tools of psychosocial biology in health care research 20

Tools of psychosocial biology inhealth care research

Andrew Steptoe

Introduction

Much health care research involves measurement of biological function. This chap-ter is concerned with the assessment of biological function in psychosocial researchon health. The psychosocial factors studied in health care research include featuresof the individual’s sociodemographic position (age, gender, ethnicity, socio-economic status etc.), aspects of social experience (such as social networks and socialsupport), exposure to different types of adversity (e.g. work stress, life events,informal care-giving), family and neighbourhood factors (such as family conflictand neighbourhood social cohesion), and psychological parameters (such as depres-sion, anxiety and hostility). Psychosocial studies in health care research are con-cerned with the contribution of these factors to disease causation, progression andprognosis, and the impact of psychosocial interventions on disease status and qualityof life. The biological measures assessed in psychosocial research fall into threecategories:

1 Biological indicators of disease states, such as blood pressure in hypertension, orairways resistance in bronchial asthma. The biological measures in this categoryare direct markers of the physiological dysfunction constituting the disease state,and are used in research on the role of psychosocial factors in the aetiology orprognosis of disease states, and in studies evaluating the impact of psychosocialinterventions.

2 Biological markers of processes involved in the aetiology of disease. Examplesinclude measures of vascular endothelial function, fibrinogen and C-reactiveprotein in the study of coronary heart disease, the concentration of CD4lymphocytes in the study of HIV/AIDS, and the measurement of metabolites ofcorticosteroids and catecholamines in the investigation of obesity and insulinresistance. The biological parameters assessed in such studies are more distal todisease states than those assessed in Category 1, but nevertheless provide objectiveinformation concerning the impact of psychosocial factors on precursors ofdisease or on underlying physiological dysfunctions.

3 Non-specific biological markers of stress-related activation or resistance to dis-ease. Biological measures in this category include heart rate, blood pressure, sweatgland activity, the stress hormones cortisol, adrenaline and noradrenaline, circulat-ing lymphocyte numbers and activity, and concentrations of inflammatorycytokines and immunoglobulins. These biological variables are not assessed as

20

Tools of psychosocial biology in health care research. In 'Handbook of Health Research Methods' edited by A. Bowling and S. Ebrahim. Open University Press, New York, 2005.

markers of specific disease states, but as more general indicators of psycho-biological activation and resistance.

Biological measures are used extensively in animal research on psychosocial fac-tors, but this chapter will be limited to work on humans. Biological measures areassessed in three types of psychosocial study: epidemiological surveys, naturalisticmonitoring studies and experimental studies. Each of these applications will bedescribed, along with their strengths and limitations. I also discuss issues of practicalmeasurement and interpretation for several common measures. This chapter isnot exhaustive, since different areas of health care research can involve complexand sophisticated biological assessments related to specific medical conditions(Cacioppo et al. 2000).

I have also not included discussion of biological correlates or indicators of healthbehaviours in this chapter. Examples include the measurement of cotinine orexhaled carbon monoxide in the study of smoking, the use of accelerometers forassessing physical activity and the monitoring of biomarkers for alcohol intake andspecific dietary nutrients. The purpose of such measures is to provide more object-ive information about health-related activities than is available with self-reportassessments. These measures are all important, but give rise to issues of validity andreliability that go beyond psychosocial research.

Theoretical foundations of psychosocial biology

Biological processes underpin all actions and behaviours, ensuring appropriate sup-plies of energy to working muscle, brain and other tissues, and the maintenance ofbodily functions. The body’s defence mechanisms are continuously active, provid-ing physical and chemical barriers against pathogens and invading micro-organisms,and responding to potential dangers. The foundations of psychosocial biology lie inthe disturbance of these regulatory processes, and in the notion that psychosocialfactors influence bodily functions and thereby lead to increased or decreased risk ofillness. These influences are mediated through psychobiological pathways: thepathways through which central nervous system function activates the autonomicnervous system, neuroendocrine and immunological responses (Steptoe 1998).

There is a substantial research literature on the relationship between psychosocialfactors and biological function, much of which derives from the study of the effectsof acute and chronic stress (Weiner 1992; Fink 2000; Ader et al. 2001). Biologicalstress responses occur both under conditions of extreme psychosocial adversity(such as a natural disaster or assault), and when longer-term but lower-leveldemands exceed the individual’s capacity to cope (Steptoe and Ayers in press).Physiological stress responses encompass most of the main organ systems and regula-tory processes of the body, including respiration and cardiovascular function, waterbalance, glucose metabolism and energy supply, blood clotting, inflammation andimmune defences. These multiple components are controlled through the auto-nomic nervous system and neuroendocrine circuitry. During stressful encounters,the sympathetic branch of the autonomic nervous system is activated in concertwith catecholamines such as adrenaline released from the medulla of the adrenalglands. The opposing parasympathetic branch of the autonomic nervous systembecomes more active under conditions of conservation or behavioural withdrawal.The hypothalamic-pituitary-adrenocortical (HPA) axis is the most important neu-roendocrine pathway in stress responses, although other neurotransmitters andhormone systems are also involved. Activation of the HPA axis leads to the releaseof cortisol in humans (corticosterone in rodents) from the adrenal cortex. The

472 Multidisciplinary research measurement

sympathoadrenal and HPA axes are interdependent, and are controlled by com-plex feedback loops involving the central nervous system (Sapolsky et al. 2000).Even short-term stressors can trigger gene expression of enzymes regulatingneuroendocrine function (Sabban and Kvetnansky 2001).

