SOCIAL AND BEHAVIORAL STRESS

There are both short-term and long-term perspectives for studying the health impact of stress. In the short term, stress generally refers to events or circumstances that elicit an emotional (usually negative) response. Episodes of stress can affect the immune system and therefore susceptibility to disease (Segerstrom and Miller, 2004). Recently, social scientists have begun to study the cumulative effects of repeated episodes of a wider range of social-psychological, physical, and environmental stressors. This literature emphasizes links between stress and such chronic diseases as cardiovascular disease and diabetes.

We have a lot to learn about the sources of stress. For example, the Whitehall study quickly showed that the concept of “executive stress” was not consistent with the gradient they observed; the executives at the top of the hierarchy had the lowest risk of several major causes of death. This raises the question of whether we know real stress when we see it or whether we can identify sources of chronic stress.

Ironically, biomarkers can help us recognize stressful circumstances. They provide one type of evidence about which individuals in an organization or society are most stressed, what kinds of social situations are most stressful, what kinds of social organizations relieve stress, and how individuals differ in their ability to handle stress. Much of the early research on this topic relied on measures associated with cardiovascular activation, which generally reflects reactions to recent events (up to 24 hours). Few of these studies were population-based. For example, catecholamines (epinephrine and norepinephrine) excretion rates (James and Brown, 1997) were used to identify which individuals in Otmoor villages in Oxfordshire were most stressed (Harrison, 1981) and which occupations were most stressful (Jenner, Harrison, Day, Huizinga, and Salzano, 1982). However, as James and Brown have noted, these measures reflect “the complexity of the circumstance, including the subject's perceptions and cognitive state, the nature of the social situation, potential food, stimulant and exercise effects, and the ambient environmental conditions, such as temperature or altitude” (James and Brown, 1997, p. 329).

Recent research has emphasized biomarkers that reflect longer term damage to the body's regulatory systems (see below). For example, chronic elevation of C-reactive protein (CRP) is a good marker for systemic infection and is closely associated with coronary heart disease (Danesh et al., 2004). CRP levels have been used to examine the long-term effect of stress associated with socioeconomic status in Whitehall II (Owen, Poulton, Hay, Mohamed-Ali, and Steptoe, 2003), in Chicago (McDade, Hawkley, and Cacioppo, 2006), and in the United States as a whole (Alley et al., 2006).

Psychologists studying the effects of stress have proposed five categories of stress (Segerstrom and Miller, 2004). Most of their research has involved episodes of acute stress. Acute time-limited stressors involve artificial experiences produced in laboratory experiments, such as tasks involving mental arithmetic. Brief naturalistic stressors involve such events as students taking SAT exams. Although the effect of these brief stressors on immune function may cumulate to create serious health problems in the long run, they will be hard to document in large longitudinal surveys (Graham, Christian, and Kiecolt-Glaser, 2006).

Social scientists are more apt to be interested in sources of stress that have longer effects. Stressful event sequences involve a single event that has a number of consequences for one's life. Examples include natural disasters and the loss of a spouse or child. Although the effects may last for months or years, they can be expected to subside over time. Chronic stressors usually require restructuring social roles or personal identity. These include becoming a caregiver for someone with dementia or a traumatic injury that causes permanent disability (Graham et al., 2006). The final category is distant stressors, traumatic experiences, such as sexual assaults in the past, that can have long-term psychological effects.

Studies of the long-term effect of stress on health, such as those related to occupational or race/ethnic relations, may involve frequent (perhaps daily) exposure to a variety of stressors, some of which don't fit neatly into these categories. For example, we might want to include unreasonable expectations, distrust, or lack of respect, which may reflect the general social environment, rather than simply the effects of single incidents.

There is no single biomarker that will identify all of the social and psychological circumstances that we consider stressful. In particular, it appears that different types of psychological stress can have very different effects on biomarkers. For example, within the category of stressful event sequences, the loss of a spouse is generally associated with a decline in natural inflammatory responses, which is not apparent following a breast biopsy (Segerstrom and Miller, 2004).¹

The complex relationship between social environments and stress is apparent from animal models. Sapolsky (2005) has reviewed the literature on the physiological effects of social hierarchies in primates. He concludes that “no consensus exists as to whether dominant or subordinate animals are more physiologically ‘stressed.’ … Rank means different things in different societies and populations.” He offers a number of generalizations.

