCDHSC, thank you for providing me with professors that opened so many doors for me (on occasion, literally) and treated me with respect despite my sometimes clumsy curiosity (Dr. Lorna Moore,
Dr. David Tracer, Dr. Susan Dreisbach, Dr. Steve Koesterand Dr. Craig Janes). Thank you to Abby Fitch and Jennifer Hageman for their enduring technical support, to the folks at the University of Colorado Altitude Research Center both past and present to Dr. Vaughn Browne for assisting with the collection of samples and data in La Paz and El Alto, Lucilla Poston for assisting me in the development phase of this project, Jalena and Jost Klawitter and Uwe Christians at the CORE Lab and to Dr. Joe McCord and Swapan Bose for allowing me to invade their laboratory space. A mojito-filled thank you to Cohort 10 in particular to Megan Wilson, thank you for your friendship and advice. To my dissertation committee (Dr. Lorna Moore, Dr. David Tracer, Dr. Joe McCord and Dr.
Henry Galan), thank you. In particular, I would like to say thank you to Lorna, for investing your time, for your guidance and for offering me so many wonderful opportunities including tagging on to the mamacita project. To the acquaintances that taught me things cant be learned in books, but that somehow helped to steer the course of my motivation, thank you. I would also like to acknowledge the American Heart Association (pre-doctorai fellowship: #0610129Z) and the National Science Foundation (dissertation improvement grant: #BCS-064719) for their financial support of this project.

1. Introduction..............................................................1
1.1 Project Summary and Research Statement....................................1
1.2 Significance..............................................................4
1.3 Overall Hypothesis and Specific Aims......................................7
2. Theoretical and Empirical Background....................................11
2.1 Reduced Fetal Growth and Preeclampsia: Definitions, Epidemiology and
2.1.1 Measuring Fetal Growth..................................................11 Birth Weight.............................................................11 Fetal Biometry...........................................................11 SGA or Intrauterine Growth Restriction (IUGR)............................12
2.1.2 Epidemiology.............................................................13 Reduced Fetal Growth: SGA................................................13 Preeclampsia.............................................................14
2.1.3 Etiology.................................................................16
2.2 Ecological Context: Bolivia..............................................16
2.2.1 A Highland Ecology.......................................................16
2.2.2 Socioeconomic, Health and Nutritional Status.............................18

2.3 High Altitude: The Challenges and History of Permanent Residence at
High Altitude...........................................................20
2.3.1 Physiological Challenges Posed by High-Altitude Exposure................21
2.3.2 Historical Accounts.....................................................25
2.3.3 Peopling of Highland Regions............................................27 Himalayas...............................................................27 Andean Plateau..........................................................28
2.4 Adaptation to High Altitude.............................................29
2.4.1 Cultural and Behavioral Adaptations to High Altitude....................30
2.4.2 Biological Adaptations to High Altitude.................................31 Ventilation and Respiratory Volumes.....................................31 Pulmonary Circulation...................................................34 Hematological Adaptation................................................36
2.5 A Tale of Two Pregnancies...............................................38
2.5.1 Extent and Variation of Birth Weight and Fetal Growth Reduction
with Hypoxia............................................................38 Hypoxia-Associated Fetal Growth and Birth Weight Reduction..............38
2.5 1.2 Variability in Hypoxia-Associated Fetal Growth and Birth Weight
2.5.2 Human Pregnancy.........................................................43
2.5.3 Maternal Adaptation to Normoxic Pregnancy...............................44 Establishment of Placental Perfusion....................................44 Systemic Maternal Adaptation to Pregnancy...............................45

2.5.4 Pregnancy in Chronic Hypoxia..........................................47
2.6 Oxidative Stress.......................................................50
2.6.1 Definition of Oxidants and Oxidative Stress............................50
2.6.2 Generation of Oxidant Species..........................................51 Mitochondrial Respiration..............................................52 NAD(P)H Oxidases.......................................................52 Xanthine Oxidase.......................................................53
2.6.3 Antioxidant Defenses and Assessment of Oxidative Stress................53 Enzymatic Antioxidants.................................................54 Assessment of Oxidative Stress.........................................55
2.6.4 Hypoxia-Induced Oxidative Stress.......................................56
2.6.5 Oxidative Stress: A Usual Suspect......................................58 Evidence of Oxidative Stress in Preeclampsia and SGA...................58 Evidence of Diminished Antioxidant Capacity in Preeclampsia
and SGA................................................................59
2.6.6 ROS and Vascular Function..............................................63
2.6.7 ROS and the Regulation of HIF1.........................................65
3. Research Design and Methodology........................................67
3.1 Research Setting.......................................................67
3.2 Subjects and Subject Recruitment.......................................67
3.3 Data Collection........................................................68
3.3.1 Maternal Attributes....................................................69
3.3.2 Fetal and Neonatal Characteristics.....................................70

3.3.3 Blood Sample Collection...................................................71 Assessment of Oxidative Stress............................................71 Assessment of Antioxidant Status.........................................72 Endothelial Damage and Vasoconstriction...................................74
3.3.4 Uterine Artery Blood Flow Characteristics.................................75
3.3.5 Ethical Considerations....................................................77
3.3.6 Statistical analyses......................................................78
4. Primary Research Findings.................................................80
4.1 Maternal Attributes.....................................................82
4.2 Fetal Biometry............................................................86
4.3. Delivery and Newborn Characteristics......................................86
4.4 U.terine Artery Blood Flow Characteristics...............................93
4.5 Oxidative Status and Markers of Endothelial Damage......................95
4.5.1 Isoprostanes (8-iso-PGF2)................................................95
4.5.2 Leptin....................................................................98
4.6 Antioxidant capacity: effects of pregnancy, altitude and ancestry.........99
4.6.1 Catalase..................................................................99
4.6.2 Superoxide dismutase (SOD)...............................................101
4.7 Establishing a Link: Oxidative Stress, Antioxidant Capacity and
Reduced Fetal Growth....................................................106
4.7.1 Endothelin...............................................................106
4.7.2 Nitric Oxide Metabolites.................................................107

4.7.3 Endothelial Damage, Antioxidants and Oxidative Stress.................107
4.7.4 Redox Status, Endothelial Damage and Mean Arterial
4.7.5 Redox Status, Endothelial Damage and Uterine Artery
Blood Flow.............................................................110
4.7.6 Fetal Growth and Oxidative Stress......................................116
4.7.7 Fetal Growth and Antioxidant Status....................................120
4.7.8 Fetal Growth, Endothelin and Nitric Oxide Metabolites..................122
4.8 Summary................................................................123
5. Discussion.............................................................124
5.1 Maternal Attributes and Fetal Growth...................................125
5.2 Uterine Artery Blood Flow Characteristics..............................129
5.3 Markers of Oxidative Stress, Antioxidant Status and Endothelial
5.3.1 Endothelin.............................................................132
5.3.2 NO metabolites.........................................................134
5.3.3 Leptin.................................................................137
5.3.4 Isoprostanes (8-isoPGF2)..............................................139
5.3.5 Endogenous Antioxidant Activity: Catalase and SOD......................141
5.3.6 Summary................................................................149

A. English Consent Form..............................................152
B. Spanish Consent Form..............................................158

1.1 Hypothesis schema............................................................8
2.1 Oxygen cascade..............................................................23
4.1 Uterine artery blood flow (UAVF) at low and high altitude...................94
4.2 Maternal catalase (CAT) activity at low and high altitude..................100
4.3 Maternal SOD activity at low and high altitude.............................102
4.4 Maternal isoprostanes and uterine artery (UA) pulsatility index............111
4.5 Maternal isoprostanes and uterine artery (UA) resistance index.............111
4.6 Maternal isoprostanes and uterine artery (UA) diameter.....................112
4.7 Maternal isoprostanes and uterine artery (UA) blood flow...................112
4.8 Maternal endothelin (EDN1) and uterine artery (UA) pulsatility index.......113
4.9 Maternal endothelin (EDN1) and uterine artery (UA) resistance index........113
4.10 Maternal endothelin (EDN1) and uterine artery (UA) flow velocity...........114
4.11 Maternal endothelin (EDN1) and uterine artery (UA) diameter................114
4.12 Maternal endothelin (EDN1) and uterine artery (UA) volumetric blood flow...115
4.13 Maternal isoprostanes and umbilical artery S/D ratios......................118
4.14 Maternal SOD levels in SGA and normal pregnancy............................121
4.15 Mechanism by which superoxide (02*) decreases the bioavailability
of nitric oxide (NO')......................................................145

3.1 Subject numbers............................................................68
4.1 Maternal attributes........................................................84
4.2 Fetal biometry.............................................................89
4.3 Delivery and newborn characteristics.......................................91
4.4 Stepwise multiple linear regression of maternal and environmental
independent variables (x) in relation to birthweight (y)..................92
4.5 Uterine artery blood flow characteristics..................................96
4.6 Markers of oxidative stress and endothelial damage........................104

1. Introduction
1.1 Project Summary and Research Statement
Given the evolutionary consequence of maternal and/or fetal mortality, pregnancy and the perinatal period are sensitive time points for examining adaptation. High-altitude (>2500 m) exposure presents a great physiological challenge, primarily due to the need to maintain adequate oxygen delivery despite a hypoxic environment. The chronic hypoxia of high-altitude residence presents an even greater challenge during pregnancy, when vast maternal physiological adaptations are required to meet both maternal and fetal oxygen and nutrient demands. Over the past 20 years, studies have consistently demonstrated that high altitude compromises fetal growth, decreasing birth weight an average of 121g per 1000m elevation rise, and increases the incidence of preeclampsia approximately 3-fold.1 The effect of chronic hypoxia on birth weight is as great or greater than that associated with low maternal weight gain during pregnancy, smoking or primiparity1, and is primarily due to reduced fetal growth during the third trimester2,3 rather than preterm delivery.4 The increased incidence of preeclampsia contributes to, but does not explain, the magnitude of hypoxia-associated birth weight reduction.1,5'7 Likewise, differences in socioeconomic status do not explain the reductive effect of altitude on birth weight.3,8 Thus, hypoxia acts independently to reduce fetal growth.
All populations studied to date demonstrate an altitude-associated reduction in birth weight, however the magnitude of the effect varies considerably.9,10 Specifically, populations with an extensive history of high-altitude residence are protected against altitude-associated reductions in birth weight relative to newcomer groups.11'15 For instance, Andeans and

Tibetans, having resided at high altitudes for more than 10,000 and 20,000 years respectively, show one-third the birth weight reduction of European and Han Chinese" populations who have only recently established permanent residence in high-altitude regions (Europeans: <400 years in South America and <150 years in North America; Han: -60 years in western China.10,15 Differences in health care access, socioeconomic indices or maternal stature do not appear to explain the observed differences in birth weight between these populations;13 this suggests that population-specific attributes, likely genetic and/or behavioral characteristics, are responsible for the protection of fetal growth afforded by multigenerational high-altitude ancestry in these populations.
The physiological mechanisms by which hypoxia acts to restrict fetal growth are not well understood, however studies conducted over the past 20-years demonstrate that hypoxia interferes with maternal vascular adaptation to pregnancy, diminishing uterine artery diameter, uteroplacental blood flow16,17 and subsequently, the delivery of nutrients and oxygen required for normal fetal growth.16,18'20 In accordance with the observation that multigenerational high-altitude ancestry protects fetal growth, recent studies have demonstrated that uterine artery blood flow near term is three-fold greater and a larger proportion of lower extremity blood flow is directed towards the uteroplacental circulation in Andean versus European women living at high altitude (a3600m).21 The protective effect of high-altitude ancestry on fetal growth is likely due to differences in the response of the maternal vasculature to pregnancy rather than variation in oxygenation since no ancestry-dependent differences in arterial oxygen content are apparent.2124 The mechanisms by which ancestry-

associated variation in maternal vascular response to high-altitude pregnancy arise have yet to be determined.
A growing body of literature calls attention to the potential significance of oxidative stress in the maternal vascular associated with preeclampsia and small-for-gestational age births (SGA).1 It has long been suspected that uteroplacental ischemia (i.e. inadequate blood flow) results in the release of one or more toxic factors from the feto-placental circulation causing maternal vascular dysfunction and hypertension associated with preeclampsia;25 what exactly this substance is and how it causes maternal endothelial dysfunction has, until recently, not been well understood. Increasing evidence suggests that the toxic" factor(s) may be reactive oxygen species (ROS), leading to oxidative stress and endothelial damage associated with altered maternal vascular function and impaired uteroplacental oxygen delivery.2627
While still the subject of some debate, studies have demonstrated that hypoxia, either by reduced oxygen tension and/or impaired perfusion, increases the generation of ROS and causes oxidative stress.28'32 Recent research suggests that some constituent or product, likely ROS, of the endothelium participates in the inhibition of vasodilation in hypoxia.33 ROS are thought to compromise the vascular endothelium, in part, by decreasing the relaxing capacity of resistance vessels in response to endothelium-dependant mechanisms in a manner similar to that observed in preeclampsia at low altitude.26,34'36 Thus, if the production of ROS is enhanced at high altitude it seems likely that ROS may play a critical role in the

altitude-associated compromise of vascular function during pregnancy and, as a result, contribute to the 3-fold increase of preeclampsia and SGA at high altitude.
This study seeks to determine whether oxidative stress and/or reduced antioxidant capacity contributes to the increased incidence of preeclampsia and SGA at high altitude. In addition, it examines whether differences in oxidative stress and/or antioxidant capacity (i.e. redox status) are associated with the protection afforded by high-altitude ancestry.
1.2 Significance
The proposed research questions are significant to the fields of public health, physiology and anthropology. SGA occurs in 7-10% of all pregnancies and is associated with a 4-fold increase in stillbirth and an 8- to 20-fold increase in neonatal mortality, depending on the degree of SGA.37 In recent years there has been an emphasis on the importance of the perinatal environment and especially birth weight on susceptibility not only to poor health outcomes in the neonatal period but also to disease in later life. For example, low birth weight increases the risk of coronary heart disease, stroke, hypertension and type 2 diabetes in later life38"41, with the smallest infants being at the greatest risk for poor health outcome 42 Preeclampsia increases the risk of SGA, accounts for >25% of all pre-term births, raises perinatal mortality 5-fold and is a leading cause of maternal mortality in both the developed and developing world.43 Further, women who develop preeclampsia are at an increased risk for hypertension and cardiovascular disease later in life.44 Thus, one significant aspect of the

present study is that it seeks to determine factors that contribute to the etiology of maternal and fetal conditions known to increase both short- and long-term morbidity and mortality.
Although endothelial dysfunction is considered to be central in the development of both preeclampsia and SGA, the causative factor(s) responsible for damaging the vascular endothelium are not known. However, recent studies implicate a role for oxidative stress (i.e. when oxidant production outpaces antioxidant defenses) in the endothelial dysfunction characteristic of these disorders at low altitude.27 Since no studies have addressed the contribution of oxidative stress to the increased incidence of preeclampsia and SGA observed at high altitude, this research provides a novel approach to investigate the physiological mechanisms underlying the increased incidence of preeclampsia and SGA at high altitude. Thus, these studies are of importance from a clinical perspective as well as of significance for the 140 million high-altitude persons worldwide, who constitute the largest single group at risk for SGA.6
The comparative, ecological approach used to address the role of antioxidants and oxidative stress in preeclampsia and SGA at high-altitude synthesizes aspects of cultural and biological anthropology. The means by which humans adapt to environmental stress are fundamental concepts in anthropology and have clinical implications. The term adaptation" has several different meanings depending upon the field in which it is used. For example, physiologists use the term adaptation" to describe acclimation of an individual organism to a given environment. In contrast, in evolutionary terms adaptation refers to alterations in allelic

frequencies in a given population due to forces of natural selection. In this dissertation the term adaptation refers to cultural, behavioral and/or biological mechanisms that increase the survivability and reproduction of viable offspring in a specific environment. Comparative studies of multi-generational high-altitude populations and recent migrant populations suggest that Andeans are physiologically unique in ways that appear to contribute to their capacity to permanently live and successfully reproduce at high altitudes.45'47 Thus, Andeans appear to be more adapted" to high-altitude life than their European high-altitude counterparts. Some of the most convincing evidence of adaptive advantage in multi-generational versus recent high-altitude populations is that of protected fetal growth.10,12,14 Since reduced fetal growth is highly associated with increased infant mortality and disease in later life, pregnancy and the perinatal period are important focal points for examining selective advantage.
Finally, this study addresses maternal and child health complications in an underserved population in a developing country. Although this study was not developed to extensively address the relationship between pregnancy complications and socioeconomic or cultural factors of the subject populations, socioeconomic indices and other demographic variables were evaluated for their potential contribution to differences in fetal growth birth weight. This study does not encompass the wide range of non-biological" factors that may potentially affect fetal growth, however it does permit a broad view of the influence of income, education, health care access on fetal growth and other complications of pregnancy. In Bolivia, Andean populations are, on average, severely economically disadvantaged, have poorer health outcomes, less education and reduced access to health care relative to their European

counterparts. In conjunction with studies designed to identify socioeconomic, cultural and genetic factors contributing to population differences in susceptibility to preeclampsia and/or SGA, these lines of inquiry have direct implications for improvements in health care delivery and the prevention of the consequences of poor health in both developed and developing countries. Given that 90% of poor outcomes of preeclampsia occur in developing countries,48 it is particularly of concern for countries such as Bolivia.
1.3 Overall Hypothesis and Specific Aims
Our overall hypothesis is that oxidative stress contributes to the increased incidence of altitude-associated increase in preeclampsia and SGA by causing endothelial damage and diminishing uteroplacental blood flow; enhanced antioxidative capacity contributes to the protection afforded by high-altitude ancestry (Figure 1.1).

