close

Вход

Забыли?

вход по аккаунту

?

Composite estimates of physiological stress age and diabetes in American Samoans.

код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 133:1028–1034 (2007)
Composite Estimates of Physiological Stress, Age, and
Diabetes in American Samoans
Douglas E. Crews
Department of Anthropology, The Ohio State University, Columbus, OH
School of Public Health, The Ohio State University, Columbus, OH
KEY WORDS
aging; allostatic load; senescence
ABSTRACT
Composite estimates of physiological
stress such as allostatic load (AL) were developed to help
assess cumulative impacts of psychosocial and physical
stressors on the body. Physiological responses to stress
generally accelerate somatic wear-and-tear and chronic
degenerative conditions (CDCs). Following McEwen
(Neuropsychopharmacology 22 (1999) 108–124) and
others, primary physiological mediators of somatic stress
responses include glucocorticoids (cortisol), catecholamines (adrenaline and noradrenaline), and serum dihydroepiandosterone-sulfate (DHEA-S). Conversely, blood
pressure (BP), serum HDL and total cholesterol, glycated
hemoglobin (HbA1c), and waist/hip (w/h) ratio are modulated by such hormones, thereby acting as secondary
mediators of stress response. When these risk factors are
aggregated into a composite score, higher stress loads
are associated with increased risks for days of school/
work missed, functional losses, morbidity, and mortality
in US samples. To examine stress loads in American
Samoans, data on all 6 secondary mediators along with
estimates of body habitus (i.e. height, weight, circumferences, skinfolds) and physiology (i.e. fasting insulin,
LDLc, triglycerides, fasting glucose) were measured on
273 individuals residing on Tutuila Island in 1992. Four
combinations of these physiological factors were used to
determine composite estimates of stress. These were
then assessed by sex for associations with age and the
presence of diabetes. Composite estimates of stress load
were higher in Samoan women than men. Associations
with age tended to be low and negative in men, but positive in women, appearing to reflect cultural circumstances and population history. Stress load scores also were
higher among those with diabetes than those without
among both men and women. These results suggest that
composite estimates of stress may be useful for assessing
future risks of CDC’s and the senescent processes that
may underlie them in cross-cultural research. Am J Phys
Anthropol 133:1028–1034, 2007. V 2007 Wiley-Liss, Inc.
How human somatic heterogeneity structures reproduction, survival, health, life expectancy, and life span
within and across populations is of interest to all biomedical researchers. Examinations of biological and physiological variation often focus on normative data juxtaposed with population and individual genetic, morphological, functional and physiological differences (Damon
1975; Borkan et al., 1982; Borkan 1986; Crews et al.,
2004), age-related dysfunction (Johnson and Wolinsky,
1994), or development of chronic degenerative conditions
(CDCs) (Crews, 1994). Stress responses and their variability across individuals are long-standing foci for examining population heterogeneity in socio-behavioral factors, physiology, senescence, aging, and chronic diseases
(James et al., 1996; Bindon et al., 1997; James and
Brown 1997; Brown et al., 1998; Goodman and Leatherman 1999; Panter-Brick and Worthman 1999; Decker
2000; McDade 2001; Crews 2003; Ice et al., 2003). Given
their holistic perspectives, anthropologically trained
human biologists see stress responses and stressors as
arising from multiple domains—most commonly classified as genes, environment, and culture (Baker, 1984).
Variation in many factors may independently or jointly
induce stress responses. Nutrition, poverty, growth and
development, disease, senescence, social circumstances,
socio-cultural and interpersonal interactions, socio-economic status (SES), and global capitalism are frequently
examined by biological anthropologists (Goodman et al.,
1992; Bindon et al., 1997; Flinn and England 1997;
Goodman and Leatherman 1999; Decker 2000; Ellison
2001; McDade 2001; Crews 2003). These, along with physi-
ological alterations (e.g., vascular compliance, lung injuries) and biophysical losses (e.g., cell dysfunction and
death) commonly examined in biomedical settings produce both psychosocial and physiological stress. Variability in many of these aspects of ecological circumstances
is peculiar to humankind’s culturally patterned behaviors. Proposed methods for assessing life-long stress and
progressive functional loss have included age of dysfunction onset, normative aging, biological/functional age
estimates, activities of daily living (ADLs), somatotyping,
biological age profiles, frailty and other composite scores
(Sheldon et al., 1954; Katz et al., 1963; Damon 1975;
Borkan et al., 1982; Shock 1985; Borkan 1986; Arking
2006). Most such scales are thought to rate an individual’s decline from some submaximal, but optimum, average functional level. Commonly, these are scaled to mean
function within a sample at a specified age. Inherent correlations with age make most such assessments about as
accurate at predicting disability, CDCs, and death as one
would be using age alone (Borkan et al., 1982; Borkan
1986). Here, I use a method commonly described as allo-
C 2007
V
WILEY-LISS, INC.
C
Correspondence to: Douglas. E. Crews, Department of Anthropology, The Ohio State University, 124 West 17th Avenue, 244 Lord
Hall, Columbus OH 43210-1364, USA. E-mail: crews.8@osu.edu
Received 13 July 2006; accepted 22 February 2007
DOI 10.1002/ajpa.20612
Published online 9 May 2007 in Wiley InterScience
(www.interscience.wiley.com).