1. In social groups with a relatively fixed hierarchy, subordinate individuals feel stress from subjugation. However, when structures are more fluid, dominant individuals tend to be more stressed, since they must struggle to maintain their dominance. This suggests that results based on the Whitehall study that high-status jobs are associated with less stress may not be generalizable across a wide range of social organizations. In particular, a study of men in the Otmoor villages in Oxfordshire found higher levels of adrenaline associated with high social class (professional and manual versus nonmanual labor) (Jenner, Reynolds, and Harrison, 1980; Harrison, 1981).
2. Subordination is more stressful in species with greatest inequality in access to resources.
3. Coping strategies and social support including the presence of kin can ameliorate the effects of stressful circumstances.
4. Subordinates can develop social strategies for avoiding interactions with dominants.

These observations about primate societies are quite consistent with hypotheses that social scientists are beginning to study. For example, some discussions of the effects of social inequality have been framed in terms of stress (Geronimus, Bound, Waidmann, Colen, and Steffick, 2001), and researchers have used biomarkers to examine the importance of social support (Seeman et al., 2002) and social ties (Seeman et al., 2004).

BIOLOGICAL ROBUSTNESS

The challenge for social scientists is to identify a manageable number of biomarkers that capture the most important features of the aging process and the development of chronic diseases. Two related concepts—biological robustness and allostatic load—provide a theoretical framework for selecting biomarkers for social science research.

Systems analysis provides a useful way of thinking about how an organism manages the numerous changes in chemistry required to deal with regular and irregular challenges that result from the interactions with the environment. The challenges come from the physical environment, stressful situations, persistent anxiety, and such behaviors as diet and smoking. This approach involves consideration of feedback loops, redundancy, and diversity of function. A central theme is the concept of robustness. Kitano states: “[r]obustness is a property that allows a system to maintain its functions despite external and internal perturbations. It is one of the fundamental and ubiquitously observed systems-level phenomena that cannot be understood by looking at the individual components” (2004, p. 826).

Over time, the functioning of biological systems leads to the accumulation of damage that reduces robustness. Initially this can result in a delayed return to equilibrium following a challenge. Longer periods spent in disequilibrium cause additional damage. The rate of accumulation is determined, in part, by the frequency and severity of challenges as well as genetic and developmental factors that determine initial robustness. Eventually this process can lead to a state of chronic, severe disequilibrium. The concept of robust response to challenge is central to many ways of thinking about the aging process.

Proper functioning of biological systems does not rely on robust maintenance of a single equilibrium point (homeostasis). Instead, the appropriate balance point changes to minimize energy requirements (and expected requirements) in given circumstances. For example, when there is a perception of a need for defensiveness and vigilance, the mobilization of energy requirements is dominated by the sympathetic nervous system. At other times, the parasympathetic system imposes cognitive control that can better balance resources. Therefore, the appropriate balance point depends, in part, on perceptions about the environment.

These perceptions are managed by the prefrontal cortex of the brain, which provides the likely link between social and psychological stress and health. Thayer and Friedman note that “in modern society … inhibition [of the sympathetic system], delayed response, and cognitive flexibility are vital for successful adjustment and self-regulation.” However, such emotional states as worry and anxiety, reinforced by feedback loops, can lead to “prolonged prefrontal inactivity [that] can lead to hypervigilance, defensiveness, and perseveration” (Thayer and Friedman, 2004, p. 577). They postulate that the resulting imbalance between the sympathetic and parasympathetic systems may be the link between persistent feelings of stress and the development of chronic diseases.

Social science research on the relationship between stress and health need not involve complex modeling of the biological linkages between cognition and other biological systems. Instead, we can use biomarkers that reflect important aspects of these systems as intermediate variables. Most research using biomarkers has examined associations with single biomarkers. However, McEwen and colleagues have described a more holistic approach that links social and psychological events to long-term health outcomes (McEwen and Stellar, 1993; Seeman, Singer, Rowe, Horwitz, and McEwen, 1997; McEwen, 1998). They use the term “allostasis” much like robustness to refer to a system's maintenance of stability.² The term “allostatic load” (AL) then refers to the cumulated damage resulting from repeated cycles of allostasis. McEwen and Stellar defined it as “the strain on the body produced by repeated ups and downs of physiologic response, as well as by the elevated activity of physiologic systems under challenge, and changes in metabolism … that can predispose the organism to disease” (McEwen et al., 1993, p. 2094).

Insulin resistance and type II diabetes offer a simple example. Insulin resistance involves reduced sensitivity of cells to insulin-stimulated glucose absorption. This leads to an increased secretion of insulin and a slower response to increases in serum glucose levels (glucose intolerance).³ As the condition worsens over a period of years, the body loses its ability to produce enough insulin to stabilize serum glucose levels, which results in diabetes, a chronic disequilibrium in glucose levels. Excess serum glucose leads to the glycosylation of proteins that causes damage to numerous other biological systems (especially the circulatory system). One result is a reduction in the ability of the pancreas to produce insulin.⁴ This simple model of insulin response to changes in serum glucose levels illustrates the progression from poor short-term responsiveness to chronic instability that develops in many systems in response to challenges.