The acute activation of biological variables elicited by stress is adapted for thesupport of vigorous physical activity and defence (fight or flight). Risk to healtharises for two reasons. First, the circumstances in everyday life that provoke psycho-biological responses are typically not ones in which vigorous activity is required;having an abrasive interaction with family members, or being frustrated at work byequipment breakdown do not call for high levels of energy expenditure. Con-sequently, the intense mobilization of biological responses is inappropriate. Second,the psychosocial factors that elicit biological responses are ubiquitous in people’slives, so repeated activation is common. In many areas of health care research it islow level, repeated, or chronic disturbance of function that is more significant thanlarge reactions to severe but rare stressful events.

In recent years, the concept of allostatic load has become influential in under-standing the chronic dysregulation of biological responses caused by psychosocialfactors. This construct was introduced by Sterling and subsequently developed byMcEwen (1998; McEwen and Wingfield 2003). It refers to the cumulative cost tothe body of the process of achieving stability of physiological systems in the face ofenvironmental challenge. Central to allostatic load is alteration in the activity ofmediators such as the HPA axis and the sympathoadrenal system, resulting in imbal-ances that can increase risk to health.

A very simplified schematic description of possible manifestations of disturbedpsychobiological function is shown in Figure 20.1. The graph on the left representsthe normal pattern of response, with a biological system such as blood pressure orcortisol increasing in the face of challenge, then decreasing back to reference levelsfollowing termination of the stimulus. The graphs on the right summarize threetypes of disturbance to this response pattern (illustrated with the dotted line). Ingraph A, the response is heightened above normal levels, perhaps as a consequenceof concurrent psychosocial adversity. For example, our group has shown that peoplereporting low control at work produce heightened fibrinogen responses to demand-ing behavioural tasks (Steptoe et al. 2003a), and relationships between neuroendo-crine responses and social support have been described (Uchino et al. 1996).Heightened responsivity may in turn be associated with disease risk. Thus it hasbeen found that high blood pressure stress reactivity, coupled with exposure toelevated life stress or work demands, predicts increased risk of hypertension andprogression of subclinical atherosclerosis (Everson et al. 1997; Light et al. 1999).High levels of cortisol promote increased concentration of lipids in the circulation,accumulation of abdominal fat, impaired fertility and decalcification of bone(Weiner 1992).

In graph B, the magnitude of responses is not altered, but restitution of homoeo-stasis is impaired. This leads to prolongation of responses and delays in post-stressrecovery. An example of this pattern comes from a study of men and womenrecruited from the Whitehall II epidemiological cohort to investigate the psycho-biological concomitants of low socioeconomic position. The magnitude of bio-logical responses to behavioural tasks did not differ with socioeconomic positiondefined by a grade of employment. However, post-stress recovery in blood pressureand in heart rate variability was impaired in lower socioeconomic status participants(Steptoe et al. 2002). Additionally, lower-status individuals showed prolongedincreases in prothrombotic factors (plasma viscosity and Factor VIII) followingtermination of the behavioural challenge (Steptoe et al. 2003b). Recovery in corti-sol and catecholamines can be delayed for several hours after stressful work (Sluiter

Tools of psychosocial biology 473

et al. 2000). Heightened or prolonged cortisol responses may also lead to downregu-lation of immune function, rendering the individual more vulnerable to infection(Vedhara et al. 1999a).

Graph C illustrates a third manifestation of allostatic load, with reduced respon-sivity in the biological parameter, due either to changes in receptor sensitivity ordepletion of physiological mediators. The best studied example is hypocortisolism,or suppressed cortisol output and responsivity (Heim et al. 2000a). Several clinicalconditions appear to be associated with low levels of cortisol including chronicfatigue syndrome, bronchial asthma and rheumatoid arthritis. Since cortisol sup-presses inflammatory responses, low levels can lead to overactivity of the immunesystem in autoimmune disease. The circumstances in which these different disturb-ances of psychobiological responses are elicited depend on a combination of thenature and duration of the challenge, and the life stage of the individual (McEwenand Wingfield 2003). Low cortisol has also been recorded in several studies ofpost-traumatic stress disorder (Yehuda 2002), although this pattern has not beenreplicated in some population-based studies (Young et al. 2004).

Figure 20.1 Schematic outline of patterns of biological response to behaviouralor social challenge. The left panel shows the normal response pattern, and on theright three types of disturbed response are illustrated, as described in the text. Thevertical axis scale is arbitrary, and could refer to a variety of responses (bloodpressure, cortisol, muscle tension, immune reaction, etc.). The solid bar in eachschematic illustrates the period of exposure to challenge.


Research paradigms

There is no single ‘ideal’ feasible study of psychosocial biology that can definitivelyprove or falsify the involvement of psychobiological dysfunctions in a particularhealth outcome. Instead, the case depends on the convergence and aggregation ofdata from different types of study, each with its strengths and limitations.

Epidemiological studies

Epidemiological studies provide the core method of establishing the contributionof psychosocial factors to the development of disease, and are also used to identifythe biological mediators of these associations. For example, lower socioeconomicstatus is associated with heightened fibrinogen, vascular inflammation and inci-dence of the metabolic syndrome, which in turn increase risk of coronary heartdisease (Brunner et al. 1997; Hemingway et al. 2003). High levels of anger andhostility have been related to blood pressure in population studies, and may increaserisk of hypertension (Jorgensen et al. 1996; Yan et al. 2003). Biological and psycho-social measures from large samples can be obtained at relatively low cost, prospectivestudy designs can be employed and potential confounders can be taken into accountstatistically. However, the biological measures in epidemiological studies are gener-ally recorded on a single occasion under resting conditions (in a clinic or medicaloffice) that are not typical of everyday life. Such studies can provide limited infor-mation about the dynamics of biological response, or the consequences of bio-logical activation. Factors such as time of day and nutritional state affect manymeasures used in psychosocial biology, and may need to be recorded to permitaccurate interpretation.

