Chronic mountain sickness at 3100 meters in North America

Material Information

Chronic mountain sickness at 3100 meters in North America multifactoral analysis of a community study
Asmus, Ingrid V
Publication Date:
Physical Description:
189 leaves : illustrations ; 28 cm

Thesis/Dissertation Information

Doctorate ( Doctor of philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Health and Behavioral Sciences, CU Denver
Degree Disciplines:
Health and behavioral sciences


Subjects / Keywords:
Mountain sickness -- Colorado -- Leadville ( lcsh )
Mountain sickness ( fast )
Colorado -- Leadville ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 177-189).
General Note:
Department of Health and Behavioral Sciences
Statement of Responsibility:
by Ingrid V. Asmus.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
51782598 ( OCLC )
LD1190.L566 2002d .A75 ( lcc )

Full Text
Ingrid V. Asmus
B.A., University of Colorado, 1972
B.S., University of Colorado Health Sciences Center, 1973
M.A., University of Colorado, 1982
M.A., University of Colorado, 1994
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor or Philosophy
Health and Behavioral Sciences
: t* ;
I .K *
L. -

2002 by Ingrid V. Asmus
All rights reserved

This thesis for the Doctor of Philosophy
degree by
Ingrid V. Asmus
has been approved
Dama Dufour

Asmus, Ingrid V. (Ph.D., Health and Behavioral Sciences)
Chronic Mountain Sickness at 3100 Meters in North America: Multifactoral Analysis
of a Community Study.
Thesis directed by Professor Loma G. Moore
Chronic Mountain Sickness (CMS) is a poorly understood and potentially fatal
malady which develops in some high altitude residents after years of life at high
altitude (> 2,500 m). Identified among all high altitude populations so far studied,
signs include excessive polycythemia, hypoventilation and low arterial oxygen
saturation. Research in the Andes and in Tibet indicates that prevalence increases with
increasing altitude and that CMS is more common in men than in women. The CMS
prevalence in Colorado is unknown.
Clinical diagnosis of CMS is based on excessive polycythemia. CMS has long
been recognized in Leadville, CO (3100 m), the highest township in the US. To
establish CMS prevalence, age and sex distribution, and to evaluate characteristics of
cases and unaffected individuals, a community survey was carried out in Leadville.
Study participants (140 men and 148 women) were > 20 years old and had a
minimum of six months residence at 3100 m. Residence history, health and
occupational background and symptoms were evaluated in all participants. The
normal range of hematocrit values was selected to encompass 95% of the population;
CMS was defined by values above this range (56.0% for men, and 51.1% for
women). Prevalence rates were 8.95% among men, 8.6% among women. While
rates were highest in oldest age groups, cases also occurred well before age 50.
Women with CMS were more likely to be post-menopausal than their healthy
counterparts. CMS cases were more likely to have been bom at high altitude than
unaffected individuals. Cases differed from healthy individuals in having higher age,
BMI, diastolic BP and erythropoietin, longer history of smoking and residence at high
altitude, and lower Sa02 and pulmonary function. The CMS symptom score, as
developed in Peru, was not useful in this population. Higher than expected
prevalence rates in Colorado suggest that CMS may have a substantial impact on the
health of high altitude residents, and requires further investigation.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
a Grindlay Moore

I dedicate this thesis especially to the people of Lake County, Colorado, and beyond
them, to the residents of high altitudes worldwide.

Many people and organizations contributed to my ability to complete this thesis. First,
my thanks to my advisor, Loma Moore, for giving me the benefit of her knowledge
and experience. My gratitude also goes to Jack Reeves, Stacy Zamudio, Dama Dufour
and Susan Niermeyer, for their time, guidance, support and interest as members of my
My thanks are also due to the University of Colorado Graduate Research Opportunity
Program and especially to the National Science Foundation for providing funding to
make this research possible.
Staff of Saint Vincents General Hospital in Leadville, Colorado gave me the
opportunity to use their facility for the actual research. In addition, they were
interested in the research, and many were willing participants in the study itself. The
involvement of the people of Lake County and Leadville, Colorado as study
participants was vital to completion of the project, and is gratefully acknowledged. In
addition, their interest in the study, and stories of the ways that CMS has affected their
lives provided new insights on the malady.
Finally, support from my brother Ernest and my niece Kristen made it possible for me
to be absent from family responsibilities at home in order to do fieldwork, confident
that all would be well in their capable hands. My sister Sigrid provided much needed
editing of the final text.

. xi
1. INTRODUCTION...................................................1
Review of Literature........................................2
Control of Ventilation...............................3
Respiratory Function.................................4
Sleep Disordered Breathing...........................5
Control of Red Blood Cell Production.................8
Menopause and CMS Risks for Women....................9
Regional and Epidemiological Information...................10
Methodological Approaches to the Study of CMS............. 12
Medical Model.......................................12
Epidemiological Model...............................13
Biological Anthropology Model ......................14
Overview of Research Design............................... 18
Research Questions.........................................20
Synthesis: Expected Significance and
Relation to Long Term Goals ...............................24

The Site.....................................................26
Local History................................................27
The Population...............................................28
Medical Background ..........................................29
Materials and Methods........................................32
Study Location and Questionnaire......................32
Study Measurements....................................35
Statistics ..................................................40
3. RESULTS.........................................................41
Establishing the Normal Range of Variation...................42
Study Participant Characteristics.....................42
Identifcation of Participants in Nominal Good Health.43
Hemoglobin and Hematocrit Normal Values...............47
CMS Prevalence by Sex and Age................................54
Characteristics of CMS Cases and
Normocythemic Men and Women..................................59
Anecdotal Reports from Participants
in the Lake County CMS Study..........................94
CMS Prevalence among Women, Relative
to Menopausal Status, Use of Hormone
Replacement Therapy and Birth
Weights of Infants...........................................99

Estimated Prevalence among Pre-
and Postmenopausal Women...........................99
Characteristics of Postmenopausal Women
Using and Not Using Hormone Replacement
Therapy (HRT)......................................102
Birth Weights of Infants Bom to Women Participants_104
Relationships between Variables Associated with CMS......104
4. DISCUSSION AND CONCLUSIONS...................................113
Prevalence: Who is Affected by CMS?......................115
CMS Prevalence in Lake County and Worldwide........115
Age and CMS........................................122
Sex and CMS........................................124
Signs and Symptoms: What Happens
to People Who Have CMS?................................. 129
Signs of CMS.......................................129
Symptoms of CMS and the CMS Symptom Score..........135
Health Outcomes and Conditions Existing with CMS___138
Characteristics of Individuals With and Without CMS......140
Duration of High-Altitude Residence................140
Occupational and Behavioral Factors and CMS........142
Weight, BMI and Neck Circumference.................142
Smoking History and Pack-Years of Smoking..........143
Explanatory Hypotheses for CMS...........................143

Age and Duration of Residence.....................144
Chemosensitivity, Respiratory Control
and Sleep Apnea in CMS............................145
Pulmonary Function and CMS........................146
Erythropoietin, Control of Erythropoiesis and CMS.147
Menopausal Status and CMS.........................154
New Explanatory Hypotheses........................154
Relationships, Causal Pathways, and
Directions for Future Research...........................160
Recommendations for Future CMS Research...........166
STUDY.................................................... 169
B. QUESTIONNAIRE............................................171

3.1 Hemoglobin by Age for Men in Good Health........................48
3.2 Hemoglobin by Age for Women in Good Health......................49
3.3 Hematocrit by Age for Men in Good Health........................50
3.4 Hematocrit by Age for Women in Good Health .....................51
3.5 Histogram of Hematocrit Values for Men in Good Health ..........52
3.6 Histogram of Hematocrit Values for Women in Good Health.........53
3.7 Age Distribution of Men Study Participants Compared to
Male Lake County Residents......................................56
3.8 Age Distribution of Women Study Participants Compared to
Female Lake County Residents....................................57
3.9 Incidence of New CMS Cases Identified Among Men
Living at 3100 m, by Age Group..................................61
3.10 Prevalence of CMS in Men, by Age Group..........................62
3.11 Incidence of New CMS Cases Identified Among Women
Livingat3100m, by AgeGroup......................................63
3.12 CMS Prevalence among Women, by Age Group........................ 64
3.13 Duration of Residence at High Altitude by Age
for Normocythemic Men and Men with CMS..........................67
3.14 Duration of Residence at High Altitude by Age
for Normocythemic Women and Women with CMS......................68
3.15 Age on Arrival for Normocythemic Men and Women, and
Men and Women with CMS..........................................70
3.16 FVC in Liters (BTPS) by Age for Normocythemic Men
and Men with CMS................................................74

3.17 FVC in liters (BTPS) by Age for Normocythemic Women
and Women with CMS..............................................75
3.18 Frequency Distribution of % Expected FVC Values for
Normocythemic Men and Men with CMS..............................77
3.19 Frequency Distribution of % Expected FVC Values for
Normocythemic Women and Women with CMS..........................78
3.20 Regression of Sa02 by Age for Normocythemic Men
and Men with CMS................................................80
3.21 Regression of Sa02 by Age for Normocythemic
Women and Women with CMS........................................81
3.22 Regression of Hemoglobin to Hematocrit for All Study
Participants, Male and Female, Normocythemic and CMS Cases......84
3.23 Linear Regression: COhgb / hgb Ratio Independent, Log Erythropoietin
Dependent for All Participants.........................................87
3.24 Linear Regression: COhgb / hgb Ratio Independent, Hematocrit
Dependent for All Participants..................................88
3.25 Polynomial Regression of Log Erythropoietin to Log Ferritin
for Men with CMS and Normocythemic Men..........................90
3.26 Polynomial Regression of Log Erythropoietin to Log Ferritin
for Women with CMS and Normocythemic Women......................91

1.1 CMS prevalence.......................................................11
3.1 Sex, Age, and Ethnic Comparison of Overall Adult Lake
County Population with all CMS Study Participants, by
Number (n = ), Percentage ( % ) and Age Group.......................42
3.2 Reasons for Exclusion from the Category 'Nominal Good Health......44
3.3 Age Distribution for Men and Women in Nominal
Good Health at 3100 m (n = 222).....................................45
3.4 Characteri sti cs of Men and Women in Good Health at 3100 m..........46
3.5 Mean Hematocrit, z-score, and Normal Range for Men and Women
in Nominal Good Health...............................................54
3.6 Mens Estimated CMS prevalence (corrected for age)...................58
3.7 Womens Estimated CMS prevalence (corrected for age).................58
3.8 CMS Prevalence Worldwide, Including Lake County......................59
3.9 Age Distribution for Lake County Men and Women
with and without CMS (actual count).................................60
3.10 Age, Duration of Residence, and Age on Arrival at 3100 m.............66
3.11 Height (m), Body Mass Index (BMI), Neck Circumference,
and Exposure to Smoke or Mining.....................................71
3.12 Concurrent Exposures to Smoking and Mining among Men.................72
3.13 Pulmonary Function Test Results......................................76
3.14 Hemoglobin, Hematocrit, hct/hgb Ratio, Cohgb/hgb Ratio,
Ca02, Erythropoietin, Ferritin and the CMS Symptom Score............82
3.15 Distribution of COhgb/hgb Ratios among All Participants
by CMS and Smoking Status ..........................................85

3.16 Erythropoietin Concentrations in mU/ml for All Study Participants........85
3.17 Ferritin in ng/ml for All Study Participants.............................89
3.18 Correlations between CMS Symptom Score and Other Variables...............92
3.19 Heart Rate, Blood Pressure and Related Factors...........................93
3.20 Characteristics of Pre- and Postmenopausal Women, with
Two Comparison Groups of Similarly Aged Men (All Individuals
were in Nominal Good Health and Living at 3100 m)..................100
3.21 Postmenopausal Women in Nominal Good Health,
with and without Hormone Replacement Therapy (HRT).................103
3.22 Correlations between Pulmonary Function Measured by
% Expected FEV, or % Expected FVC, and Variables
Related to Pulmonary Function, Chemosensitiveity, etc..............105
3.23 Correlations between Sa02 and Related Variables.........................106
3.24 Correlations between Log Erythropoietin and Related Variables...........107
3.25 Correlations between Hematocrit and Related Variables...................108
3.26 Correlations between Diastolic Blood Pressure and Related Variables ... 109
3.27 Hematocrit Dependent Multiple Regression
for Normocythemic Men, r = 0.203...................................110
3.28 Hematocrit Dependent Multiple Regression
for Men with CMS, r2 = 0.290.......................................Ill
3.29 Hematocrit Dependent Multiple Regression
for Normocythemic Women, r2 = 0.115................................Ill
3.3 0 Hematocrit Dependent Multiple Regression
for Women with CMS, r2 = 0.292.....................................Ill
3.31 Hematocrit Dependent Multiple Regression
for All Study Participants, r = 0.302..............................112
3.32 CMS Status Dependent Logistic Regression
for All Study Participants.........................................112

Chronic Mountain Sickness (CMS) is a poorly understood malady which
develops in some persons after years of life at altitudes of 2500 m or above. Found in
all populated high altitude regions so far studied, CMS is identified by the presence
of thick blood, or too many red blood cells. Excessive red blood cells may strain
the heart and impede circulation and so contribute substantially to the development of
heart disease, stroke, or blood clots in affected individuals.
This chapter will begin with a brief review of literature, a historical
background on CMS studies and a basic description of CMS, including individuals
likely to be affected, and the signs, symptoms and consequences of the malady. A
summary of current thought on CMS follows. While this literature provides useful
information on the nature and origins of CMS, it also leaves many questions
unanswered. These gaps may have developed because the approaches and research
goals reflect methodological assumptions about the nature of high altitude populations
and what kinds of responses might occur to the stresses of high altitude hypoxia.
Research goals for this study, described at the end of the chapter, were developed
with the intention of addressing these questions, or at least building a foundation for
better understanding of human response to chronic hypoxia.