This “glucose-centric” view of insulin resistance is grossly oversimplified. For example, the insulin resistance (or metabolic) syndrome is generally defined as including elevated insulin levels, glucose intolerance, obesity (especially a concentration of fat around the waist), elevated blood pressure, elevated serum triglycerides, and decreased high-density lipoproteins (HDL).⁵ The last four are also commonly recognized risk factors for another major health problem: heart disease. The symptoms of the metabolic syndrome are related through complex feedback loops involving adiposity, triglycerides, insulin, leptin, interleukin-6 (IL-6), and tumor-necrosis factor α (TNFα) (Kitano et al., 2004). IL-6 and TNFα involve stimulation of the immune system, which plays an important role in the feedback loops that lead to chronic insulin resistance and which may contribute to the development of vascular disease. Their release is attenuated by the parasympathetic nervous system, thus providing an important link to emotional stress (Thayer and Friedman, 2004). This might explain, in part, the association of insulin resistance with affective disorders (including depression) and Alzheimer's disease in addition to diabetes and heart disease (Rasgon and Jarvik, 2004).

SELECTING THE MOST USEFUL BIOMARKERS

The list of biomarkers that could be included in large social science surveys is expanding rapidly (McDade et al., 2006). Singer, Ryff, and Seeman state that the operationalization of allostatic load is still in its infancy, and they provide guidelines for selecting biomarkers that provide insights into the concept. They propose that “any operationalization of the concept of AL should signal pending onset of diverse kinds of disability and disease … [and] should represent the biological signature of cumulative antecedent challenges” (Singer, Ryff, and Seeman, 2005, pp. 113-114). Therefore, we want both leading indicators as well as measures of cumulative damage. To limit the number of indicators, we need measures associated with multiple health conditions.

We can see aspects of these criteria in the 10 measures included in the original operationalization of allostatic load (Seeman et al., 1997). The list included six items that are generally part of the metabolic syndrome: systolic and diastolic blood pressure, HDL, the ratio of total cholesterol to HDL, glycosylated hemoglobin, and waist-hip ratio (a measure of abdominal obesity). Three are chemical messengers that are primary mediators of allostatic load: cortisol (a major stress hormone) and its antagonist, DHEAS, as well as two measures of sympathetic nervous system activity, epinephrine, and norepinephrine.

Different measures are used to measure different stages of the accumulation of allostatic load. We have already seen that the metabolic syndrome is associated with multiple important diseases, including diabetes and heart disease. Among the elements of the metabolic syndrome, the waist-hip ratio is a risk factor or precursor of disease, whereas glycosylated hemoglobin measures the cumulative effect of serum glucose levels over the past few months. The Whitehall II study included a fasting glucose tolerance test that directly measures the body's response to challenge.

Singer, Ryff, and Seeman note that the complex dynamics of the immune system “suggest[s] that a large battery of the immune measures be incorporated in an allostatic load scoring scheme. On the other hand, because our larger objective is to use biological indicators that reflect possible malfunction across multiple systems … it is important to have a limited set of biomarkers in place for each system, but use those that are sensitive to broadly based dysregulation” (Singer et al., 2005, pp. 130-131, emphasis in original).

In this context, they discuss the value of adding IL-6 to the list of biomarkers. Trauma, infection, fever, and stressful experiences are all associated with elevated levels of IL-6 (Singer et al., 2005). Elevated levels are also associated with a wide range of health conditions and with several forms of disability and poor health, including depression.⁶ The immune system is tightly integrated with the nervous system and the endocrine system (Cacioppo, Berntson, Sheridan, and McClintock, 2000; Glaser and Kiecolt-Glaser, 2005), which means that immune markers are associated with broader dysregulation. Measurements of Il-6 were included in the Social Environment and Biomarkers of Aging Study in Taiwan (Goldman, Glei, Seplaki, Liu, and Weinstein, 2005), a nested case-control study in Whitehall II (Brunner et al., 2002) and the Health ABC Study (Koster et al., 2006).