Naturalistic monitoring studies

The second kind of study in psychosocial biology involves the sampling of bio-logical variables during everyday life. Such studies take many forms, from recordingsduring challenging tasks such as parachuting or speaking in public, to repeatedmeasures of blood pressure or salivary cortisol over an ordinary day. Some of thesetechniques are extensions of methods used in clinical investigation, such as ‘Holter’monitoring of the electrocardiogram in patients with coronary disease, and the useof ambulatory blood pressure monitors for evaluating hypertension. The purpose ofthese methods in psychosocial research is to assess biological activity under naturalconditions and to examine the covariation between everyday activities, emotionsand biology. For example, one study of people with bronchial asthma and non-asthmatic controls involved repeated spirometric measurements and mood assess-ments several times a day for three weeks (Ritz and Steptoe 2000). Negative moodstates were associated with reduced forced expiratory volume in asthmatic partici-pants but not in controls. Multiple samples of saliva have been obtained to measurethe profile of cortisol release over the day, showing that cortisol increases inresponse to stressful daily events (van Eck et al. 1996). Measurements of muscletension from surface electrodes have been made in supermarket cashiers, and haveshown heightened trapezius muscle tension during work that is associated withcomplaints of neck and shoulder pain (Lundberg et al. 1999).

Naturalistic monitoring methods have the advantage of ecological validity, evalu-ating biological activity in real life rather than the artificial conditions of a labora-tory or clinic. Associations between psychosocial factors and biological responsesmay be observed that are not detectable when single measures are taken under


clinical conditions. But naturalistic methods also have limitations. First, the range ofbiological markers that can be assessed is relatively small in comparison with themore sophisticated possibilities available in the clinic. Second, the measurementtechniques need to be relatively unobtrusive, so as not to interfere with ongoingactivities. There have been some heroic developments in naturalistic monitoring,such as the development of portable radioactivity detectors focused over the heart,and mounted in vest-like garments, that have been used to assess the impact ofmental stress on cardiac function in patients with coronary artery disease (Burg et al.1993). Much of the data on circadian rhythms of cortisol secretion have involvedvenepuncture every two hours for 24 hours, or the periodic withdrawal of bloodfrom an indwelling catheter (Van Cauter et al. 2000). There is a danger that suchmethods are so stressful in themselves that they will obscure any association betweenpsychosocial factors and biological responses, and certainly they can cause sleepdisturbance (Jarrett et al. 1984). Measures of hormones such as cortisol, dehydroepian-drosterone (DHEA), testosterone, prolactin and estrogen in saliva overcome many ofthese problems. Third, there are several extrinsic factors that influence biologicalfunction that need to be taken into account, including cigarette smoking, food andcaffeine intake, sleep and physical activity. These factors have to be monitored innaturalistic studies and taken into account statistically in analysis. Multilevel modellinghas become the method of choice in analyses of these data (Schwartz and Stone 1998).

Mental stress testing

The third research method involves monitoring biological responses to standardizedpsychological or social stimuli. A wide range of mental stress tests are employed,including cognitive and problem solving tasks, simulated public speaking, upsettingfilms and interpersonal conflict tasks. Mental stress testing is typically carried out ina laboratory or clinic, but similar methods have been applied in home settings,particularly with children and elderly groups. A mental stress testing session involvesa period of rest so that baseline levels of physiological function can be established,followed by a stress or challenge period that may last anything from five minutes tothree hours. Further biological measures are obtained during the challenge periodand for some time afterwards, depending on the dynamics of the measure underinvestigation. For example, blood pressure and heart rate respond within 1–2 min-utes of onset of stress, while cortisol in saliva and blood may not peak for 30minutes, and inflammatory cytokines such as interleukin (IL) 6 continued to rise forat least two hours. It is possible to obtain measures from several people simul-taneously, if equipment is available. An important series of studies by Kiecolt-Glaseret al. involved measurement of cardiovascular, endocrine and immune functionfrom couples during discussion of areas of conflict. Hostile and negative behaviourselicited heightened biological responses in both men and women, and therewere also striking differences in the response patterns of husbands and wives(Kiecolt-Glaser and Newton 2001). Interestingly, the magnitude of adrenaline andnoradrenaline responses during the conflict task were found to predict troubledmarriages and divorce ten years later (Kiecolt-Glaser et al. 2003a).

The value of mental stress testing is that responses to psychosocial stimuli can bemonitored under environmentally controlled conditions, reducing many of thesources of bias and individual difference that might otherwise be present. Experi-mental designs can be used with randomization to different conditions (such as lowand high stress controllability), and sophisticated biological measures are possible.There are two major limitations. The first is that the stimuli used are often arbitraryand divorced from everyday life; few people spend much of their lives carrying outmental arithmetic or problem-solving under time pressure. Studies using more


ecologically valid challenges such as interpersonal conflict tasks remain in theminority. Second, the stimuli used are brief, so that only acute biological responsesare recorded. Chronic challenges may elicit different response patterns because ofhabituation and adaptation. The generalizability of biological adjustments thereforeremains uncertain in many cases.

Intervention studies

One of the principle methods of determining causality in biomedical research is tomodify putative causal pathways and assess impact on outcome. Biological measuresare used extensively in intervention studies to assess the impact of lifestyle or cogni-tive-behavioural treatments on disease processes (e.g. Tuomilehto et al. 2001). It isalso known that psychological treatments such as relaxation training have effects onneuroendocrine activity, muscle tension and blood pressure that are opposite tothose induced by stress (Antoni 2003). However, intervention methods have notbeen extensively used to test causal models, for instance by evaluating the effects ofchanges in putative biological mediators of psychosocial influence. Such anapproach has been used to test mechanisms in animal research, applyingbeta-adrenergic pharmacological blockade to demonstrate that sympathetic acti-vation mediates the impact of social stress on atherosclerosis in primates (Kaplan etal. 1991). But intervention approaches have yet to be exploited fully inpsychobiological research in many health care settings.