Review of Literature
Chronic Mountain Sickness (CMS) was originally identified in 1928 by
Carlos Monge Medrano in Peru (67, 124, 130). CMS represents a loss of, or a failure
to maintain adaptation to high altitude hypoxia in the face of prolonged residence at
high altitude, especially in combination with the physiological processes of age and
the passage of time (2). Young to middle-aged and especially older men living in
mining towns at altitudes above 4000 m. were thought to be the most common
victims. Signs include low ventilation, low arterial oxygen saturation, marked
cyanosis, edema of face and legs, elevated pulmonary arterial pressure and especially
excessive red blood cell volume (erythrocytosis) (83, 135). Symptoms included
shortness of breath, palpitations, headache, dizziness, tinnitus, numbness and tingling
in hands and feet, prominent blood vessels in eyes, loss of memory, fatigue and
insomnia (4,84). A notable characteristic of CMS is that both signs and symptoms
disappeared or resolved when the victim descended to lower altitude, indicating that
the malady was reversible and altitude dependent (30, 31,70,96, 124). For affected
individuals who did not descend to lower altitude, outcomes are less clear. CMS can
be treated by phlebotomy (17,19, 67, 129), or by use of drugs which stimulate
respiration or limit erythropoiesis (51), but the long term results of either are
unknown. By analogy with polycythemia vera, likely outcomes of CMS are
increased risk of heart disease, stroke and blood (85, 118).
Because of the relationship between chronic high altitude hypoxia and CMS,
most theories proposed to explain its origins have focused on changes in respiratory
drive or function. These include age-related physiological changes, sleep disordered

breathing or apnea and, in women, hormonal changes associated with the menopause.
Also open to question is control of red blood cell production (erythropoiesis) since it
is not clear how excessive erythropoiesis is initiated or maintained. By contributing to
development of low arterial oxygen saturation and excessive production of red blood
cells, several of these factors might be involved in development of CMS.
Control of Ventilation
Hyperventilation is one of the initial human responses to hypoxia at high altitude, an
acclimatory response mediated by peripheral chemoreceptors in the carotid body (21,
35,54,105, 106). Individuals and populations that adapt well to high altitude appear
able to maintain this hyperventilation over time (105,106). However, in some
individuals, hyperventilation diminishes with continued exposure to hypoxia. This is
referred to as a blunted hypoxic ventilatory response or blunted HVR. Many long
term residents of high altitude demonstrate this type of blunted chemosensitivity by
failing to respond to acute hypoxia with an increase in ventilation (69, 105, 110).
Blunted HVR is found in CMS cases, but is also present in many healthy high altitude
residents (35,54, 111). Because blunted HVR cannot be used to differentiate
between CMS cases and normocythemic individuals, it does not appear to cause CMS
(52), though it may of course contribute to associated problems.
In some individuals, another response to hypoxia is elevated blood pressure in
the lungs, or pulmonary hypertension (PH) (84,95, 124). On acute exposure, this is
associated with increased risk of development of high altitude pulmonary edema
(HAPE) (101); with chronic exposure to hypoxia, pulmonary hypertension may lead

to right ventricular hypertrophy (RVH), as the heart struggles to circulate blood in the
lungs. Pulmonary hypertension and RVH have been documented in a number of
CMS cases (40, 69. 84, 86, 87, 135, 136), presumably made more severe by
increased blood viscosity, but the sequence of events is not clear: does pulmonary
hypertension contribute to the development of CMS, or is it an end result of the
Respiratory Function
Inadequate ventilation (hypoventilation) may have other causes as well.
Individuals with CMS tend to be heavier than unaffected people, and obesity has been
linked to hypoventilation (100). Mismatches between ventilation in the alveoli and
blood circulation in the lung, known as ventilation -perfusion inequality, may worsen
Sa02 and contribute to the development of CMS (17,52). Both age-related declines
in pulmonary function (107, 132) and the presence of respiratory disease (22,32,40,
49,58) can limit pulmonary function or decrease Sa02 by widening the alveolar-
arterial oxygen gradient (3). Chronic obstructive pulmonary disease (COPD),
comprising asthma, chronic bronchitis or obstructive emphysema appear to increase
the risk of CMS compared to acute or upper respiratory tract disease (48,52,58).
The degree to which pulmonary disease at high altitude has been confounded with
CMS is not clear; even at sea level, severe COPD may result in the development of
polycythemia. The terms primary and secondary CMS have been used to
differentiate between the two (2), with secondary CMS suggesting polycythemia
secondary to pulmonary disease, and primary CMS indicating polycythemia in the

absence of pulmonary disease or other recognized causes of hypoxemia. Whether
primary CMS occurs at all has been subject to some debate (87). In any case, the
range of pulmonary function in CMS cases and unaffected individuals resident at high
altitude has not been well documented. As a result, it is not clear whether pulmonary
disease in high altitude residents inevitably results in CMS, or not, and it is essential
to include evaluation of pulmonary function in this study to resolve the question.
Hypoventilation caused by pulmonary disease will likely result in low arterial
oxygen pressure or saturation especially during waking hours; oxygen saturation may
also fall further during sleep (23,35, 50, 51). Whether chronic or intermittent, low
oxygen saturation should induce greater production of erythropoietin, and so trigger
greater production of red blood cells (42,91). However, some CMS cases appear to
have normal pulmonary function and maintain a normal SaO, during the day, yet they
are still polycythemic.
Sleep Disordered Breathing
These cases could be a result of sleep disordered breathing, which is also
common among CMS cases. Kryger et al. (50,51) showed that low arterial oxygen
saturation during sleep is a potent trigger for CMS; CMS cases treated with medroxy
progesterone acetate, a synthetic progesterone and respiratory stimulant, had fewer
and less severe sleep desaturations and diminished polycythemia. Sun et al. (113,
114) found that during the day, male CMS cases maintained an arterial oxygen content
similar to that of healthy controls. However, during sleep, arterial oxygen saturation
and content were lower and more variable among cases than controls, a result of more

episodes of hypoventilation and apnea, and longer episodes of all types. Such
nocturnal desaturations are likely to be severe enough to trigger production of
erythropoietin and red blood cells. In men, the likelihood of sleep apneas increases
with age (65); in women, sleep disordered breathing becomes more common after
menopause (89). Obesity may also contribute to sleep disordered breathing and
obstructive apneas (100).
Aging at high altitude has been suggested as a cause of CMS (15,56, 59, 68,
107, 109). At altitude, increasing age is associated with a progressive decline in
arterial oxygen saturation (SaO: )(109) due to diminished chemoreceptor sensitivity to
hypoxia, worsening respiratory function or both (15). Accompanying the age-
associated decline in Sa02, a progressive increase in hemoglobin levels has been
reported in some studies (56, 57, 59, 68, 132). Age distribution of CMS cases
among men and women in Cerro de Pasco, Peru (4300m) indicated that male miners
(n = 3386) showed a smooth increase in CMS prevalence with age, from a low of
6.8% among 20 29 year olds, to a high of 33.7% in the 60 69 year age
category(69). The age distribution among women(59), based on a sample of 152
women between 30 and 54 years of age, showed consistently lower prevalence in
premenopausal women and then a more rapid increase in prevalence following the
menopause. These data suggest that while increasing age is a risk factor for CMS, it
can and does occur in young adults.

A closely related variable, the amount of time an individual must spend at high
altitude before CMS can develop, has been poorly documented. Peruvian studies
have required 10 years residence at 4300 m for inclusion in prevalence studies, and
found ample numbers of CMS cases. Pei et al. (84) gave 15 years as the mean
duration of high altitude residence before the onset of symptoms, but note that the
range was from 1 to 30 years in a sample of 17 cases, and other studies have reported
a few CMS cases in recent immigrants to high altitude.
In any case, while advanced age and prolonged altitude exposure may increase
risks of CMS, they are neither necessary nor sufficient to explain susceptibility to
CMS. Many older people living at high altitude do not develop CMS, even when they
have been there since birth or childhood. Other individuals who were born at high
altitude, or immigrated there later in life, have developed CMS by the time they
reached their 20s or 30s, at a time when they should theoretically be in peak
physiological condition (84). Consequently, it appears that while time- and age-
related changes in ventilatory, circulatory, erythropoietic, and hormonal systems may
contribute to the development of CMS, they do not necessarily cause the malady. As
emphasized by Arking et al.(3), age-related changes are based on a foundation of
individual genetics, gender, and phenotypic changes due to life experiences and
exposures, so age alone cannot explain all variations.

Control of Red Blood Cell Production
Regardless of the cause of low arterial oxygen saturation (hypoxemia), and
whether it is constant or episodic, blood or tissue hypoxia is sensed by the kidney,
which responds by producing erythropoietin, a hormone that stimulates the bone
marrow to produce more red blood cells (RBC) (22, 24,42, 46, 119). Since RBCs
represent the largest proportion of cells in the blood, the result is an increase in the
total cell mass of the blood, described as polycythemia, or more accurately as
excessive erythrocytosis, since the increased cell volume is a result of increased
numbers of RBCs relative to other blood cells.
The role of erythropoietin (epo) in the development of excessive polycythemia
in CMS has not been clear, as epo levels in CMS cases may be low, normal or high
(18,55). If erythropoietin concentrations do not drive the excessive polycythemia of
CMS, as suggested by existing data, then it is important to determine what other
factors may be responsible in order to understand the process. RBC production is
complex, and there may be feedback loops between different steps. Epo targets the
bone marrow to trigger the production of RBCs, but other factors are also required,
including iron, folate, vitamin B12, and protein (24,47). Iron deficiency is common
world-wide (104), and responsible for many cases of anemia. However, no studies
have reported results of iron, folate or B12 assays in CMS cases or healthy high
altitude residents, so it is impossible to know whether deficiencies in any of these
might play a part in CMS.
Increases in red blood cell numbers and volume have long been considered to
be normal, and adaptive, responses to altitude (5), since more RBCs and

consequently higher hemoglobin increase the oxygen carrying capacity of the blood.
It is not clear, however, whether extreme or excessive erythrocytosis actually
enhances fitness for high altitude, especially when exposure is chronic. Animals such
as yak or llamas and alpacas, which are generally considered to be well adapted to
high altitude, do not display marked polycythemia in response to altitude (22,36).
Also, regardless of what factors stimulate red cell production, and whatever its
benefits, polycythemia increases blood viscosity. As hematocrits rise above about
55%, the result is increased cardiac workload and ultimately RVH (22,34,99). In
concert with the low Sa02 often seen in CMS cases, this contributes to poor cerebral
blood flow (85, 118) and systemic oxygen delivery (4, 32, 34, 99, 112), and may
trigger such symptoms as headache, tinnitus, numbness and tingling in hands and
feet, breathlessness and fatigue (85).
Menopause and CMS Risks for Women
The low proportion of women among reported CMS cases suggested to
Peruvian researchers that women may be protected against CMS until the passage of
menopause (59,60). This hypothesis was based on knowledge that progesterone is a
respiratory stimulant both alone and in combination with estrogen (37,89,94,116),
while estrogen inhibits erythropoiesis by inactivating epo() and by limiting the bone
marrow progenitor cells response to epo(47,88, 119, 122). Results of research
undertaken to test this hypothesis suggest that premenopausal hormone levels may
provide a degree of protection against development of CMS in premenopausal
women, since CMS prevalence was consistently low among premenopausal women,

followed by a rapid increase following the menopause. Still, these changes in female
hormones with menopause are also based on a foundation of individual variation.
Some premenopausal women develop CMS, and some postmenopausal women
remain free of it. While sex hormones may be actively involved in individual
responses to hypoxia, including ventilation and erythropoiesis, by themselves they do
not offer a convincing explanation for the existence of CMS in women.
Regional and Epidemiological Information
For many years there was speculation about whether CMS existed in high-
altitude populations in the Himalayan region, but little reliable information. Results
from research done on the Tibetan plateau became available to the West in the late
1980s. These have confirmed the presence of CMS on the Tibetan plateau (40,84,
112, 113, 135-7). Prevalence data from this research provides a broader perspective
on the characteristics and distribution of CMS for men and women of different
ethnicities and at different altitude ranges, especially when prevalence data from Peru
and Bolivia is included. Several of these studies (26,77) noted that rural populations
tended to have lower CMS rates than urban populations of the same ethnicity and
living at the same altitudes. Available prevalence data are summarized in Table 1.1,
Despite gaps in this information, several observations can be made.
Prevalence increases with altitude, and there are marked differences in CMS
prevalence between population groups residing at the same altitude (46,54,101).
Where gender-based prevalence data are available, men appear to be about twice as

likely to develop CMS as women, though the difference seems to diminish with
increasing altitude.
Table 1.1: CMS Prevalence (%)
(59, 69, 121, 135) Moderate Altitude 2260-2800 m High Altitude 3050-3800 m Very High Altitude 4000-5200 m
Population: Male Female Male Female Male Female
Tibetans n= 12,385 0.0 0.0 1.1 0.6 2.0 1.6
Chinese (Han) n=13,233 1.4 0.7 5.2 1.5 10.1 6.6
Peruvians n=3028 no data no data no data no data 15.6 8.8
Bolivians n = 1,115,403 no data no data 5.2 no data no data
Prevalence information of this kind is vital since it provides a broader
perspective on CMS, especially in the face of equivocal results of hypothesis testing
for causes of CMS. These prevalence estimates indicate that CMS may represent a
substantial risk to the health of high altitude residents, and so justify continuing
research into its causes, prevention, diagnosis and treatment. Under these
circumstances, the absence of data for North America is distressing. The
epidemiological approach used to develop these estimates has the potential to provide
a database for each high altitude population, including background information and
measurements on both healthy residents and CMS cases. This would make it possible
to evaluate both the amount of variation in, and the relationships between, variables of
known or potential relevance to high altitude response. Ultimately such a method
would facilitate the process of developing testable hypotheses about the origins of

Methodological Approaches to the Study of CMS
Medical Model
Much work on CMS has been based on a medical paradigm which held that
CMS was a result of some pathological difference or change resulting in loss of high
altitude adaptation in a few individuals from a population that was generally well
adapted (70). This belief grew from the assumption that normal function of an
organism should result in health, and only pathology would result in disease. As a
result, the research goal was to identify the pathological difference or change
responsible for CMS in individuals, and on hypothesis testing to evaluate the roles of
specific factors. Because of this narrow focus, data reported for individual CMS
cases was often limited to symptoms and basic signs such as hematocrit (Hct), body
mass index (BMI), or the presence of right ventricular hypertrophy (RVH). Worse,
little information was available about corresponding values among the normal
population, making comparison difficult and so limiting the value of any conclusions
that might be reached.
However, hypothesis testing has failed to provide causal explanations, and
instead yielded more questions. CMS cases could not easily be differentiated from
normal healthy individuals on the basis of age, HVR, or erythropoietin concentrations
(15, 18, 52, 55), even though each of these factors appears likely to play some role in
CMS. This outcome suggests that the basic assumptions and paradigms which had
formed the basis for CMS research need to be re-evaluated.