Antioxidant status
Oxidative stress
Endothelial damage
Normal maternal vascular adaptation to pregnancy
UA blood flow
UA blood flow
Figure 1.1 Hypothesis schema

Our specific aims are to:
Aim 1. Determine whether markers of oxidative stress
a. increase with pregnancy
b. increase with chronic hypoxia
c. are elevated during pregnancy at high versus low altitude
d. are elevated during pregnancy in European versus Andean high-altitude residents.
Hypothesis: Oxidative stress increases with pregnancy at high altitude to a greater degree in European than Andean women.
Aim 2: Determine whether antioxidant capacity
a. decreases with pregnancy
b. decreases with chronic hypoxia
c. is lower during pregnancy at high versus low altitude
d. is lower during pregnancy in European versus Andean high-altitude residents.
Hypothesis: Pregnancy decreases antioxidant capacity at high -altitude in European but not Andean women.
Aim 3: Determine whether oxidative stress or reduced antioxidant status contributes to altitude-associated preeclampsia and/or SGA via
a. increasing circulating markers of endothelial damage
b. increasing circulating vasoconstrictors relative to vasodilators

c. diminishing uterine artery blood flow
d. increasing the pregnancy complications of preeclampsia and/or SGA Hypothesis: Oxidative stress increases altitude-associated SGA and/or PE as the result of endothelial damage and diminished bioavailability of circulating vasodilators, which leads to a reduction in UA blood flow.

2. Theoretical and Empirical Background
2.1 Reduced Fetal Growth and Preeclampsia: Definitions, Epidemiology and Etiology
This section will review current definitions, epidemiology, risk factors and etiological theories of preeclampsia and SGA.
2.1.1 Measuring Fetal Growth Birth Weight
Birth weight, adjusted for gestational age, is commonly used to determine whether fetal somatic growth in utero was adequate. The use of this measure has several advantages in that the measurement of weight at birth is generally included in routine labor and delivery protocols in hospital or clinic settings, it can be obtained with simple, inexpensive equipment and is attainable even in the most rural or remote locations. In these respects, birth weight, when controlled for gestational age, is one of the most reliable, and accessible, measures of fetal growth for comparative study.49 Fetal Biometry
Human fetal development may be divided into four primary periods of growth including slow (0 to 15-16 weeks, <10g per week), accelerating (16-17 to 26-27 weeks, 85g per week), maximal (26-27 to 37-38 weeks, 200g per week) and decelerating (37-38 to 44 weeks, ~70g per week) stages.50 Ultrasonographic measures of head circumference, abdominal circumference, femur length, biparietal diameter and occipitofrontal diameter are compared to standard growth trajectories in order to detect insufficient fetal growth. From a clinical perspective this method is advantageous in that it enables early detection of insufficient development, can describe the period during which growth restriction began and can

distinguish between asymmetric and symmetric growth restriction. Typically, asymmetric fetal growth restriction, or growth restriction in which brain development is preserved at the expense of visceral growth (i.e. brain sparing), can be detected between 28 and 32 weeks of gestation. Symmetric growth restriction may be considered a more severe form of growth restriction, occurring earlier in gestation and limiting both brain and visceral growth. Thus, fetal biometry, used in concert with gestational-age-adjusted birth weight substantiates, and further describes, patterns of fetal growth. SGA or Intrauterine Growth Restriction (IUGRJ
Current literature uses two primary terms to describe insufficient fetal growth, SGA and IUGR. The former refers not to a specific disease entity, but rather to attenuated fetal growth resulting from suboptimal intrauterine conditions, thereby reducing maximal growth potential and birth weight.51 SGA refers to birth weight below a predetermined cut-off point or percentile for a given gestational age and sex based on sea-level criteria rather than the pattern of intrauterine growth.52 Both of these indices pose some difficulty in interpretation. Given that birth weight distribution is strikingly normal, some infants will be born with normal birth weights even though they may not have reached their genetically determined growth potential due to suboptimal intrauterine conditions. Further, no protocol is currently available to identify whether, for a given infant, fetal growth is optimized to its full genetic potential. In the case of SGA, the percentile below which an infant is considered to be SGA is variable; the 5th, 10th and 15th percentile have all been suggested. Further, since literature indicates that birth weight distribution curves may differ between populations the use of any standardized curve to compare multiple populations may over or under diagnose SGA.

While the terms IUGR and SGA are not synonymous, they are overlapping in the sense that an infant who is IUGR may or may not also be SGA, and vice versa. Throughout this thesis the term SGA will be used to describe insufficient fetal growth and will include all infants born below the 10th percentile for a given gestational age and sex; this term was chosen since birth weight falls with ascending altitude in all populations studied to date it is reasonable to assume that the hypoxia of chronic high altitude residence diminishes fetal growth potential1. In order to address these concerns, this study has included multiple measures to assess fetal growth, including birth weight, fetal biometry, and a measure ofcategorized insufficient growth, SGA.
2.1.2 Epidemiology Reduced Fetal Growth: SGA
SGA occurs in 7-10% of all pregnancies and is associated with a 4-fold increase in stillbirth and an 8- to 20-fold increase in neonatal mortality, depending on the degree of SGA.37 A population-based study conducted in Germany demonstrated that among preterm infants with very low birth weights (<1500g) those born SGA had twice the mortality rate of preterm infants of normal weight for their gestational age.53 The etiology of SGA is not singular, and likely involves a combination of genetic, physiological, immunological, infectious, environmental (e.g. hypoxia), behavioral (e.g. toxic insults such as smoking or drug use) and social factors (e.g. low socioeconomic status). Inadequate delivery of nutrients and oxygen to the fetus appears to be the proximate mechanism by which fetal growth is reduced. Risk factors for SGA include maternal undernutrition,54 multiple births,55 smoking,56 altitude,1

congenital infections,57,58 preeclampsia9 and maternal pathologies including diabetes, collagen vascular disease and other conditions associated with altered vascular function. Of particular importance for the present study, although preeclampsia is a risk factor for diminished fetal growth, it is not a prerequisite. The risk of preeclampsia increases approximately 3-fold at high versus low altitude and contributes to, but cannot account for, the magnitude of birth weight reduction.1,7 9 Preeclampsia
Hypertensive disorders during pregnancy include chronic hypertension that predates pregnancy, pregnancy-induced hypertension with or without proteinuria (preeclampsia or gestational hypertension, respectively). Preeclampsia is defined as > 2 BP readings that are >140 mmHg systolic and/or > 90 mmHg diastolic at least six hours apart after week 20 of pregnancy accompanied by proteinuria (either > 300mg per day or urinary protein to creatinine ratio > 30mg/mmol).59 The hallmark maternal manifestations of this systemic disorder include hypertension, proteinuria, endothelial dysfunction and edema. Additional symptoms may include thrombocytopenia (low platelets) or liver function abnormalities.
Globally, preeclampsia is a leading cause of maternal and perinatal mortality and premature delivery. Preeclampsia occurs in 0.4%-2.8% of all pregnancies in developed countries; however the incidence and severity of outcome is markedly higher among developing countries60 where > 90% of the most serious outcomes occur.48 It is highly likely that the incidence of preeclampsia is underestimated in developing countries since diagnostic capacity and access to health care is generally lower than that available in the developed

world. However, even among developed countries with comparatively low maternal mortality rates, the proportion of maternal deaths attributed to preeclampsia-related complications is high, accounting for up to two-thirds of maternal mortality in the United Kingdom.61
Preeclampsia increases the risk of SGA, accounts for >25% of all pre-term births, raises perinatal mortality 5-fold43 and is associated with the highest rates of maternal and infant morbidity and mortality of all pregnancy complications.61 Further, an extensive epidemiological study revealed that women who had experienced preeclampsia had an excess risk of long-term mortality (over 24-36 years follow-up), much of which was attributed to cardiovascular disease, relative to women who were not diagnosed with preeclampsia.44 Whether preeclampsia acts as the precipitating factor to the development of future cardiovascular disease or whether women who develop preeclampsia are predisposed to cardiovascular disease is unknown. However, suggesting that preeclampsia may initiate vascular dysfunction the risk of long-term mortality increased progressively with the number of preeclamptic episodes; compared to women without preeclampsia the relative risk of mortality increased 2-fold with one preeclamptic delivery to more than 6-fold with three or more preeclamptic episodes.44
Risk factors for preeclampsia include behavioral, genetic and environmental factors including chronic hypertension62, familial or personal history of preeclampsia63 nulliparity,64 increasing maternal age,64obesity,65diabetes,66 African ancestry67 and altitude9,68 Further, the recent identification of several genes associated with increased risk of preeclampsia suggests a
69 72
genetic component may also be involved.

2.1.3 Etiology
Despite the severity of preeclampsia- and SGA-related health outcomes to both the mother and infant, the mechanisms underlying these disorders have not been clearly identified. Most literature describes preeclampsia using a two-stage model in which poor placental perfusion precedes and results in the maternal syndrome.25 Early in pregnancy cytotrophoblasts invade the uterine spiral arteries and progressively replace vascular endothelial cells, medial elastic tissue, smooth muscle and neural tissue thereby creating a low-resistance vascular system. In preeclampsia, cytotrophoblast invasion of the spiral arteries is shallow resulting in inadequate remodeling, higher vascular resistance and diminished uteroplacental blood flow. Maternal endothelial and vascular dysfunction are thought to be due to the release of some toxic factor from the placenta64,73 in response to reduced placental perfusion. Current literature suggests that reactive oxygen species may play a principle role in endothelial damage and vascular dysfunction associated with preeclampsia and SGA.26,27,74-77 As reviewed, SGA and preeclampsia increase maternal and infant mortality and morbidity rates, and contribute to cardiovascular disease in later life. The prevalence and poor outcome of low birth weight and preeclampsia are elevated in developing countries, such as Bolivia.
2.2 Ecological Context: Bolivia
2.2.1 A Highland Ecology
The Andean altiplano (or high-plains) stretches nearly 4800 km, occupying nearly 250,000 square km. The portion of the altiplano within the boundaries of Bolivia averages approximately 3800 m in altitude and is defined by two mountain ranges, the Cordillera Occidental to the west and the Cordillera Oriental to the east. To the east of the Cordillera Oriental the topography shifts rapidly and dramatically. To the northeast of the Cordillera

Oriental, the terrain rapidly descends through densely forested and moist slopes, eventually reaching the Llanos de Mamore and the Amazon River basin. To the southeast of the Cordillera Oriental the descent into the eastern lowlands is less abrupt and, in terms of landscape, quite different. The southern portion of the Bolivian lowlands primarily consists of the Gran Chaco, a semi-arid savanna, which houses the urban center of Santa Cruz (416 m). Although two-thirds of Bolivia's territory consists of tropical or semi-tropical lowlands and, the basins and valleys of eastern Bolivia are in many respects more hospitable for human settlement than the altiplano, early human habitation was concentrated in intra-mountain valleys and highlands. This was primarily due their strategic location between prosperous valley towns and coastal trade centers, accessible mineral deposits, reasonably fertile soils and the warm microclimate provided by Lake Titicaca.78 Although Bolivia has participated in the global market system since the time of the Spanish Conquest in the 16th century traditional vertical agriculture exchange still dominates a large part of the highland economy. Thus nearly half of the population (44%) still resides in the altiplano while the remainder live in lowland areas.
The 16th century infusion of European-derived populations into the highland and lowland areas of Bolivia resulted in a blending of pre-Columbian and Spanish culture and peoples, which distinguishes modern day Bolivia.78 Today the indigenous population is 3.6 million, while the remaining 5.4 million are Mestizo (Andean-European mixture) or are of European decent. Of all countries in the western hemisphere, Bolivia has the largest, highest and longest-resident high-altitude population. Andean and European-derived groups now live across a range of altitudes from near sea-level to altitudes above 4,000m,81 providing a

unique opportunity for comparative study.
2.2.2 Socioeconomic, Health and Nutritional Status
With a per capita gross national income of approximately 2000 USD in 2001 and 960 USD in
82 83
2004, Bolivia is one of the poorest countries in Latin America. The majority of Bolivians (63%) live below the national poverty line, however poverty rates are higher in rural versus urban areas with 76% and 47% impoverishment, respectively. According to the 1992 National Population and Housing Census 70% of all children lived in extreme poverty and did not attend school. The same survey indicated that indigenous populations were 40% more likely to be impoverished than other groups and that the birth of each additional child increased the risk of poverty by an additional 6.5%. Nearly half of the population lives in the altiplano and 38% reside in rural areas,83 84 each locality tending to be predominately
indigenous and at higher risk of ill-health and malnutrition.
Nationally, maternal mortality rates hover around 420 deaths per 100,000 live births, a rate more than double that of Latin America as a whole. According to the Population Reference Bureau, in 2003 Bolivia bore the greatest burden of infant mortality (<1 yr) in the Western hemisphere with a rate of 61 deaths per 1,000 live births, a rate nearly double that of neighboring Peru and six-times that of Chile.82 In 2002 the Pan American Health Organization (PAHO) reported highly divergent maternal mortality rates between urban (274/100,000 live births), rural (524/100,000 live births) and rural-altiplano areas (602/100,000 live births). These differences are likely attributable to higher poverty rates,

reduced access to health care, a high percentage of home births (57%) and lower maternal educational attainment in rural-altitude areas.85,86 Infant mortality and the proportion of infants born at <2500g increases from sea level to high altitude in both rural and urban areas.8'86-88
Sixty percent of mothers obtain s5 years of formal education while fathers fare slightly better with an average of approximately 7 years of schooling.89 Birth intervals are short, with almost half of births occurring less than 2 years after the preceding one, and in 1998 only half of mothers had prenatal or delivery care. Lack of education, short birth intervals and lack of health care access likely contribute to high rates of malnutrition in Bolivia.
While extensive or recent accounts of nutritional status in Bolivia are sparse at best, it is clear from available sources that mal- and under-nutrition remain important public health issues, particularly in highland and rural areas. Demographic and Health Surveys indicate that between 1989 and 1994 the proportion of undernourished children has decreased in Bolivia as a whole, however in underprivileged-altiplano areas, such as the Department of Potosi, rates have increased. In 1998 30% of Bolivian children aged 0-60 months were stunted and nearly one-third of children under 5 years of age were chronically malnourished,91 representing the highest rate in Latin America. The prevalence of childhood stunting is greater among indigenous (50.5%) than non-indigenous populations, in rural (37.2%) versus urban (18.5%) areas, in lower socioeconomic strata (lowest decile: 42.2% vs. highest: 9.7%) and in the altiplano.92-94 In some Aymara and Quechua communities the prevalence of childhood malnutrition is twice the national average.85