COMPOSITE ESTIMATES OF STRESS IN AMERICAN SAMOANS
static load (AL) to estimate associations of physiological
components of stress response with age and diabetes in
American Samoans. This is accomplished by combining a
series of known risk factors for chronic degenerative conditions (CDCs) into an estimate of stress load.
THEORETICAL BACKGROUND
Pragmatic assessment scales used to determine nursing and residential care needs offer some of the best
assessments of physiological dysfunction currently available. Abilities to complete Activities of Daily Living
(ADLs) (Katz et al., 1963) and Instrumental Activities of
Daily Living (IADLs) (Johnson and Wolinsky, 1994) are
predictive of hospitalization, mortality, and remaining
life span in both the aged and the disabled. Indices of
frailty that include aspects of ADLs/IADLs, along with
biomedical assessments such as gait and balance, also
are useful for assessing individual capabilities and predicting undesirable outcomes (Johnson and Wolinsky,
1994). Allostatic load (AL) is a new addition to such composite measures of stress (McEwen and Stellar, 1993;
Seeman et al., 1997; Stewart, 2006). It was designed to
investigate the cumulative physiological impact of physical and psychosocial stress on hospitalization, morbidity,
and mortality from CDCs (Sterling and Eyer, 1988; Seeman et al., 1997, 2001; McEwen, 1999, 2000; Schulken,
2003; Stewart, 2006).
Allostasis is the dynamic interplay of bodily systems
in response to a changing environment, reflecting an
organism’s continual altering of its physiological milieu
to maintain its internal stability (Sterling and Eyer,
1988). All body systems constantly respond to stressors.
Constant response to changing stressors by the autonomic nervous system, the HPA axis, and pulmonary,
cardiovascular, and immune systems leads to continual
variation in adaptive physiological milieus (Sterling and
Eyer, 1988; Schulken, 2003). AL is a conceptual contrast
to the static/constant set points denoted by homeostasis
(Berntson and Cacioppo, 2000). Allostasis is the process
of reacting to environmental, physical, and social
changes/stressors by constantly altering physiological
states. Implicit in this model is a broad definition of
stressors and stress as both external and internal processes that require physiological defense responses. Allostatic responses are closely compatible with anthropological views of individual adaptability to environmental
changes requiring constant physiological adjustments to
maintain the soma. Physiological responses to constant
internal and external stressors produce somatic strain—
an allostatic load (AL), ‘‘. . . the cumulative somatic strain
created partly by elevated activity of physiological systems attempting to remain at optimum function, partly
through wear-and-tear these cause, and partly because
not all ill effects may be halted . . .’’ (McEwen and Stellar,
1993; see also McEwen, 1999; Schulken, 2003). In many
ways AL recalls somatic wear-and-tear as originally
proposed by Weismann (1881–1889) (1993), and recently
revisited by Masoro (1996) as a model of somatic senescence. AL also conforms to current models of senescence
poisting an immortal germline carried through time by
a series of disposable somas (Kirkwood and Austad,
2000).
As originally presented by McEwen (1999, 2000), McEwen and Stellar (1993), and Seeman et al., (1997, 2001,
2004), AL may be estimated using 10 physiological measures described as primary and secondary mediators of
1029
TABLE 1. The components of allostatic load
after McEwen (1999)
Secondary mediators of stress
Systolic and diastolic blood pressure
Waist/hip ratio
HDL-cholesterol and total-cholesterol
Glycated hemoglobin
Primary mediators of stress
Serum dihydroepiandosterone sulfate
Overnight urinary cortisol, adrenaline, noradrenaline
stress (Table 1). Their primary modulators included 4
physiologically active molecules secreted in response to
stress. Secondary modulators were 6 physiological phenotypes representing unwanted responses to the primary
mediators that are associated with increased risks for
CDCs (Table 1) Since its conceptualization and first
report (McEwen and Stellar, 1993), numerous physiological measures have been used to estimate AL (Stewart,
2006). These include serotonin, fibrinogen, C-reactive
protein (CRP), creatanine, albumin, and thrombin
(Stewart, 2006). In its original application to the MacArthur Foundation’s Studies of Successful Aging sample of
1,189 persons aged 70–79 in 1997 (Seeman et al., 2004),
AL at baseline was correlated with future mortality
risks, cognitive and physical functioning (Seeman et al.,
2001), and socio-economic differences (Seeman et al.,
2004). Later, AL was found to predict days of school
missed in children (Johnston-Brooks et al., 1998), development of addictive behaviors in adults (Schulken,
2003), and memory declines in elders (Seeman et al.,
1997). In fact, in the MacAuthur cohort, AL was a stronger predictor of future senescent decline and mortality
than was Syndrome X (Seeman et al., 2001). AL also
may be the physiological cause of mortality usually
attributed to psychosocial mechanisms.
The specific assessments used by McEwen (1999,
2000) and Seeman et al., (1997, 2001, 2004) may not be
the most important for determining stress loads across
different populations (McEwen and Stellar, 1993; McEwen, 1999, 2000; Schulken, 2003; Stewart, 2006). Additional physiological measures, genetic markers, and
social and economic factors likely will be needed to fully
specify composite stress indexes (McEwen, 1999, 2000),
and these likely will vary across socio-cultural settings.
Human biologists commonly included all secondary
measures proposed by McEwen, Seeman and colleagues
for assessing AL in cross cultural studies of health,
CDC’s, growth, adaptability and physiological variation.