One fact which suggests the need for reassessment is that CMS signs and
symptoms disappear when a case descends to lower altitude (30,31, 96). The
reversible nature of CMS suggests that the disease is not merely a result of isolated
pathology of a single body system. Instead, it is likely to be due to a gap between
what is needed to maintain health in the high altitude environment, with its specific
constraints and challenges, and the range of adaptability or flexibility with which a
particular organism can respond. The range of responses in any individual will be a
product of genetics, life experiences, and behavioral choices; in other words, each
individuals response will be shaped by their phenotype. Rural /urban differences in
CMS prevalence (26,77) among members of the same population living at the same
altitude reinforce the conclusion that an individuals ability to respond positively to the
stress of high altitude residence is likely to be shaped by many factors, some of which
are likely to be occupational or behavioral. Under these circumstances, it seems
unlikely that CMS is a result of some single, unique cause or pathology. There may
be no magic bullet to explain the origin of CMS.
Epidemiological Model
Under these circumstances, an epidemiological or descriptive approach seems
appropriate (28). The goal in this case is to develop a database containing information
on many factors relevant to the life, health, and high altitude experience of residents,
both those with CMS and those without. Analysis of such a database for a specific
high altitude population should make it possible to see whether outcomes of CMS
research elsewhere can be replicated, and so evaluate associations between signs and

symptoms and the presence of CMS. The absence of matching data on individuals
unaffected by CMS in previous work has made it difficult to make any inferences
about causality (67). Future work, building on a community-wide database, will
make it possible to build up inferences from the patterns of association between
specific characteristics and CMS. The epidemiological method of choice in this
situation is a case-control study. Case-control studies can be completed fairly quickly
and allow comparison of people with and without CMS on the basis of a wide variety
of factors likely to be associated with health at high altitude. From this more stable
foundation, causal hypotheses can be generated for future research.
Biological Anthropology Model
Another discipline that has held a sustained interest in CMS is Biological
Anthropology (25), and this paradigm provides a very different perspective. Here,
the central issue is: how does adaptation to environmental stressors, and hence
evolution, take place? Upon arrival in a stressful environment, an individual may be
able to acclimatize, that is, respond rapidly (behaviorally and/or physiologically) in
ways which enable him or her to stay, live a healthy life, and reproduce successfully
in that environment. Alternatively, the individual might become ill, and either choose
to leave, stay and suffer potentially debilitating problems, or die. In themselves, these
outcomes are not evolution, but simply the results of an interaction between
environmental stresses, and each individuals own pre-existing physiological
adaptability. However, they do set the stage for evolutionary change in the
populations gene pool by providing a selection between people who can successfully
respond to the challenges of high altitude residence and whose genes are likely to stay

in the gene pool, and those who can't. For newcomers, an initial ability to acclimatize
to high altitude is a prerequisite to residence.
Over time, there will always be a continuum of genetic characteristics in a
resident population. This variability would be maintained by the diversity of the
group, enhanced by relationships with other nearby groups, and mutations. In
addition, behavioral choices, technological changes and individual health experience
would vary, resulting in phenotypic differences. Some of these variations might
enhance individual fitness, improving chances of survival, reproduction, and
beneficial responses to the environment, while other variations might diminish
fitness, leading to illness, or a more limited repertoire of effective responses to the
environment. Selection will act on the range of individual variation present in the
group. Eventually, as a result of this selection, the population as a whole will tend to
adapt: to become more fit and able to survive, reproduce, and otherwise respond
effectively to that specific environment.
From the perspective of biological anthropology, illness in this situation is due
not so much to pathology as to a poor fit between the constraints posed by the
environment, and the organisms ability to respond in beneficial ways. Instead of a
strict, and fixed, dichotomy between healthy and pathologic responses, individuals in
a population present a continuum of responses. The relationship between the
environment and an individuals characteristics is also likely to be dynamic, changing
with time, age, life experience, and with seasonal or other environmental fluctuations.
With its emphasis on evolutionary change and human variability, biological
anthropology suggest that human populations vary in the amount of time they have

spent at high altitude, and that as a result, few human populations would be
completely adapted to high altitude (73,81). Populations which have had a prolonged
unbroken residence at high altitude would have been subject to selection for many
generations. They would be likely to have more beneficial traits, and fewer deleterious
ones, than newcomers to the environment. From this perspective, the proportion or
prevalence of CMS cases becomes an indicator of the adaptive fitness of a population
living at high altitude.
As shown in Table 1.1, data from the Tibetan plateau indicates that the Tibetan
population has the lowest prevalence of any population at any altitude, while the Han
(Chinese) population has higher CMS prevalence than the Tibetans at all altitudes.
Data for High Altitude (3050-3800 m)(121), and Very High Altitude (4000 5200
m)(59,69), in the Andes show prevalence rates higher than those among the Han at
Very High Altitude, but similar to them at High Altitude. This prevalence distribution
fits with the assumption that Tibetan populations have the longest history of sustained
high altitude residence, and hence, more complete high altitude adaptation, while the
Han population had limited high altitude fitness. The Andean population has had a
moderately long record of residence at high altitude, but interpretation of its level of
adaptive fitness for altitude is complicated by unknown degrees of population
admixture and extremely high mortality rates during and following the Spanish
conquest (81). These factors make it impossible to determine to what degree current
high altitude residents are descended from pre colonial high altitude residents, or
whether through processes of genetic drift in the decimated native population, any

existing genetic adaptations to high altitude have been transmitted to current
It is important to recognize that physiological responses to high altitude
(including CMS or the factors which contribute to its development) are dependent on,
and essentially continuous with each individuals innate physiological characteristics
and responses. As a consequence, these characteristics (whether genetic or
phenotypic) would be selected against as they came to affect the well-being of an
individual (3), which might manifest in different forms at different stages of life.
Thus, CMS and other difficulties such as preecclampsia, intra-uterine growth
retardation, subacute infantile mountain sickness and even acute mountain sickness
could all represent a pattern of systematic failures which highlight the weak points of
the human organism in coping with the hypoxia of high altitude.
In theory, CMS is a malady of age at high altitude. If selection against the
traits responsible for CMS did not take place until age 50, when an individual actually
developed it, that individual might already have reproduced, and their illness or death
would have comparatively little effect on the populations gene pool. On the other
hand, if the factors which increase risks of CMS are present, and actually or
potentially deleterious at all stages of life, then it becomes easier to understand how
selection against them could take place earlier in life, resulting in a more rapid or
marked change in gene frequencies. At this time, this idea is just speculation, but its
potential explanatory value is sufficiently great that it is worth mentioning. In order to
evaluate whether the presence of CMS might be associated with a greater lifetime risk
of altitude realted problems, it will be important to gather data not merely on current

symptoms and physiological characteristics, but on the life and health experiences of
affected and unaffected individuals.
In conclusion, it is important to recognize that each methodological approach
has its own costs and benefits. The medical paradigm has an excellent depth of focus,
and is able to examine a health problem down to the molecular level, but the drawback
is that it looses breadth of perspective on the problem. Epidemiology is not designed
to focus deeply on specific illnesses, but it compensates for this limitation with a wide
perspective and the ability to relate parts and whole, or causes and effects. Biological
Anthropology, by contrast, provides a very different viewpoint on the relationships
between individual characteristics and responses, population genetics, the
environment and health. This perspective reminds us that illness may be due to a
poor fit between the individual and his or her environment, rather than pathology.
Under such circumstances, the poor fit could manifest in different ways at different
life stages, and be based on phenotype, making inclusion of data on occupation,
behavior and medical history vital.
Overview of Research Design
In designing this research, I felt that the primary need in current CMS studies
was to develop a database on the Leadville and Lake County, CO population.
Variables were selected on the basis of data from previous CMS studies, as well as
factors known to be associated with pulmonary function, control of erythropoiesis,
occurrence of sleep disordered breathing and related issues in sea level populations.

Insights provided by a review of medical, epidemiological and anthropological
perspectives on high altitude response suggested additional varibles. In order to
achieve the goal, the following study criteria were developed:
A community survey to provide a representative basis for estimation of residence
history, health experience, occupational and behavioral variables and other
characteristics for the population as a whole.
A broad range of variables was selected for evaluation in all participants, both
CMS cases and normals. These included measured values for height, weight,
neck size, blood pressure, Sa02, pulmonary function testing, assay of
hemoglobin and hematocrit, erythropoietin and ferritin. Additional information
was solicited in the questionnaire on age, duration of residence at high altitude,
CMS symptom score, medical and occupational history, smoking history,
anecdotal information, and, for women, menopausal status and use of birth
control or hormone replacement therapy (HRT).
Inclusive selection of subjects: to qualify for participation, individuals needed only
to be aged 20 years or older, and to have lived in Lake County for a minimum of
6 months.
This approach was designed to evaluate a representative cross-section of the
Lake County, CO population, to demonstrate how the population as a whole could be
characterized in terms of study variables, and which variables were useful in
differentiating between CMS cases and unaffected persons. Rather than assuming
that individuals who did not have CMS were adapted, and need not be considered,

data on all variables was gathered on all participants, making it possible to compare
individuals with and without CMS in terms of exposures. With this database in place,
much can be done.
Research Questions
The following questions or issues will be addressed in this study:
1. What is the CMS prevalence among men and women at 3100 m. in Colorado?
Given the absence of prevalence data for high altitude residents in North America
shown in Figure 1.1, a primary goal of this research was to estimate CMS prevalence
in Leadville, CO. Work by Kryger beginning in the 1970s (49) showed that CMS, or
excessive erythrocytosis of altitude, was present in Leadville and Lake County.
Members of this population have lived at high altitude for a relatively short time (a
maximum of 5 generations), and should provide a useful comparison with
populations in South American and the Himalayas, which have experienced many
generations of hypoxic stress.
The study will establish normal ranges for hemoglobin and hematocrit for the
adult population of Lake County, CO at 3100m, an essential prerequisite for
identifying excessive or abnormal values. In turn, information on normal ranges will
be used to identify new cases of CMS in the surveyed population of Lake County.
Only these new cases will be used to estimate prevalence among men and women.
Prevalence information is needed to allow comparison with other high altitude groups,

and to generate estimates of the risk of excessive polycythemia for the approximately
360,000 persons living above 2800 m (9000 ft) in Colorado (13).
2. Evaluate how this population, and the variables included in research, compare to
the results of previous work. Specifically, in what areas do data from this study
support outcomes of previous research, or differ from them? Are there study
outcomes which suggest new perspectives for CMS research? Several areas of
emphasis are described below.
2a. Age distribution of CMS cases has been documented only in Peru (59,69). This
data indicates that some people in their 20s and 30s are affected by CMS, as well as
those in their 40s and 50s. If people are susceptible to CMS at all ages, then health
and financial consequences of CMS for high altitude residents are likely to be more
Differences in age distribution of CMS cases among men (69) and women
(59) in the Peruvian research support the hypothesis that premenopausal women may
be protected by effects of sex hormones on ventilation and erythropoiesis against the
development of CMS, as compared to postmenopausal women, and perhaps to men.
However, it is important to confirm prevalence and age distribution in the Colorado
population. Occupational exposures and smoking may have differential effects on
prevalence outcomes, and the choice of criteria for CMS identification will also affect
apparent prevalence.
The amount of time an individual must spend at high altitude to become
susceptible to CMS is also unclear. Since onset of CMS may be insidious, cases may

have been polycythemic for some time before the problem was identified,
complicating understanding of exposure time before onset of disease. Chinese studies
have provided information on duration of high altitude exposure before the
development of CMS in Han (Chinese) residents, with mean exposure times of 15
years (84) and 21 years (137), but both cases age ranges were very wide. Is there a
required threshold of exposure time? Is a long duration of high altitude exposure
simply a proxy for age? Or are there other factors involved? Results from this study
will document the amount of time required before identification of CMS in natives and
recent immigrants, and so help to guide future research.
2b. What is the range of pulmonary function in individuals with and without CMS?
Many early studies were done in high altitude mining towns, where pulmonary
disease is a familiar consequence of inhalation and accumulation of mining dust in the
lungs. Consequently, many of the CMS cases were likely to have pulmonary disease,
but pulmonary function was not well documented in the research. As a result, there
has been debate about whether CMS is simply polycythemia secondary to pulmonary
disease, complicated by altitude but not necessarily caused by it, and whether CMS
ever develops in people with healthy pulmonary function. While basic pulmonary
function tests (PFTs) are not foolproof, and may be normal in the presence of sub-
clinical lung disease, they do at least provide a means to screen for frank pulmonary
disease, and so come to some conclusions about whether pure or primary CMS
might actually occur.
Also of interest is information on actual ranges of pulmonary function in CMS
cases and normals. It is not clear whether all individuals with poor pulmonary

function develop CMS. Documentation of the presence of individuals with poor
pulmonary function who remain normocythemic in this high altitude population would
lend support to the theory that development of CMS is at least partially independent
of pulmonary function, and more likely to be due to aberrant control of
erythropoiesis, or some other factor.
2c. Erythropoietin (epo) levels vary widely in CMS cases (18,55), obscuring
relationships between erythropoietic drive, control of erythropoiesis and
polycythemia. This study will evaluate epo levels in CMS cases and unaffected
individuals, and seek to determine whether other variables associated with epo levels
and the outcome variable of hematocrit may explain some of the variation.
2d. Do pre- and post-menopausal women differ in their susceptibility to CMS?
Together, the respiratory stimulation(37) and inhibition of erythropoiesis by
progesterone and estrogen(47,88,117) may protect against the development of CMS
in pre-menopausal women. Lower female hormone levels and relatively higher
androgen levels (83) characteristic of post-menopausal women could increase risks of
oxygen desaturation and trigger development of CMS. Leon-Velarde et al. (59)
showed that Sa02 levels were lower and hematocrits higher in post-menopausal than
pre-menopausal women, suggesting that menopause may be a risk factor for the
development of CMS in women. Work in Tibet also suggests a similar pattern (135).
The possibility that women may be protected against CMS during their premenopausal
years could be validated in two ways. First, are there significant differences in
prevalence between newly-identified CMS cases among pre- and post-menopausal
women? And second, what is the age distribution of cases among women? Are there

lower prevalence rates during the 20s, 30s and 40s, followed by an rapid increase
in prevalence following menopause? Both methods will be used in this study.
To make this possible, the questionnaire requested information from women
on their menopausal status and use of birth control or hormone replacement, including
pharmaceutical names and dosages. Evaluation of Sa02, hematocrit, hemoglobin and
other factors in these women will contribute to an improved understanding of the
effects of endogenous and exogenous hormones on factors crucial to the development
of CMS.
3. Use the data from a sample representative of the Lake County population to
develop a more complete picture of the nature and characteristics of this pariticular
high altitude population. What characterizes the population as a whole, and how can
CMS cases be differentiated from their normal counterparts in terms of study
variables? By study design, hemoglobin and hematocrit will differ significantly
between CMS cases and their normocythemic counterparts. Therefore, all other
variables can be evaluated in terms of their association with normal (for this altitude)
or excessive polycythemia. Based on this information, new hypotheses regarding the
origins of CMS can be generated.
Synthesis: Expected Significance and Relation to
Long-term Goals
With a world population of 6 Billion, and rising, there are already about 140
million individuals living above 2500 m (84). Assuming a CMS prevalence rate of