Combinations of methods

I have described the main research methods in psychosocial biology, but there is anincreasing trend towards combining the different methods. For example, naturalisticmonitoring of salivary cortisol has been included in epidemiological populationstudies such as the CARDIA (Coronary Artery Risk Development in YoungAdults) and Whitehall II surveys. Mental stress testing has been extended into theepidemiological framework, allowing the impact of individual differences in stressresponsivity on disease progression to be evaluated (Everson et al. 1997; Carroll et al.2003). Mental stress testing and naturalistic monitoring have been combined toevaluate the extent to which individual differences in acute blood pressure stressreactivity generalize to everyday life situations (Kamarck et al. 2000). This trend isvery welcome, and is likely to grow with advances in instrumentation of technologyand biological assay of techniques.

Measurement and interpretation of biological variables

This section provides an overview of some of the principle biological variablesmeasured in psychosocial studies. Space prevents detailed evaluation of all themethods available, so the aim is to focus on general measures of neuroendocrinefunction cardiovascular activity, inflammatory processes, immune function andmusculoskeletal activity. For each set of measures, I outline why they are assessed,how they are collected and processed, and what factors need to be taken in accountin health care research settings to ensure accurate interpretation.

Neuroendocrine factors

Cortisol

Cortisol is a steroid glucocorticoid hormone produced by the adrenal cortex underthe control of adrenocorticotropic hormone (ACTH) and the HPA axis. It acts on


almost all the nucleated cells in the body. It is implicated in a variety of conditionsstudied in health care research including depression, disturbances of cognitive func-tion, obesity (particularly abdominal or central obesity), inhibition of growth andfertility, hypertension, Type II diabetes, inflammatory, and autoimmune conditions(McEwen et al. 1997; Lupien et al. 1999; Bjorntorp 2001; Wolf 2003). There arechanges in cortisol with age which may have implications for the maintenance ofmemory function, particularly in old age (Lupien and Lepage 2001). Low socio-economic status has been associated with heightened cortisol levels both in adultsand children (Lupien et al. 2000; Steptoe et al. 2003c). Many aspects of the ways inwhich cortisol dysfunction affects disease risk are still poorly understood. The bio-logical actions of cortisol depend not only on output, but on the ability of cortisolto bind to glucocorticoid receptors. The relative importance of changes in secretionas opposed to alterations in uptake and tissue clearance is unclear, and impairedsignalling may prove critical in many situations (Raison and Miller 2003).

Cortisol can be assessed in blood, urine and saliva. For many years, urine was thepreferred vehicle for non-invasive assessments, and it provides an integrated measureof the secretion of cortisol and its metabolites over several hours. For example,increased secretion of urinary cortisol metabolites collected over a 24-hour periodwas recently observed in individuals with the metabolic syndrome (Brunner et al.2002). But salivary cortisol has become the method of choice in much psycho-biological research for a number of reasons (Kirschbaum and Hellhammer 2000):

• First, the method of collection is less complicated and embarrassing than forurinary measurement. The individual spits into a test tube, or gently chews adental roll for a couple of minutes until it is saturated with saliva.

• Second, samples are stable over several days at room temperature and can be sentthrough the post. This means that samples can be collected by people in theirown homes or at work over several days without an investigator being involved.

• Third, levels in saliva respond quickly, and are more sensitive to psychosocialexperience than are integrated urinary assessments. For example, several studieshave failed to show that urinary cortisol is higher on work than leisure days, andthis has cast doubt on the impact of work stress on neuroendocrine function (e.g.Pollard et al. 1996). But salivary measures indicate higher levels on work days,with differences in the magnitude of secretions associated with work stress factorssuch as low job control (Kunz-Ebrecht et al. 2004a, 2004b).

• Fourth, methods of analysis are relatively straightforward and inexpensive, withestablished techniques using enzyme-linked immunosorbent assay (ELISA),radioimmunoassay, and immunofluorescence being available in many laborator-ies. The level of free cortisol in saliva is only a fraction of that in blood, but itnevertheless correlates highly with plasma and serum concentrations.

Cortisol shows a pronounced circadian rhythm. Figure 20.2 summarizes dataaveraged from 163 middle-aged men and women over a working day. Cortisollevels are high early in the morning, and decline over the day, with a small peakassociated with eating lunch. Data can be analysed using point comparisons, aver-ages, or by calculating the slope of decline across the day (Stone et al. 2001). Add-itionally, there is typically an increase in cortisol over the first 20–30 minutes afterwaking. The cortisol waking response is of great interest in itself, since it has beenassociated with several psychosocial factors (Clow et al. 2004). In particular, it hasbeen shown to be larger in more depressed individuals, those experiencing chroniclife stress, and in more lonely people (Pruessner et al. 2003; Steptoe et al. 2004). It isinversely associated with socioeconomic position, and positively with work stress(Kunz-Ebrecht et al. 2004a). The evidence is inconsistent about whether it is associ-ated with time of waking or quality of sleep. Cortisol levels can be affected by


smoking, medication, use of oral contraceptives, food consumption and body massindex, so these factors need to be controlled. The assessment of cortisol iscompromised in people taking corticosteroid medication.

Experimental studies of acute stress responses have shown increases in cortisolfollowing demanding tasks such as simulated public speaking (Biondi and Picardi1999). However, it is striking that cortisol increases are frequently not observed withbehavioural challenges that elicit responses in blood pressure and other variables.The reason may be that stimulation needs to be quite intense to produce reliablecortisol responses. The time course of cortisol increases is also longer than that ofmany biological variables, so that peak increases may not emerge until 20–40 min-utes following the onset of stimulation. Increased cortisol responsivity to laboratorystress has been positively correlated with waist/hip ratio (Epel et al. 2000) and withthe experience of chronic stress (Biondi and Picardi 1999), though some studieshave described attenuated responses in stressed groups (Kristenson et al. 1998).Evidence from animal research indicates that adversity in early life promotes corti-costeroid stress responsivity in adult animals (Meaney 2001), but evidence inhumans is limited to date (Heim et al. 2000b). In contrast, lower than normalsalivary cortisol responses to challenging tasks have been recorded from childrenwith autoimmune conditions such as atopic dermatitis (Buske-Kirschbaum et al.1997).