In 1998 56% of children under 5 years of age were iron deficient, with poor and high-altitude regions being the most affected (poor: 72%; high-altitude regions: Potosi 70%, Oruro 69%, La Paz 58%).79 In 1991 19.5% of children living in rural areas of the highland plateau were severely vitamin A deficient (retinol levels < 20^g/dl) while 48.3% were marginally deficient (retinol levels <30pg/dl). Iodine deficiency was reduced from 60% to 4.7% in the general population between 1981 and 1994.79,85
The above review demonstrates that little information is available regarding nutritional status among Bolivian populations and that indigenous and high-altitude populations are more impoverished, have less access to health care and are at higher risk of poor health outcome and malnutrition than other groups. Taking available socioeconomic, nutritional and geographic differences into account, Andean women would intuitively be expected to deliver IUGR infants more often than their European counterparts, however current data suggest the opposite.10,12
2.3 High Altitude: The Challenges and History of Permenant Residence at High Altitude
High altitude, primarily due to the pervasive effects and near-inescapability of hypoxia, poses one of the greatest threats to human survival. However, over 140 million people live at high altitude and, as archeological evidence suggests, humans first made a presence in highland regions more than 20,000 years ago. This section will review the physiological challenges

posed by exposure to high altitude and will outline current knowledge regarding the duration of human highland occupation.
2.3.1 Physiological Challenges Posed by High-Altitude Exposure
High altitude poses one of the greatest ecological threats to human survival.47 Ambient temperature drops approximately 1C every 150 m rise and, since water vapor pressure is depressed at reduced temperatures, absolute humidity falls. For example, a temperature shift from +20C to -20C diminishes water vapor pressure from 17mmHg to only 1mmHg, an effect that contributes to excessive water loss through ventilation at altitude. Further, exposure to solar radiation increases markedly with altitude, since there is less atmospheric water vapor to absorb direct solar radiation as well as that reflected from the earth, which is exaggerated by snow cover. These generally cold and dry conditions not only pose direct physiological stress but can also limit cultivation and travel for resource acquisition.
However, the most physiologically pervasive effect of high altitude is that of hypoxia, primarily because its effects cannot truly be ameliorated through behavioral means, other than emigration. Hypoxia, a decrease in tissue oxygen supply below normal levels, limits the ability to maintain metabolic activity necessary for survival. Hypoxia may be categorized into subtypes based upon the underlying etiology of reduced tissue oxygenation; anemic hypoxia (hypoxia due to decreased concentration of functional hemoglobin or a reduced number of red blood cells, as seen in anemia and hemorrhage), ischemic hypoxia (tissue hypoxia characterized by tissue oligemia caused by arteriolar obstruction or vasoconstriction), oxygen-affinity hypoxia (hypoxia due to reduced ability of hemoglobin to release oxygen),

stagnant hypoxia (tissue hypoxia characterized by intravascular stasis due to impairment of venous outflow or decreased arterial flow) and hypoxic hypoxia (hypoxia resulting from a defective mechanism of oxygenation in the lung, low atmospheric oxygen tension (as is present at high altitude), abnormal pulmonary function, airway obstruction, or a right-to-left shunt in the heart).
To understand the magnitude of oxygen deprivation experienced at high altitude, it is useful to follow the compartments through which oxygen moves in order to reach metabolically active tissues. This illustration, often referred to as the "oxygen cascade", begins with the oxygen tension of ambient air, to that of moist inspired air, to the alveoli and terminating in the arterial blood (Figure 2.1). Since the fraction of 02 in ambient air (F|02; 20.9%) is independent of altitude the decline in barometric pressure (PB) with altitude reduces the partial pressure of inspired oxygen (P|02), which may be calculated from the barometric pressure and F|02 using the equation: ambient P02 = PB F|02.

For example, at sea level (PB ~ 760 mmHg, 101.3 kPa) ambient P02 is approximately 160
mmHg (31.33 kPa), whereas at 5800m (PB ~380 mmHg, 50.7 kPa) ambient P02 is reduced to approximately 80 mmHg (10.7 kPa), or half that of sea level values. As the ambient air moves through the nose, mouth, larynx and trachea it is warmed to body temperature and moistened by water vapor in the respiratory tract. Thus, there is a significant drop from the ambient
to inspired P02 due to the displacement of about 10mmHg P02 due to the water vapor pressure in the respiratory tract. Since PH2o at body temperature remains constant at ~47 mmHg the fraction of ambient P02 displaced rises as PB falls such that at 5800m the loss is twice that experienced at sea level. As inspired air moves to the alveolar branches there is a further loss of P02that is dependent upon metabolic rate (i.e. carbon dioxide exchange and oxygen uptake) and ventilation. Demonstrating that increased ventilation is a critical aspect of maintaining P02 at high altitude, a doubling of ventilation halves the drop in P02 of inspired air to that of the alveoli. The drop in P02 across the alveolar-capillary membrane is generally between 6-10 mmHg
02 compartments Factors contributing
to drop P02
Ambient P02
Inspired P02
* Altitude
* H20 vapor
Alveolar P02
Metabolic rate Ventilatory rate
Arterial P02
Ventilation perfusion ratio
Mixed venous P02
Cardiac output Oxygen carrying capacity Metabolic rate
Figure 2.1 Oxygen cascade

and is approximately equal at all altitudes. Thus, the cumulative loss of P02 is greater at high altitude, resulting in a reduction in arterial P02 relative to sea level. As will be discussed in Section 2.4, in response to hypoxia a series of physiological changes occur which serve to diminish the loss of arterial P02 (Pa02).
Oxygen is transported from the lungs to peripheral tissues by hemoglobin, a protein that reversibly binds 02 found within erythrocytes. Two terms that are commonly used in the literature to describe oxygen transport are oxygen carrying capacity and oxygen saturation. The former refers to the ability of hemoglobin to transport oxygen; several factors influence oxygen carrying capacity, including temperature, pH, PC02 and circulating levels of 2,3-diphosphogycerate (2, 3-DPG), a metabolite that regulates the dissociation of oxygen from hemoglobin. Oxygen saturation (Sa02) is the percentage of hemoglobin that is bound to oxygen, and is equal to the amount of Pa02 in the blood divided by oxygen carrying capacity. Oxygenation of arterial blood is often described in relation to the oxyhemoglobin dissociation curve, which plots the sigmoid relationship between Pa02 and Sa02. With increasing altitude Pa02 decreases and thus so does Sa02. For example, at an altitude of 5800m arterial P02 is approximately 38 mmHg (5.1 kPa) in comparison to 95 mmHg (12.6 kPa) at sea level; this difference translates into a 32% drop in Sa02. The elbow or inflection point of the curve, where Pa02 is approximately 60-70 mmHg, marks a point of clinical interest in that below this point oxygen saturation decreases exponentially with further decrements in Pa02 95 This critical point corresponds to an altitude of approximately (2500 m), and defines the altitude threshold considered to be relevant in the investigation of physiological response and/or

adaptation to high altitude. As such, the term altitude in this manuscript refers to elevations greater than or equal to 2500m or 8500ft.
2.3.2 Historical Accounts
The acute physiological challenge of an oxygen-deprived environment was documented as long ago as 37BC when a Chinese official stated,
"From Pe-shan southwards there are four to five kingdoms not attached to China. The Chinese Commision will in such circumstances be left to starve among the hills and valleys.. Again on passing the Great Headache Mountain, the Little Headache Mountain, the Red Land, and the Fever Slope, mens bodies become feverish, they lose colour and are attacked with headache and vomiting; the asses and cattle being all in like condition.96 (cited in Ward, 2000).
Accounts of acute mountain sickness are nearly global. In Tibet it is referred to as damigiri or dam (breath seizing), dugri (poison of the mountain) or ladrak (poison of the pass). To the Moguls it is referred to as yas, tunk (wakhi and badakhshi), esh (turki) or bish-ka-hawa. In the Andes it is called puna, soroche, mareo or veta. In each case, the state being named consisted of some combination of the following symptoms headache, vomiting, exhaustion, difficulty sleeping, aphasia or swelling of the hands and feet.

Historical accounts made during the 16th century Spanish Conquest of South America reveal population-level differences in reproductive success at altitude. In 1639 Antonio de la Calancha described the prohibitive effect of hypoxia on reproductive success among Spanish living in Potosi (4100m), now part of Bolivia:
In Potosi all children born of Spanish parents died either at birth or within a fortnight thereafter... 97
Father Cobo, a 16th century missionary made similar observations in Jauja, the first colonial capital of Peru, which he described as sterile, a place where horses, pigs and cows could not be raised.98 Further, Cobo also noted that infant mortality appeared to be proportional to the amount of pure-blooded Spanish ancestry; his account indicated that the indigenous high-altitude populations
...are healthiest and where they (reproduce) the most prolifically is in these same cold air-tempers, which is quite the reverse of what happens to the children of the Spaniards, most of whom when born in such regions do not survive.98
These historical accounts, despite the uncertainty of their accuracy, illustrate the challenge of high altitude poses to reproductive success. As will be discussed, more recent accounts and scientific inquiries indicate similar findings. However, despite the challenges posed by high altitude, more than 140 million persons live at high altitude. The four primary highland

regions capable of supporting permanent human habitation include the Tibetan plateau and Himalayan mountain range, the South American altiplano and Andean mountain range, the North American Rocky Mountain region and the Ethiopian highlands. The antiquity of human settlement in these highland regions is unclear however archaeological evidence estimates that humans moved into the Himalayan highlands between 7,000 (permanent residence) and 50,000 (visitation) years ago and into the South American altiplano approximatley 10,000 years ago.
2.3.3 Peopling of Highland Regions
It is important to acknowledge that the duration of time human populations have made permanent residence at high altitude in the regions of the world discussed here, the Himalayas and the Andes, is extensive and thus, likely sufficient for genetic adaptation to have occurred to their high altitude environments. For this reason, this section will discuss the timing of human movement into and permanent residence in highland regions. Himalayas
The precise timing of human movement onto the Tibetan plateau is, thus far, somewhat controversial. Artifacts including stone flakes and fossil remains of Homo, archaic and modern Homo sapiens, indicate that hominids lived along the plateau's fringes in the Pliocene, at least 2 million years ago.99-101 Stone tool artifacts including flakes, scrapers and microliths cross-dated with similar tools assigned to the Middle to Upper Paleolithic period found in northern regions of Asia suggest that humans began visiting the higher steppes of

the plateau between 25,000 and 50,000 years ago.102 103 Quartzite core and flake tools found at 3100 m104 and human hand and footprints impressed into a travertine deposit at 4200 m105 suggest that humans lived, at least temporarily, on the plateau between 18,000 and 20,000 years ago.106 However, farming implements dated in situ suggest that human occupation of the Tibetan highlands did not occur until much later, approximately 7000 years ago.107 While upon first appearance these data seem mutually exclusive, it is possible that humans settled these highland regions more than once due to climactic variations or otherwise. Based on available ecological, geographical and archaeological data Madsen, et al. (2006) suggest that the peopling of the Tibetan Plateau occurred via a three-step process beginning with temporary occupation by mobile foragers of lower plateau regions (<3000 m) as early as 40,000 to 25,000 years ago (14C dating).108 This, Madsen suggests, was followed by more permanent settlements in these areas and recurrent foraging and short-term occupation on the middle to upper elevations of the plateau between 25,000 and 10,000 years ago (14C dating) or immediately before and after the last glacial maximum. According to this model the third step, permanent habitation of the upper-level plateau, likely did not occur until the early Neolithic period (~ 7000 years ago).108 The above evidence clearly indicates a long history of human activity in the highlands of the Himalayas, however the exact timing of permanent habitation in these regions and whether current inhabitants in these regions are descendants of these early highland occupants remains unknown. Andean Plateau
Available resources suggest that the peopling of the Andean Plateau occurred somewhat later than that of high-altitude regions in the Himalayas.109,110 Artifacts found along the

shorelines of modern day Peru indicate permanent human settlement in South America approximately 11,000 to 13,000 years ago.111'112 Suggesting that these coastal populations began moving into the highland regions of the Andes during the this same period, obsidian of highland origin was found at some of the coastal sites situated along an easy travel route to the higher elevation.111 Further, an extensive review of obsidian artifacts suggests that major highland deposits of the volcanic glass (southern Peru and northern Bolivia) were extensively exploited nearly ~10,000 years ago.113 After this time documentation of human settlement at high altitude in the Andes is well defined. Currently, nearly seven million persons reside on the a/f/p/ano,114 some communities can be found as high as 5,200m.115 Thus, given the success and extensive duration of highland populations in the Himalayas and the Andes the challenges posed by altitude environments must have been minimized by behavioral and biological means (genetic adaptation, developmental responses). In the case of high altitude, the recent migration of lowland populations, including the movement of European populations into the Andes (- 500 years ago) and Rocky Mountains (~150 years ago) and the Han Chinese" migration into the Tibetan Plateau (~60 years ago), permits the examination of relationships between environmental pressures and biological variation as evidence for such adaptations.
2.4 Adaptation to High Altitude
Adaptation is a broad term used to describe features, forms, functions or behaviors (e.g. social patterns or subsistence strategies) that enhance survival and reproduction in a given environment. Biological adaptations can be categorized along a temporal continuum, ranging from short-term physiological responses in a single individual to population-level genetic

variability. Acute (minutes, days)- or long-term (weeks, months) exposure to a given environment such as hypoxia elicits adaptive physiological responses such as the hypoxic ventilatory response or increased erythropoesis. Developmental adaptation, or phenotypic plasticity, refers to altered phenotype as a result of continual exposure (years) to a given environmental stressor. Traits acquired through developmental adaptation are not heritable, although the capacity for such adaptation may be conserved over time. In contrast to developmental adaptations, which functions within the individual, genetic adaptation refers to increasing frequency of genes that enhance survival and reproductive success in a given environment within a population. Unlike traits acquired through developmental adaptation, genetic traits can be inherited over time. While it is only those traits that confer some adaptive benefit that are of relevance for understanding the processes of natural selection and environmental adaptation, it is important to recognize that not all genetic or physiological variability between populations reflects adaptive advantage.
This section will review apparent behavioral and biological adaptations to high altitude, highlighting differences between populations of extensive high-altitude ancestry and newcomers.
2.4.1 Cultural and Behavioral Adaptations to High Altitude
Cultural innovations ameliorate, or at least modify, the negative effects of ecological stress, including that of high-altitude. In order to meet energy demands Andeans have focused their food production efforts on indigenous high-altitude animals (llama, vicuna and guinea pig) and plants, including over 2,500 varieties of tubers (ex. oca, mashua and ulloco), quinoa and

other grains (ex. tarwi and kiwicha).116 Vertical trade between the highland regions and the more fertile, lowland regions of Bolivia has supplemented high-altitude diets with fruits, beans, peanuts, wheat and other consumable products (e.g. coca) grown at lower elevations.117 Further, foodstuffs (e g. potatoes, chuho or meats, charqui) are preserved by freeze-drying for later consumption.117 Historically, and likely currently in rural high-altitude regions, labor delegation to adolescent family members, who use less energy to complete the same amount of work, preferential distribution of food to those performing the most work, temporary out-migration of adolescent and adult family members, and engagement in less-energy intensive activities serve to maximize seasonal variation in caloric availability.117118 To avert increased energy expenditure associated with cold temperatures housing and clothing design are such that heat gain is maximized (e g. dark clothing to absorb radiant heat) and heat loss is minimized (e.g. layered clothing).119 Thus, while some of the negative effects of high-altitude residence are remediable by cultural means the hypoxia of high altitude, posing one of the most potent ecological threats to human survival, can only truly be avoided by emigration.
2.4.2 Biological Adaptation to High Altitude Ventilation and Respiratory Volumes
The hypoxic ventilatory response increases alveolar ventilation in response to hypoxia and is usually initiated when P|02 falls below 100 mmHg.120 The hypercapnic ventilatory response increases sensitivity to alveolar PC02 (PaC02)47 such that the inhibitory effect of low PaC02 (caused by the hypoxic ventilatory response) on ventilation is averted at high relative to low

altitude. The hypercapnic ventilatory response does not appear to differ by duration of high-
altitude residence 47,121'123
Historically, the consensus has been that ventilation is blunted in high altitude natives (e.g. Andean124, Peruvians and Tibetans125) relative to acclimatized newcomers, suggesting that these high-altitude groups lack hypoxic ventilatory sensitivity. However, evidence for such a response remains controversial.126 For example, shifting PA02 from 200 to 30 mmHg increased ventilation significantly in lowlanders, but not in Sherpas. However, more recent comparisons of Tibetan highland natives and acclimatized newcomers demonstrate that at 4243 m ventilation is equivalent between groups, or when controlled for body size greater among the Tibetans.127 Similarly, at 3800m minute ventilation is greater and PetC02 reduced in Tibetans than acclimatized Han Chinese at the similar altitudes.128 129 Thus, in these latter studies high altitude natives neither hypoventilate nor have blunted a ventilatory response to hypoxia relative to acclimatized newcomer groups. Among Andeans, ventilatory sensitivity to hypoxia is less than that of Tibetans, as indicated by lower PetC02 levels at similar end-tidal P02, and less than acclimatized newcomers.128,129 As noted by Reeves, et al. interpretations of these results should consider that the newcomer groups used for comparison likely demonstrate a great variability in ventilatory response to hypoxia,130 and in some cases, where mountaineers were used as the comparison group, they likely have a more brisk hypoxic ventilatory response response than normal"131. Suggesting that ventilatory sensitivity to hypoxia may also be, in part, a developmental response, Colorado residents born and living at altitude (3100m) also demonstrate a blunted hypoxic ventilatory
response, such that ventilatory response to hypoxia is 10% that of sea level controls.