4%, (a conservative estimate, based on prevalence rates for Han (Chinese) living in
the high altitude range of 3050 3800 m., and a prevalence estimate of 5.2% for the
city of LaPaz, Bolivia at 3,400 m.(121)), 5.6 million of these people may be
suffering from CMS (81). Of these, about 360,000 persons live above 2800 m (9000
ft) in Colorado (77), and based on the same 4% prevalence estimate for CMS,
14,400 people may already be affected. These numbers are likely to grow if upward
migration, as is occurring in the United States and worldwide, continues to bring
low-altitude natives into the high altitude environment. Also, given the aging
population in the United States, if CMS is age-related, we can expect to see more
cases developing in the future. The research presented in this thesis is was done in the
hope that it would contribute to a deeper understanding of the processes involved in
adaptation to the stresses of life at high altitude, and of the ways in which these
processes may go awry. Ultimately, such understanding should contribute to
knowledge of the causes of CMS, how it can best be diagnosed and treated, or better
yet, prevented, so that people who choose to live in the high country can do so in
good health.

This chapter covers a range of information specific to this study, beginning with the
altitude, environment, history, and population of Lake County, Colorado, the site of
the study. The criteria for participation, and how study participants were selected or
located follows. A summary of the questions and measurements employed in the
study is then given, along with information on the reasons these particular variables
were selected, and with descriptions of the ways they were measured or evaluated. A
brief statement on statistical methods concludes the chapter.
The Site
For a researcher concerned with the prevalence of Chronic Mountain Sickness (CMS)
in North America, comparatively few sites meet study criteria. The need for a large,
diverse high-altitude population, including men and women of all ages, with varying
health and occupational backgrounds and lengths of residence, meant that the site had
to be a community of some permanence. Moreover, because the prevalence of this
malady increases with altitude, it is most easily studied at the highest possible altitude.
Few locations of this kind exist in the United States; however, the two criteria do meet
in the town of Leadville, county seat of Lake County, Colorado, the highest
incorporated township in the United States at 10,152 feet (3100 meters). As a result,

it is no accident that a number of other research projects concerned with the health
consequences of high-altitude residence have taken place in Leadville. As a few of
these (49-52) were focused on CMS, or excessive polycythemia of high altitude, it
was certain the malady was recognized among the people of Leadville.
Lake County has a total area of 383 square miles (1017 square kilometers),
and there are only two population centers, Leadville and the much smaller Twin
Lakes. The County as a whole ranges from an altitude of 14,433 feet (4441 meters)
at the summit of Mount Elbert, Colorados highest peak, to about 9,000 feet (2770
meters) in the drainage of the Arkansas River. The countys mean altitude is about
10,152 feet (3100 meters), providing a relatively large area within an altitude range
roughly equivalent to the high altitude (3050-3800 meter) range used in the studies by
Wu et al (135) in China, shown in Table 1.1. Most of the population lives at altitudes
between 9,000 feet (2770 meters) and 10,400 feet (3200 meters); some of the older
residents lived for a time in the mining town of Climax above Fremont Pass at 11,000
feet (3385 meters) before the town was disbanded and moved in 1960. Many worked
at the Climax Molybdenum mine until it was closed in about 1994 (123). Study
participants were drawn from residents of the entire county.
Local History
Like many high-altitude communities in Colorado and worldwide, Leadville owes its
existence to mining. First settled in 1860 by gold miners, silver and then
molybdenum became the economic base in Leadville (13). The city and region have
endured several cycles of boom and bust. In the 1980s, production at the Climax
Molybdenum Mine (a subsidiary of Phelps Dodge) began to slow, and the mine was

closed by 1994 (123). With this economic mainstay gone, many people left the area.
Still, with a few other mines still working and tourism or ski area employment
offering a precarious living, the county had 8680 permanent residents in 1998 (16).
Most residents work locally in the outdoor recreation industry, or in tourism. Many
use Leadville as a bedroom community while they travel to work at nearby resorts
such as Vail, where residential housing is prohibitively expensive. A few people have
retired to Lake County; some work for the schools, hospital or county, or are
employed at the local Timberline Campus of the Colorado Mountain College. Given
the high altitude and short growing season, there is little or no agriculture, though
there are a few ranches in the valleys surrounding the upper stretches of the Arkansas
The Population
The 1998 Census Estimate issued by the Colorado Department of Demography was
used in all study-related population estimates and calculations (16). This estimate
was used because it was the best available resource during the time research for this
study was actually being carried out. Also, Lake County is appealing results of the
recently available 2000 Census because they felt it was falsely low, but there has not
yet been time for a recheck. For these reasons, I have continued to use the 1998
estimated census in all calculations related to this research.
The estimated population of Lake County was 8,680 in 1998 (16). Of this
total, 32%, or 2,771 individuals, were age 19 or under, while 68% or 5,909 people
were aged 20 or over. These individuals vary in their duration of residence from days
to a lifetime. A few are descendants of early settlers, and have several generations

history of residence. Because CMS is more typically seen in older residents, only
individuals age 20 or older were included in the study. This decision also had the
benefit of reducing uncertainty by ruling out the chance of confounding normal
developmental changes in hemoglobin and hematocrit occurring during childhood
with those that occur among adults as a result of altitude or age.
According to the 1998 Census estimate, the Lake County population was
roughly 76% Caucasian and 24% Hispanic (16); most Hispanics were originally of
Mexican descent. Both groups include individuals who arrived in Leadville recently
as well as those whose ancestors came a generation or more ago.
Medical Background
The International Classification of Diseases (133) does not include CMS as a
diagnosis; instead it is covered under Section #289.0: Secondary polycythemia due
to high altitude. Because the medical system in North America, specifically
including Leadville, follows this system of classification, individuals who have
unusually high red blood cell (RBC) mass compared to the relative polycythemia
already existing at 3100 meters are diagnosed as having excessive polycythemia. This
is generally measured by the hematocrit, or the proportion of RBCs in a volume of
blood. Functionally, this has meant that at St. Vincents Hospital in Leadville, men
with hematocrits in excess of 55%, and women with hematocrits above about 53% are
diagnosed with excessive polycythemia, and are then treated by routine phlebotomy
whenever their hematocrit rises above that level (personal communication, John Pema
MD). These cut-off values were not based on a population -wide set of values, nor
do they represent some number of standard deviations above the mean.

The decision made by the International CMS Consensus Group at the 2001
International Hypoxia Symposium (Jasper, Canada)(97) to use red blood cell counts,
(or, by proxy, hemoglobin and hematocrit) alone to identify CMS means that all
polycythemia cases that have been identified and treated in Leadville since at least the
1970s meet these current criteria for CMS. In this study, the only criterion used to
identify CMS cases was excessive erythrocytosis or polycythemia, as measured by
hematocrit. Data on many other characteristics known or suspected to be related to the
presence of excessive polycythemia were collected on all participants.
The inclusion criteria for subjects in this study were age 20 years or over, and
residence in Lake County for a minimum of six months. The latter requirement was
used to assure that the processes associated with long-term acclimatization were
completed before subjects participated in the study. In order to acquire a
representative sample of the community as study participants, letters were sent to 274
households, or approximately 10% of the ~ 2800 households in Lake County.
Addresses were selected at random from the 1998 Polk City Directory: Leadville
Colorado, including Lake County (1998) (90), which is updated annually, and
includes all residences and businesses, even if they have unlisted numbers or do not
have phones. The letters briefly described the study, and invited participation by
anyone in the household who met age and minimum length of residence criteria.
Individuals from 75 of these 274 randomly selected households participated, a
response rate of 75/274, or 27.4%. The 88 participating individuals from these 75

households represented 30.5% of the total study population (88 individuals from
randomly selected households/ a total of 288 study participants)
The initial goal was to obtain a completely random sample but while the
response rate was respectable, the number of individuals who chose to participate was
too small to provide adequate statistical power. Mailing letters to randomly selected
Lake County residences and contacting these households to attempt to enroll residents
in the study had proven to be a time consuming process for which reliable assistance
was unavailable. To circumvent this problem, I decided to contact organizations and
employers in Leadville directly. These sites were given flyers which described the
study, its local relevance and criteria for participation, and provided information on
how individuals could become involved, with the request that the flyers be posted in
central places. Flyers were also displayed at the library, city and county offices, the
senior center, community hospital, supermarkets, restaurants, stores and other
meeting places. In response to this approach, 180 additional individuals volunteered,
for a total of 268 participants. These individuals represented a broad cross-section of
the community, from the mayor to individuals patronizing church soup kitchens, from
lifetime residents to recent immigrants, and from endurance athletes to elderly veterans
of Leadvilles boom days.
Finally, it was important to include individuals from among those who had
already received a diagnosis of excessive polycythemia, equivalent to CMS by the
decision of the CMS Consensus Group in 2001 (97), in order to examine their
responses and to evaluate how they compared to the rest of the study population and
to the newly diagnosed cases identified during the study. At the investigators

request, St. Vincents General Hospital in Leadville notified the approximately 120
individuals who were under routine treatment for excessive polycythemia about the
study in progress. These letters provided information about the study, and included
stamped, postcards, pre-addressed to the investigator, which interested individuals
could return to request participation. Through this outreach, 20 previously diagnosed
CMS cases volunteered (16 men and 4 women) and were included in the study,
bringing the total number of participants to 288, including 148 women and 140 men.
Materials and Methods
Study Location and Questionnaire
Studies were performed from July through November of 1998 at St. Vincents
General Hospital in Leadville, Colorado (3100 meters, barometric pressure averaging
about 532 mm Hg by published research [95]). On arrival at the study site, all
participants gave informed consent to the study procedures, as approved by the
University of Colorado at Denver Human Research Committee, (see appendix).
All participants completed a written questionnaire (see text of this document in
appendix), which included background questions on age, gender, ethnic origins,
subjective health status, and a brief medical history. To establish the extent to which
participants had been exposed to high altitude hypoxia, questions asked how long
they had lived in Lake County, and the number of days spent at altitudes lower than
6,000 feet (1648 meters) during the preceding six months. Additional questions
covered potential occupational and behavioral risk factors. Occupational exposure was

a categorical variable, simply indicating whether participants had been exposed to
mining dust, smelter scale, smoke, or similar conditions, or not. Years of work
exposure was a numerical variable, suggesting a relative dose of occupational
exposure. Participants were also asked if they currently smoked tobacco, had smoked
in the past but quit, or had never smoked. For all those who had ever smoked,
information was requested on the number of years they smoked, and the number of
packs per day; this information provided an estimation of their smoking exposure or
dose. Pack-years are defined as the average number of packs of cigarettes per day *
the number of years smoked. For example, someone who has smoked 2 packs per
day for 15 years would be rated as having smoked for 30 pack-years. Both years of
smoking and the number of packs per day can separately, as well as synergistically,
increase risks for pulmonary disease and possibly risk of secondary polycythemia as
well. One question covered alcohol use, as alcohol is known to increases risks of
sleep-disordered breathing (SDB) and polycythemia among men and post-menopausal
women (78, 115).
To screen for sleep-disordered breathing and sleep apnea, participants were
asked whether they had been told they snore, and if so, how frequently and how
loudly. They were also asked if they had been told they sometimes stopped breathing
for brief periods while sleeping, or if they woke gasping for breath.
The CMS score, as developed in Peru by Arregui, Leon-Velarde et al.(4,58)
was included in the questionnaire. In Peruvian research, the CMS score was designed
to screen for CMS on the basis of symptoms commonly associated with
polycythemia. Common symptoms, or those which might also occur with a variety of