DHEA and DHEA sulphate

DHEA and its sulphate are the most abundant steroid hormones in the humancirculation. They have marked effects on the central nervous system, having anantiglucocorticoid function, and an impact on memory and emotional behaviour inanimal studies (Wolf and Kirschbaum, 1999). Animal studies also indicate that

Figure 20.2 Mean levels of free cortisol sampled in saliva from 163 middle-agedmen and women over a working day and evening. Timing of samples is indicatedalong the horizontal axis. Error bars are 95 per cent confidence intervals.Source: Steptoe et al. (2003c)


DHEA administration reduces body fat and risk of diabetes mellitus and coronaryheart disease, while enhancing immune function (Herbert 1995). The concentra-tion of DHEA and DHEA sulphate increases during adolescence, reaching a peakbetween ages 20 and 30. It then declines markedly with age, such that levels inpeople aged 70–80 years are less than 20 per cent of those in young adults. Thereare, however, important individual differences in this age-related decline, and peoplewith smaller decrements have less deterioration in memory and functional capacityin old age (Berkman et al. 1993). Low DHEA predicted increased mortality in theRancho Bernardo study and in a French cohort of men, but not women (Mazat etal. 2001). Some investigators argue that the ratio of cortisol to DHEA is crucial.Low DHEA and an elevated cortisol/DHEA ratio has been reported in depressedpatients (Young et al. 2002). Research by Goodyer et al. (2001) has shown thatdepressive disorder in adolescence is associated with high cortisol and low DHEA,and that these hormones predict subsequent depression in high-risk individualsindependently of life events, long-term difficulties and premorbid symptom levels.Low DHEA is a component of allostatic load, which has been shown to predictmortality and functional decline in the MacArthur studies of successful ageing(Seeman et al. 2001). All these findings might suggest that DHEA administrationwould reverse age-related decline in function. Unfortunately, trails of DHEAsupplementation in old age have yet to demonstrate that effects of low levels can bereversed (Huppert and Van Niekerk 2001).

DHEA and its sulphate can be measured in blood and saliva. There is a diurnalrhythm in output, but this is less marked than for cortisol. Many of the same factorsthat are relevant for cortisol collection and interpretation apply to DHEA.

Catecholamines

The catecholamines noradrenaline (norepinephrine), adrenaline (epinephrine) anddopamine are released from nerve terminals and the adrenal glands into the blood,urine and cerebrospinal fluid. Most noradrenaline is produced by sympatheticnerves and activates adrenergic receptors locally, so only a small fraction enters thebloodstream. Adrenaline by contrast is produced by the adrenal medulla and acts as ahormone, being distributed by the bloodstream to adrenergic receptors in manyareas. Catecholamines are measured as indices of sympathoadrenal activity in psy-chosocial biology, and can be assessed at present in blood and urine but not in saliva.

Urinary measures of catecholamines and their metabolites can be used to providean indication of sympthoadrenal activity, and have demonstrated associations withpsychosocial factors such as work stress (Frankenhaeuser et al. 1989), and withprecursors of disease such as the metabolic syndrome (Brunner et al. 2002). Urinaryconcentrations of adrenaline and noradrenaline collected over 12 hours were com-ponents of the index of allostatic load used to predict functional decline in theMacArthur studies of successful ageing (Karlamangla et al. 2002). But studies ofresponsivity to psychosocial stimuli require measures that change over shorter timeperiods.

Plasma noradrenaline and adrenaline are often measured for this purpose, yet faceseveral technical and conceptual problems. The most accurate assessments arecarried out with high-pressure liquid chromatography, but reproducibility and sen-sitivity is far from perfect. A more important point is that the noradrenaline concen-tration measured in venous blood samples is strongly determined by overflow fromlocal tissues (Hjemdahl 1993). Thus about half of the catecholamine extracted fromblood samples taken from the forearm is derived from the muscles of the arm itself.Sympathetic nervous system activity is highly differentiated, particularly understressful conditions, with levels of activity in different tissues varying widely. This


means that venous samples are not good indicators of general ‘sympathetic tone’.This problem has yet to be resolved. More precise measures of sympathetic activityinclude direct microneurographic assessments from nerves, and measures of radio-actively-labelled noradrenaline. For example, Esler et al. (1989) showed that mentalstress elicited no change in venous noradrenaline, while radiotracer techniquesdemonstrated large increases in noradrenergic activity in the heart and kidney.Unfortunately, these are technically demanding methods that cannot be widelyused in health care research at present (Grassi and Esler 1999).

Other hormones

Insulin-like growth factor-1 (IGF-1) is produced in response to growth hormonereleased by the pituitary gland. IGF-1 has a variety of effects in the central nervoussystem, being involved in the differentiation of neurons, the release of neuro-transmitters and stimulation of dendritic growth (Schneider et al. 2003). Thedecrease in hippocampal neurogenesis with advancing age is mediated by reducedIGF-1 among other factors. Both cross-sectional and prospective studies haveshown positive associations between IGF-1 and cognitive function. Age-relateddeclines in IGF-1 are associated with deleterious changes in body composition,while studies of growth hormone administration in older adults have documentedrises in IGF-1 together with increased bone mineral density content and reducedfat mass (Lamberts et al. 1997). But effects on disease are poorly understood.Numerous studies have shown adverse effects of IGF-1 on atherosclerosis, glucosemetabolism and diabetic vascular lesions, and high IGF-1 is associated withincreased risk of some cancers (Hankinson et al. 1998; Sandhu et al. 2002). On theother hand, low rather than high circulating IGF-1 has been related prospectivelywith the incidence of coronary heart disease (Juul et al. 2002).