Similar criteria can be used to select genes for social science research on health.⁷ For example, a relatively common allele of the ApoE gene has been shown to be associated with increased risk of ischemic heart disease and Alzheimer disease (Farrer et al., 1997; Eichner et al., 2002). In addition, both ApoE genotype and serum concentration of ApoE are involved in regulating the acquired immune response (Colton, Brown, and Vitek, 2005). Similarly, the central role of IL-6 in the immune system has led to a number of studies of mortality associated with a common allele of the gene for IL-6 (Christensen, Johnson, and Vaupel, 2006).

Several studies have examined the association of various causes of morbidity or mortality with individual biomarkers or subindexes containing some of the elements of allostatic load (Seeman, McEwen, Rowe, and Singer, 2001; Goldman et al., 2005; Chapter 3 in this volume). This type of analysis is important for testing the value of including specific biomarkers, and it may help to understand the specific mechanisms at work in a population. However, the potential value of an index of allostatic load is that it doesn't simply measure a number of important biological systems. Rather it should recognize that these systems are not independent. Although the early signs of chronic dysregulation may first appear in one system or another, examining biomarkers of separate systems risks missing the forest for the trees. Although early reports are encouraging, the value of indexes of biomarkers based on the theory of allostatic load is still uncertain.

BIOMARKERS IN A LIFE-CYCLE CONTEXT

Aging is a process. Much recent research is based on the hypothesis that many common chronic health problems, such as cardiovascular disease, diabetes, and dementia, can result from the cumulated effects of repeated or sustained social and physical challenges (Garruto, Little, James, and Brown, 1999). Yet socioeconomic characteristics can change over time and environmental or social change (e.g., moving to a new location, becoming unemployed or widowed) is itself associated with increased psychological stress. There is evidence that the negative effects of stress begin in utero (Graham et al., 2006). Documenting a cohort's history of early life conditions, social and psychological stress, and the appearance of dysfunction of several systems would require a longitudinal study of extraordinary duration. For this reason, it will always be necessary to piece together results from different studies to create synthetic cohorts. A good example of this approach is work by Seeman and colleagues that compares results from the Wisconsin Longitudinal Study (ages 58 and 59) and the MacArthur Study of Successful Aging (initially 70-79) (Seeman et al., 2002).

Figure 8-1 suggests how a life-cycle approach could piece together results from a number of longitudinal surveys. The list of studies could be expanded to include research at younger ages on the short-term effects of stress at different parts of the life cycle (Lundberg, 2005). Among adolescents and young adults, research should focus on the prevalence of risk factors and on measures of short-term responses to stress. For example, some studies on this topic suggest that responsiveness to stress at younger ages may interact with genetic predisposition to cause problems of hypertension at later ages (Light et al., 1999). The Whitehall II study covers a middle-aged group and provides evidence about both the response to acute stress and the effects of chronic stress. It included nested case-control studies to examine differences in heart rate, blood pressure, and cortisol levels through the work day that can be compared with simpler case-control studies (Matthews et al., 2000; Steptoe et al., 2003). They have also examined heart rate variability (Hemingway et al., 2005), which Thayer and Friedman emphasize as an important marker for parasympathetic nervous system (Thayer and Friedman, 2004). At the older ages, research can focus on chronic stress and chronic dysregulation.

FIGURE 8-1

A life-cycle depiction of the ages included in six studies that employed biomarkers. The black rectangles show the ages included in the baseline and the grey indicate ages included in the follow-up surveys.

The life-cycle approaches used in the social sciences parallel the more biological “life history” approach to studying development and aging. This approach is based on the fact that available resources must be allocated among three processes: growth, maintenance, and reproduction. It has been used to examine the development of the immune system at early ages that can have serious implications for mortality at the older ages (McDade, 2005). It is also the basis for the “disposable soma” theory of the evolution of aging (Kirkwood and Austad, 2000).

Life-cycle models have been used extensively to study a wide range of topics in sociology and demography. However, the simplicity in diagrams of such models is deceiving. Issues of causal ordering and feedback loops raise some of the most serious modeling and statistical issues facing social scientists (National Research Council, 2001b). In addition, there are risks in combining results from different types of samples in different populations. At this point, we may not know enough about what is important to be able to tell when results from studies in one population can reliably be extrapolated to other populations.