other health problems, included dizziness, physical weakness or fatigue, mental
fatigue or difficulty thinking, anorexia or lack of appetite, and muscle or joint pains.
Each of these fairly general symptoms was given a score based on how frequently it
occurred: 0 points if the symptom occurred never or very seldom, 1 point if it
occurred once or twice per month, and 2 points if it occurred three or more times per
month. Symptoms moderately specific for CMS included breathlessness or heart
palpitations, difficulty sleeping or frequent waking, cyanosis, prominent blood
vessels in the eyes and/or large veins in hands and feet, and numbness and tingling
(paresthesias) of hands and feet. Each of these moderately specific symptoms received
0,2 or 4 points depending on frequency. The two most specific symptoms of CMS,
according to the Peruvian instrument, were defined as headache and tinnitus, and each
of these highly specific symptoms received 0,3 or 6 points based on frequency.
Points for all categories were totaled to produce a CMS score with a numerical
scale from 0 to 40 to indicate the number and frequency of CMS symptoms. This
same method of quantification was used in this study, In the Peruvian work,
symptom scores greater than the mean score two standard deviations, that is, greater
than a score of 21, were defined as high and were associated with increased likelihood
of polycythemia. It was anticipated that the CMS score would provide a useful
method of screening for polycythemia in the Lake County population as well.
In order to better investigate CMS prevalence among women in the Lake
County study, the questionnaire included additional questions covering menopausal
status and use of birth control pills or hormone replacement therapy (HRT). These
were included to help clarify issues regarding CMS prevalence among women before

and after the menopause. Altitude residence is known to be associated with low infant
birth weights (77). Previous research has shown that maternal ventilation and SaQ,
are higher in women who have higher birth weight infants (ibid.). To test the
possibility that women who have lower birth weight infants might be at greater risk
for future development of CMS, women who had delivered babies while living in
Lake County were asked for information on the birth weight of each of their babies.
Study Measurements
Height and Weight. After completion of the questionnaire, height and weight
were measured using standard anthropometric techniques. Height is needed to
determine expected volumes in pulmonary function testing by spirometry, and both
height and weight are needed to calculate body mass index (BMI) as a measure of
obesity, where BMI = (weight in kilograms / height in meters2)(106). Neck
circumference was measured in all subjects because neck sizes greater than 18 inches
(45.7 cm) in men, or 15 inches (38.1 cm) in women, are predictors for sleep-
disordered breathing (SDB) or obstructive sleep apnea (Robert Ballard, personal
communication). Since SDB and sleep apnea appear to be common in CMS cases
(50,113, 114), and because the measurement is easy to accomplish, this element was
added to determine if neck size would be of any value in predicting or recognizing risk
factors for CMS.
Arterial oxygen saturation (SaOj. Sa02 and pulse rates were monitored in
comfortably seated subjects using a Biox 3740 Oximeter with a flex II probe
(Ohmeda, Louisville, CO), during the last 2.5 minutes of a 5 minute test period. The

probe was clipped to a finger on either hand, and readings were taken every 30
seconds. The six values obtained during the last 2.5 minutes were averaged to provide
each individuals average saturation.
Pulmonary Function Testing. A portable recording spirometer (Breon #2400,
Spirometer Inc., Waltham, MA) was used for all pulmonary function tests. Tests
done included measured volumes for the forced expiratory volume in 1 second(FEV,)
and forced vital capacity (FVC) and the calculated ratio FEVj / FVC. All study
measurements were made on seated, rather than standing, subjects, as recommended
in the manual Clinical Pulmonary Function-Testing p. 12 (80).
Forced vital capacity (FVC) is a test in which subjects take as deep a breath as
they can, and then exhale as forcefully and quickly as they can into the spirometer
mouthpiece. The spirometer graphs the volume of air in liters, as well as graphing the
relationship between the duration of exhalation and volume of air. The forced
expiratory volume in 1 second (FEV,) is shown on this graph as the volume of air
exhaled during the first second. These tests were completed in triplicate, and the best
effort was accepted and used to measure FEV1 and FVC volumes in liters.
Volumes measured were at ambient temperature and pressure, dry (ATPD),
and the volumes needed to be corrected to body temperature and pressure, saturated
(BTPS) to make it possible to compare them with published values expected for
individuals in good health. Volumes were corrected to BTPS using the equation:
V(BTPS) = V(obs) (273 +37 / 273 + t) (Pb- Ptto / Pb- 47),
where V = volume, V(obs) = observed spirometer volume, t = spirometer temperature
(22*C or 72Y in the hospital setting), Pb = Barometric pressure in mmHg at 3100
meters, Ph2o= vapor pressure of water at 22*C, (20 mmHg) and 47 = vapor pressure

of water at 37C (49). Barometric pressure in Leadville from mid-June through mid-
November averaged 532 mmHg, as reported by Reeves et al, 1993 (95). Calculation
of the formula results in V(BTPS) = V(obs) 1.109, and all observed FVC and FEV,
volumes were corrected to BTPS by this factor. The volumes corrected to BTPS are
reported as Actual FEV, and FVC, and the FEV, / FVC ratio was calculated from
these BTPS volumes.
Pulmonary function test results vary not only by individual constitution, but
also with age, gender, and height. Hence, the results of each participants pulmonary
function test, corrected to BTPS, were then evaluated using the Intermountain
Thoracic Societys Manual, Clinical Pulmonary FunctionTesting (80). This manual
provides pulmonary function values expected for men and women in good health,
standardized for height and age. These values, termed expected values, are then
compared to the actual BTPS volumes or ratios obtained in pulmonary function testing
to produce a percentage of expected values. For example: an Actual FVC of 4.8 liters
/ Expected FVC of 5.2 liters for a healthy person of the same gender, age and height
(from the manual) 100 = 92.3 % Expected FVC. In general, if the percentage
expected value is over 80%, it is considered to indicate (relatively) healthy pulmonary
function. Both the absolute (BTPS) values for FEV,, FVC and FEV,/FVC
measurements, and the % expected FEV,, FVC, and FEV,/FVC values were used in
the statistical analyses.
Blood Pressure. Blood pressure (BP) was taken manually by the investigator
with sphygmomanometer and stethoscope. BP has been reported to be significantly
higher among CMS cases than healthy individuals in several studies in Peru (57,86)
and on the Tibetan Plateau (84, 136). Other studies (21) have reported no difference

in BP between CMS cases and healthy individuals. For this study, considering that
the potential for increased blood volume and changes in cardiac function in CMS
cases, when combined with the status of pulmonary or systemic vascular tone (34,
106), and the effect these could have on blood flow and oxygen delivery,
measurement of BP should be mandatory in this and future CMS studies.
Hematology and Biochemistry. Blood was drawn by venipuncture from the
antecubital vein. Samples included a 7 ml vaccutainer (Becton-Dickinson, Franklin
Lakes, NJ) with 12.6 mg KjEDTA to prevent coagulation for hematological analysis
for hemoglobin and hematocrit analysis, and two 10 ml vaccutainer tubes for serum
samples intended for erythropoietin, ferritin, and hormone assays. Duplicate
hematocrit values were obtained immediately after the blood draw by the
microcentrifuge method. The lavender-top tubes were then refrigerated and
transported to Denver for duplicate assay of hemoglobin, carboxyhemoglobin
(COhgb), and methemoglobin (Methgb) by hemoglobinometer (OSM 3, Radiometer,
Westlake, OH). The carboxyhemoglobin / hemoglobin ratio, used routinely (38,134)
to evaluate severity of carbon monoxide toxicity, was calculated as: COHgb g/dl /
total Hgb g/dl 100.
Evaluation of Tests. Hematocrit and hemoglobin concentrations are the most
frequently used diagnostic signs in the evaluation of CMS, and as such they were
important in evaluating the normal range present in this population and consequently
the characteristics of abnormal or excessive polycythemia. The ranges for
hemoglobin and hematocrit used at Leadvilles St. Vincents General Hospital
Laboratory were not used to define the normal range for this study. Separate measures
were made to rule out both differences in methodology and equipment, and the chance

that normal values grounded on data gathered by a hospital-based laboratory might
differ from values in the non-hospitalized population.
Arterial oxygen content (Ca02) was calculated from Sa02 and hemoglobin
(Hgb) data, as: (hgb 1.36) Sa02 = Ca02. The OSM3 Hemoglobinometer used for
Hgb assay also provides data for carboxyhemoglobin (COHgb) and methemoglobin
(Methgb) concentrations. This made it possible to calculate a corrected Ca02 value by
subtracting COHgb and Methgb values from the total Hgb as shown below.
[(Total Hgb COHgb) Methgb] 1.36 Sa02 = Ca02 corrected for the presence of
COHgb and Methgb. This correction is important because neither COHgb nor
Methgb is capable of transporting oxygen, and for individuals with high COHgb or
Methgb levels the Ca02 would appear to be falsely elevated.
Sample Processing. Red-top tubes were allowed to clot for approximately one
hour, then were spun in a refrigerated centrifuge at 2500 rpm for 10 minutes. The
serum was separated and aliquoted into microcentrifuge tubes, which were
immediately sealed and frozen in liquid nitrogen. These frozen samples were
transported to the Denver laboratory on dry ice, and stored at minus 70T ( minus
56CC). Samples for assay of erythropoietin and ferritin were shipped frozen to Ramco
Laboratories, Houston, Texas, where they were analyzed by radio immuno-assay
(RIA). This laboratory provided a normal sea level range for erythropoietin of 10 to
20 mU/ml, and for ferritin of 20 to 300 ng/ml. Frozen serum specimens are being
maintained for future assay of sex hormone concentrations. When these results are
available, they will be evaluated in analyses designed to assess the contribution of sex
hormone levels to both respiratory control and erythropoiesis.

All data are reported as mean one standard error of the mean (SEM) in the text,
tables, and figures. Characteristics of normocythemic and polycythemic men and
women were compared using unpaired Students t-tests or chi-square tests. Multiple
regression, using a forward stepwise procedure, and analysis of variance were used
to evaluate the relationships between variables, and the degree to which specific
variables contributed to the outcome of CMS. Comparisons were considered
significant when p < 0.05.

This chapter presents the results of research studies on CMS undertaken at Leadville,
Colorado, with the discussion set out in five sections. First, the normal ranges for
key variables known to be associated with CMS are presented in order to define a
group of participants in nominal good health. Second, estimates of CMS prevalence,
the sex and age distribution of cases among residents of Lake County, Colorado, are
provided. Third, the findings for the normal and CMS groups in Lake County are
then used to examine a range of factors that relate to the presence of polycythemia.
The fourth section discusses the potential that identification of new CMS cases among
premenopausal and postmenopausal women offers, providing an opportunity to
determine whether prevalence rates for CMS differ between premenopausal and
postmenopausal women in this population. Also discussed for this population is the
additional element of hormone replacement therapy (HRT); its availability in Lake
County has made it possible to evaluate the potential effects of exogenous sex
hormones on the presence of CMS and factors related to it among postmenopausal
women. Included in this section is information on birth weights of infants bom in
Lake County to women study participants with and without CMS. The fifth section
summarizes these data and findings and suggests relationships among specific factors
that point to possible causal pathways for the development of CMS.

Establishing the Normal Range of Variation
Study Participant Characteristics
Table 3.1 presents basic characteristics of the study participants as well as those of all
Lake County residents over age 20 (n = 5909; females = 2898 or 49.0%, males =
3011 or 51.0%, using data supplied in the Colorado Department of Demography 1998
estimate (16).
Table 3.1 Sex, Age, and Ethnic Comparison of Overall
Adult Lake County Population with all CMS Study Participants,
by Number (n = ), Percentage (%), and Age Group
Age Groups (years) All Female Lake County Residents n = 2898 49.0 % Female Study Participants n= 148 51.4% All Male Lake County Residents n = 3011 51.0% Male Study Participants n= 140 48.6%
20-29 n = 608 21.0% n= 23 15.5% n = 653 21.7% n= 16 11.4%
30-39 n= 832 28.7% n= 32 21.6% n= 793 26.3% n= 29 20.7%
40-49 n= 727 25.1% n = 40 27.0% n= 805 26.7% n= 43 30.7%
50-59 n= 320 11.0% n = 31 20.9% n = 363 12.0% n= 18 12.9%
60-69 n = 227 7.8% n= 12 8.1% n= 297 8.8% n= 20 14.2%
70+ n= 184 6.4% n= 10 6.7% n= 150 5.0% n = 14 10.0%
Totals n = 2898 49.0% n= 148 51.4% n = 3011 51.0% n= 140 48.6%
Mean age ( SEM) 42.2 Years (approximate) 45.6 1.2 Range 20 88 43.1 Years (approximate) 47.4 1.3 Range 20-87
Ethnicity (%) 76% White 24% Hispanic 82% White 18% Hispanic 76% White 24% Hispanic 90% White 10% Hispanic
The study population represents all age groups present among adult men and women
in Lake County. As the table shows, age distribution is similar between participants

and residents; however among men the representation of Hispanics is lower in the
participant population than in the resident population. Even small differences in age
and sex ratios as exist between the total and study populations could cause inaccurate
estimates of CMS prevalence. Hence, prevalence data for CMS were age-adjusted for
both men and women in order to control for differences in age distribution between
the participant sample and the aggregate Lake County population.
Identification of Participants in Nominal Good Health
Individuals with heart or lung disease may develop polycythemia even at sea level. In
order to define normal ranges for hematocrit in healthy individuals in the Lake County
population, it was necessary to screen participants on the basis of medical history and
self-assessment of health (both included in the questionnaire) as well as study
measurements. Only results from those individuals in nominal good health, based on
the following criteria, were used to establish normal values for hematocrit. Thus,
participants whose results for one or more pulmonary function tests (FEV,, FVC or
FEV,/FVC) were below 80% of the values expected for persons of their gender, age,
and height were excluded from the pool of individuals in designated as being in
nominal good health. Also excluded were individuals who suffered from severe
scoliosis, osteoporosis, or osteomalacia, because these conditions have been
suggested as possible risk factors for CMS (70, 130). Further qualifications
necessary for individuals to be included in the pool of participants considered to be in
nominal good health required that persons identify themselves as being in Good
Health on item 3 of the questionnaire, and that they gave a medical history on item 17
free of diagnosis or treatment for high blood pressure (measured pressures greater

than 140/90 mmHg), stroke, clotting problems, heart or lung disease, or a pre-
existing diagnosis of CMS. Self-reported data of this kind is subject to error, but
obtaining further verification under the terms of this study was not practicable.
Individuals with controlled diabetes were included if no specific health problems as
described above were found. Investigator and subject assessment of nominal good
health agreed in 90% of the cases. The reasons for which individuals were excluded
from the group in nominal good health, and from use of their hematocrit data to define
normal values in this population are summarized in Table 3.2.
Table 3.2. Reasons for Exclusion from the Category "Nominal Good Health
Number of Individuals Excluded Reasons for Exclusion
12 Medical history of heart disease
8 Medical history of lung disease and poor PFTs
3 Medical history of stroke or clotting problems
18 Previous diagnosis of polycythemia / CMS
10 Pulmonary Function Tests (PFTs) < 80% of expected
13 High blood pressure > 140/90
2 Severe osteoporosis / scoliosis
Total number excluded: 66 (22 women, 44 men)
Of the original 288 participants, 22 women and 44 men were excluded from
the category nominal good health. Age and sex distributions for the remaining 222
participants who were considered to be in nominal good health are shown in Table
3.3. Women predominated in this subset of the study participants, as they did in the
subject group as a whole, in part because all individuals in good health over age 70
were women. Age distribution of participants included in the category nominal good
health are given in Table 3.3.