The gonadal hormones testosterone and oestrogen have also been studied inhuman psychosocial research. There is ample evidence from animal studies thatstressful conditions and social adversity impair gonadal function and reproductiveviability. Testosterone levels have been positively related to aggression and violentbehaviour in some but not all investigations, and sustained production of testoster-one in older men is associated with maintenance of muscle strength and functionalwell-being (Lamberts et al. 1997; Charney 2004). There is evidence that oestrogenblunts HPA axis and catecholamine stress responses in perimenopausal women(Komesaroff et al. 1999), and effects on mood have been described (McEwen 2002).Additionally, Taylor et al. (2000) have proposed that oxytocin is released as part ofthe stress response in women, and may mediate sex-specific coping behaviours inthe face of adversity.

Cardiovascular measures

Blood pressure

The need for rigour in the measurement procedure and standardization of condi-tions when measuring blood pressure cannot be overemphasized. At the least, tworeadings should be obtained, preferably with a validated electronic blood pressuremonitor, after the person has been seated quietly in a chair for five minutes, withfeet on the floor and arm supported at heart level. An important issue for psycho-social studies is the ‘white coat’ effect, or the tendency of some people to havehigher blood pressure when measured by a physician in the clinic than at home.White coat hypertension, where the blood pressure is sufficiently elevated whenmeasured clinically to warrant a diagnosis, is thought to occur in at least 20 per cent


of patients (Pickering et al. 1988). There is little evidence for white coat effectsbeing associated with particular physiological profiles, and exaggerated responsivityto other stimuli is often not present. It is possible that white coat effects are specificconditioned responses, with the negative emotional concomitants of measurementin the clinic leading to a classically conditioned response in some individuals. Oneway of overcoming this difficulty in psychosocial studies is to carry out self-monitoring or ambulatory blood pressure monitoring, as described below.

There is an extensive literature concerning blood pressure responses to mentalstress testing, and the use of this method to investigate the role of psychologicalstress in the development of essential hypertension and coronary heart disease (Step-toe 1997). This has stimulated detailed methodological examinations of such issuesas the duration of baselines, the blood pressure measurement technique, the controlof physical activity and the effects of smoking and hormonal status, that go wellbeyond the scope of this chapter (Schneiderman et al. 1989). The clinical signifi-cance of acute blood pressure stress responses has been a particular concern, and hasled to the gradual accumulation of prospective data that tend to confirm thatheightened stress responsivity predicts future hypertension and cardiovascular dis-ease risk (Steptoe and Willemson 2002). New technologies, such as the Finapresinstrument which assesses blood pressure on a beat by beat basis from the finger, areproviding more refined insights into the ways in which psychosocial factors affectthe cardiovascular system.

Ambulatory blood pressure and self-measurement

Ambulatory monitoring devices consist of an arm cuff, a signal detection deviceand a portable pump. They work in the same way as standard blood pressure moni-tors, except that the equipment is miniaturized and portable, and can be wornbeneath clothing. The cuff is programmed to inflate periodically, with intervals of15–60 minutes being used in different studies, so that a profile of blood pressureover the day can be built up. Ambulatory monitors are often used at night as well,and provide useful information about blood pressure ‘dipping’ and subsequentmorning surges (Kario et al. 2003).

Clinically, the use of ambulatory monitors has increased because of the evidencethat they provide better risk assessments than do conventional methods (Pickering2002). In psychosocial studies, ambulatory devices allow the impact of daily lifeexperience to be measured, and effects are frequently observed that are not detect-able with conventional methods. A good example is in the study of work stress. Therelationship between clinical blood pressure and work stress is very inconsistent, butit has now been found in several investigations that work stress is associated withsmall elevations in blood pressure recorded over the working day (Steenland et al.2000). Figure 20.3 illustrates results from a recent investigation of 200 male andfemale participants in the Whitehall II study who carried out ambulatory bloodpressure monitoring every 20 minutes over a working day. Low control at work wasassociated with higher systolic and diastolic blood pressure both during the day andin the evening after work, and these effects were independent of age, gender,socioeconomic position, smoking, body mass, and physical activity (Steptoe andWillemsen in press). Ambulatory blood pressure has been related to many otherpsychosocial factors as well, including marital conflict, social support, hostility andmood (Schwartz et al. 1994; Baker et al. 1999). Care has to be taken in the interpret-ation of ambulatory results, since physical activity, smoking, consumption of coffeeand alcohol can all affect blood pressure values. Multilevel modelling has beenintroduced into the analysis of ambulatory blood pressure in order to tease out theindependent contribution of psychosocial factors (Schwartz et al. 1994).


Self-measurement of blood pressure is a less expensive method, and has increasedgreatly in popularity with the availability of reliable instruments. It has beenendorsed in the guidelines of the respected Joint National Committee on preven-tion, detection, evaluation and treatment of high blood pressure as an aid to man-agement (Chobanian et al. 2003), and can equally well be used in psychosocialstudies. For example, Evans and Steptoe (2001) carried out a study in which nursesand accountants performed self-measurement of blood pressure and heart rate onwork and leisure days. Blood pressure was typically lower on leisure days, whileheart rates were reduced in people who had high social support.