A FEW WORDS OF CAUTION

The extensive literature on the relationship between the ApoE genotype and mortality suggests two important lessons that apply to all research on biomarkers. First, type 2 errors may be more frequent than type 1 errors in some types of research and are potentially more serious. The search for genes associated with longevity has led to many (apparently) false positives. However, there have also been false negatives in longitudinal studies of single genes (Bader, Zuliani, Kostner, and Fellin, 1998; Juva et al., 2000). Social scientists tend to obsess about avoiding false positives (p-values) and we are rarely concerned with the risk of false negatives (power). This attitude seems to be heightened in biomedical research. One reason is that clinicians have to be cautious in applying new research findings since the cost of type 1 errors can have serious effects. A second factor is that in genetic epidemiology scans of the genome generally involve a large number of multiple comparisons, which increases the risk of type 2 errors. However, in certain circumstances, false negatives may outnumber false positives. For example, standard power calculations set the risk of a type 1 error at 0.05 and the risk of a type 2 error at 0.20. False negatives may also be more damaging than false positives because negative results tend to discourage further research. This emphasizes the need for replication and multiple tests of good hypotheses even when they don't pan out immediately.⁸

Second, at this stage of research using biomarkers, there is a serious risk of unwarranted generalization. The use of biomarkers in social science research is still quite new. We are eager to draw conclusions and to base the design of new studies on previous results. However, a true negative in one population might be a true positive in another population. This can happen for both biological and sociocultural reasons. For example, the ApoE e34 genotype is not associated with excess risk of death in African Americans but it is associated with excess risk in Europeans (Ewbank, 2007). If research on the topic had started with studies of African Americans, this line of research might have ended after only one or two studies. Similar problems can arise with social variables. As noted above, similar social positions (e.g., manager) or events (e.g., widowhood) may be associated with very different levels of stress in different social settings. As scientists, we are trained to detect patterns and draw conclusions. However, we have to be cautious when results are not confirmed in subsequent studies—especially when the studies are of populations that are biologically and socially as different as Whitehall and Taiwan.

CONCLUDING COMMENTS

The use of biomarkers in social science research holds great promise. Biomarkers are quantifiable intermediate variables that can act between the complex and heterogeneous social factors and the equally diverse and complex health outcomes. Chang and colleagues note that “it is too early for us to answer the really big question: Does the inclusion of biomarkers in household surveys help us to understand SES differences in health, particularly with regard to the role of stressful experience?” (Chapter 3 in this volume). It is still early, but I am convinced that biomarkers will enable us to greatly expand our understanding of the effects of social factors on health. We are still testing new, promising biomarkers, and we are just beginning to get the data needed to sort out the effects of various kinds of stress at different ages. Work is still needed to better operationalize and test the concept of allostatic load.

We have learned a lot in the past 10 years. First, a number of researchers have gained experience collecting samples and working with medical labs that can analyze them. This is a fundamental step and one that required a significant amount of work. Second, there are now data on several biomarkers in very different populations. We are just beginning to see comparable analyses of similar data in different settings. We have expanded the list of biomarkers that have been tested, and we have moved beyond the list of 10 initially used to measure allostatic load. In addition to these lessons, the new data from Taiwan should teach us to be cautious about drawing conclusions from cross-sectional data (see Chapter 3). Similarly, the Whitehall II study has demonstrated the advantages of working in a defined, structured population in which it is possible to collect a wider range of biomarkers and the social structure is easier to define. It is not necessary for all studies to be nationally representative.

The experience thus far also offers lessons to guide future research. First, research at the older ages needs to be combined with studies of the early signs of reduced robustness in younger adults. The research on younger ages will require using biological challenges, like two-hour glucose tolerance tests, that are difficult to carry out in large population studies. Second, we need to be aware that some biomarkers might be associated with stress in different ways at different ages. There may also be different associations with different types of stress. This will complicate comparisons among studies until we have a sufficient amount of data to begin disentangling these relationships.

Finally, I think the most encouraging development in the past 10 years is the emergence of a cohort of young researchers who feel that biomarkers are central to what we do and how we do it. The pioneering research makes it easy to imagine what this approach has to offer and provides valuable models for future research.

ACKNOWLEDGMENTS

This research was supported by grant number R01-AG-016683 from the National Institute on Aging.