Table 3.3. Age Distribution for Men and Women
in Nominal Good Health at 3100m (n 222)
Ages 20-29 30-39 40-49 50-59 60-69 70+ Totals
Men 16 25 35 10 10 0 96
Women 21 31 36 25 7 6 126
Table 3.4 summarizes data for the 222 individuals in good health. In each
gender, mean values for height, body mass index (BMI), blood pressure, and
pulmonary function tests were generally unremarkable. Men and women are similar
in their smoking patterns, with approximately 1/2 of each group being current or
former smokers and the remainder never smoked. For both men and women the mean
number of pack-years (equivalent to the number of packs of cigarettes smoked per day
times the number of years smoked) was relatively low. Twenty-one percent of the
men had occupational exposures due primarily to mining dust, as well as to smelter
ash or smoke, but the mean number of years of exposure was 2.9. Womens dust
exposure was very low, both in terms of categorical exposure to mining, and the
number of years exposed. Mean arterial oxygen saturation (Sa02) was low compared
to sea-level standards, but comparable to results for populations at similar altitudes
(80). For these healthy men, FVC and FEVj averaged about 108% of the values
expected for age and height; while for women in good health the mean was closer to
114% of expected values.

Table 3.4. Characteristics of Men and Women in Good Health at 3100 m
Men n = 96 Women n = 126
Age (years) 41.8 1.2 43.6 1.2
Age range 20-69 20-85
Duration of residence (years) 14.0 1.6 17.7 1.6
Range 0.5 62 0.5-83
CMS Symptom Score 9.1 0.7 11.6 0.6
Height (m) 1.8 0.0 1.6 0.0
Body mass index (BMI) 25.4 0.4 26.1 0.5
Neck circumference (cm) 41.2 0.6 34.3 0.2
Systolic BP (mmHg) 125.0 0.9 121.1 1.0
Diastolic BP (mmHg) 74.7 0.6 73.5 0.5
Heart rate, beats / min. 77.4 1.2 78.3 0.9
Actual FEVj (liters) BTPS 4.6 0.1 3.3 0.1
Actual FVC (liters) BTPS 5.7 0.1 4.0 0.1
Actual FEV,/FVC (%) 80.8 0.6 83.7 0.5
FEVj as % expected 107.6 1.5 114.3 1.3
FVC as % expected 107.9 1.6 113.5 1.3
FEV,/FVC as % expected 100.1 0.7 101.7 0.6
Current or past smokers (%) 55.8 50.0
Never smoked (%) 44.2 50.0
Pack-years of smoking 7.1 1.1 7.9 2.6
Occupational dust exposure (percentage) 21.1 0.8
Years of mining, or other exposure 2.9 0.7 0.09 0.1
SaO, (percentage) 95% confidence interval (two tailed) 92.8 0.2 89.6 96.0 93.5 0.15 90.3 96.7
Hemoglobin g/dl 95% confidence interval (one tailed) 17.2 0.1 15.2 - 18.8 15.2 0.1 13.1 17.0
Hematocrit % 95% confidence interval (one tailed) 50.4 0.3 43.6 56.0 45.9 0.3 39.7-51.1
Carboxyhemoglobin (COhgb) g/dl 1.2 0.20 0.8 1.3
Values are given as means SEM.

Hemoglobin and Hematocrit Normal Values
By study design, and in accord with the current recommendations of the
International CMS Study Group (97), the only criterion used to identify CMS cases in
this study was excessive erythrocytosis or polycythemia as measured by hematocrit
(hct). This seemed appropriate for several reasons. Beyond the International CMS
Study Group criterion, viscosity problems may be important in the syndrome of
CMS, and hematocrit is directly correlated with blood viscosity (34,85), while
hemoglobin (hgb) values are not. Finally, it was important to be as inclusive as
possible in identifying CMS cases in order to be able to address the wide range of
factors that may relate to the etiology or longitudinal development of CMS. Although
hemoglobin has been used as a criterion for defining CMS in several studies (56,59),
its use was not considered in this study, because although they are often closely
related, the relationship between hgb and hct may vary, given that differences in red
blood cell volume and hemoglobin concentration per red blood cell may exist.
Also of concern was the existence of age-related increases in hct and hgb, as
have been reported from Peru (56, 57). As shown in Figure 3.1 for men and Figure
3.2 for women, hemoglobin actually falls modestly with age among men (p = 0.015),
and women (p = 0.051). As no significant association between hct and age was
found, in either men (fig. 3.3) or women (fig. 3.4) in this study, there was no need to
define age-specific normal values for hematocrit in either men or women.
Having decided to use hct measures as the central factor in identifying CMS
cases, the next issue was how to clearly determine where to set the cutoff point
between the normal range for hct, and excessive hct levels. As had been expected, the
distribution of hematocrit in men (fig. 3.5) and women (fig. 3.6) was normal by the

Figure 3.1 Hemoglobin by Age for
Men in Good Health
k* u. Age (years)
A Men in good Health
Figure 3.1: For men in nominal 'good health', (n = 96),
hemoglobin concentrations fall modestly with increasing
age (y = 18.236 0.024x, r2 = .087, p = .0036).

Figure 3.2 Hemoglobin by Age for
Women in Good Health
Age (years)
O Women in good Health
Figure 3.2: For women in nominal 'good health' (n = 126),
hemoglobin concentrations fall modestly with increasing age
(y = -0.016x + 15.941, r2 = .04, p = .02).

Hematocrit (%)
Figure 3.3 Hematocrit by Age
for Men in Good Health
Age (years)
A Men in Good Health
Figure 3.3: In this hematocrit-dependent regression for Men in
nominal good health', hematocrit remains stable with increasing age.
y = -0.036x + 51.882, r2 =.016, p = .22.

Hematocrit (96)
Figure 3.4: Hematocrit by Age for
Women in Good Health
Age (years)
O Women in 'Good Health'
Figure 3.4: For women in nominal 'good health' (n = 126),
hematocrit remains stable with increasing age in this hematocrit -
dependent regression (y = -0.024x + 46.975, r2 = .01, p = .26).

Figure 3.5 Histogram of Hematocrit
Values for Men in Good Health
Figure 3.5: The hematocrit is normally distributed among
Lake County men in nominal good health, (n = 96).

Figure 3.6 Histogram of Hematocrit Values for
Women in Good Health
35 40 45 50 55 60
Hematocrit (%)
Women in nominal good health
Figure 3.6: The hematocrit is normally distributed among
Lake County women in nominal 'good health', (n = 126).

Shapiro-Wilk W statistic, p = NS. Most published studies on CMS have used the
mean value 2 standard deviations to define the normal range. When this method is
used, 95.45% of the distribution will be in this range, with the remaining 4.55%
divided between 2.275% below the lower end of the normal distribution, and
2.275% beyond the high end. As a result, only the top 2.275% of values are used to
identify the top of the normal range; in consequence, polycythemia cases might be
under-identified, while individuals with abnormally low hgb or hct values were of
little interest in the study of CMS. Instead, it seemed appropriate to consider the
entire top 5% of the normal distribution as above the normal range, and so make it
possible to evaluate CMS cases in greater depth. Using hct cutoff values defined by
the mean hct value plus the standard deviation a z-score of 1.645 resulted in 95% of
sample values below the cutoff of 56.0% for men, and 51.1% for women, as shown
in Table 3.5.
Table 3.5. Mean Hematocrit, z-score, and Normal Range for Men and Women in
Nominal Good Healt l
Mean Hematocrit Standard Deviation * z-score Upper Limit of Normal Range Normal Range
Men 50.38% (n = 96) 3.40 * 1.645 55.97% 41.7-56.0%
Women: 45.90% (n = 126) 3.14 * 1.645 51.06% 37.4-51.1 %
CMS Prevalence by Sex and Age
CMS prevalence estimates in this study were developed based solely on evaluation of
those study participants who had never been previously diagnosed with polycythemia,
using hematocrit cutoff values greater than 56.0% for men and 51.1% for women to

define polycythemia. The study found 12 new cases in men and 12 new cases in
women, yielding prevalence estimates of 9.04% for men and 8.33% for women.
The age distribution of the study participants differs from that of the Lake
County population when compared with related data in the Colorado Department of
Demographys 1998 estimate for Lake County. As shown in Figure 3.7 for men, the
20 29 and 30 39 year old age groups are underrepresented in the study population,
while the 40 49 and 60 69 year age groups are over-represented, as compared to
the proportions present in the county population. The representation for the 50-59
and 70+ age groups is similar in the study population and the county population. In
Figure 3.8 for women, the 20 29,30 39, and 60 69 year old age groups are
underrepresented, the 40 49 and 50 59 year age groups are over-represented, and
representation for the 70+ group of the study population is similar in proportion to
that present in the county population as a whole.
In effect, these findings indicate that the study population is somewhat older
than the Lake County population as a whole. Thus, since the development of CMS is
thought to be associated with increasing age (68,107), a prevalence estimate for the
Lake County population based on the number and age distribution of new cases in the
study population could overestimate the prevalence. Age adjustment was therefore
used to correct the estimate, and was carried out as follows.
The percent prevalence for each ten-year age group, shown in Tables 3.6 and
3.7 below, was multiplied by the proportion of the entire Lake County population in
that age group, to produce an age-corrected estimate of the total number of cases in
that age group. The sum of these estimates is the total number of CMS cases likely
present in men and women in Lake County as a whole. For men, this was 270 cases

Figure 3.7: Age Distribution of Men
Study Participants Compared to
Male Lake County Residents
Lake County Residents
Figure 3.7: The age distribution of Lake County men provided by the 1998
Colorado estimated census is shown by filled boxes and unbroken lines.
Compared to this census data and shown as crossed circles and dashed
lines, proportions of male study participants were under- represented in
the 20 29 and 30 39 year age groups, over-represented in the 40-49
and 60 69 age groups, and similar in the 50 59 and 70+ groups as
compared to the age distribution of men according to the census.

Figure 3.8 Age Distribution of Women Study
Participants Compared to Female
Lake County Residents
-------- Lake County Residents
Figure 3.8: The age distribution of Lake County women given by the
1998 Colorado estimated census is shown by filled boxes and unbroken
lines. Compared to this census data, and as shown by crossed circles and
dashed lines, lower proportions of study participants were found in the 20
- 29, 30 39 and 60 69 year age groups, while there were higher
proportions of study participants in the 40 49 and 50 -59 age groups.
The proportion of women in the 70 + age group was similar for study
participants and county residents.

270 cases /3011, and for women 251.0 cases / 2898, yielding age-corrected
prevalence estimates of 9.0% for men and 8.6% for women respectively; (Tables 3.6
and 3.7).
Table 3.6. Mens Estimated CMS Prevalence (corrected for age)
Age Groups (years) Number of Study Participants in Age Group Number New Cases in Age Group Prevalence in Age Group (%) Total Male Lake County Population in Age Group Estimated Number of CMS Cases per Age Group (corrected for age)
20-29 16 0 0.0 653 0.0
30-39 29 3 10.3 793 82.0
40-49 39 4 10.3 805 82.59
50-59 14 1 7.1 363 25.92
60-69 18 1 5.6 267 14.85
70+ 7 3 42.8 150 64.27
Totals 124 12 3011 269.63
CMS Among Par NotAg< Prevalence Male Study 12/122 = ticipants 9.84% ;-Corrected Estimated CMS Prevalence Among All 269.6/3011 = Lake County Men 9.0% Corrected for Age
Table 3.7. Womens Estimated CMS Prevalence (corrected for age)
Age Groups (years) Number of Study Participants in Age Group Number New Cases in Age Group Prevalence in Age Group (%) Total Female Lake County Population in Age Group Estimated Number of CMS Cases per Age Group (corrected for age)
20-29 23 1 4.35 608 26.45
30-39 32 3 9.37 832 77.95
40-49 40 3 7.50 727 54.52
50-59 28 1 3.57 320 11.84
60-69 12 1 8.33 227 18.91
70+ 9 3 33.33 184 61.33
Totals 144 12 2898 251.0
CMS Prevalence Among Female 12/144 = Study Participants 8.38% Not A ge-Corrected Estimated C Amc Lake Cou Correct* VIS Prevalence >ngAll 251.0/2898 = nty Women 8.6% ;d for Age

When the age-corrected prevalence estimates for each sex, as shown in Tables
3.6 and 3.7 above, are inserted into Table 1.1, the results in Table 3.8 are produced.
Table 3.8. CMS Prevalence Worldwide, Including Lake County (46, 54, 108)
Moderate Altitude 2260 2800 m High Altitude 3050 3800 m Very High Altitude 4000 5200 m
Population Male Female Male Female Male Female
Tibetans n = 12,385 0.0 0.0 1.1 0.6 2.0 1.6
Chinese (Han) n = 13,233 1.4 0.7 5.2 1.5 10.1 6.6
Peruvians n = 365 no data no data no data no data 15.6 8.8
Bolivians, n= 1,115,000 no data no data 5.2 no data no data
Lake County n = 288 no data no data 9.0 8.6 no data no data
Characteristics of CMS Cases and Normocythemic
Men and Women
In order to better understand the characteristics of CMS cases, in both males and
females, the investigator included all individuals with CMS, whether identified during
the study (12 men and 12 women) or previously diagnosed by Leadville physicians
and currently under treatment (18 men and 4 women), in the analyses that follow.
This provided CMS sample sizes of 30 men and 16 women. Table 3.9 provides
information on the gender, CMS status, and age distributions of all study participants.
Tables following 3.9 summarize the characteristics of normocythemic men and
women, as compared to their CMS study counterparts. After general descriptions of
the results shown in these tables, characteristics of special interest are further
illustrated in separate graphs for men and women.