Often primary modulators of AL or reasonable proxies
also are included. Multiple additional likely primary
determinants (e.g., CRP, immunoglobins, insulin) and
possible secondary modulators (e.g., skinfolds, body circumferences, blood chemistries) of allostasis also are often available. Here, distributions of composite assessments of stress and their possible associations with sex,
age, and diabetes are examined in a Samoan sample to
determine their usefulness in a non-Western population.
Four combinations of physiological measures are used to
estimate composite stress loads in this sample of 273
American Samoans (125 men, 148 women). All 6 secondary mediators suggested by McEwen (1999, 2000), along
with measurements of fasting insulin, triglycerides, fasting glucose, low-density lipoprotein cholesterol (LDLc),
height, weight, body mass index (BMI), relative fat pat-
American Journal of Physical Anthropology—DOI 10.1002/ajpa
1030
D.E. CREWS
TABLE 2. Means, standard deviations, ranges, and P-values comparing means between men and women and quartile cut-points for
risk factors used to compute stress load composites for American Samoansa
Quartile cut-points
Mean (s.d.)
Age
Systolic blood pressure
Diastolic blood pressure
Waist-hip ratio
Glycated hemoglobin
Total cholesterol
High density cholesterol
Low density cholesterol
Triglycerides
Fasting serum glucose
Fasting serum insulin
Body mass index
Relative fat pattern index
Triceps skinfolds
Subscapular skinfolds
a
55.5
141.2
85.1
91.0
10.0
193.1
28.4
138.8
148.8
158
167
28.8
0.50
35.9
50.6
(9.8)
(24.8)
(16.1)
(7.1)
(6.0)
(37.8)
(8.4)
(63.0)
(37.8)
(76)
(206)
(5.8)
(0.04)
(16)
(13)
Range
P
Men
Women
All
35–88
88–236
10–140
72–112
5.0–90.3
87–325
12–70
114–999
87–325
34–492
14–2495
14–51
0.31–0.72
6–60
8–69
0.385
0.593
0.028
0.000
0.455
0.451
0.006
0.017
0.000
0.078
0.972
0.048
0.310
0.000
0.000
63.7
154.0
96.0
99.0
10.2
212.0
20.6
154.0
200.0
211.0
169.0
31.2
0.53
32.2
57.0
62.1
156.0
92.0
91.5
10.6
222.0
24.6
163.0
137.0
179.0
214.0
33.1
0.50
55.3
60.0
62.6
155.0
93.0
96.0
10.6
218.0
22.7
159.0
175.0
188.0
189.0
32.0
0.50
49.2
60.0
Quartile cut-point for HDL-c is the upper bound of the 1st. P ¼ 0.000 is <0.0005; N ¼ 273 (men, 125, women, 148).
tern index, and triceps and subscapular skinfolds (SF)
(Table 2) are available for this sample. Fasting insulin
has recently been recognized as a possible primary modulator of calorie restriction, stress responses, and senescence in rodents and humans (Rincon et al., 2005; Stewart, 2006).
In the composite stress load model, individuals having
the same total score may have very different risk profiles. Additionally, an individual’s score changes over
their life course. Over a single day, an individual may
show a range of stress load scores. Because blood pressure and lipids/insulin/glucose are determined while
resting and fasting respectively, these measures do not
capture diurnal variability. Stress load composites are
not meant to capture all variability in dynamic physiological regulation or to suggest that all individuals are
equal in their responses to similar stressors (Berntson
and Cacioppo, 2000). Allodynamic regulation of the physiological milieu fluctuates with patterns of primary
responses, compensatory alternations by secondary systems as primaries fail, and systemic losses as responses
to stressors continue over time (Berntson and Cacioppo,
2000; Stewart, 2006).
METHODS AND SAMPLES
Sample
American Samoa is a group of five volcanic islands
located in the South Pacific *15 degrees below the
Equator and just east of the International Dateline. The
origin of the first inhabitants of Samoa and other Polynesian islands has long been a topic of debate. The most
widely held theory asserts that Austronesian-speaking
people from S.E. Asia reached Samoa and Tonga around
3,300 BPE carrying with them a culture known as Lapita, characterized by red slip pottery with dentate design
(Bellwood, 1989). Since 1976, American Samoans have
taken part in what is known as the Samoans Study Project (Baker et al., 1986). In 1992, a research team from
The Ohio State University and University of AlabamaTuscaloosa collected data from 273 American Samoans
(125 men, 148 women) on the island of Tutuila (Bindon
et al., 1997; Crews et al., 2004). Ages of these 273 participants ranged from 35 to 88 years. Approximately half
the sample was randomly selected, while the other half
was part of the longitudinal study begun in 1976. Additional aspects of data collection and sampling strategies
are detailed elsewhere (Bindon et al., 1997; Crews et al.,
2004). Participants came from a range of villages and
towns and were recruited using the 1976 registry and on
an opportunistic basis. Measures of SES, anthropometrics, and socio-cultural variability in this sample tend to
match those reported for the general population (Bindon
et al., 1997).