Whether or not these differences in hypoxic ventilatory response have functional implications
is unclear. However, a brisk HVR response to acute hypoxic exposure provides a degree of
protection against high-altitude pulmonary edema and is associated with increased
exercise capacity at altitude. Thus, while it appears that a brisk hypoxic ventilatory response
appears to be beneficial in some populations, the apparent blunting of ventilation with
hypoxia seen in Andeans may reflect the presence of alternative methods for increasing the
availability or utilization of oxygen.
Vital capacity, the volume of air expired after a maximal inspiration, is greater in Tibetans (5080 ml_) than in Han (4280 mL) residing at the same altitudes, despite similar ages, heights and weights.134 In Andeans total lung capacity was 500mL higher at altitude (4540m) than at sea level, mostly due to an increase in residual volume.135 Even though respiratory volumes differ slightly between high- and low-altitude natives the magnitude of difference is small and likely has little to no effect on oxygen delivery or performance at high altitude.47 Suggesting that highland ancestry is advantageous for physical performance at altitude, among healthy young Peruvian men of Quechua-Spanish ancestry, the extent of Spanish admixture was positively associated the reduction of maximal oxygen uptake from sea level to high-altitude (4,338m).136 Consistent with these observations, during exercise at altitude Tibetans maintained higher oxygen saturations during exercise than persons of lowland ancestry due
to a narrower alveolar-arterial (A-a) diffusion gradient.

Infants born at high altitude have higher thoracic compliance than their low altitude counterparts. Adults at high versus low altitude demonstrate increased thoracic blood volume and an increased residual volume as a percentage of total lung capacity. As suggested by Ward, et al., higher lung capacity due to thoracic compliance may increase diffusion surface area, and, in conjunction with increased blood volume may facilitate diffusing capacity.47 Indeed, at altitude Tibetan highlanders have greater pulmonary diffusing capacity compared to Han.139 The fact that pulmonary diffusion capacity is greater in children raised at high versus low altitude,140 in Colorado and South America47 suggests the involvement of both genetic and developmental adaptation.
In summary, while ancestry-dependent differences in ventilation and hypoxic ventilatory sensitivity exist it is not clear whether these differences are genetic, developmental or both. Further, unknown is whether the differences observed are of adaptive significance. Pulmonary Circulation
In contrast to its effect in the systemic circulation, hypoxia has strong vasoconstrictive effects in the pulmonary circulation. The critical stimulus for the hypoxic pulmonary pressor response (HPPR) is a reduction in alveolar oxygen tension rather than Pa02 in the pulmonary vessels themselves. As a result of the HPPR, pulmonary arterial pressure increases with high-altitude exposure. Crossbreed studies in cattle indicate that a muted HPPR is determined by a single, dominant gene that is inherited in a classic Mendelian manner.141 Although the HPPR causes a more uniform distribution of blood flow throughout the lung and

increases ventilation-perfusion ratio slightly, its overall effect on gas exchange is relatively insignificant. Further, the HPPR has been associated with the development of pulmonary hypertension, right ventricular hypertrophy, high-altitude pulmonary edema (HAPE) and, in cattle, Briskett's disease47 suggesting that it is a maladaptive response.
In lowlanders and Andeans, hypoxic exposure results in increased pulmonary artery pressure and right ventricular hypertrophy via the HPPR.47 It has been suggested that the involution of the muscular coating in the pulmonary arteries that normally occurs after birth is partial, or absent, in Andeans born at high altitude and that this may contribute to increased pulmonary pressure at altitude.47 In contrast, Tibetans' appear protected from hypoxia-induced increases in pulmonary artery pressure or resistance.142
In Lhasa, Tibet (3658m) electrocardiographic (ECG) abnormalities suggestive of right ventricular hypertrophy were found to be present in about one-quarter of healthy Tibetan natives, one third of healthy Han (ethnic Chinese) persons and in half of patients with chronic mountain sickness or Monges disease. Interestingly, 90% of the chronic mountain sickness cases were individuals of Han ancestry. Suggesting that duration of altitude residence likely plays a role in the development of pulmonary hypertension, Han who had immigrated to Lhasa as children were more likely to present with ECG evidence of right ventricular hypertrophy than those who migrated as adults. Similar data was obtained in the Andes, where ECG studies among native Peruvians of life-long residence at 4540m showed evidence of pulmonary hypertension less often (18%) than persons who had migrated from

lower altitudes (28%). In North America 30% of children living at high altitude have ECG results suggestive of pulmonary hypertension, a rate greater than that seen in South America or in the Himalayas, suggesting that highland ancestry is protective against pulmonary hypertension. Hematological Adaptation
One of the most well known markers of physiological adaptation to hypoxia is that of accelerated erythropoesis. Some studies indicate that hemoglobin levels are inversely related to duration of high-altitude residence (in generations), suggesting that highland populations have better oxygen delivery and/or ventilation (i.e. less hypoxic stimulus for erythropoesis), or that they are hemodiluted relative to newcomers. For instance, hemoglobin levels were reported to be 3-4 g/dL higher in Andeans than Tibetans living at the same altitudes.143 During pregnancy at 3,658m hemoglobin levels were reported to be 2g/dL greater in Han than Tibetan women,24 however our recent data indicates no difference between Andean and European women during pregnancy or in the non-pregnant state at high altitude.144
Suggesting that populations of high- versus low-altitude ancestry may possess metabolic adaptations that could be protective against ROS-induced tissue damage are studies indicating the differential expression of several proteins (glutahione-S-transferase P1-1, A2-enoyl-CoA-hydratase, NADH-ubiquinone oxidoreductase, glyceraldehyde-3-phosphate dehydrogenase, lactate dehydrogenase and myoglobin) between highland born and living and second-generation lowland born and living Tibetans, and lowland bom and living

Nepalis.145 Relative to lowland Nepalis, glutahione-S-transferase P1-1 isoform, an enzyme that conjugates reduced GSH to electrphilic acceptors (detoxification), was 380% and 50% over-expressed in highland and lowland Tibetans, respectively. Likewise, A2-enoyl-CoA-hydratase and myoglobin were upregulated in both Tibetan groups. NADH-ubiquinone oxidoreductase (complex I of the respiratory chain), an enzyme which couples the oxidation of NADH and the reduction of ubiquinone, was slightly up-regulated in highland, but not lowland, Tibetans. Glyceraldehyde-3-phosphate dehydrogenase and lactate dehydrogenase were slightly down-regulated in highland Tibetans. Absent from this study was a highland Nepali comparison group, which would enable the determination of whether or not observed differences are due to simply due to altitude exposure or are intrinsic to each population group or the result of permanent highland residence.
In summary, Tibetans and Andeans appear to have a unique, but different, physiology relative to populations of low-altitude ancestry that appears to confer protection against some of the negative consequences associated with the hypoxia of high altitude (hypoxia-tolerant phenotypes"). As will be discussed below, one of the most poignant examples of such variation is that of protected fetal growth. Since population survival depends on reproductive success, pregnancy and perinatal development are logical points to begin looking for characteristics suggestive of high-altitude adaptation (behavioral and/or biological). Thus, while comprising a small portion of the worlds population, residents of high altitude are in a unique position to help define the genetic, physiological as well as culturally-related factors by which human populations can respond to hypoxic stress, among the most challenging health threats known.

2.5 A Tale of Two Pregnancies
Pregnancy presents a unique physiological and anatomical challenge in humans due to the long duration of gestation, the proportionally large cranial and body size of infants at birth and upright human posture. In addition to the mechanical challenge of delivery posed by upright posture and proportionally large infant size, the ability to deliver an adequate supply of oxygenated blood to the uteroplacental circulation requires significant adaptations in maternal physiology.
This section will first review physiological factors affecting uteroplacental blood flow and arterial oxygenation during normal and chronically hypoxic pregnancy. The magnitude and population-level variation of hypoxia-associated birth weight and fetal growth reduction will then be discussed. Differences that may help to explain protection from SGA and preeclampsia among multigenerational high-altitude populations at high altitude will be highlighted.
2.5.1 Extent and Variation of Birth Weight and Fetal Growth Reduction with Hypoxia Hypoxia-Associated Fetal Growth and Birth Weight Reduction
The persistent hypoxia of permanent residence at high altitude potently decreases fetal growth, decreasing birth weight an average of 121g per 1000 m elevation rise when the relationship is expressed in a linear fashion.1 In reality, birth weight declines in a curvilinear fashion with increasing altitude such that beyond 2000m the effect of altitude becomes greater.146 The effect of high-altitude residence on birth weight is as great or greater than that associated with low maternal weight gain, smoking, or primiparity1 and acts at the population

level, shifting the entire birth weight distribution curve to the left4 such that the proportion of low birth weight infants (<2500g) increases.1
Reduced fetal growth during the third trimester,2,3 rather than preterm delivery, is responsible for diminished birth weight at altitude.4 High-altitude residence in Colorado (> 2744m) decreases gestational age by approximately 0.5 weeks, an insufficient duration to explain the observed reduction in birth weight.1 Moreover, high-altitude residence nearly triples the risk of SGA, which by definition controls for gestational age9,147 Suggesting that hypoxia is a separable cause of reduced fetal growth at altitude, Krampl, et al (2000) demonstrated that among a group of ethnically and socio-economically similar women fetal growth follows a lower trajectory at altitude (Cerro de Pasco, Peru; 4300m) relative to sea level (Lima, Peru; sea level), with head and abdominal circumference, biparietal and occipitofrontal diameter and femur length being smaller at altitude.3 The effect of altitude became apparent between 25 and 29 weeks of gestation and persisted until delivery.3 Similarly, preliminary data from studies conducted in Leadville, Colorado (3100m) demonstrate that relative to 1609m a measurable decline in fetal head circumference becomes apparent at 30 weeks gestation and reduced abdominal circumference and femur length are evident by 36 weeks of gestation.
Beyond the effect of chronic hypoxia on fetal growth, high-altitude residence increases the incidence of preeclampsia approximately 3-fold, an effect which contributes to but does not explain the magnitude of hypoxia-associated birth weight reduction.1,s'7 In Colorado, a survey

of over 200 women indicates that preeclampsia is more than 5-times as frequent at 3100m (16%) than at 1260m (3%).7 Additional pregnancy and perinatal complications are more common at high altitude, including gestational hypertension (1.7-fold),148 fetal and neonatal distress increased (7-fold), early arterial oxygen desaturation, perinatal acute respiratory infections and pulmonary hypertension.150
Recent studies indicate that indices of maternal under-nutrition (i.e. maternal stature and skinfold measurements) are not strongly correlated with the reduction of birth weight with altitude, either within or between populations.15 However, chronic altitude exposure does reduce maternal glucose levels, as women residing at 4300m in Peru demonstrate circulating glucose levels that are ~10% less than their low-altitude counterparts, both during pregnancy and in the non-pregnant state.151 This may be due, in part, to differences in glucose utilization or insulin-sensitivity between high- and low-altitude groups, or impaired glucose transport. Variability in Hypoxia-Associated Fetal Growth and Birth Weight Reduction
Hypoxia exerts a depressant effect on fetal growth and birth weight in all populations studied to date.1'9'10,12'153'154 However, the magnitude of altitude-associated birth weight decline is variable between populations9,10,12,15,155156 such that long-term, multi-generational high-altitude residents are protected relative to newcomers.10-15,156
Nearly 25 years ago, Haas indicated that infants born in La Paz to high-altitude born European women were 120g lighter than high-altitude Andean women, a difference that

increased to 191 g after controlling for maternal stature, indices of nutritional status, iron status and hemoglobin concentration.13 Haas study did not include a sea level comparison group, which limited the ability to determine whether the effect of hypoxia on birth weight was different between Andeans and Europeans. However, a more recent and extensive review of births to Tibetan, Andean, European and Han Chinese" women across a range of altitudes indicates that the trajectory of birth weight decline with altitude is dependant upon population ancestry.155 Specifically, Tibetans and Andeans, having resided at high altitudes for more than 10,000 and 20,000 years respectively, show one-third the birth weight reduction of European and Han populations who have only recently established permanent residence in high-altitude regions (Europeans: <400 years in South America and <150 years in North America; Han: -50 years in western China).155 Suggesting that the protection afforded by high-altitude ancestry is augmented with increasing altitude, infants born to Tibetan women were 310g and 530g heavier than those born to Han women at altitudes ranging from 2700-3000m and 3000-3800m, respectively.15 Further, Andean Amerindians and Tibetans experience less altitude-associated SGA than do Europeans and Han living at the same altitudes.12'13'15'156
Further, although Han women lived at the highest altitude site, Nachu (4800 m) no infants were born there during the two-year study period because they migrated to low altitude for delivery.15 The protection of birth weight among the Tibetans is likely indicative of adaptive advantage given that perinatal and postnatal mortality estimates were greater in Han than Tibetans,15 which is likely due, in part, to the inability of Han infants to achieve adequate

oxygen saturation at delivery and during the early months life157 or possibly due to complications associated with perinatal pulmonary hypertension.158
Likewise, a medical records review examining 3551 consecutive deliveries at low (300 m, Santa Cruz), intermediate (2500 m, Cochabamba) or high altitude (3200-4100 m, La Paz and 3700 m, Oruro) in Bolivia demonstrated that birth weight decreased and the proportion of infants born SGA increased in all ancestry groups (Andean, European or Mestizo). However, Andeans weighed more and were less often SGA than Mestizos (Andean-European admixture) or Europeans at high altitude. After accounting for the influences of maternal hypertensive complications of pregnancy, parity, body weight, and number of prenatal visits, European relative to Andean ancestry increased the frequency of SGA at high altitude nearly 5-fold. Similar to previous studies in Colorado, altitude increased the incidence of
7 12 159
preeclampsia and gestational hypertension in Bolivian populations. '
Despite the known contribution of health care access, socioeconomic indices and nutritional status to fetal growth, these factors do not appear to explain the aforementioned differences in birth weight among these populations.12,13'15'86,156 In support of an evolutionary explanation, life-long high-altitude residence does not appear to confer protection against the altitude-associated reduction in birth weight or altered maternal vascular adjustment to
159 160