Figure 20.3 Mean systolic pressure (upper graph) and diastolic pressure (lowergraph) in men and women reporting low job control (solid lines) and high jobcontrol (dotted lines), over the working day. Data are averaged into four timeperiods over the day, and are adjusted for gender, employment grade, age, body massindex, smoking status and concomitant physical activity. Error bars are standarderror of the mean.Source: Steptoe and Willemson (in press)


Heart rate variability

Heart rate variability has become an important indicator of autonomic nervoussystem control over the heart. The heart rate varies naturally with the breathingcycle and other modulators. Variability can be assessed in the time domain bymeasuring the difference between short and long interbeat intervals, or in thefrequency domain with power spectrum analysis. Low heart rate variability is amarker of reduced parasympathetic and high sympathetic control, and is a predictorof all cause mortality, and the prognosis of coronary heart disease and diabetes(Bigger et al. 1993; Dekker et al. 2000). Heart rate variability decreases with age, andthe maintenance of higher heart rate variability may be an indicator of successfulageing. Studies in healthy populations have shown that heart rate variability isreduced during acute stress, and is associated with factors such as social isolation andpsychological distress (Hemingway et al. 2001). It has been postulated that reducedheart rate variability partly mediates the association between depression and acutemyocardial infarction (Carney et al. 2001), and low heart rate variability has alsobeen observed in the metabolic syndrome (Brunner et al. 2002). It should, however,be pointed out that heart rate and heart rate variability are typically inverselycorrelated. It is unclear in some psychosocial studies whether the more complicatedvariability measures provide sufficient additional information over and above thatobtained with heart rate assessments to warrant the extra work required.

Measures of inflammation

Inflammatory markers have been introduced into psychosocial studies only inrecent years. The measures assessed include C-reactive protein, fibrinogen and theproinflammatory cytokines IL-6 and tumour necrosis factor a (TNFα). Inflamma-tion is involved in many of the diseases studied in health care research, includingcoronary heart disease, hypertension, diabetes, some cancers, rheumatoid condi-tions, osteoporosis, multiple sclerosis and periodontal disease. Inflammatory markershave been associated with all cause mortality and death from cardiovascular diseasein several cohort studies (Danesh et al. 1998; Ridker et al. 2000). Proinflammatorycytokines regulate the major drivers of postnatal growth, namely growth hormoneand IGF-1. They are also relevant to ageing and overproduction of IL-6 is associatedwith frailty and functional decline in old age. Cytokines play a major role ininducing sickness behaviour (Dantzer 2001), and may be responsible for the mul-tiple symptoms associated with certain treatments for cancer (Cleeland et al.2003).

There is growing evidence for an influence of psychosocial factors on inflamma-tory processes. Fibrinogen concentration is inversely associated with socio-economic status, and has also been related to low control at work, effort-rewardimbalance, social isolation and hostility (Brunner et al. 1996; Wamala et al. 1999).Chronic stressors such as caring for a dementing relative stimulate more rapidincreases in IL-6 with age (Kiecolt-Glaser et al. 2003b). Clinical depression andnon-clinical depressed mood have been associated with higher IL-6, TNFα and C-reactive protein concentrations (Penninx et al. 2003). The magnitude of IL-6 andC-reactive protein responses to the acute stress of surgery is inversely related to rateof recovery, while chronic stress has been found to predict the magnitude of localinflammation in wounds (Glaser et al. 1999). Several other forms of stress such asacademic examination stimulate heightened inflammatory cytokines and C-reactive protein, and the IL-6 response to acute mental stress is greater in people oflower socioeconomic status (Brydon et al. 2004). Inflammatory processes are alsoaffected by body weight, smoking, physical activity and alcohol consumption, and


interactions between adiposity and depression in the prediction of IL-6 have beendescribed (Miller et al. 2003).

Inflammatory markers are assessed with relatively standard biochemical assays ofblood samples, and commercial kits are available. In the quantification of C-reactiveprotein, so-called high sensitivity assays are required. Plasma or serum levels of thesesubstances are typically measured, but in addition, inflammatory cytokines can bequantified by stimulating the mononuclear cells with the mitogen lipopolysac-charide. This procedure requires culturing cells and incubation of samples, so ismore complicated than the assessment of plasma levels.

Immunological measures

The interdisciplinary field of psychoneuroimmunology has become very promin-ent over the last 25 years (Ader et al. 2001). Much of this research is focused on basicphysiology, the regulation of the immune system, and on the interplay between thebrain and immune responses. But psychoneuroimmunology is also relevant tohealth care research for a number of reasons. First, the evidence indicates thatpsychosocial factors can modulate immune defences against pathogens such asbacteria, fungi and viruses, and can therefore increase or reduce resistance todisease (Kiecolt-Glaser et al. 2002). Second, there are disorders in which disturb-ances of immune function are a central feature, notably HIV/AIDS. Psychologicalcharacteristics and social factors are thought to be responsible in part for the largevariations in the rate of progression of HIV in different individuals, and theseinfluences may be mediated by central nervous system factors (Antoni 2003). Third,variations in clinical status, exacerbations of control, or flares in conditions such asrheumatoid arthritis, diabetes mellitus and systemic lupus erythematosus areassociated with psychosocial factors (Da Costa et al. 1999; Zautra et al. 1999).Psychoneuroimmunological pathways may be responsible for these effects.

Measurement of immune function

The measurement of immune function in psychosocial research is complicated, andonly a brief summary is given here (Vedhara et al. 1999b; Ader et al. 2001). Perhapsthe most basic method is to count white blood cells and sub-sets of lymphocytes inthe circulation. This is done using flow cytometry, with specific antigens detectingcluster designation (CD) markers on different immune cells. For example, cellsmarked with CD4+ are T helper cells, CD8+ are cytotoxic T cells, CD19+ is a Bcell marker and CD16/CD56 designates natural killer cells. Lymphocyte counts aremodified by acute and chronic stress, and in conditions such as depression (Zorrillaet al. 2001). A limitation to this method is that cell numbers do not necessarilycorrelate with cell function. Various functional assays are therefore used. Theseinclude measures of cytotoxicity, in which the ability of natural killer cells or T cellsto destroy target cells is measured. Another functional measure is the assessment oflymphocyte proliferation in response to stimuli such as the lectin phytohemag-glutinin, where greater proliferation implies that the lymphocytes are more respon-sive to the presence of potentially dangerous substances. An early finding inpsychoneuroimmunology was that proliferative responses were impaired in peoplewho had recently suffered a personal bereavement, a group that is known to be atraised risk for various illnesses (Bartrop et al. 1977). Natural killer cell cytotoxicityhas been shown to be impaired in a range of stressful conditions, but to be positivelyassociated with social support (Uchino et al. 1996; Zorrilla et al. 2001).