Data
Phenotypic and genotypic variation among American
Samoans has been widely reported (Crews, 1988, 1989;
Crews et al., 1991, 1993, 2004; Bindon et al., 1993, 1997;
Barley et al., 1994; Crews and Harper, 1998). All participants in our 1992 research protocols were assessed for
age, systolic (SBP) and diastolic blood pressure (DBP),
waist and hip circumferences, height, weight, fasting
glucose, triceps, suprailiac and sub-scapular skinfolds
(SF), total-cholesterol (total-c), triglycerides, glycated hemoglobin (HbAl-c), high density lipoprotein cholesterol
(HDLc), LDLc, and fasting insulin and glucose by our
research team (Table 2) (Bindon et al., 1997; Crews
et al., 2004). Body mass index (BMI ¼ weight (kg)/height
(m2), the relative fat pattern index (RFPI ¼ subscapular
SF/subscapular SF + suprailiac SF), and waist/hip ratio
(w/h) were calculated from these data. Systolic (SBP), diastolic (DBP) blood pressure, total-c HDL-c, glycated hemoglobin, serum glucose, and w/h represent secondary
modulators of stress as defined by McEwen (1999). Additionally, triglycerides, LDLc, fasting glucose, RFPI, triceps and subscapular SFs are plausible secondary modulators, and several have been used as such in estimates
of AL (reviewed by Stewart (2006)). Fasting insulin represents a possible primary modulator of stress responses.
In addition to these data, information on sex, age, and
type II diabetes were available. Diabetes was assessed
by a fasting or post-load glucose level either at or above
140 mg/dl. All tests were performed by the research
team using a Glucometer31 tested for reliability on a
daily basis. Further details of measurement protocols
are presented in Fitton and Crews (1994). Age was
determined from birth certificates and/or driver’s
licenses; sex was recorded during the interview. All risk
American Journal of Physical Anthropology—DOI 10.1002/ajpa
1031
COMPOSITE ESTIMATES OF STRESS IN AMERICAN SAMOANS
TABLE 3. Composite estimates of stress load in 273 American Samoans based upon a selection of physiological factors from Table 2
Total
Men (125)
Women (148)
Model
Mean
s.d.
Range
Mean
s.d.
Range
Mean
s.d.
Range
1
2
3
4
1.66
1.19
3.19
2.33
1.38
1.13
2.19
1.90
0–6
0–5
0–9
0–8
1.70
1.22
2.63
2.37
1.38
1.11
1.95
1.87
0–6
0–4
0–7
0–8
1.74
1.24
3.56
2.39
1.39
1.07
2.24
1.98
0–5
0–4
0–9
0–8
TABLE 4. Dependence of physiological stress load composites on age in American Samoans
Total
Men (125)
Women (148)
Model
b
P
R
b
P
R
b
P
R
1
2
3
4
0.020
0.048
0.144
0.002
0.741
0.438
0.020
0.976
0.001
0.002
0.017
0.000
0.115
0.099
0.190
0.147
0.216
0.286
0.041
0.113
0.005
0.001
0.028
0.022
0.141
0.181
0.046
0.124
0.092
0.030
0.588
0.140
0.013
0.026
0.005
0.008
factors and the presence of diabetes were assessed during the same 24-h period.
A composite model of psycho-physiological and physical stress is examined here. Following McEwen (2000),
stress scores for individuals are calculated by scoring all
measures as 1 for the highest quartile of risk and 0 for
the 3 lower risk quartiles. These are then summed
across all measures resulting in a score for each individual that theoretically ranges from zero to the number of
variables examined. In general, the highest quartile is
scored 1, because for all measures, except HDLc, highest
risks occur therein. For HDLc the lowest quartile is
scored 1.
Given the availability of multiple physiological measures for this Samoan sample, a variety of semihierarchical composite estimates of stress loads are examined.
Rationales for these estimates include: Model 1—all secondary modulators proposed by McEwen and Stellar
(1993), along with fasting insulin as a primary mediator;
Model 2—drops both glycated hemoglobin (HbAlc) and
fasting insulin to examine associations of the remaining
composite score with age and diabetes; Model 3—combines Model 1 with additional measures of body habitus:
triceps and subscapular skinfolds, BMI, and RFPI to
examine relationships of stress to obesity and body form
(habitus); Model 4—combines Model 1 with additional
lipids (LDLc and triglycerides) and fasting glucose to
examine lipids/glucose as elevators of physiological
stress. After being estimated, composite stress load
scores are examined for associations with chronological
age and Type II diabetes. Analysis of age and diabetes
are completed for the total sample and for men and
women separately. Based upon observed statistically significant differences between men and women in means
and quartile cut-points for the risk factors examined,
stress loads within each sex are determined using sexspecific cut-points (Table 2).
variable representing diabetes. The suggestion from this
methodology is that diabetes is the independent factor
determining current stress load; however physiologically
diabetes may just as well be the dependent variable that
is explained by variable stress loads. Given the cross-sectional data available for analysis, either interpretation is
valid. The former interpretation is used to structure the
discussion; that is stress load scores are viewed as being
predicted by diabetes. An alternative analytic strategy
would be to estimate odds ratios for diabetes based upon
variability in individual stress loads using logistic
regression. However, given the available data and the
nature of the composite method used (simple addition of
nonstandardized scores without weighting), simple linear
regression is appropriate and easier to interpret. All statistical analyses used SPSSPC. Results are presented as
beta-coefficients with associated P-values and the variation explained by age (R). For diabetes, results are presented as average stress scores in those with and without, along with associated P- and R-values.