Physiological studies examining maternal adaptations to pregnancy suggest that fetal growth is limited at high altitude, in part, by reduced uteroplacental blood flow. However, neither the mechanisms by which hypoxia influences uteroplacental blood flow nor the factors conferring protection against altitude-associated SGA among long-resident high-altitude populations are well understood.
2.5.2 Human Pregnancy
A brief description of uteroplacental vascular anatomy will serve to orient the following discussion of maternal adaptation to pregnancy. The uterine arteries supply the majority of blood flow to the uterus, although anastomotic branches from the ovarian arteries also contribute a small proportion of the blood supply. The afferent uterine artery splits into an arcuate wreath, sending smaller vessels backwards into the outer muscular layer of the uterus and forward towards the inner layers of the uterine wall.161 The larger branches of the arcuate arteries, the radial arteries, supply the endometrium and continue towards the uterine lumen where, after passing through the myometrium, they are referred to as basal arteries and, more peripherally still, the spiral arteries. The spiral arteries supply the intervillous space with oxygen and nutrient rich blood to support fetal metabolic demands.
The placenta serves as the conduit for the exchange of nutrients and waste products between the maternal and fetal circulations in eutherian mammals. Placentation in humans can be distinguished from that of our haplorhine primate cousins by the extensive invasion of fetal trophoblast cells into maternal spiral arteries and the early and complete implantation of the blastocyst into the uterine stroma (implantation occurs later and is shallower in other

primates studied thus far).162 It has been suggested that the human species has modified features of our haplorhine heritage affecting the uteroplacental circulation to overcome the biomechanical and gravitational constraints posed by the osteological reconfiguration necessary for the evolution of bipedalism.162 Further, anatomical alterations associated with upright posture, including the situation of internal organs (e g. uterus, bladder) within the pelvic cavity and muscular tone of the abdominal wall, results in the compression of the uterus, its supplying vessels and the vena cava.163 Additionally, sympathetic tone is chronically enhanced in humans in order to maintain adequate cerebral perfusion, however this characteristic also impairs venous return and decreases cardiac output.162 164 The unique physiological attributes of human pregnancy are likely a response to the challenges posed by bipedalism in order to permit adequate blood supply to the utereoplacental circulation.
2.5.3 Maternal Adaptations to Normoxic Pregnancy Establishment of Placental Perfusion
The chorionic villous cytotrophoblast cells proliferate and differentiate into syncytiotrophoblast cells, which form the syncytial layer lining the villous placenta, and extravillous trophoblast cells, which invade the uterine parenchyma and spiral arteries and are thus in direct contact with maternal uterine tissue. Interestingly, oxygen partially regulates the phenotype of villous cytotrophoblast cells, such that hypoxia stimulates proliferation but limits invasion into the spiral arteries.165'166 As the extravillous trophoblasts cells invade the spiral arteries they erode the vascular smooth muscle causing a loss of vasoreactivity and enhanced dilatation of the peripheral portions of the spiral arteries so that the vessels obtain a funnel-like shape, a process that is complete by 20 weeks of gestation.167 The spiral arteries then perfuse the

intervillous space and bathe the fetal vessels with oxygen and nutrient rich blood. In preeclampsia, and in some cases of SGA, trophoblast invasion is shallow, inhibiting the remodeling of spiral arteries beyond decidual portions of the vessels. This results in the retention of contractile response of the vessels,168 the maintenance of elevated vascular resistance and reduced uteroplacental blood flow. It has been proposed that the unusually high vascular conductance of the placenta indicates that the central portions of the radial and arcuate arteries dictate uteroplacental placental blood flow.161 However, pregnancy also initiates a profound cascade of systemic maternal physiological adaptations that are integral in permitting adequate oxygen and nutrient delivery to support the metabolic demands of both the mother and the growing fetus. Systemic Maternal Adaptations to Pregnancy
As a result of reduced maternal systemic vascular resistance, plasma volume expansion and increased red cell mass, stroke volume and heart rate and an increase in heart rate, pregnancy results in a nearly 40% increase in cardiac output. Maternal systemic vascular resistance falls with pregnancy,171 likely reflecting changes in circulating levels of vasoactive and angiogenic factors172'174 early in pregnancy. Specifically, production of the vasodilators estrogen, progestserone and nitric oxide (NO')172 increase while circulating levels of the vasoconstrictor endothelin-1 (EDN1) decrease.173 The reduction in peripheral vascular resistance decreases renal perfusion thereby augmenting the production of aldosterone and, in turn, renal sodium and water reabsorption. Elevated aldosterone and pregnancy-associated reductions in the osmotic thresholds for thirst and arginine-vasopressin release contribute to plasma volume expansion during pregnancy.175'177 Inadequate plasma

volume expansion during pregnancy is associated with the development of preeclampsia and fetal growth restriction.164
Alveolar ventilation and tidal volume increase with pregnancy largely due to the stimulatory effect of progesterone on the medullary respiratory centers,178 increased carotid body ventilatory sensitivity to hypoxia and greater central nervous system translation of carotid body signals mediated by estrogen.179 The primary physiological effect of these ventilatory changes is a reduction in PaC02, despite the additional production of C02 from the fetal metabolism. Increased ventilation under normal, sea-level conditions has a negligible effect on oxygen delivery to peripheral tissue since Sa02 is already near maximal.17,23 Since the increase in plasma volume occurs earlier in pregnancy and is larger than the rise in red cell mass, hemoglobin levels decline, producing the characteristic hemodilutional anemia of pregnancy.181'183 Therefore, increased blood flow, rather than oxygen content, is the primary mechanism increasing oxygen and nutrient delivery to the uteroplacental circulation during pregnancy.
The bilateral uterine and ovarian arteries supply the uteroplacental circulation, with two-thirds of total uteroplacental blood flow being supplied by the uterine arteries while the ovarian arteries supply the remaining third. Thus, uterine artery (UA) blood flow may be used to approximate changes in total uteroplacental blood flow. During pregnancy uterine blood flow increases 40-fold184 as a result of increased flow velocity and enlargement of the vessel diameter. The luminal diameter of the uterine vessel increases as a result of increased growth (i.e. hyperplasia and cellular hypertrophy,185,186 alterations in vasoreactivity (i.e.

enhanced response to vasodilators, reduced response to vasoconstrictive substances),187 increased vasodilation in response to flow188 and changes in the active and passive properties of the vascular wall (i.e. reduced myogenic tone and greater distensibility.189 In addition, the proportion of common iliac blood flow distributed to the uterine artery versus the external iliac artery increases with pregnancy,184 demonstrating that pelvic blood flow is redistributed during pregnancy to favor the uteroplacental circulation.
2.5.4 Pregnancy in Chronic Hypoxia
The chronic hypoxia of high-altitude residence interferes with maternal vascular adjustment to pregnancy, attenuating the normal expansion of UA diameter and blood flow and altering patterns of blood flow redistribution seen during pregnancy at low altitude.16,21 Moreover, it has been demonstrated that the protection afforded by multigenerational high-altitude residence against altitude-associated reductions in fetal growth likely involves the retention of normal maternal vascular adjustments to pregnancy despite the hypoxic environment.21
Hypoxia decreases the pregnancy-induced rise in cardiac output, likely due to lower absolute blood volume and/or a failure to decrease peripheral vascular resistance. Failure to
decrease peripheral vascular resistance during high altitude pregnancy may be attributed to increased availability of vasoconstrictors, decreased levels of vasodilators, .limited vascular remodeling194 and/or greater myogenic tone,189 the intrinsic vessel response to transmural pressure or stretch.

The pregnancy-stimulated rise in ventilation raises maternal Sa02 saturation at high altitude and helps to preserve arterial 02 content near sea-level values.17,195 Greater pregnancy-associated hypoxic ventilatory sensitivity, enhanced ventilation and higher arterial oxygen content are positively associated with birth weight,15,22'24 indicating the potential importance of higher maternal 02 pressure and/or content for fetal growth. However, at altitude UA diameter and blood flow near term is reduced relative to low altitude however, 02 content is similar, indicating that diminished uteroplacental blood flow likely plays a more central role in altitude-associated SGA.16,17 Similarly, among women residing at high altitude in Colorado the proportion of lower extremity blood flow distributed towards the uteroplacental circulation is lower in comparison to their low-altitude counterparts. In support of an upstream regulation of diminished uteroplacental blood flow at altitude, uteroplacental impedance parameters, which are used to demonstrate placental insufficiency and adverse pregnancy outcome including SGA or preeclampsia, are reduced at high altitude.
Comparisons of maternal vascular adjustment to pregnancy made between multigenerational high-altitude populations and recent migrant populations to high altitude reveal quite a different story, one that likely explains, at least in part, the variable incidence of SGA between these populations. Among Andean women residing at £3600m in Bolivia volumetric UA flow increased progressively across pregnancy due to an increase in UA vessel diameter and flow velocity.21 By the third trimester 02 delivery to the uteroplacental circulation via the UA among Andean women was more than 3-fold greater than European women residing at the same altitude.21 This difference can be attributed to a larger increase in UA diameter with pregnancy (2-fold difference between populations), reduced vascular resistance in the UA

and greater UA volumetric flow.21 Likewise, Tibetan women, whose infants are also protected from altitude-associated SGA, have higher UA blood flow velocity and a greater distribution of lower extremity blood flow towards the uteroplacental circulation relative to their Han high-altitude counterparts near term.15,24,156 Thus, despite ambient hypoxia, vascular adjustment to pregnancy in Andean and Tibetan women mirrors that seen in healthy, low-altitude pregnancy or, in other words, they maintain a low-altitude phenotype with regard to vascular adjustments during pregnancy. The mechanisms by which these population differences in vascular response to hypoxia during pregnancy arise are not known.

2.6 Oxidative Stress
This section will first review the definition of oxidative stress and describe the mechanisms by which oxidant species are generated. Then, literature documenting the effects of hypoxia and pregnancy on oxidative status will be reviewed. Finally, the mechanisms by which altered redox status may contribute to the development of preeclampsia and/or SGA will be presented.
2.6.1 Definition of Oxidants and Oxidative Stress
Oxygen is not only a requisite for life, but also an accomplice in cell death and tissue damage resulting from oxidative stress. Oxidative stress occurs when there is an imbalance between the production of ROS and antioxidative capacity. The majority of biological free radicals fit into the broader category of ROS, which includes oxygen-containing free radicals or molecules with an unpaired electron (e.g. hydroxyl radical (HO*), superoxide anion radical (02') and NO) as well as their non-radical intermediates (e.g. hydrogen peroxide (H202), hypoclorous acid (HOCI) and the highly reactive peroxynitrite anion (ONOO*)). ROS play an integral role in several biochemical processes including intracellular communication and apoptosis,74,199 regulation of gene expression including that of hypoxia-inducible factor 1a (HIF1a)200 and intracellular defense against microorganisms. Despite the essential role of ROS in normal physiological processes, when the production of ROS outpaces the capacity of antioxidant defense201'205 oxidative damage to tissues ensues.
Oxidative stress has been associated the development of numerous pathologies in humans, including preeclampsia and fetal growth restriction.27,76,206 Since the syndrome of

preeclampsia disappears after delivery, it has long been suspected that uteroplacental ischemia results in the release of a toxic factor from the placenta causing maternal vascular dysfunction and hypertension associated with preeclampsia; what exactly this substance is and how it causes maternal endothelial dysfunction has, until recently, not been well understood. Increasing evidence suggests that the toxic factor(s) may be ROS, leading to oxidative stress and the endothelial damage associated with altered maternal vascular function and impaired uteroplacental oxygen delivery.26,27 The mechanisms by which ROS may contribute to the development of vascular dysfunction will be discussed in detail below.
2.6.2 Generation of Oxidant Species
Atmospheric oxygen is composed of two oxygen atoms bonded to one another. The molecule, however, contains two unpaired electrons, defining it as a diradical. While relatively unreactive, oxygen has a high potential for electron transfer reactions that result in the production of oxidant species capable of cellular and tissue damage. Reactive oxygen species are primarily generated via aerobic respiration and substrate oxidation,207 nicotinamide adenine dinucleotide phoshphate (NAD(P)H) oxidase activity 208 209 and xanthine oxioreductase (XOR)210,211 The most common biological free radicals are the peroxide radicals, nitrogen radicals and superoxide.212 Superoxide is relatively non-toxic but becomes more so when it reacts with either NO' or to H202 to form the highly reactive compounds peroxynitrite (ONOO') or hydroxyl radical (OH*), respectively. This section will provide a brief overview of the mechanisms by which ROS are generated by these processes.
51 Mitochondrial Respiration
Aerobic respiration involves a series of oxidation-reduction reactions in which molecular oxygen serves as the final electron acceptor for cytochrome-c oxidase. Free radical intermediates generated as a normal part of aerobic respiration may donate an unpaired electron to molecular oxygen thereby forming additional free radicals. Moreover, a small percentage of electrons passing through the electron transport chain redox complexes, primarily complex I (NADH coenzyme Q reductase), complex II (succinate dehydrogenase) or complex III (ubiquinol Cyt c reductase), leak" onto molecular oxygen to form free radicals213 The leakage" of electrons or incomplete reduction of molecular oxygen yields 02', H202 or the highly reactive OH* when one, two or three electrons are donated, respectively214"218 Mitochondrial generation of 02" appears to be accelerated in several pathological states including hyperglycemia hyperleptinemia ischemia-reperfusion or hypoxia. NAD(P)H Oxidases
NAD(P)H oxidases, membrane-associated enzymes involved in oxygen sensing, are a major source of superoxide and H202 in neutrophils, cardiomyocytes, vascular endothelial cells and vascular smooth muscle.208,209 NAD(P)H oxidase catalyzes the transfer of electrons from intracellular NADPH (the reduced form of NADP+) across the plasma membrane to univalently reduce molecular oxygen to create superoxide (02') or hydroperoxyl radical (H02'), its protonated form. A number of factors including increased cellular oxygen uptake, angiotensin II, shear stress, neutrophil activation and cytokines enhance NAD(P)H oxidase activity.208 The role of neutrophil activation and enhanced ROS generation in preeclampsia is

supported by a recent study indicating that neutrophils isolated from third-trimester preeclamptic women have increased sensitivity to agonist stimulation and produce more ROS than neutrophils isolated from non-preeclamptic women.222 Further, NAD(P)H oxidases have been implicated in the etiology of vascular dysfunction, including that of pulmonary hypertension at high altitude223, preeclampsia, chronic hypertension and atherosclerosis208 Xanthine Oxidase
'^H^raxide is also generated via the XOR enzymes, which exist in two forms, xanthine
and xanthine oxidase (XO)210'211 XDH and XO serve the same
^^ttje oxidation of hypoxanthine to xanthine and xanthine to 2-6.3 Antioxidant (hat XDH uses NAD+ as the electron
^^ctron recipient, thereby generating
Several defense mechanisms exist to protSt^^^chemia 0r in the decompose, remove or capture free radicals). Oxidants may be nd.XO is