REFERENCES

1. Alley DE, Seeman TE, Kim JK, Karlamangla A, Hu P, Crimmins EM. Socioeconomic status and C-reactive protein levels in the U.S. population: NHANES IV. Brain, Behavior, and Immunity. 2006;20(5):498–504. [PubMed: 16330181]
2. Bader G, Zuliani G, Kostner GM, Fellin R. Apolipoprotein E polymorphism is not associated with longevity or disability in a sample of Italian octo- and nonagenarians. Gerontology. 1998;44(5):293–299. [PubMed: 9693262]
3. Brunner EJ, Hemingway H, Walker BR, Page M, Clarke P, Juneja M, Shipley MJ, Kumari M, Andrew R, Seckl JR, Papadopoulos A, Checkley S, Rumley A, Lowe GDO, Stansfeld SA, Marmot MG. Adrenocortical, autonomic, and inflammatory causes of the metabolic syndrome: Nested case-control study. Circulation. 2002;106(21):2659–2665. [PubMed: 12438290]
4. Cacioppo JT, Berntson GG, Sheridan JF, McClintock MK. Multilevel integrative analyses of human behavior: Social neuroscience and the complementing nature of social and biological approaches. Psychological Bulletin. 2000;126(6):829–843. [PubMed: 11107878]
5. Christensen K, Johnson TE, Vaupel JW. The quest for genetic determinants of human longevity: Challenges and insights. Nature Reviews Genetics. 2006;7(6):436–448. [PMC free article: PMC2726954] [PubMed: 16708071]
6. Colton CA, Brown CM, Vitek MP. Sex steroids, APOE genotype, and the innate immune system. Neurobiology of Aging. 2005;26(3):363–372. [PubMed: 15639315]
7. Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A, Lowe GDO, Pepys MB, Gudnason V. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. New England Journal of Medicine. 2004;350(14):1387–1397. [PubMed: 15070788]
8. Eichner JE, Dunn ST, Perveen G, Thompson DM, Stewart KE, Stoehla BC. Apolipoprotein E polymorphism and cardiovascular disease: A HuGE review. American Journal of Epidemiology. 2002;155(6):487–495. [PubMed: 11882522]
9. Ewbank DC. Differences in the association between APOE genotype and mortality across populations. Journals of Gerontology Series A: Biological Sciences and Medical Sciences. 2007;62A(8):899–907. [PubMed: 17702883]
10. Farrer LA, Cupples LA, Haines JL, Hyman B, Kukull WA, Mayeux R, Myers RH, Pericak-Vance MA, Risch N, Van Duijn CM. Effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease. Journal of the American Medical Association. 1997;278(16):1349–1356. [PubMed: 9343467]
11. Garruto RM, Little MA, James GD, Brown DE. Natural experimental models: The global search for biomedical paradigms among traditional, modernizing, and modern populations. Proceedings of the National Academy of Sciences of the United States of America. 1999;96(18):10536–10543. [PMC free article: PMC17924] [PubMed: 10468644]
12. Geronimus AT, Bound J, Waidmann TA, Colen C, Steffick D. Inequality in life expectancy, functional status, and active life expectancy across selected black and white populations in the United States. Demography. 2001;38(2):227–251. [PubMed: 11392910]
13. Glaser R, Kiecolt-Glaser JK. Science and society: Stress-induced immune dysfunction: Implications for health. Nature Reviews Immunology. 2005;5(3):243–251. [PubMed: 15738954]
14. Goldman N, Glei DA, Seplaki C, Liu I, Weinstein M. Perceived stress and physiological dysregulation in older adults. Stress. 2005;8(2):95–105. [PubMed: 16019601]
15. Graham JE, Christian LM, Kiecolt-Glaser JK. Stress, age, and immune function: Toward a lifespan approach. Journal of Behavioral Medicine. 2006;29(4):389–400. [PMC free article: PMC2805089] [PubMed: 16715331]
16. Harrison GA. Catecholamine excretion rates in relation to life-styles in the male population of Otmoor, Oxfordshire. Annals of Human Biology. 1981;8(3):197–209. [PubMed: 7259097]
17. Hemingway H, Shipley M, Brunner E, Britton A, Malik M, Marmot M. Does autonomic function link social position to coronary risk?: The Whitehall II Study. Circulation. 2005;111(23):3071–3077. [PubMed: 15939818]
18. James GD, Brown DE. The biological stress response and lifestyle: Catecholamines and blood pressure. Annual Review of Anthropology. 1997;26:313–335.
19. Jenner DA, Harrison GA, Day JA, Huizinga J, Salzano FM. Inter-population comparisons of urinary catecholamines: A pilot study. Annals of Human Biology. 1982;9(6):579–582. [PubMed: 7181448]
20. Jenner DA, Reynolds V, Harrison GA. Catecholamine excretion rates and occupation. Ergonomics. 1980;23(3):237–246. [PubMed: 7428768]
21. Juva K, Verkkoniemi A, Viramo P, Polvikoski T, Kainulainen K, Kontula K, Sulkava R. APOE epsilon4 does not predict mortality, cognitive decline, or dementia in the oldest old. Neurology. 2000;54(2):412–415. [PubMed: 10668704]
22. Kahn R. The metabolic syndrome (emperor) wears no clothes. Diabetes Care. 2006;29(7):1693–1696. [PubMed: 17065717]
23. Kahn R, Buse J, Ferrannini E, Stern M. The metabolic syndrome: Time for a critical appraisal: Joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care. 2005;28(9):2289–2304. [PubMed: 16123508]
24. Kirkwood TBL, Austad SN. Why do we age? Nature. 2000;408(6809):233–238. [PubMed: 11089980]
25. Kitano H. Biological robustness. Nature Reviews Genetics. 2004;5(11):826–837. [PubMed: 15520792]
26. Kitano H, Oda K, Kimura T, Matsuoka Y, Csete M, Doyle J, Muramatsu M. Metabolic syndrome and robustness tradeoffs. Diabetes. 2004;53(Suppl. 3):S6–S15. [PubMed: 15561923]
27. Koster A, Bosma H, Penninx BWJH, Newman AB, Harris TB, van Eijk JThM, Kempen GIJM, Simonsick EM, Johnson KC, Rooks RN, Ayonayon HN, Rubin SM, Kritchevsky SB., for the Health ABC Study. Association of inflammatory markers with socioeconomic status. Journals of Gerontology Series A: Biological Sciences and Medical Sciences. 2006;61(3):284–290. [PubMed: 16567379]