RESULTS
Estimates of stress load
For all risk factors, this sample is statistically similar
to results from other American Samoan samples (Table
2, Bindon et al., 1997). Average age is 55.5 years (Table
2). Data presented in Table 2 provide the basis for calculating stress load composites (Table 3 ). Men show significantly higher DBP, w/h, and triglycerides than women,
while women show significantly higher HDLc, LDLc,
BMI, and triceps and subscapular skinfolds than men.
Men and women tend to show similar average stress
scores. The sole exception is Model 3, which includes all
body habitus measures. Here women show a higher average score. For all later analyses, sex-specific stress
load estimates are examined. Differences by sex are not
subject to statistical tests.
Statistical methods
Stress scores for men and women are presented to
illustrate differences by sex. However, since these estimates are based upon quartiles observed within each sex
independently, differences are not submitted to a statistical test. Regression is used to estimate the dependence
of stress loads on chronological age. Finally, differences
in stress load among those with and without type II diabetes are compared using regression on a categorical
Age and stress load
Estimates of stress load are poorly associated with age
when the total sample is examined (Table 4). Among
men, stress load scores show negative and generally nonsignificant associations with age. Only Model 3 is statistically significant, but therein age explains only about
3% of the total variation. Among women, most estimates
of stress load are positively associated with age. Age
American Journal of Physical Anthropology—DOI 10.1002/ajpa
1032
D.E. CREWS
TABLE 5. Differences in physiological stress load composites among 273 American Samoans with and without (W/O)
Type II diabetes
Total
Men (125)
Women (148)
Model
With
W/O
P
R
With
W/O
P
R
With
W/O
P
R
1
2
3
4
2.00
1.30
3.71
3.01
1.46
1.18
2.96
1.87
0.001
0.380
0.005
0.0001
0.034
0.000
0.025
0.084
1.93
1.26
3.15
2.91
1.62
1.28
2.36
2.08
0.214
0.913
0.028
0.014
0.005
0.008
0.032
0.043
2.15
1.51
4.00
3.22
1.31
1.00
3.24
1.62
0.0002
0.0038
0.0376
0.0000
0.087
0.050
0.023
0.158
explains almost 3% of the total variance in stress load
for Model 2 without HbAlc or insulin included. Age also
shows a borderline association with Model 1, but
explains only 1% of the total variance.
Diabetes and stress load
Associations of stress load estimates with the presence
or absence of type II diabetes were examined by comparing subgroups with and without diabetes (Table 5). In
the total sample and among both men and women, stress
load estimates generally were higher among those with,
than those without type II diabetes. In the total sample,
all stress load estimates, except Model 2 without fasting
insulin and glucose included, showed significant associations with having diabetes. This consistent association
was likely due to the combination of both sexes and
using the sample total distribution to determine quartile
cut-points. Among men only 2 models were significantly
associated with diabetes, while among women all were
(Table 5). Type II diabetes explains about 3–4% of the
total variation in stress estimates among men when either all body habitus or all lipid/glucose measures are
included. However, among women type II diabetes
explains almost 9% of the total variance in Model 1, and
16% in Model 4 with all lipid and metabolic variables
included. Interestingly, only 2% of total variance in
Model 3, with all body habitus variables included, is
explained by diabetes. Among women, the total variance
in Model 2 explained by the presence or absence of diabetes was about 5% even though neither insulin nor glucose was included in the estimate (Table 5). Model 2
showed almost no association with diabetes among men.
DISCUSSION
Regardless of the method used for measurement,
stress load composites tend to show predictive value for
diabetes morbidity among women and to some degree
among men. In addition, Samoans show a broad range,
from zero to about the number of variables minus 1
included in a model, for stress load estimates (Table 3).
With only 7 mediators included, the average was 1.66,
but ranged from 0 to 6. With 9 mediators it rose to 3.19
and ranged from 0 to 9. Stress load composites tend to
be negatively associated with age in men, but both positively and negatively associated in women (Table 4).
Conversely, stress scores are associated with type II diabetes in both sexes (Tables 5), and remain so for women
even without fasting insulin, glucose, and HbAlc
included in the composite estimate. Because both diabetes and all associated risk factors were assessed at the
same time, temporality of causal order cannot be determined from these analyses. Causality may flow either
way. Allostatic load represents a variety of disruptions in
homeostatic functions that may themselves aggravate or
lead to CDCs; on the other hand, CDCs, particularly diabetes, disrupt allostasis and may lead to elevations in
stress load composites. It is likely that what we are
observing is a self-perpetuating system of mutual aggravation.
As measured here, stress load composites show variable and significant associations with diabetes. Differences between the sexes are consistent, and stress loads
show stronger associations with diabetes in women. Age
is negatively associated with stress load estimates in Samoan men (Table 4), and explains little variance (0.1%
to 5.9%). Among women, these associations are more
positive (Table 4). When the score is based upon fewer
secondary mediators, age is a significant predictor of
stress load composites among women. Among men, only
the score based upon lipids is significantly associated
with age. This suggests that older Samoan men have
experienced lower cumulative physiological and physical
stress loads than younger, while younger women may
show less cumulative physiological stress than older
women do. Contrary to their low associations with age,
stress load composites are closely associated with type II
diabetes. These associations are both independent of current glucose metabolism as measured here and stronger
in women. To some degree, differential results for women
and men may be influenced by gender-based differences
in acculturation and participation in nontraditional lifestyles that relate to population history and biocultural
interactions. These likely include early and continuous
participation in wage labor and more migration of men
to towns to participate in the market economy, and a
greater disruption of women’s traditional ways of life, by
piped water, kerosene stoves, and nontraditional commodities among other factors (Baker et al., 1986; Crews,
1988, 1989; Bindon and Crews, 1993; Bindon et al.,
1997).