supported by a recent study indicating that neutrophils isolated from third-trimester preeclamptic women have increased sensitivity to agonist stimulation and produce more ROS than neutrophils isolated from non-preeclamptic women. Further, NAD(P)H oxidases have been implicated in the etiology of vascular dysfunction, including that of pulmonary hypertension at high altitude preeclampsia, chronic hypertension and atherosclerosis. Xanthine Oxidase
Superoxide is also generated via the XOR enzymes, which exist in two forms, xanthine dehydrogenase (XDH) and xanthine oxidase (XO).210'211 XDH and XO serve the same enzymatic function, catalyzing the oxidation of hypoxanthine to xanthine and xanthine to urate. The relevant difference between isoforms is that XDH uses NAD+ as the electron acceptor in the reaction, whereas XO uses 02 as an electron recipient, thereby generating superoxide.210 XDH can be converted to XO during periods of hypoxia, ischemia or in the presence of proinflammatory mediators such as tumor necrosis factor-a.224 Indeed, XO is
responsible for the free radical production associated with ischemia-reperfusion.221 However,
superoxide production by XO is far less than less that of NADPH oxidase.
2.6.3 Antioxidant Defenses and Assessment of Oxidative Stress
Several defense mechanisms exist to protect against cellular oxidative damage (i.e. to decompose, remove or capture free radicals). Oxidants may be non-enzymatically

decomposed by transition metals, removed via disproportionation reactions into less reactive components or be captured by antioxidant molecules and/or enzymes.225 While the relationship between dietary antioxidants and altitude-associated pregnancy complications is of significant interest, this study focuses on the role of endogenous antioxidant enzymes, namely superoxide dismutase (SOD) and catalase. This approach was taken for several reasons. First, in the process of detoxifying oxidant species, dietary antioxidants such as those used in the VIP trial are stoiciometrically consumed. In contrast, enzymatic antioxidants such as SOD and catalase simply act as catalysts of dismutation reactions, rather than substrates. Thus, the capacity of enzymatic antioxidants to detoxify oxidants is greater than that of ascorbic acid or tocopherol, for example. Secondly, recent evidence has demonstrated that supplementation with dietary antioxidants does not diminish the likelihood of preeclampsia, and in some cases has a negative effect on fetal growth and health outcome. The negative result for dietary antioxidants in the protection against preeclampsia does not negate the potential for antioxidants to be used as part of a treatment or prevention protocol; the study results may be an issue of dose or timing of supplementation. Enzymatic Antioxidants
Superoxide dismutase, an enzymatic antioxidant, catalyzes the dismutation reaction of 02'", producing hydrogen peroxide and oxygen. Three types of SOD, based on composition and location, exist in mammalian cells. Copper zinc-SOD (CuZn-SOD; SOD 1) is confined to the mitochondrial intermembrane space, nucleus, and cytosol. Manganese SOD (Mn-SOD; SOD
2) exists within the mitochondrial matrix, sequestering a portion of the oxidant radicals formed

by respiratory chain. Extracellular SOD (EC-SOD or SOD 3) is secreted across the mitochondrial plasma membrane and binds to cell surfaces.226 SOD catalyzes the reaction of 02" molecules to form H202 and 02. Catalase, a stable, ubiquitous, haem-containing enzyme whose primary function is to catalyze the decomposition of H202 to water and oxygen, is known to have one of the most rapid turnover rates of all enzymes. The highest concentrations of the primarily intracellular enzyme, catalase, are found in liver cells and erythrocytes. Assessment of Oxidative Stress
Several biochemical markers may be used to assess oxidative stress, the most common of which are products of lipid peroxidation. Lipid peroxidation ensues when ROS interact with the polyunsaturated fatty acids in membranes or lipoproteins, converting the fatty acids into lipid peroxides and secondary metabolites. Direct assessment of oxidative stress is notoriously difficult to quantify, as ROS and other oxidant species are extremely short-lived. Thus, markers of oxidative stress, such as those used in this study (i.e. isoprostanes), are used to demonstrate the presence or absence of an imbalance between ROS and antioxidant capacity. Since leptin increases lipid peroxidation, we also discuss the effect of pregnancy, altitude and ancestry on maternal leptin levels, although it is not a marker of oxidative stress, it is considered to be a precursor.
Despite extensive, redundant and powerful antioxidant defense systems at times the production of ROS exceeds antioxidant capacity, resulting in oxidative stress. Of particular importance for the present research, it is increasingly appreciated that 02, H202 and

oxidative stress influence endothelial function and phenotype and that they are key players in physiological regulation and in disease pathophysiology.227 While the cellular mechanisms responsible for the production of ROS are well elucidated in current literature the environmental and physiological forces driving these responses and the involvement of ROS in the etiology of vascular dysfunction and associated pathologies are less straightforward.
2.6.4 Hypoxia-Induced Oxidative Stress
It is well established that exposure to hyperoxia increases the production of ROS. This has been demonstrated in multiple cell lines, lung tissue, Saccharomyces cerevisiae (yeast) and isolated mitochondria.228'233 However, whether hypoxia elicits increased ROS production or oxidative stress remains somewhat controversial;234,235 in fact, while the vast majority of studies indicate that hypoxia does enhance ROS production some research groups report diminished ROS with hypoxic exposure (reviewed in Chandel, 2007).
Several lines of evidence support an increase in ROS production and oxidative stress as a result of hypoxic exposure. For example, hypoxia increases ROS formation in a variety of cell types including endothelial cells236, hepatocytes237, and cardiomyocytes238. Likewise, hypoxia has been shown to augment oxidation reactions in various cell lines and tissues239"243 and to cause oxidative lesions to mammalian nuclear DNA and specific gene regions (e.g. vascular endothelial growth factor (VEGF) in lung endothelial cells).244,245 In intact rat models, a direct relationship was found between the degree of hypoxia and ROS production.193
Human studies demonstrate that inspiratory and hypobaric hypoxia increase markers of

oxidative stress.30-32 Acute hypobaric hypoxia (4h at 5500m) increases markers of lipid peroxidation, protein oxidation and oxidized glutathione,32 while chronic altitude exposure (4300m) increased markers of lipid peroxidation.31 Of particular interest for our study, research suggests that increased oxidative stress may be associated with conditions of impaired adaptation to high-altitude including Monge's disease and pulmonary hypertension 31 High-altitude residents with excessive erythrocytosis (chronic mountain sickness or Monges disease) have greater oxidative stress than their healthy high-altitude counterparts as measured by plasma TBARS and urinary F(2)-isoprostane. Thus, although it has not been established whether oxidative stress in a cause or effect of chronic mountain sickness, lipid peroxidation products increase with altitude and individuals with impaired adaptation to hypoxia appear more affected than healthy individuals.
Thus, the current consensus is that both hyperoxia and hypoxia increase ROS production, possibly to different degrees, from different sites and with divergent consequences.226 It has been suggested that enhanced ROS production due to hyperoxia may trigger cell death and senescence while lower, but still elevated, levels of ROS as a result of hypoxia may serve to increase the transcription of the hypoxia-dependent gene regions.226 In support of this speculation, under chronic low oxygen concentrations (3%), HIF1a is stabilized in parallel with an increase in mitochondrial superoxide.43 Of note, the hypoxia-associated increase in mitochondrial ROS production and stabilization of HIF1a was abolished with the addition of a mitochondrial specific antioxidant.28
2.6.5 Oxidative Stress: A Usual Suspect

Current literature strongly links placental and maternal oxidative stress to the development of preeclampsia at low altitude. Increasing evidence suggests that uteroplacental
ischemia results in the release of ROS from the placental circulation, and that these ROS contribute to vascular dysfunction characteristic to preeclampsia via oxidative damage to the vascular endothelium. It has also been hypothesized that oxidative stress in the
placenta triggers increased apoptosis and, subsequently, augmented shedding of syncytiotrophoblast fragments into the maternal circulation that leads to a strong maternal inflammatory response and vascular dysfunction.246 It is suspected that poor placental perfusion may initiate the free radical process thereby increasing lipid peroxidation.247'249
Abundant, biologically plausible evidence indicating that oxidative stress, as demonstrated in both placental tissue and in the maternal circulation, is involved in the etiology of preeclampsia and suggests that the disorder may be amenable to antioxidant therapies. It should be noted that it has not been established whether oxidative stress in preeclampsia is a cause or effect of the pathology, or both. Evidence of Oxidative Stress in Preeclampsia and SGA
Several lines of evidence demonstrate that oxidative processes are augmented in placental tissues obtained from preeclamptic, relative to normotensive, women.248,250,251 Lipid hydroperoxide concentrations are higher in deciduas basilis tissue isolated from women with preeclampsia relative to their normotensive counterparts pregnancies.252 Levels of free isoprostane (8-iso-PGF2a), which is formed exclusively by free radical initiated peroxidation of

arachidonic acid and has vasoconstrictive properties, are 2-fold greater in placental tissue obtained from preeclamptic versus normotensive women.254 Increased nitrotyrosine immunostaining, a footprint of oxidative damage to proteins, has been demonstrated in placental villous vascular endothelium, surrounding vascular smooth muscle and villous stroma in pregnancies complicated by preeclampsia or SGA relative to normal pregnancies.255
Leptin, a hypoxia-inducible protein that increases mitochondrial 02' production220 and lipid peroxidation, increases dramatically in the maternal plasma in normal pregnancy, an effect that is exaggerated in preeclampsia256 and SGA, particularly SGA associated with altered vascular function. This difference in has been attributed, in part, to the fact that in normal pregnancy ROS are counter-balanced by an adequate increase in antioxidant activity.257 Thus, studies have sought to determine whether higher maternal antioxidant capacity decreases the risk of preeclampsia, if lower antioxidative capacity increases risk, and/or whether antioxidant supplementation reduces the incidence of preeclampsia in high-risk populations. Evidence of Diminished Antioxidant Capacity in Preeclampsia and SGA
Evaluations of antioxidative capacity in the maternal circulation, whether by dietary antioxidant concentrations, antioxidant enzyme activity or total antioxidant capacity, indicate that maternal antioxidant capacity is reduced in preeclampsia.258'261 Antioxidant capacity may

be limited in preeclampsia and SGA either by low intake of dietary antioxidants and/or by decreased activity or production of endogenous antioxidant defenses.
The first study to cite decreased mean ascorbate levels during pregnancy was conducted in 1964.262 Since that time several studies have demonstrated similar results in cross-sectional and longitudinal study designs alike. Ascorbate reserves have been shown to be 50% lower in preeclampsia relative to normal pregnancy.43 Prospective, longitudinal studies (week 16 to term) report abnormally low plasma ascorbate concentrations prior to the onset of clinical disease, which suggests that hypertension does not initiate the decline in plasma ascorbate.206 Chappell, et al. suggest that this difference likely reflects increased metabolic ascorbate consumption, but may result from variable dietary intake as wel.206
Plasma tocopherol (gamma-tocopherol and a-tocopherol) levels have been shown to be higher263, lower264 and equivalent43,206 in preeclamptic versus normotensive pregnancy. Since vitamin E is carried on lipoproteins this variation may reflect different methods of correction, or non-correction, for plasma lipids.35 However, as described by Chappell, et al. vitamin E supplementation may decrease lipid peroxidation in women at high risk for preeclampsia.206 Specifically, plasma markers of lipid peroxidation (8-epi-PGF2J in women at high-risk for preeclampsia (history of preeclampsia, UA resistance 295th percentile at 20 weeks) supplemented with vitamin E (400IU/d) and vitamin C (1000mg/d) were equivalent to levels in low-risk women and inversely related to a-tocopherol concentrations.206 This effect of vitamin

E has been attributed to its ability to inhibit NAD(P)H oxidase, reduce placental apoptosis, and inhibit leukocyte and endothelial activation.
Carotenoids and selenium have also been hypothesized to be protective against the development of preeclampsia. Case-control studies have demonstrated that women with plasma p-carotene concentrations in the highest versus lowest quartile had a 50% lower risk of preeclampsia (OR = 0.50, 95% Cl: 0.25 to 1.00).265 Serum selenium levels were shown to be markedly lower weeks 28-39 of pregnancy in preeclamptic than non-pregnant or healthy pregnant women, and to be inversely proportional to placental malondialdehyde (MDA) levels 266. In support of the relationship between selenium deficiency and hypertensive response during pregnancy are studies indicating selenium-free versus normal diets elicit pregnancy-specific increases in systolic blood pressure (116.45.2 mmHg vs 108 8.8 mmHg; p<0.05) and proteinuria (9.68 2.12 pg/ml vs 5.93 1.59 pg/ml) in rats.267 The pups born to rats fed the selenium-free versus a normal diet were significantly lighter at birth despite similar placental weight and litter size.
In comparison to the amount of literature regarding dietary antioxidants and maternal complications of pregnancy, relatively little attention has been paid to the role of antioxidant enzymes in the protection against the development of these disorders. SOD and glutathione peroxidase have been shown to be lower in preeclamptic women.266,251 Consistent with these results, hypertensive, relative to normotensive, patients demonstrated a higher GSSG (oxidized glutathione) to GSH (reduced glutathione) ratio, suggesting oxidative imbalance

and/or altered glutathione peroxidase activity.268 In contrast, maternal erythrocyte catalase activity was elevated and inversely proportional to birth weight.266
Placental tissue obtained at delivery from women residing at either high (3100m) or medium altitude (1600m) was found to have reduced endogenous antioxidant activity, as determined by SOD, glutathione peroxidase, thioredoxin reductase and thioredoxin.269 Despite reduced enzymatic antioxidant activity, the placental tissue obtained from women at high altitude did not appear to have been exposed to greater oxidative stress, as assessed via lipid peroxidation or protein carbonyl formation, than that of low altitude women.269 However, indicating augmented nitrative stress in the trophoblast at altitude, nitrotyrosine staining in the synciotrophoblast and the extravillous trophoblasts was greater in high- than low-altitude
placental tissue. It is possible that samples obtained from 1600 m are not representative of low altitude, although pregnancy outcome is not altered at this altitude relative to sea level. Further, it is possible that markers of oxidative stress and/or antioxidant status obtained in the placental tissue at delivery are not representative of earlier time points in pregnancy when vascular remodeling and growth in the placental and maternal vessels is occurring.
The literature clearly demonstrates increased oxidative stress at altitude and with pregnancy. Since the risk of preeclampsia and intrauterine growth restriction are greatly increased at high altitude and are accompanied by diminished uterine artery blood flow, it seems likely that ROS and oxidative stress may contribute to the increased incidence of these disorders.

2.6.6 ROS and Vascular Function
Despite promoting processes that improve blood flow and oxygen delivery such as neovasularization and angiogenesis,270 accumulating evidence suggests that ROS play a causal role in the development of several vascular disorders including essential hypertension and preeclampsia, the microvascular damage seen in ischemia-reperfusion, hemorrhagic shock-resuscitation, diabetes and hypercholesterolemia193,271 as well as hypoxia-induced inflammation.
Reactive oxygen species are thought to initiate endothelial dysfunction through a variety of mechanisms including oxidative damage to endothelial tissue, enhancing the production of proinflammatory mediators and decreasing the bioavailability of NO. ROS signals originating from the mitochondrial electron transport chain during hypoxia trigger a series of responses that regulate the proinflammatory cytokines, including the activation of IL-6, and increases leukocyte adherence, which increases endothelial permeability as is apparent in preeclampsia.236
ROS are thought to compromise the vascular endothelium, in part, by decreasing the relaxing capacity of resistance vessels in response to endothelium-dependant mechanisms in a manner similar to that observed in preeclampsia at low altitude.26'34-36 With a half-life of only a few seconds, the 02'" radical is unstable in aqueous solution. Superoxide can react with itself, albeit rather slowly (rate constant: 8 X 104mol-1 s-1) to form H202 and 02. In the presence of SOD this reaction proceeds more rapidly (rate constant: 2 X 109 mol-1 s-1).