Table 3.9. Age Distribution for Lake County Men and Women
with and without CMS (actual count)____________________
Age Group 20-29 30-9 40-49 50-59 60-69 70+ Totals
Normocythemic men 16 26 34 13 17 4 110
Men with CMS 0 3 9 5 3 10 30
Normocythemic women 22 29 37 28 10 6 132
Women with CMS 1 3 3 3 2 4 16
Total / Age group 39 61 83 49 32 24 288
Figures 3.9 and 3.11 show incidence of new cases of CMS among men and
women by age group, compared to those participants who had never been identified
as CMS cases. Figures 3.10 and 3.12 show total CMS prevalence for men and
women, respectively by age group, and include all participants, both CMS cases
identified during the study, and those previously diagnosed and under treatment by
local physicians, as well as their normocythemic participants. Figures 3.9 and 3.11,
showing incidence for men and women, are very similar, with comparatively low
CMS rates (< 10%) during the first four to five decades of adult life, and then
substantially higher incidence in the groups age 70 years and over (Mens incidence
rate = 43%, womens = 33%). When all CMS cases, both those diagnosed prior to
the study, and those identifed during the study are included as in Figure 3.10 for men
and 3.12, the figures show CMS prevalence rates. In Figure 3.10, men show a
pattern of steady increase in prevalence with age, with the exception of the 60-69
year old age group, which has an unexpectedly low prevalence. Among men 70 and
over, the prevalence rate was over 70%, the high proportion reflecting an inflation of
this population caused by the inclusion of 16 previously diagnosed CMS cases, 7 of

Prevalence by percent within each Age Group
Figure 3.9: Incidence of New CMS
Cases Identified Among Men Living at
3100 m, by Age Group

20 29 30 39 40 49 50 59 60 69 70 +
Normocythemic Men Age Groups
B Men with CMS
Figure 3.9: This figure shows all male study participants who had
never been identifed as CMS or excessive polycythemia cases prior
to this study (n = 122). These men are divided by age group, and
the proportion of new CMS cases and normocythemic men in each
age group is shown. CMS incidence appears to be 10% or less in
each age group of the first five decades of adulthood, with slightly
higher incidence during the 30's and 40's, then appears to increase
rapidly in men over 70 years.

Figure 3.10: Prevalence of CMS in Men,
by Age Group
_ 20 29 30 39 40 49 50 59 60 69 70 +
FI Normocythemic Men
H Men with CMS Age Groups
Figure 3.10: In this figure, the total number of men in each age group
represents 100%. This total is divided proportionally between CMS cases
(both those identifed during the study, and those previously diagnosed)
and normocythemic men to indicate the relative prevalence of CMS by age
group, and so the age distribution of CMS cases in this population. With
the exception of the low proportion of CMS cases in the 60 69 year age
group, CMS prevalence among men appears to increase steadily with age.
However, the prevalence of cases in the 70+ age groups was inflated by
inclusion of 7 previously diagnosed CMS cases, while there was only one
previously diagnosed case in the 60 69 age group.

Figure 3.11: Incidence of New CMS
Cases Identified Among Women Living
at 3100 m, by Age Group
20 29 30-39 40 49 50 59 60 69 70 +
C3 Normocythemic Women Age Groups
H Women with CMS
Figure 3.11: This figure shows those female study participants
who had never been identifed as CMS or excessive polycythemia
cases prior to this study (n = 144). These women are divided by
age group, and the proportion of new CMS cases and
normocythemic women in each age group is shown. Among
these women, CMS incidence is under 10% during the first five
decades of adulthood, but appears to be somewhat greater during
the 30s and 40's, then increases rapidly in women aged above
70 years.

Figure 3.12: CMS Prevalence among Women,
by Age Group
20 29 30 39 40 49 50 59 60 69
Age Groups
Normocythemic Women
70 +
B Women with CMS
Figure 3.12: In this figure, the total number of women in each age
group represents 100%. That total is divided proportionally between
CMS cases (including both those identifed during the study, and
previosly identifed cases) and normocythemic women to show the the
relative prevalence of CMS by age group, and the age distribution of
CMS in this population. Among the women of Lake County, CMS
appears to have a low prevalence during the first 4 decades of adult
life, and then to increase rapidly in prevalence after age 60.

whom were over 70. In contrast, women show consistently low prevalence in the
four youngest age groups. Prevalence increases in the last two age groups, reaching
40% in the 70+ age group. The degree to which out-migration has affected these
outcomes is unknown, but the somewhat lower prevalence rate among women in the
70+ age group, compared to men, is likely to be due (at least in part) to the larger
proportion of women surviving into their eighties in the population. Prevalence rates
by age group (shown in figures 3.10 and 3.12) result in CMS age distributions
among men and women in Lake County that are generally similar to that described for
men and women of Cerro de Pasco, Peru (59, 69).
Participants who were able to remember the age at which they were originally
diagnosed with polycythemia gave ages between 18 and 71 years, with many
diagnosed during their mid-thirties. This is described more fully at the end of this
section under Case and Family Histories.
Age. CMS cases were significantly (p < 0.05) older than normocythemic
Lake County residents. While the age range of cases was higher than that of normal
residents, it is important to note that not all CMS cases were found in older age
groups. Ages, as given in Table 3.10, were those present at the time of the study, not
age of diagnosis. The youngest male case was 30 years old, and the youngest female
case 24 years old, indicating that CMS can and does affect young adults.
Duration of Residence. CMS cases have lived at 3100 m longer than
normocythemic individuals. Figures 3.13 for men and 3.14 for women provide linear
regressions of duration of residence as a function of age, with separate regression
lines for all normocythemic individuals and all CMS cases. Such plotting highlights

Table 3.10. Age, Duration of Residence, and Age on Arrival at 3100 m
Men Women
Normocythemic n= 110 CMS n =30 Normocythemic n = 132 CMS n =16
Age (years) 44.7 1.3 57.3 2.91 44.6 1.2 54.1 4.6*
Age range 20-81 30-87 20-85 24-88
Duration of Residence (years) at 3100 m 15.9 1.6 41.7 3.8+ 17.5 1.5 31.8 5.5*
Range 0.5 67 2-77 0.5-83 1-72
Age on Arrival at 3100 m: Bom at 3100 m 10/110 or 9.1% 10/30 or 33.3%+ 11/132 or 8.3% 7/16 or 43.7%+
Infancy to Age 10 7/110 or 6.4% 5/30 or 16.7% 13/132 or 9.8% 0
Ages 10 to 20 6/110 or 5.4% 2/30 or 6.7% 34/132 or 25.8% 0
Over Age 20 86/110 or 78.2% 13/30 or 43.3% 96/132 or 72.7% 9/16 or 56.2%*
Values are given as means SEM. Indicates significant difference (p < .05) and +
indicates significant difference (p < 0.001) between Normocythemic and CMS.
the diversity seen in length of residence at high altitude among normocythemic
individuals and CMS cases.
For many of those identified as CMS cases, residence has been lifelong. In
Figures 3.13 and 3.14 these high-altitude natives are indicated on the figures by the
diagonal line marked Lifelong Residents on which age = duration of residence. It is
also important to note that some of the CMS cases newly identified in this study had
been at high altitude for relatively short periods of time. New CMS cases included 5
men who had been at high altitude for 0.5, 2, 5,7, and 11 years. Five women with
residence histories of 1,2,4, 10, and 11 years at high altitude had also developed
CMS. Other cases, especially among women, had lived at altitude for about the same

Duration of Residence at High Altitude (Years)
Figure 3.13: Duration of Residence
at High Altitude, by Age for
Normocythemic Men, and Men with CMS
Men w i th CMS Age (Years)
---X Normocythemic Men
Figure 3.13: Duration of Residence at High Altitude by Age. Men with
CMS (y = 0.860x 7.960, r2 = 0.406, p < .001) have lived at high altitude
significantly longer than normocythemic men (y = 0.576x 9.683, r2 =
0.221, p < .001), and are also more likely to have been bom at high
altitude (p < .001).

Figure 3.14: Duration of Residence at High Altitude
by Age for Normocythemic Women
and Women with CMS
Women with CMS Age (Years)
---x Normocythemic Women
Figure 3.14: Duration of Residence at High Altitude by Age. Women with
CMS (y = 0.724x 6.753, r^ = 0.333, p = .02) have lived at high altitude
for significantly longer than normocythemic women (y = 0.623x 10.182,
tt = 0.236, p < .0001), and are also more likely to have been bom at high
altitude (p < .001).

number of years as their normocythemic counterparts. Thus, prolonged residence at
3100 meters does not appear to be a prerequisite for the development of polycythemia.
Normocythemic men and women showed an increase in both duration of residence at
high altitude and in variability in duration of residence, with age.
Analysis of anecdotal information and questionnaire responses
providedinformation on how long 19 / 30 male and 15 /16 female CMS cases had
lived in Lake County before the diagnosis of CMS was made. These 19 men had
been in Lake County (whether by birth or migration) for an average of 33.4 years
(range 0.6 77 years); the 15 women averaged 31.6 years residence (range 1 72)
before CMS was identified. In all other cases (11/30 men and 1 / 16 women) age
and duration of residence shown in the figures indicate years of residence recorded at
the time of the study.
Age on Arrival at 3100 m. A high proportion of CMS cases were bom in
Leadville. In fact, 43% of the women (7/16) and 33% of the men (10 / 30) with
CMS were bom at high altitude, a number greater (p < 0.001) than the 9% of
normocythemic men and 8% women (Figure 3.15). However, the majority of both
CMS cases and normocythemic persons had arrived at 3100 meters as adults.

Figure 3.15: Age on Arrival for Normocythemic Men
and Women,
and Men and Women with CMS
Women with CMS Normocythemic Women n
n = 16 - 132
Age on Arrival
at 3100 m.:
Birth to age 10
10 to age 20
Over 20 years
Men with CMS Normocythemic Men
n = 30 n=110
Figure 3.15: Men and women with CMS are more likely to have been bom at
high altitude than their normocythemic counterparts.

Height and Body Mass Index (BMI). As indicated in Table 3.11, height did
not differ between normocythemic men and women or men and women with CMS.
However, BMI was significantly greater in men and women with CMS, indicating
that cases were about the same height as normals, but heavier. Persons with body
mass indexes from 25.0 to 29.9 are classified as overweight, and those with BMIs of
> 30.0 as obese (107). BMI classifications placed 52.5% of normocythemic
individuals and 76.1% of those with CMS in the overweight category; some were
Table 3.11. Height (m), Body Mass Index (BMI), Neck Circumference,
and Exposure to Smoke or Mining____________________________________
Men Women
Normocythemic n= 110 CMS n =30 Normocythemic n= 132 CMS n= 16
Height (m) 1.80 0.008 1.77 0.012 1.637 0.006 1.651 0.024
Body mass index (BMI) 25.2 0.4 28.6 Lit 26.8 =t 0.5 29.0 1.8
Neck circumference (cm) 41.2 0.5 45.1 1.4* 34.5 0.2 36.1 0.8*
Current or past smokers (%) 60.5 76.6* 49.2 68. 8
Never smoked (%) 39.5 23.3* 50.8 31.3
Pack-years of smoking 9.1 1.2 19.9 4.1$ 7.9 2.4 10.4 5.2
Occupational Exposure to Mining Dust (%) 26.6% 46.7% * 1.5% 0%
Years Exposure Mining Dust 3.1 6.7 * 0.1 0.0
Values are given as means SE. Indicates Significant difference (p < 0.05), and $
indicates significant difference (p < 0.001) between normocythemic and CMS cases.
Neck Circumference. Average neck circumference was larger for CMS cases
than for normocythemic individuals, although the mean neck circumference for CMS

cases did not equal or exceed the measurements that suggest increased risk of sleep
disordered breathing, > 47.7 cm in men, and >38.1 cm in women, according to
Robert Ballard, MD (personal communication, 1997).
Smoking History and Pack-Years of Smoking. The questionnaire categorized
smoking history three ways: never smoked, smoked in the past, or current smoker
(Table 3.11). Male CMS cases were significantly more likely to be current smokers
or to have smoked in the past than normocythemic men, and male CMS cases
averaged twice as many pack-years as normocythemic men. Female CMS cases did
not differ from normocythemic women in smoking history or pack years.
Occupational Dust Exposures. Male CMS cases were more likely to have had
occupational dust exposures, typically to mining dust, and also had longer exposures
than normocythemic men. Only two women had worked as miners and both were
normocythemic. There were no differences between female CMS cases and
normocythemic individuals in terms of occupational exposures or years of work
Table 3.12. Concurrent Exposures to Smoking and Mining among Men
All Male Participants n= 140 Normocythemic Men n= 109 Men with CMS n = 30
Never smoked and never worked as miner 37/109 or 33.9% 4 / 30 or 13.3%
Never smoked but did work as miner 6 / 109 or 5.5% 3 / 30 or 10.0%
Smoked but never worked as miner 22 / 109 or 20.2% 12 / 30 or 40.0%
Smoked and worked as miner 43 / 109 or 39.4% 11 / 30 or 36.7%
Those men who had both smoked and worked as miners, and those did smoke
even though they never worked in the mines, appeared to be at highest risk for CMS.