With a sample of 273 participants, the analyses
reported here have sufficient statistical power for estimating stress load composites and their variation by sex,
age, and diabetes. When analyses are extended to compare those with to those without diabetes by sex, sample
sizes approach their minimums. Samoans show wide
variation for all measures used in this analysis (Table 2).
This, in addition to the large sample, decreases the likelihood that any one individual will fall into the high-risk
quartile for all measures. These cross-sectional data suggest that stress load may not change greatly with age in
Samoans, perhaps indicating that all middle-aged Samoans share population-wide elevated risks for CDCs. Well
integrated into cosmopolitan ways of life, the American
Samoan population tends to share physiological and
socio-cultural risks for CDCs common to the U.S. population. The Samoans however experienced these cultural
and environmental alterations relatively recently, and
risk profiles continue to increase among middle-aged
individuals, while being lower in the oldest segments
American Journal of Physical Anthropology—DOI 10.1002/ajpa
COMPOSITE ESTIMATES OF STRESS IN AMERICAN SAMOANS
(Bindon et al., 1997). Modernization and cosmopolitan
lifestyles came late but fast to Samoa. During World
War II, the Samoan Islands went from an out-of-the-way
South Pacific paradise to the center of allied naval forces
and logistics for the Pacific war. Culture change and
physiological responses were rapid in American Samoa
during and following the war (Baker et al., 1986). Major
changes in the economy and opportunities for wage labor
greatly altered traditional subsistence and socio-cultural
patterns. Only later, during the 1960s and 70s, as more
Samoans participated in formal schooling and nonlabor
occupations, did patterns of physical activity decline
greatly and dependence on nonlocal resources increase.
At this time Samoan body sizes and proportions also
began increasing precipitously. The oldest Samoans in
our sample are those who grew and developed before
these massive changes, the middle-aged experienced
many of these changes, and the youngest are physically
the largest Samoans ever.
Observed associations with type II diabetes support suggestions that composite assessments of stress load provide
a useful quantitative measure of stress and a predictive
tool for assessing risks for CDCs, and, by extension, senescent alterations cross culturally. Associations between
stress and diabetes might be expected in this analysis
since many of the risk factors used herein also are used to
define Syndrome X (the metabolic syndrome). It could even
be suggested that AL and Syndrome X are inseparable,
such that AL simply is an assessment of the latter (see
also Seeman et al., 2004). However, non-Syndrome X physiological variables have been incorporated into AL calculations in recent years, without reducing associations with
outcomes (Stewart 2006). If AL was simply a proxy for
Syndrome X, AL should be highly associated with age, due
to the latter’s strong association with age across multiple
populations. However, results presented here suggest that
stress loads respond to socio-cultural changes and may
even reflect biocultural interactions, but are not closely
associated with age in this sample of Samoans. The consistency and ease of replicating allostatic load assessments in
U.S. samples also may in part reflect similar widely shared
cultural attributes. These results suggest that multiple
other CDCs associated with cardiovascular, pulmonary,
and hormonal function may be usefully examined using
this stress load model.
Estimates of stress load reported here were based
mainly on secondary modulators of stress responses,
except for insulin. They provide data to formulate
hypotheses for future studies of stress in Samoans. For
one, risk factor composites appear to be a useful tool for
aggregating aspects of physiological variation into a single score across cultures. Addition of socio-cultural factors (e.g., SES, lifestyle) may improve the accuracy of
such composites. Similarly, given genetic influences on
multiple aspects of glucose metabolism, autonomic, HPA,
and immune response, addition of genotypes known to
affect lipid, glucose, and hormone metabolism also
should improve measurement of the cumulative strain of
physiological systems attempting to remain at optimum
function. Associations with age suggest composite stress
load estimates may provide a useful assessment scale for
some aspects of senescence. All current evolutionary theories of senescence including antagonistic pleiotropy,
age-specific gene action, and thrifty/pleiotropic alleles
(Williams, 1957; Crews, 2003) are compatible with the
model of allostatic load. In general, populations integrating into cosmopolitan life tend to show increases in both
1033
primary and secondary mediators of stress response
(McGarvey et al., 1994; Jenner et al., 1987; James et al.,
1996; Flinn and England, 1997; James and Brown, 1997;
Panter-Brick and Worthman, 1999). Differential associations by sex, suggest that gender-based socio-cultural
variability influences physiological function, somatic variability, and stress responses in Samoans. Exactly which
socio-cultural factors may account for such variance is
unclear, although obesity, hypertension, and diabetes
mortality were more common among more traditionalliving Samoan women in the 1970–1980 samples (Baker
et al., 1986; Crews, 1989; McGarvey et al., 1994).
American Samoans are ever more integrated into
global markets and cosmopolitan ways of life (Bindon
et al., 1997). The Samoan sample examined here represents
the range of physiological variation and socio-economic
levels available on Tutuila, American Samoa (Bindon
et al., 1997). They represent a socio-cultural system and
lifestyles different from those of the general U.S. population. Stress load composites based upon the AL model
appear useful for assessing differential risks for disease
in this population. These results also suggest caution.