28. Light KC, Girdler SS, Sherwood A, Bragdon EE, Brownley KA, West SG, Hinderliter AL. High stress responsivity predicts later blood pressure only in combination with positive family history and high life stress. Hypertension. 1999;33:1458–1464. [PubMed: 10373233]
29. Lundberg U. Stress hormones in health and illness: The roles of work and gender. Psychoneuroendocrinology. 2005;30(10):1017–1021. [PubMed: 15963652]
30. Matthews KA, Raikkonen K, Everson SA, Flory JD, Marco CA, Owens JF, Lloyd CF. Do the daily experiences of healthy men and women vary according to occupational prestige and work strain? Psychosomatic Medicine. 2000;62(3):346–353. [PubMed: 10845348]
31. McDade TW. Life history, maintenance, and the early origins of immune function. American Journal of Human Biology. 2005;17(1):81–94. [PubMed: 15612049]
32. McDade TW, Hawkley LC, Cacioppo JT. Psychosocial and behavioral predictors of inflammation in middle-aged and older adults: The Chicago health, aging, and social relations study. Psychosomatic Medicine. 2006;68(3):376–381. [PubMed: 16738067]
33. McEwen BS. Protective and damaging effects of stress mediators. New England Journal of Medicine. 1998;338(3):171–179. [PubMed: 9428819]
34. McEwen BS. The neurobiology of stress: From serendipity to clinical relevance. Brain Research. 2000;886(1-2):172–189. [PubMed: 11119695]
35. McEwen BS, Stellar E. Stress and the individual: Mechanisms leading to disease. Archives of Internal Medicine. 1993;153(18):2093–2101. [PubMed: 8379800]
36. National Research Council; Commission on Behavioral and Social Sciences and Education. Cells and surveys: Should biological measures be included in social science research? Committee on PopulationFinch CE, Vaupel JW, Kinsella K, editors. Washington, DC: National Academy Press; 2001. [PubMed: 23166967]
37. National Research Council; Board on Behavioral, Cognitive, and Sensory Sciences. Commission on Behavioral and Social Sciences and Education. New horizons in health: An integrative approach. Committee on Future Directions for Behavioral and Social Sciences Research at the National Institutes of HealthSinger BH, Ryff CD, editors. Washington, DC: National Academy Press; 2001.
38. Owen N, Poulton T, Hay FC, Mohamed-Ali V, Steptoe A. Socioeconomic status, C-reactive protein, immune factors, and responses to acute mental stress: Special issue on psychological risk factors and immune system involvement in cardiovascular disease. Brain, Behavior, and Immunity. 2003;17(4):286–295. [PubMed: 12831831]
39. Perez-Stable EJ, Marin BV, Marin G, Brody DJ, Benowitz NL. Apparent underreporting of cigarette consumption among Mexican American smokers. American Journal of Public Health. 1990;80(9):1057–1061. [PMC free article: PMC1404846] [PubMed: 2382741]
40. Rasgon N, Jarvik L. Insulin resistance, affective disorders, and Alzheimer's disease: Review and hypothesis. Journals of Gerontology Series A: Biological Sciences and Medical Sciences. 2004;59(2):M178–M183. [PubMed: 14999034]
41. Sapolsky RM. The influence of social hierarchy on primate health. Science. 2005;308(5722):648–652. [PubMed: 15860617]
42. Seeman TE, Glei DA, Goldman N, Weinstein M, Singer B, Lin HY. Social relationships and allostatic load in Taiwanese elderly and near elderly. Social Science and Medicine. 2004;59(11):2245–2257. [PubMed: 15450701]
43. Seeman TE, McEwen BS, Rowe JW, Singer BH. Allostatic load as a marker of cumulative biological risk: MacArthur studies of successful aging. Proceedings of the National Academy of Sciences, USA. 2001;98(8):4770–4775. [PMC free article: PMC31909] [PubMed: 11287659]
44. Seeman TE, Singer BH, Rowe JW, Horwitz RI, McEwen BS. Price of adaptation: Allostatic load and its health consequences. MacArthur studies of successful aging. Archives of Internal Medicine. 1997;157(19):2259–2268. [PubMed: 9343003]
45. Seeman TE, Singer BH, Ryff CD, Dienberg Love G, Levy-Storms L. Social relationships, gender, and allostatic load across two age cohorts. Psychosomatic Medicine. 2002;64(3):395–406. [PubMed: 12021414]
46. Segerstrom SC, Lubach GR, Coe CL. Identifying immune traits and biobehavioral correlates: Generalizability and reliability of immune responses in rhesus macaques. Brain, Behavior, and Immunity. 2006;20(4):349–358. [PubMed: 16293393]
47. Segerstrom SC, Miller GE. Psychological stress and the human immune system: A meta-analytic study of 30 years of inquiry. Psychological Bulletin. 2004;130(4):601–630. [PMC free article: PMC1361287] [PubMed: 15250815]
48. Singer B, Ryff CD, Seeman TE. Operationalizing allostatic load. In: Shulkin J, editor. Allostasis, homeostasis, and the costs of physiological adaptation. Cambridge, England: Cambridge University Press; 2005. pp. 113–149.
49. Steptoe A, Kunz-Ebrecht S, Owen N, Feldman PJ, Willemsen G, Kirschbaum C, Marmot M. Socioeconomic status and stress-related biological responses over the working day. Psychosomatic Medicine. 2003;65(3):461–470. [PubMed: 12764220]
50. Straub RH, Cutolo M, Zietz B, Scholmerich J. The process of aging changes the interplay of the immune, endocrine, and nervous systems. Mechanisms of Ageing and Development. 2001;122(14):1591–1611. [PubMed: 11511399]
51. Thayer JF, Friedman BH. National Research Council; Panel on Race, Ethnicity and Health in Later Life. Committee on Population, Division of Behavioral and Social Sciences and Education. A neurovisceral integration model of health disparities in aging. In: Anderson NB, Bulatao R, Cohen B, editors. Critical perspectives on racial and ethnic differences in health in late life. Washington, DC: The National Academies Press; 2004. pp. 567–603. [PubMed: 20669464]