However, NO' reacts even more avidly with 02" (rate constant: 7 X 109 mol '1 s'1) to generate peroxynitrite, thereby decreasing NO' bioavailability227 by sequestering it in tissues. Hypoxia-induced 02*' generation occurs in parallel with lower bioavailability of NO',272 suggesting that either NO' generation is diminished in hypoxia or that NO is being 'consumed' by 02' due to increased production of ONOO'. Evidence strongly supports that NO' generation is not impaired in hypoxia, but rather is depleted by increased ROS production 193. Recent research suggests that some endothelium-derived constituent or product, likely ROS, participates in the inhibition of vasodilation in hypoxia.33
NO-induced vasodilation plays a critical role in normal maternal vascular adaptation to pregnancy. Chronic treatment with the nitric oxide synthase (NOS) inhibitor L(G)-nitro-L-arginine methyl ester (L-NAME) produces a preeclampsia-like syndrome in rats, characterized by hypertension and reduced fetal growth.273 Interestingly, the administration of an NOS inhibitor N-nitro-L-arginine (L-NNA) decreases vasodilation of guinea pig uterine arteries in response to flow vasodilation, whereas L-NNA increased flow vasodilation in UA from chronically hypoxic animals.188 The authors suggest that this response to NOS inhibition may be due to the reduced ROS formation (NO' + 02").
Using total circulating NO' metabolites (NOx; nitrates, nitrites and nitrosamine) as a proxy for NO' bioavailability, preliminary data from our research group indicates that NO' bioavailability is lower during pregnancy at high than low altitude and is greater among Andean than European women during high-altitude pregnancy. These results are consistent with the possibility that hypoxia reduces NO bioavailability by decreasing NO* production or, as suggested above, increasing ROS formation. The latter suggestion is supported by data from

studies in non-pregnant humans as well as in pregnant ewes indicating that NO' production is elevated at high altitude.187,274277 Thus, if the production of ROS is enhanced at high altitude it seems likely that reduced NO' bioavailability and endothelial damage may play critical roles in the vascular dysfunction associated with preeclampsia and SGA.
2.6.7 ROS and the Regulation of HIF1
Hypoxia initiates a 'transcriptional response that serves to maintain tissue oxygenation and increase glycolytic capacity despite reduced oxygen availability.278 Several genes are known to be regulated, at least in part, by hypoxia including several involved in angiogenesis, erythropoesis, vasoaction and glucose transport. Each of these genes contains the signature hypoxia response element (HRE; core consensus sequence 5-CGTG-3). The regulation of these genes in response to changes in 02 tension is mediated primarily by hypoxia-inducible transcription factors, HIF1a and HIF2a, which are composed of 02-sensitive (HIF1a) and constitutively expressed (HIFip/ARNT) sub-units.279,280 Under normoxic conditions, HIF1a is proline hydroxylated, which facilitates its recognition by von Hippel-Lindau tumor suppressor protein (VHL), an E3 ubiquitin ligase, and subsequently targets HIF1 for degradation.281'284 Conversely, under hypoxic conditions HIF1a is preserved (i.e. HIF1a which has not undergone proline hydroxylation goes unrecognized by VHL protein). Thus, hypoxia permits the accumulation of HIF-a, enabling it to interact with ARNT, bind to the HRE and begin transcription.285 The 02 sensor(s) upstream of the HIFs remains unknown, however there is considerable evidence that the mitochondria and ROS released from the cytosol may

serve as an 02 signal for the HIFs241 and subsequently, augment the transcription of hypoxia-regulated genes.
While the conditions under which ROS modulate HIF1a activation have yet to be clarified, it is well accepted that ROS levels have an impact on both the hypoxic and non-hypoxic regulation of HIF1a.287 Not surprisingly, antioxidant systems also play a role in the ROS modulation of HIF1a. For instance, in pulmonary artery muscle cells treatment with the antioxidant ascorbate prevented ROS production, the accumulation of HIF1a and activation of the HIF pathway induced by the agonists thrombin and CoCI288 Likewise, ascorbate ablated HIF1a protein expression and inhibited HIF-1 dependent VEGF expression in aortic smooth muscle cells.289 Catalase, an endogenous antioxidant enzyme, has been shown to inhibit agonist-induced (i.e thrombin, CoCI2, amgiotensin II) HIF1a levels.288,290 Likewise, under hypoxic conditions the overexpression of SOD diminished HIF1a levels in Hep3B cells. Thus, the potential role of ROS and/or reduced antioxidant capacity in the control of HIF-mediated gene expression presents another possible mechanism by which redox status may influence vascular function, beyond direct cytotoxic effects.

3. Research Design and Methodology
This section includes the general study approach, followed by a detailed description of each study aim, specific study preparations, ethical considerations, and statistical methods.
3.1 Research Setting
The study was conducted in Bolivia, the country in the western hemisphere with the largest, highest and longest-resident high-altitude population as well as a recent influx of European-derived populations to both the high-altitude and low-altitude locales.81 The high-altitude studies were conducted in clinics and hospitals located in La Paz or El Alto (3600 4100m); the low-altitude studies were conducted in Santa Cruz (300m). In each of these locations, physicians working at private clinics (CEMES in La Paz and Clinica Sirani in Santa Cruz) as well as at public hospitals belonging to the Caja Nacional de Salud (CNS, the nations largest health-care provider to insured workers) and clinics Clinica CIES (Centro Integral de Educacion Sexual; Santa Cruz) and Clinica Grupo Medico Solidario (Santa Cruz) were informed of the study and invited to refer subjects to study coordinators.
3.2 Subjects and Subject Recruitment
The low-altitude sample consisted of 22 Andean and 32 European women measured at three time points; non-pregnant (NP), 20 and 36 weeks (Table 3.1). The high-altitude sample consisted of 19 Andean and 22 European women in the NP state, 19 Andean and 17 European women at 20 weeks and 47 Andean and 24 European women at 36 weeks. All women at low altitude and the European women at high altitude were a subset of a larger,

prospective investigation, and thus are longitudinal samples. The high-altitude Andean sample is cross-sectional.
Women were required to be between the ages of 14 and 45, self-identified as being of Andean or European ancestry, residing at low or high altitude (a 1 year), receiving prenatal care, have a singleton pregnancy, of good general health, and free of known risk factors for adverse pregnancy outcome (e.g., chronic hypertension, diabetes).
Table 3.1. Subject numbers
Pregnancy group
Ancestry Altitude Non-pregnant 20 weeks 36 weeks
Andean Low 22 22 22
High 19 19 47
European Low 32 32 32
High 22 17 24
3.3 Data Collection
Maternal interviews were conducted to obtain a full reproductive health history, educational and socioeconomic information, a general scan of health behaviors that could influence fetal growth, pregnancy outcome or the variables of interest in this investigation (e.g. smoking, alcohol consumption, exercise level and vitamin and medicine use) and to assess the degree

of high-altitude ancestry. At each of the three time points (20 weeks, 36 weeks and nonpregnant) a general clinical assessment including the measurement of blood pressure, heart rate, weight, skinfold thickness and screening for edema or proteinuria was conducted. The physical exam was followed by blood sample collection and the measurement of UA blood flow characteristics. Fetal biometry was assessed at 20 and 36 weeks of pregnancy. The pregnancy category non-pregnant consisted of women who were :3 months post-partum or nulliparous. Birth weight, gestational age, preeclampsia and other infant and maternal complications were determined using information obtained from medical records following delivery. Detailed methodological considerations for each variable are discussed below.
Whole blood was obtained at each time by routine venipuncture and prepared as needed for the measurement of oxidative stress, endothelial damage and antioxidant status.
3.3.1 Maternal Attributes
General maternal characteristics were identified by questionnaire (age, place of birth, current and childhood residence, duration of high-altitude residence, maternal and paternal surnames, marital status, gravidity, parity, complications of previous pregnancies, normal body weight, educational, health and employment history, general health behavior such as smoking, alcohol consumption, exercise levels, vitamin or medication use). Hypertension during pregnancy was defined as > 2 BP readings > 140/90 mmHg at least 6h apart in a woman who was normotensive before week 20 or postpartum; these data were obtained from both the clinical assessments and from prenatal medical records obtained after delivery. Women were classified as preeclamptic when both hypertension and proteinuria (qualitative

reading of > 1+ proteinuria or 300 mg/L in a 24-hour collection) were noted either during physical examination and/or in medical records. Women with a single elevated blood pressure with or without proteinuria were classified as normal.
3.3.2 Fetal and Neonatal Characteristics
Fetal biometry was obtained using an ATL 3000 or 5000 (La Paz and Santa Cruz, respectively) (Global Medical Imaging, NC, USA) or MicroMaxx (El Alto) (SonoSite, Inc, WA, USA) Doppler ultrasound system. Measurements included head circumference (HC), abdominal circumference (AC), biparietal diameter (BPD) and femur length (FL). Umbilical (UMB) and mid-cerebral artery (MCA) peak systolic to diastolic velocity ratios were also calculated as a measure of fetal compromise (higher values for UMB S/D ratios and lower MCA S/D ratios are indicative of fetal compromise). Biometry images were printed and stored with individual maternal records.
Birth weight, ponderal index, head and abdominal circumference, and gestational age were obtained following delivery from labor and delivery records. Gestational age was back calculated from the first day of the last menstrual period, estimated by fetal biometry obtained using ultrasound at 20 weeks292 and as a clinical estimate based on physical exam at delivery. Reported values for gestational age are those calculated using date of last menstrual period except when these differ by more than 2 weeks from the estimations determined by fetal biometry, or in cases where the date of last menstrual period was unknown. In the latter two cases, gestational age was back calculated from the date of birth

to the estimated gestational age assessed by fetal biometry at week 20. SGA was defined as a birth weight <10th percentile for a given GA and sex based on data that has been validated with perinatal mortality criteria 293 at sea-level in the United States. Ponderal index was calculated as 100 x birth weight (gm) / crown-heel length (cm)3. Newborn respiratory complications were defined as any condition requiring oxygen treatment at birth or during the first few days of life; diagnosis of respiratory distress, apnea, respiratory depression, pulmonary hypertension, or hyaline membrane disease; or newborn hypoxia as a discharge diagnosis. Babies born at £37 weeks of gestation were considered to be pre-term.
3.3.3 Blood Sample Collection
Blood sample collection for the antioxidant, oxidative stress and endothelial damage marker assays consisted of a 14 ml of blood draw (8 ml into a heparinized vacutainer and 6 ml into a serum separator vacutainer) at each study point using standard venipuncture techniques. Assessments of Oxidative Stress
Isoprostane (8-iso-PF2a) concentrations in the maternal plasma were quantified using an HPLC mass spectrometry assay.294 Validation studies using human plasma demonstrate that this method is linear for concentrations between 0.0025 and 80 (xg/L and, for samples above the lower detection limit, has an inter-day accuracy and precision of <10%. Similar to the protocol described in Haschke (2007), the protein precipitation solution (methanol/0.2 mol/L ZnS04, 7:3, v/v) contained an internal standard (8-F2t-lsoP-d4) at 4pg/L (Cayman Chemical); 500 |xL protein precipitation solution was added to an equal volume of plasma. After vortexing (1 minute) and centrifugation (13000g, 10 min, 4C), the supernatant was

transferred into an HPLC vial and stored at -80C until being placed in the autosampler at 4C. The standard curve was generated using 2.5 pg/ml, 5.0 pg/ml, 10.0 pg/ml, 25 pg/ml, 50 pg/ml, 100 pg/ml, 250 pg/ml and 500 pg/ml standards (standard in pooled human plasma).
To ensure accuracy of the measures a blank (pooled human plasma) and quality control samples of 15pg/ml, 75 pg/ml and 300 pg/ml were also quantified.
Leptin is known to induce lipid peroxidation, and thus is included in the oxidative stress section here as well as in the discussion section. Leptin was measured using a quantitative sandwich enzyme immunoassay technique (R&D Systems, Minneapolis, MN) with a sensitivity of 7.8 pg/ml, intra-assay CV of 3.2 3.3% and an inter-assay CV of 3.5 5.4% and a recovery range of 91 100%.
Total serum nitrite (nitrite and nitrate converted to nitrite) was assessed using an assay that is based on the conversion of nitrate to nitrite by nitrate reductase (R&D Systems, Minneapolis, MN). Nitrite concentrations are quantified via colorometric analysis of the azo dye obtained after reaction with the Griess reagent. Acidified N02' produces a nitrosating agent, which reacts with sulfanilic acid to produce the diazonium ion. This ion is then coupled to N-(1-naphthyl) ethylenediamine to form the chromophoric azo-derivative. Assessment of Antioxidant Status
Total SOD activity (mitochondrial SOD, Mn-SOD', cytosolic SOD, Cu, Zn-SOD and extracellular SOD, EC-SOD) was assessed by the xanthine oxidase/xanthine/cytochrome c method, which is a spectrophometric assay for quantifying the SOD activity of any solution or

cell extract.295 As described by McCord, the principle of the assay depends upon the ability of SOD to compete equally with a known standard reaction for 02". The standard reaction superoxide reacts with ferricytochrome c, under known pH, ionic strength, temperature and cytochrome c concentration, which determines the rate of disappearance of superoxide via reaction with cytochrome c;295 the reaction is represented by the equation:
Rcyi c =kcyt c [02] [cytochrome c]
where kcyt c is the rate constant for the reaction of cytochrome c with 02" (~6 x 105 M'1sec'1 at pH 7.8, 25C).296 The rate of disappearance of 02" via a catalyzed dismutation (RSod) using one-standard unit of SOD is represented by the equation:
Rsod ~ kSoD [02 ] [SOD]
where kSoD is the rate constant for the reaction of SOD with 02" (~2 x 109 M'1sec'1 at pH 7.8, 25C).297 In the presence of one-standard unit of SOD, the disappearance rate of 02" in the standard reaction (Rcyt c) is equal to that of RSod, thus:
[SOD] = kcyt c [cytochrome c] / kSoD = 3 x 10'9 M.
The assay determines the ability of the sample to inhibit the reaction between cytochrome c and 02" where greater inhibition (or faster disappearance rate of 02") means higher SOD activity in the sample of interest.