Observed physiological variation may have underlying
biocultural influences. Stress load estimates may not be
interpretable without knowledge of population history,
including recent ecological, environmental and cultural
changes. Research in human biology has frequently been
invigorated by incorporating methods from epidemiology
and ecology into studies of human stress while viewing
adaptability from a life history perspective (Bindon
et al., 1997; Ulijaszek and Hush-Ashmore, 1997; Crews,
1998, 1999; Panter-Brick and Worthman, 1999; Ellison,
2001; Ben-Shlomo and Kuh, 2002). Methods for assessing allostatic or stress load using composites of physiological variables proposed for use in epidemiologic
research should fit well within biocultural models examining life-long outcomes of stress responses. Thus, this
methodology should be useful across multiple populations and CDCs that are associated with a variety of
responses to physiological stress and stressful life events
that alter allostatic physiological responses.
ACKNOWLEDGMENTS
Many people contributed to obtaining data for the
sample reported here. I first acknowledge the time,
patience, and efforts of those Samoans who participated
in our research protocols on Tutuila, American Samoa.
Acknowledgments are due to my colleagues who aided in
data collection, James Bindon and Lori Fitton. Acknowledgments are also due to the American Samoan Public
Health nurses, particularly Tele Hill, for aiding our fieldwork, and Charles ‘‘Mic’’ Micudden. I also thank my
office assistants, Nora Smith, Rita Grefer, Tacarra Christie, and Jocyline Wantsala, for typing multiple drafts of
this manuscript.
LITERATURE CITED
Arking R. 2006. Biology of aging: observations and principles, 3rd ed.
New York: Oxford University Press.
Baker PI. 1984. The adaptive limits of human populations. Man 19:1–
14.
Baker PT, Hanna MJ, Baker TS. 1986. The changing Samoans:
behavior and health in transition. New York: Oxford University
Press.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
1034
D.E. CREWS
Barley J, Blackwood A, Carter N, Crews DE, Cruickshank JK, Jeffrey
S, Ogunlesi AG, Sahnelta GA. 1994. Angiotensin converting
enzyme insertion/deletion polymorphism-association with ethnic
origin. J Hypertens 12:955–957.
Bellwood PS. 1989. The colonization of the Pacific: some current
hypotheses. In: Hill AVS, Serjeantson SW, editors. The colonization
of the Pacific: a genetic trail. Oxford: Clarendon. p 1–59.
Ben-Shlomo Y, Kuh D. 2002. A life course approach to chronic disease
epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives. Int J Epidemiol 31:285–293.
Berntson GG, Cacioppo JT. 2000. From homeostasis to allodynamic
regulation. In: Cacioppo JT, Tassinary LG, Berntson GG, editors.
Handbook of psychophysiology, 2nd ed. New York: Cambridge University Press. p 459–481.
Bindon JR, Crews DE. 1993. Changes in some health, status characteristics of American Samoan men: a 12-year follow-up study. Am J
Hum Biol 5:31–38.
Bindon JR, Knight A, Dressler WW, Crews DE. 1997. Social context
and psychosocial influences on blood pressure among American
Samoans. Am J Hum Biol 103:7–18.
Borkan GA. 1986. Biological age assessment in adulthood. In: Bittles
AH, Collins KJ, editors. The biology of human aging. New York:
Cambridge University Press. p 81–94.
Borkan GA, Hults DE, Mayer PJ. 1982. Physical anthropological
approaches to aging. Yearbook Phys Anthropol 25:181–202.
Brown D, James GD, Nordioh L. 1998. Comparison of factors affecting
daily variation of blood pressure in Filipino-American and Caucasian nurses in Hawaii. Am J Hum Biol 106:373–383.
Crews DE. 1988. Multiple causes of death and the epidemiological
transition in American Samoa. Social Biol 35:198–213.
Crews DE. 1989. Multivariate prediction of total and cardiovascular
mortality in an obese Polynesian population. Am J Public Health
79:982–986.
Crews DE 2003. Human senescence: evolutionary and biocultural
perspectives. Cambridge: Cambridge University Press.
Crews DE Harper GJ. 1998. Renin, ANP, and ACE polymorphisms,
blood pressure and age in American Samoans. Am J Hum Biol
10:439–449.
Crews DE, Bindon JR, Kamboh MI. 1993. Apolipoprotein polymorphisms and phenotypic variability in American Samoans: preliminary data. Am J Hum Biol 5:39–48.
Crews DE, Fitton LJ, Kamboh MI, Kottke B. 2004. Population genetics of apolipoproteins A-IV, E, and the angiotensin converting
enzyme (ACE): associations with lipids and apolipoproteins. Am J
Phys Anthropol 124:364–372.
Damon A, editor. 1975. Physiological anthropology. New York: Oxford
University Press.
Decker S. 2000. Salivary cortisol and social status among Dominican
men. Hormones Behav 38:29–38.
Ellison PT, editor. 2001. Reproductive ecology and human evolution.
New York: Aldine de Gruyter.
Fitton LJ, Crews DE. 1994. Glucose measurement in field situations:
portable meters and laboratory analyzers. Coll Anthropol 18:45–
51.
Flinn M, England B. 1997. Social economics of childhood glucocorticoid stress response and health. Am J Phys Anthropol 102:33–
53.
Goodman A, Leatherman T. 1999. Building a biocultural synthesis.