Results of Pulmonary Function Tests. Pulmonary function tests (PFTs) used
in this study included Forced Expiratory Volume in 1 second (FEV,) and Forced Vital
Capacity (FVC), both measured in liters and converted to BTPS volumes, as well as
the calculated FEV,/FVC ratio. As summarized in Table 3.14, FEV1 and FVC
volumes were significantly lower for male (p < 0.001) and female (p < 0.05) CMS
cases than for their normal counterparts. Mean FVC volumes drop with age for both
CMS cases and normocythemic individuals. All but six of the CMS cases have FVC
volumes below their normocythemic counterparts. Regressions of age PFT indicate
that male (Figure 3.16) and female (Figure 3.17) CMS cases tend to have lower FVC
than normocythemic individuals. The calculated FEV,/FVC ratio was also lower for
male and female cases than for normocythemic people, as shown in Table 3.13.
When measurements are expressed as % of expected volume given the
individuals sex, age, and height, results for expected FEV, and FVC % were
significantly lower in CMS cases than in their normal counterparts, as shown in
Figures 16 and 17. CMS cases tended to average close to 100% of the expected
values, while their normal counterparts had % expected FEV 1 and FVC values that
averaged about 15% higher. Expected FEV,/FVC did not differ between CMS cases
and normocythemic individuals.
When measurements are expressed as % of expected volume given the
individuals sex, age, and height, results for expected FEV, and FVC % were
significantly lower in CMS cases than in their normal counterparts, as shown in
Figures 3.16 and 3.17. CMS cases tended to average close to 100% of the expected
values, while their normal counterparts had % expected FEV 1 and FVC values that

FVC in liters (BTPS)
Figure 3.16: FVC in Liters (BTPS)
by Age for Normocythemic Men
and Men with CMS
-x--- Normocythemic Men
Figure 3.16: Men with CMS (y = -0.054x + 7.977 r2 = 0.407,
p < .0001) have lower FVC at all ages than Normocythemic Men
(y = -0.056x + 8.618 r2 = 0.379, p < .0001), but the slope of the
age related decline in FVC is simlar.

Figure 3.17: FVC in Liters (BTPS)
by Age for Normocythemic Women
and Women with CMS
---- Women with CMS
Figure 3.17: FVC in Liters (BTPS) for normocythemic women and women
with CMS. After initially similar FVC volumes, FVC in women with
CMS (y = -0.054x + 6.611 r~ = 0.601, p < .0001) falls more rapidly,
and to lower mean volumes with age than FVC among normocythemic
women (y = -0.039x + 6.124 r2 = 0.355, p < .0001 ).

Table 3.13. Pulmonary Function Test Results
Men Women
Normocythemic n= 110 CMS n = 30 Normocythemic n= 132 CMS n = 16
BTPS FEV, (liters) 4.9 0.1 3.8 0.2$ 3.7 0.07 3.1 0.2 *
BTPS FVC (liters) 6.1 0.1 4.9 0.2$ 4.4 0.1 3.7 0.3 *
FEV,/FVC (percentage) 80.2 0.6 76.6 1.8 * 83.6 0.4 80.1 1.5 *
FEV, as % expected 116.6 1.7 98.9 4.8$ 126.4 1.4 110.1 6.4$
FVC as % expected 117.4 1.8 102.3 3.9$ 125.4 1.4 109.6 6.2?
FEV,/FVC as % expected 99.8 0.8 97.2 2.1 101.7 0.6 100.5 1.7
Sa02 (%) 92.7 0.2 89.0. 0.8$ 93.3 0.1 91.5 0.8+
* Significant Difference (p < .05);: significant difference (p < .001).
averaged about 15% higher. Expected FEV,/FVC did not differ between CMS cases
and normocythemic individuals.
These results suggest that while CMS cases may have pulmonary function that
is worse than that of their normocythemic counterparts, they do not clearly suffer
from pulmonary disease, at least as it is defined at low altitude. In order to evaluate
whether CMS cases and normocythemic individuals could be distinguished on the
basis of pulmonary function tests (PFTs), Figures 3.18 and 3.19 show the frequency
of % Expected FVC values. Percentage expected values were used for the purposes
of illustration because these control for the effects of age and height. While CMS
cases have lower mean % expected values than their normocythemic counterparts, the
range of variation in CMS cases is equivalent to that among normocythemic
individuals, or nearly so. Many CMS cases have values that are equivalent to those of

Figure 3.18: Frequency Distribution of
% Expected FVC Values for Normocythmeic
Men and Men with CMS
E3 Normocythemic Men % Expected FVC
B Men with CMS
Figure 3.18: Mean values for % Expected FVC differ between
Normocythemic Men and Men with CMS (p < .001), but the total
range of variation is similar, and there is considerable overlap.

Figure 3.19: Frequency Distribution of
% Expected FVC Values for Normocythmeic
Women and Women with CMS
Eljl} Normocythemic Women % Expected FVC
B Women with CMS
Figure 3.19: Mean values for % Expected FVC differ between
normocythemic women and women with CMS (p < .001), but the
range of variation is very similar, and values for normocythemic
women and women with CMS overlap.

their normocythemic counterparts. This suggests that the condition called primary
CMS, or altitude-related polycythemia in the absence of pulmonary disease, very
likely does exist in this population. It also suggests that poor pulmonary function at
high altitude does not inevitably result in development of CMS.
Arterial Oxygen Saturation, SaCk % Sa02 was lower in both male and female
CMS cases than in normals, as shown in regressions of Age Sa02 (Figures 3.20
for men, and 3.21 for women). The range of Sa02 levels was wide, with some male
and female CMS cases having Sa02 values above those of healthy individuals in the
same age range. In both sexes Sa02 tends to fall with age, but the drop appears to be
steeper and more rapid in people with CMS.

Figure 3.20: Regression of Sa02 by Age
for Normocythemic Men and Men with CMS
Age (years)
---*--- Normocythemic Men
- - - Men with CMS
Figure 3.20: Regression of Sa02 by Age for Men. Initial Sa09 is lower for
men with CMS (y = -0.097x + 94.566, r2 = .128, p = .052) than for
normocythemic men (y = -0.026x + 93.842, r2 = .047, p = .023). Sa07 is
markedly lower with age, and reaches lower levels, for men with CMS as
compared to normocythemic men.

Figure 3.21: Regression of Sa02 by Age
for Normocythemic Women and Women with CMS
--- - Women with CMS
Figure 3.21: Regression of Age by Sa07 for Women. Initial Sa02 was
lower for women with CMS (y = -0.065x + 95.032, r2 = .146, p = ns)
than for normocythemic women (y = -0.040x + 95.094, r2 = .109,
p = .0001), and Sa07 fell more rapidly with age for women with CMS.
However, while five of the women with CMS had high SaO0 seven had
Sa07 below 90 %, and the remaining four had Sa02 values similar to those
of normocythemic women

Hemoglobin and Hematocrit. By study design, both hgb and hct were higher
among CMS cases than in their normocythemic counterparts; see Table 3.14. This
occurred even though a few of the previously diagnosed CMS cases (5 men) had
recently had phlebotomies, and entered the study with hgb and hct in the normal
Table 3.14. Hemoglobin, Hematocrit, hct / hgb Ratio, COhgb / hgb Ratio, Ca02,
Erythropoietin, Ferritin, and CMS Symptom Score_______________________________"
Men Women
Normocythemic n = 110 CMS n = 30 Normocythemic n = 132 CMS n =16
Total Hemoglobin g/dl (hgb) 17.1 0.2 18.7 0.3+ 15.2 0.1 17.8 0.4$
Hematocrit (%) (hct) 50.0 0.3 58.6 0.9? 45.6 0.3 55.9 1.4$
Hematocrit /Hemoglobin ratio (hct/hgb) 2.93 0.01 3.15 0.04+ 3.02 0.01 3.14 0.04 *
Carboxyhemoglobin / Total Hemoglobin ratio 7.7 6.4 3.3 10.8$
Ca02 (ml O2/100 ml blood) not corrected for COHgb & Methgb 21.5 0.1 22.7 0.4$ 19.2 0.1 22.2 0.5$
Ca02 (ml O2/100ml blood) Corrected for COHgb and Methgb) 19.2 0.3 20.7 0.5 * 18.0 0.2 19.2 0.8 *
Erythropoietin mU/ml 17.2 0.8 83.4 39.7 * 17.9 0.6 31.3 6.8+
Log erythropoietin 1.2 0.01 1.5 0.09$ 1.2 0.01 1.4 0.08 *
Ferritin ng/ml 144.0 10.3 127.9 30.4 63.1 8.1 96.4 35.1
Log ferritin 2.1 0.03 1.7* 0.1$ 1.6 0.04 1.7 0.14
CMS score 9.6 0.7 10.9 1.6 12.3 0.7 11.5 1.4
Values are given as means SE. Indicates significant difference (p < .05),
$ significant difference (p < 0.001) between normocythemic and CMS cases.

Hematocrit / Hemoglobin Ratio. In the general population, the ratio of
hct / hgb is expected to average about 3(112), and for all study participants, the mean
was in fact 3.0. However, the ratio tended to be higher for CMS cases. Men with
CMS had average hct / hgb ratios of 3.15 compared to 2.9 for normocythemic men (p
< 0.001), while female CMS cases averaged 3.14 as compared to 3.02 for
normocythemic women(p < 0.05). This suggests that CMS cases, with their higher
hematocrits, have lower concentrations of hemoglobin per red blood cell than
normocythemic people. Figure 3.22 shows the relationship for normocythemic
individuals and those with CMS. At higher hematocrits, the CMS group had both
lower hgb and greater variability in hgb, than normocythemic participants.
Arterial Oxygen Content (CaO-,). Ca02 was higher (p < 0.001) for CMS
cases, male and female, than for normals. When the Ca02 was corrected for the
concentrations of carboxyhemoglobin and methemoglobin present in the blood
(neither of which contributes to transport of oxygen), CMS cases still had higher
meanCa02 values than normocythemic individuals.
Carboxyhemoglobin / Hemoglobin Ratio. At sea level, COhgb / hgb ratios for
nonsmokers average around 1% while smokers have COhgb / hgb ratios of 2 15%,
averaging about 5% for moderate smokers (134). COhgb / hgb ratios in this
population were not normally distributed; the median COhgb value for the 253 study
participants was 1.6%; normocythemic men and women with CMS had higher
median ratios, generally reflecting higher proportions of current smokers. There was
considerable variation in COhgb / hgb ratios, based primarily on smoking status, as
shown in Table 3.15. Twenty-one people, or 8.3% of the sample, had CO levels that
would be considered toxic at sea level.

Figure 3.22: Regression of Hemoglobin to Hematocrit
for All Study Participants, Male and Female,
Normocythemic and CMS Cases.
**--- Normocythemic
Individuals Hemoglobin (g/dl)
--------- Individuals
with CMS
Figure 3.22: Regression of Hemoglobin Hematocrit for all study
participants. For normocythemic men and women, indicated by x's, the
ratio of hematocrit / hemoglobin = 3.0, with a standard error of 0.01
(y = 2.142x + 13.280, r2 = .678, p < .0001). Among male and female
CMS cases, hematocrit / hemoglobin = 3.14 with a standard error of 0.03
(y = 2.317x + 15.143, r2 = .550, p < .0001), suggesting lower
hemoglobin mass / red blood cell than among normocythemic individuals.

Table 3.15. Distribution of COhgb/hgb Ratios among All Participants, by Smoking
and CMS Status
COhgb/hgb Ratio: 0-5 5-15 15-20 20-25 25-35 35-45 45+
CMS Smoker n = 9 1 0 1 2 5 0 0
Normocythemic Smoker n = 43 8 14 4 3 9 2 3
CMS Non- Smoker n = 33 27 5 1 0 0 0 0
Normocythemic Non-Smoker n = 169 154 11 1 1 1 1 0
Inteipretation of COhgb/hgb ratios: Normal Low High Border -line Toxic Low High
Erythropoietin (epo). Mean epo values obtained for both men and women in
nominal good health were in the normal range (10-20 mU/ml) provided by Ramco
Laboratories, which completed the assay. Since epo values did not differ between
men and women in good health, values for both sexes were combined in Table 3.16.
Compared with normocythemic individuals, CMS cases had higher epo
concentrations, but 50% of the CMS cases were in the normal range.
Table 3.16. Erythropoietin Concentrations in mU/ml for All Study Participants
Erythropoietin mU/ml epo <10 epo 10-20 epo 20-50 epo 50- 100 epo 100
Range: Below Normal Normal Range Above Normal Very High Extremely High
Normocythemic Men and Women n = 242 5 (2.1%) 187 (77.3%) 49 (20.2%) 1 (0.4%) 0 (0.0%)
CMS Cases, Men and Women n = 46 1 (2.2%) (47=2%) 11 (23.9%) 9 (19.6%) 3 (6.5%)
Total n = 288 6 (2.1%) 209 (72.6%) 60 (20.8%) 10 (3.4%) 3 (1.0%)

In Figure 3.23, it is apparent that log epo levels are not related to COhgb / hgb
ratios, even though oxygen delivery to the tissues is subtantially reduced by high
levels of COhgb. Despite the fact that log epo values did not respond to increased
levels of COhgb / hgb, hematocrits were still higher among individuals with higher
COhgb/hgb ratios (Figure 3.24), suggesting the existance of some mechanism for
stimulating erythropoiesis other than erythropoietin itself.
Ferritin. The distribution of ferritin results was not normal, so results were
log transformed, and log ferritin values were used in all statistical comparisons.
Ramco Laboratories, which completed the ferritin assays, provided a normal range of
20 300 ng/ml. Normocythemic men had higher ferritin concentrations than CMS
cases (Table 3.17). Low ferritin levels in male CMS cases were most likely a result
of frequent phlebotomy among the 18 male CMS cases who had undergone frequent
phlebotomy treatments since their initial diagnoses, but there were also low ferritin
values among individuals not previously treated for polycythemia. There were no
differences in ferritin concentrations between normocythemic and CMS women.
Seventeen individuals, 12 men and 5 women or 5.9% of the study population, had
excessive iron stores; of these 4 men and 1 woman were polycythemic (10.8% of the
46 CMS cases).
Individuals with low iron-storage concentrations had markedly higher epo
levels than individuals with normal ferritin values. This relationship was clearest
among CMS cases, and in normocythemic women, but individuals with CMS tended
to have markedly higher epo responses to high or low iron stores than their
normocythemic counterparts. As shown in Figures 3.25 for men, and Figure 3.26
for women, 61% of the variation in epo was accounted for by variations in ferritin