Catalase was assessed using the method of Beers and Sizer, which determines the rate of disappearance of H202 by catalase.298 The buffer solution used was prepared using 100ml of 50mM Kpi + 0.1 mM EDTA, pH 7.4 and ~240nl 30% H202, 8.8M, and its absorbancy verified to be 0.872 5% at 240X. 1ml of the buffer solution and 2\i\ of erythrocytes (100-fold dilution with HPLC-grade water) were added to a cuvette and absorbance was determined at 240 -500X. Values for catalase were determined using the following calculation:
60s x A x Cv/Sv x D/0.0436
where A is the absorbance (au/s), Cv is the total volume of solution in the cuvette (nl), Sv is the volume of sample (jil), D is the dilution factor and 0.0436 (au/s) is the absorbance of the buffer solution. Endothelial Damage and Vasoconstriction
Serum EDN1 was measured using immunoassay kits (R&D systems, Minneapolis, MN) following manufacturer specifications. The EDN1 assay is a solid phase quantitative sandwich ELISA that determines relative mass values of natural human EDN1 and has a detection limit of 0.16 pg/ml and an inter-assay CV of 10%.
3.3.4 Uterine Artery Blood Flow Characteristics
UA blood flow characteristics were measured bilaterally, in triplicate, using an Advanced Technologies Laboratory (ATL 3000 (3600 m) or 5000 (400 m)) (Global Medical Imaging,

NC, US) or MicroMaxx (3600 -4100 m) (SonoSite, WA, USA) Doppler ultrasound system. Ultrasound studies were performed while the subject was in the prone position, slightly tilted to avoid pressure on the inferior vena cava by the uterus.
The crossover point of the external iliac and uterine artery provides a convenient anatomical landmark to localize the UA. The UA is clearly distinguished from the external iliac artery by its low pulsatility index and low resistance waveform. Using color imaging, the external iliac was traced to the location of apparent crossover with the UA, at which point UA diameter measurements were made. UA diameter was measured in triplicate with color imaging, and was obtained using a longitudinal view at a high angle of insonation using the cine loop feature in order to obtain a well-articulated and parallel view of the vessel margins. Since the use of color imaging over estimates actual UA vessel diameter, values were subsequently color-corrected using the difference between with and without color measurements obtained at a similar anatomical depth in the common iliac vessel. UA flow velocity was measured in the same portion of the vessel used to obtain diameter measurements. The angle of insonation was adjusted to obtain a waveform with a sharp, clear outline at the smallest possible angle of insonation. After at least three well-defined, consecutive waveforms were obtained, the time averaged mean (TAM) velocity flow was calculated throughout the cardiac cycle. The waveforms used to determine flow velocity were also used to calculate peak -systolic velocity (PSV) and end-diastolic velocity (EDV), pulsatility index (PI; (PSV-EDV) / TAM), resistance index (Rl; (PSV-EDV) /PS V), the systolic to diastolic ratio (S/D ratio) (peak systolic/end diastolic velocity). To minimize measurement error introduced by large

angles of insonation velocity measurements were obtained at an angle of insonation as near to zero as possible and were excluded from analysis if greater than 50 degrees.
Volumetric flow was calculated using the vessel cross sectional area (jtr2) and TAM by the following equation: 60*(nr2)*TAM and is expressed in ml/min. Volumetric flow for the right and left UAs were calculated separately, summed for a single measurement and then averaged; values are expressed as mean volumetric flow (ml/min).
Upon inspection, UA blood flow values obtained using the MicroMaxx ultrasound system were systematically greater than those obtained using the ATL systems and was primarily due to variation in vessel diameter rather than flow velocity. Each ultrasound unit was internally consistent. At high altitude, UA blood flow characteristics for all of the Andean women and 2 of the 20 European subjects at 36 weeks high altitude were obtained using the Sonosite MicroMaxx Doppler ultrasound system. Upon further inspection, we found that the operator of the SonoSite unit had obtained UA diameter measurements in systole whereas, similar to our previous studies, the operator of the ATL units simply averaged three UA diameter measurements. Considering that the 2/3 of the cardiac cycle is spent in diastole, it is more likely that ATL diameters reflect diastolic diameter. Thus, the point of the cardiac cylce at which diameter measurements were taken likely accounts for some of the variation observed. Our research group has conducted serial studies of UA blood flow during pregnancy in high-altitude Andeans previously using the ATL ultrasound system. Thus, based on UA diameter and UA blood flow measurements obtained in Andean and European

women at 36 weeks from our previous studies a scaling factor was derived to account for systematic differences in diameter measurement between the Sonosite and the ATL systems.
Regression equations were obtained for the relationship between UA volume flow and diameter for data obtained from each system:
ATL: UA volume flow = 1.822 (UA diameter) 0.5972
MicroMaxx: UA volume flow = 2.443 (UA diameter) 0.6702
This time point was chosen since there were both European and Andean high-altitude subjects in whom measurements were taken using both systems. UA volumetric flow data obtained using the Sonosite system were then adjusted using the regression equation obtained for the ATL derived results. Adjusted UA volume flow values were calculated using the ATL regression equation and the Sonosite diameters. These adjusted UA flows were then used to generate adjusted UA diameters using the Sonosite regression equation. This diameter was then used to calculate an adjusted UA volume flow using the equation: 60*(nr2)*TAM.
3.3.5 Ethical Considerations
Prior to participation all women signed the informed consent form approved by the Colorado Multiple Institution Review Board (COMIRB; protocol #99-866), the Health Insurance Portability and Accountability Act (HIPPA) as well as human subjects protection forms required by Colegio Medico, the Bolivian equivalent of COMIRB, prior to participation. All subjects were assured that their participation was entirely voluntary, and that they could withdraw at anytime, for any reason. Our subject population included monolingual Spanish-

speakers, or bi-lingual Quecha-Spanish or Aymara-Spanish speakers. Thus Spanishspeaking staff at ail study sites were familiarized with consent procedures and were responsible for obtaining verbal and written consents to ensure complete understanding of the study by each participant. Copies of consent forms are included in Appendix A and B.
3.3.6 Statistical Analyses
After screening for errors in data entry, data dispersion for each variable was assessed by interquartile range (IQR; 3rd quartile (Q3) 1st quartile (Q1). Points identified as being Q3 + 3*IQR were considered to be extreme outliers and were excluded from analyses. Outliers within 3*IQR of Q1 or Q3 were retained.
Maternal attributes, fetal biometry and delivery and newborn characteristics between altitude and ancestry groups were compared using one-way analysis of variance (ANOVA) for continuous variables and chi-squared (%2) or Fisher exact tests for categorical variables, as appropriate. Categorical variables were dichotomized using a biological relevant threshold (e.g. prematurity and SGA) or were self-dichotomized (e.g. prenatal vitamin use, yes/no). Vascular parameters, endothelial damage, oxidative stress and antioxidant status across pregnancy were assessed using 2-way ANOVA using ancestry and altitude as main effects. Scheffes post-hoc tests were used to determine the source of significant differences observed. Factors known to be associated with the dependent variable of interest were included as covariates. Differences were considered to be significant when p<0.05 and

trends reported when 0.05< p<0.10 (t). Values are expressed as mean SEM for continuous variables and as percentages and 95% Cl for categorical variables.
The relationships between oxidative stress, endothelial damage, circulating levels of vasoactive factors, determinants of uteroplacental blood flow, pregnancy outcome (birth weight, preeclampsia and SGA) were assessed using simple and multiple regression models within altitude and ancestry groups. Phase one of the analyses began with a series of simple linear regression models to identify associations between independent variables and outcome variable (e.g. SOD and UA diameter; EDN1 and UA diameter). For each outcome variable, hierarchical multiple regression models (using the significant independent variables identified in phase one) were used to assess the degree to which each independent variable can explain variability in the outcome variable. Categorical independent variables were recoded according to SPSS specifications for use in regression analyses (SPSS, Chicago, IL, US).

4. Primary Research Findings
This chapter will first describe results obtained from comparisons of maternal attributes, fetal biometry, newborn and delivery characteristics and UA blood flow parameters at low and high altitude between ancestry groups and, where applicable, across pregnancy time points (sections 4.1 4.4). These data will serve to describe the subject population and to identify variations in maternal characteristics that may contribute to altitude- or ancestry-associated variability in fetal growth.
The remainder of the chapter is organized to correspond to the specific hypotheses presented on pages 17-19. To test the hypothesis that oxidative stress increases with pregnancy at high altitude to a greater degree in European than Andean women, section 4.5 will describe the effects of pregnancy, altitude and ancestry on maternal isoprostane and leptin levels. Then, to test the hypothesis that pregnancy decreases antioxidant capacity at high altitude in European but not Andean women, section 4.6 will explore the effects of pregnancy, altitude and ancestry on enzymatic antioxidant capacity (i.e maternal erythrocyte catalase and SOD).
Subsequently, to address the hypothesis that oxidative stress and/or reduced antioxidant capacity contributes to altitude-associated SGA and/or preeclampsia as the result of endothelial damage, diminished bioavailability of circulating vasodilators and reduced UA blood flow, section 4.7 explores the relationship between these variables, SGA and hypertensive disorders of pregnancy. Section 4.7 also includes analyses of the effects of pregnancy, altitude and ancestry on markers of endothelial damage and circulating

vasodilators. The chapter will conclude with a brief summary of the study findings (section 4.8).
Chi-squared tests were used to compare dichotomous variables between altitude and ancestry groups; values for these tests in the text and tables are expressed as percentages (%) with the 95% confidence interval (Cl) in brackets. Analysis of variance (ANOVA; one-, two- and three-way ANOVA) was used to compare continuous variables across time, between time points, ancestry and altitude groups; values for these tests in the text and tables are expressed as means SEM; within the text the specific ANOVA tests used are noted. One-way ANOVA (1-way) was used to evaluate differences across time points within ancestry and/or altitude groups and to identify differences between altitudes or ancestry groups within time points. Two-way ANOVA (2-way) was used to evaluate differences across time between ancestry groups at low or high altitude (time*ancestry, at low or high altitude), differences across time points between altitude groups (time*altitude), differences across altitudes between ancestry groups (altitude*ancestry) and differences across time between ancestry groups (time*ancestry). Three-way ANOVA (3-way) (main effects: time, ancestry, altitude) was used to identify whether the effect of altitude on the dependent variable of interest differed across time between ancestry groups (altitude*ancestry*time). As appropriate, Scheffes post-hoc tests were used to evaluate differences between groups identified by ANOVA tests. Linear regression was used to evaluate the nature of relationships (direction and strength of association) between variables and the percentage of variation in the dependant variable that could be assigned to the independent variable(s).

4.1 Maternal Attributes (Table 4.1)
At high altitude Andean women were younger, of shorter stature and had lower body weight, skinfold thickness (p<0.01, each; 1-way) and mean arterial pressure (p<0.001; 1-way) in the non-pregnant state relative to their European counterparts. However, there was no difference between Andean and European women residing at low altitude in age, height, nonpregnant body weight, skinfold thickness or mean arterial pressure. Andeans had lower nonpregnant body weight at high than low altitude (p<0.001; 1-way) whereas there was no effect in Europeans. Women of both ancestry groups were leaner at high than low altitude (Andean: p<0.001; European: <0.01; 1-way).
A total of 5 out of 156 women reported smoking during pregnancy. At low altitude the frequency of tobacco use was equivalent between ancestry groups (4.6% [0.8, 21.8] and 6.3% [1.7, 20.2] for Andeans and Europeans, respectively). In contrast, at high attitude more European than Andean women smoked during their pregnancy (7.7% [2.1,24.1] and 2.3% [0.6, 8.1]). The use of alcohol during pregnancy was equivalent between all ancestry and altitude groups. Nearly twice as many European (88.9% [71.9, 96.2] than Andean (47.4% [37.5, 58.1]) women reported the use of prenatal vitamins at high altitude. Reported vitamin use declined 39% with altitude among Andean women (p<0.05), whereas there was no altitude-associated difference in European women.
Monthly household income decreased with altitude in Andean women (p<0.001; 1-way), whereas the opposite was true for Europeans, who, at high altitude, had greater incomes than their low-altitude counterparts or than Andean women at high altitude (p<0.001, each; 1-

way). Similarly, Andean women at high altitude tended to have less formal education equivalent to, or beyond, secondary school (71.8% [61.4, 80.2] and 100% [85.1, 100], respectively) than Andeans at low altitude. All of the Andean women at low altitude and the European women at both altitudes, reported having secondary school education or above.
High-altitude European women reported having lived at altitudes >2500m, on average, 11.2 years less than their Andean counterparts (p<0.001; 1-way). Of the women residing at high altitude, a greater proportion of Andeans than Europeans were born at £2500m (96.5% [90.1, 98.8] and 44.4% [27.6, 62.7], respectively; p<0.001). At low altitude, there was no difference in the proportion of women born at high altitude between ancestry groups (Andean: 13.6% [4.8, 33.3] and European: 12.5% [4.8, 28.1], respectively; p=NS).
There was no difference between altitude or ancestry groups in gravidity, parity or the percentage of women who were primiparous. The proportion of women having proteinuria levels 21+ at 36 weeks was nearly 4-fold greater among European than Andean women at high altitude (32.0% [17.2, 55.5] and 8.7% [3.4, 20.3]); however, there was no difference at low altitude between ancestry groups. Altitude tended to increase the proportion of European women with proteinuria.

Table 4.1 Maternal attributes
Low High
Variable Ancestry 400m 3600-4100m p-altitude
Maternal age, yrs Andean (22) 27.95.6 (86) 25.910.7 NS
European (32) 26.96.0 (27) 30.911.1 <0.05
p-ancestry NS <0.01
Height, cm Andean (22) 158.11.0 (86) 150.510.6 <0.001
European (32) 160.71.0 (27)161.911.6 NS
p-ancestry NS <0.01
Weight, kg 20w Andean (22) 65.1 2.2 (19)61.811.9 NS
European (30) 68.82.4 (12) 65.312.7 NS
p-ancestry NS NS
36w Andean (22) 74.32.1 (47) 62.711.7 <0.01
European (31) 77.12.3 (24) 73.111.8 NS
p-ancestry NS <0.001
NP Andean (22) 66.61.8 (19) 55.511.7 <0.001
European (32) 69.112.3 (23) 64.012.0 NS
p-ancestry NS <0.01
Skinfold, 20w Andean (22) 46.03.4 (19) 34.612.9 <0.05
mm European (32) 48.412.6 (12) 43.112.9 NS
p-ancestry NS NSf
36w Andean (22) 63.014.3 (47) 40.711.8 <0.001
European (32) 63.513.4 (4) 44.217.4 NSt
p-ancestry NS NS
NP Andean (22) 64.813.2 (19) 23.911.1 <0.001
European (32) 66.113.0 (15) 48.615.7 <0.01
p-ancestry NS <0.001
MAP, mmHg 20w Andean (22) 78.011.4 (19) 69.211.5 <0.001
European (32) 78.511.3 (12) 76.411.8 NS
p-ancestry NS <0.001
MAP, 36w Andean (22)82.411.9 (46) 72.011.4 <0.001
European (32) 77.710.9 (24) 77.311.4 NS
p-ancestry <0.01 <0.05
MAP, NP Andean (22) 77.812.6 (19) 71.911.4 p=0.06
European (32) 77.311.0 (23) 78.711.7 NS
p-ancestry NS <0.01
Smoke, % yes Andean 4.6 [0.8,21.8] 2.3 [0.6, 8.1] NS
European 6.3 [1.7,20.2] 7.7 [2.1,24.1] NS
p-ancestry NS <0.001
Alcohol use, % yes Andean 22.7 [10.1,43.4] 14.0 [8.2, 22.8] NS
European 31.3 [18.0,48.6] 22.2 [10.6,40.8] NS
p-ancestry NS NS

Table 4.1. Maternal attributes (Cont)
Low High
Variable Ancestry 400m 3600-4100m p-altitude
Pntl vitamins, % yes Andean 86.3 [66.7, 95.2] 47.7 [37.5,58.1] <0.05
European 78.1 [61.3, 89.0] 88.9 [71.9, 96.2] NS
p-ancestry NS <0.05
Income, $/mo. Andean (17)28253 (86)8416 <0.001
European (24)37956 (16)12501386 <0.001
p-ancestry NS <0.001
Education > 2, % Andean 100 [85.1, 100] 71.8 [61.4, 80.2] NSt
European 100 [89.3, 100] 100 [87.1, 100] NS
p-ancestry NS NSt
Years 22500m Andean (22) 2.611.3 (85) 25.010.7 <0.001
European (32) 1.110.8 (25) 13.812.5 <0.001
p-ancestry NS <0.001
Gravidity, no. Andean (22)2.110.3 (86) 2.310.2 NS
European (32) 2.310.3 (28) 2.210.2 NS
p-ancestry NS NS
Parity, no. Andean (22) 1.710.2 (86) 2.010.2 NS
European (32) 1.910.2 (28) 1.710.1 NS
p-ancestry NS NS
Primigravid, % Andean 50.0 [30.7, 69.3] 36.1 [26.7,46.6] NS
European 37.5 [22.9, 54.8] 46.4 [29.5, 64.2] NS
p-ancestry NS NS
Proteinuria, % a1 + Andean 13.6 [4.8, 33.3] 8.7 [3.4, 20.3] NS
36 weeks European 12.5 [5.0, 28.1] 32.0 [17.2, 55.5] NSt
p-ancestry NS <0.05
Values are reported as means SEM or as percentages with 95% Cl in brackets [ ]. Oneway ANOVA was used for comparisons between continuous variables; chi-squared tests were used for comparisons between categorical variables. Abbreviations: MAP = mean arterial pressure; pntl = prenatal; p-ancestry = significance values between ancestry groups within altitude groups; p-altitude: significance values between altitude groups, within ancestry groups; t=0.05 85

Full Text


\ ......





PAGE 106

a ;;r

PAGE 109

t t

PAGE 112


PAGE 120


PAGE 127

200 0 2 0 3 0

PAGE 128

0 1

PAGE 195


PAGE 208


PAGE 210