Ann Arbor, MI: University of Michigan Press.
Goodman AH, Thomas RB, Swedlund AC Armelagos GL. 1992. Biocultural perspectives on stress in prehistoric, historical, and contemporary population research. Yearbook Phys Anthropol 35:169–
202.
Ice GH, James GD, Crews DE. 2003. Blood pressure variation in the
institutionalized elderly. Coll Anthropolo 27:47–55.
James GD, Brown D. 1997. The biological stress response and life
style: catecholamines and blood pressure. Am J Hum Biol 26:313–
335.
James GD, Broege P, Schlussei Y. 1996. Assessing cardiovascular risk
and stress-related blood pressure variability in young women
employed in wage jobs. Am J Hum Biol 8:743–749.
Jenner D, Harrison G, Prior I. 1987. Catechlomine excretion in Tokelauans living in three different communities. Hum Biol 59:165–
172.
Johnston-Brooks CH, Lewis MA, Evans GW, Whalen CK. 1998.
Chronic stress and illness in children: the role of allostatic load. Psychosom Med 60:597–603.
Johnson RJ Wolinsky FD. 1994. Gender, race and health: the
structure of health status among older adults. Gerontologist 34:
24–35.
Katz SA, Ford AB, Moskowitz RW, Jackson BA Jaffe MW. 1963. Studies of illness in the aged. The index of ADL: a standardized measure
of biological and psyhosocial function. JAMA 185:914–919.
Kirkwood TB, Austad L. 2000. Molecular gerontology—bridging the
simple and the complex. Mol Cell Gerontol 908:14–20.
Kloting N, Bluher, M. 2005. Extended longevity and insulin signaling
in adipose tissue. Exp Gerontol 40:878–883.
Masoro EJ. 1996. The biological mechanism of aging: is it still an
enigma? Age 19:141–145.
McDade T. 2001. Lifestyle incongruity, social integration, and
immune function in Samoan adolescents. Soc Sci Med 53:1351–
1362.
McEwen BS. 1999. Allostasis and allostatic load: implications for
neuropsychopharmacology. Neuropsychopharmacology 22:108–
124.
McEwen BS. 2000. Protective and damaging effects of stress mediators: central role of the brain. In: Mayer EA, Saper CB, editors. Progress in brain research, vol. 122: The biological basis for mind body
interactions . New York. Elsevier Science BV. p 25–34.
McEwen BS, Stellar E. 1993. Stress and the individual: mechanisms
leading to disease. Arch Intern Med 153:2093–2101.
McGarvey ST, Bindon JR, Crews DE, Schendel DE. 1989. Modernization and adiposity: causes and consequences. In: Little MA, Haas
JD, editors. Human population biology: a transdisciplinary science.
New York: Oxford University Press. p 263–279.
Panter-Brick C, Worthman C. 1999. Hormones, health, and behavior:
a socio-ecological and life span perspective. Cambridge: Cambridge
University Press.
Rincon M, Rudin E, Barzilai N. 2005. The insulin/IGF-1 signaling in
mammals and its relevance to human longevity. Exp Gerontol
40:873–877.
Schulken J. 2003. Rethinking homeostasis: allostatic regulation and
pathophysiology. Cambridge, MA: MIT Press.
Seeman TE, Singer BH, Rowe JW, Horowitz RI McEwen BS. 1997.
The price of adaptation—allostatic load and its consequences. Arch
Intern Med 57:2259–2268.
Seeman TE, McEwen BS, Rowe JW, Singer BH. 2001. Allostatic load
as a marker of cumulative biological risk: MacArthur studies of successful aging. Proc Natl Acad Sci USA 98:4770–4775.
Seeman TE, Crimmins E, Hvang MH, Singer B, Bacur A, Gruenewald T, Berkman LF, Reuben DB. 2004. Cumulative biological risk
and socio-economic differences in mortality: MacArthur studies of
successful aging. Soc Sci Med 58:1985–1997.
Sheldon WH, Dupertuis CW, McDermott E. 1954. Atlas of men: a
guide for somatotyping the adult male at all ages. New York:
Harper & Brothers.
Shock NW. 1985. Longitudinal studies of aging in humans. In: Finch
CE, Schneider EL, editors. Handbook of the biology of aging, 2nd
ed. New York: Van Nostrans Reinhold. p 721–743.
Sterling P, Eyer J. 1988. Allostasis: a new paradigm to explain arousal
pathology. In: Fisher S, Reason J, editors. Handbook of life stress,
cognition, and health. New York: John Wiley. p 629–649.
Stewart JA. 2006. The detrimental effects of allostasis: allostatic load
as a measure of cumulative stress. J Physiol Anthropol 25:133–145.
Ulijaszek SJ, Huss-Ashmore R. 1997. Human adaptability: past, present, and future. New York: Oxford University Press.
Williams GC. 1957. Pleiotropy, natural selection and the evolution of
senescence. Evolution 11:398–411.
Weismann A. 1993. The duration of life (a paper presented in 1881).
In: Poulton EB, Schonland S, Shipley AE, editors. Essays upon heredity and kindred biological problems. Oxford: Clarendon Press.
p 1–66.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
Документ
Категория
Без категории
Просмотров
3
Размер файла
127 Кб
Теги
physiological, estimates, diabetes, american, age, compositum, stress, samoans
1/--страниц
Пожаловаться на содержимое документа