Footnotes

1

Segerstrom and Miller provide an excellent meta-analysis of the effects of various types of stressors on a wide range of immune indicators, including 24 articles on the effects of chronic stress and 44 articles on the effects of important life events (Segerstrom and Miller, 2004).

2

McEwen points out that the terms “homeostasis” and “allostasis” are sometimes used in the narrow sense of stable levels of those functions immediately necessary to sustain life (e.g., temperature, pH, oxygen tension) and at other times they are expanded to include systems that are more variable (e.g., hormone levels, heart rate, blood pressure). He prefers to limit homeostasis to mean stability in those systems that are most essential to life. He uses the term allostasis in the broader sense to include those systems that maintain homeostasis by responding to changing environment (McEwen, 2000).

3

An alternative view is that this progression begins with inadequate production of insulin that only becomes a problem when there is decreased sensitivity to insulin. In either case, the initial result is a decreased robustness in the body's ability to deal with a sudden increase in serum glucose.

4

For a complete description of the metabolic syndrome from the perspective of system analysis and robustness, see Kitano (2004).

5

There is extensive evidence of statistical associations among the components of this syndrome, but the biological and clinical significance of these associations and the exact list of elements of the syndrome are still hotly debated (Kahn, Buse, Ferrannini, and Stern, 2005; Kahn, 2006).

6

Note that these characteristics diminish its usefulness in a clinical context.

7

Genetic material collected in social science surveys can be used to perform a large number of genetic tests. They can therefore contribute important information on genotype frequencies for a large number of genes and may be useful for studying genetic risk factors for disease. However, only a few genotypes might be important enough to include in analyses of the associations between social factors and health.

8

Many biomarkers measure normal levels of hormones that fluctuate over time. In many cases, it may be that the easiest way to increase the power of a test is to carry out multiple measurements on each individual (Jenner et al., 1980; Segerstrom, Lubach, and Coe, 2006).