A different set of assays are used to quantify humoral immunity and levels ofimmunoglobulin. Antibodies are produced by B cells in response to exposure to


antigens, and are grouped in major classes such as immunoglobulin A (IgA), whichis secreted in saliva and tears and protects mucosal surfaces, and immunoglobulin E(IgE) which is involved in the release of histamine and is implicated in conditionssuch as bronchial asthma. Immunoglobulins can be assayed from blood or salivausing ELISA techniques, and levels have been found to vary with psychosocialfactors such as mood and daily stressors. However, the interpretation of responsescan be problematic. Some investigators assess total immunoglobulin levels, but thesevalues are only moderately associated with specific antibody responses. Antibodytitres in response to specific antigens are more precise, but depend on recentexposure to the antigen.

Other methods assess immune function in vivo, rather than through processingof blood or saliva samples in the laboratory. Administration of live virus has beenused in experimental research on upper respiratory infection. In a series of studies,Cohen et al. (1991, 1997) demonstrated that stress increases the likelihood that amoderate dose of experimentally administered coronavirus will lead to infectionand a clinical cold, while social networks have a protective effect. Another modelinvolves measurement of antibody responses to attenuated viruses such as thoseused in influenza vaccination. It has been shown in some studies that antibodyresponses to vaccination are impaired in individuals experiencing chronic lifestress, and this may be indicative of weakened immune defences (Vedhara et al.1999a).

Musculoskeletal measures

There is substantial evidence linking psychosocial aspects of work with upperextremity musculoskeletal problems (Sauter and Moon 1996). Factors such as lowautonomy, lack of role clarity, low job satisfaction and high work pressure have beenassociated with pain in the neck and shoulder regions, and with hand or wristproblems. A recent systematic review confirmed these observations, but alsopointed to evidence relating upper extremity problems with psychosocial factorsoutside work such as low social support and general stress (Bongers et al. 2002).Work-related upper extremity disorders are particularly common in jobs with astatic load involving monotonous and repetitive tasks, even when physical demandsare only low or moderate. Computer data entry, cashier work in supermarkets andother outlets, routine scientific bench work and traditional assembly line work allhave these characteristics, and musculoskeletal problems are particularly commonamong women (Klumb and Lampert 2004).

Much research on musculoskeletal disorders is based on self-report or physicalexamination. However, direct measurement of muscle tension using surface elec-tromyography (EMG) provides valuable additional information. Miniaturizedtransducers and telemetric equipment are available that allow readings to beobtained from free-moving individuals. Positive correlations have been reportedbetween objectively assessed muscle tension and feelings of stress and exhaustionduring work, but correlations with pain are often not obtained (e.g. Rissen et al.2000). Studies of this kind indicate that the perception and appraisal of muscletension may be important as well as objective differences in tension. Work withchronic back pain patients has identified problems in the accuracy of discriminationof EMG levels recorded from back muscles, and a tendency to overestimate muscletension by some individuals (Flor et al. 1999). Surface electromyography is also usedextensively in headache research, with monitoring of muscles of the neck, back andforehead. A meta-analysis of studies of frontal (forehead) EMG indicated thatpatients with tension-type headache do have higher muscle tension than controlson average, but with wide variability (Wittrock 1997). This suggests that a


combined assessment of subjective and objective measures may be valuable in theinvestigation of these problems.

Conclusions

There are many applications of biological measures in health care research. Theiruse in psychosocial studies is particularly appealing, since they provide objectiveevidence for the influence of psychological and social processes on pathophysiologyand the biological mechanisms underlying disease states. There is sometimesscepticism about whether psychosocial factors are associated with genuine healthoutcomes, or only with symptom complaints and illness behaviour (Watson andPennebaker 1989). The evidence that alterations in neuroendocrine function,inflammatory responses, viral antibodies or cardiovascular activity take place allowseffects to be verified objectively. The field of psychosocial biology is continuallyexpanding, with recent research on phenomena such as vascular endothelial dys-function, cytokine gene expression and glucocorticoid receptor sensitivity sup-plementing the more established measures described here. Great care needs to betaken in the measurement and interpretation of biological measures, and their inclu-sion in a health care research project may call for additional standardization both ofthe setting and timing of assessments. However, the gain in understanding ofpsychosocial influences can be considerable, so the extra costs and effort requiredfor including biological measures is well worthwhile.

• The biological measures assessed in psychosocial research are: biologicalKeypoints indicators of disease states; biological markers of processes involved in the aeti-

ology of disease; and non-specific biological markers of stress-related activationor resistance to disease.

• There is a substantial amount of research on the relationship betweenpsychosocial factors and biological function.

• Epidemiological studies provide the core method of establishing the contribu-tion of psychosocial factors to the development of disease, and are also used toidentify the biological mediators of these associations. However, biological meas-ures in epidemiological studies have the limitation that they are generallyrecorded on a single occasion, under resting conditions that are not typical ofeveryday life.

• Naturalistic monitoring studies involve sampling biological variables duringeveryday life, from recordings during challenging tasks to repeated measures (e.g.of blood pressure over an ordinary day).

• Naturalistic monitoring methods have the advantage of ecological validity,although the range of biological markers that can be assessed is relatively small,and need to be relatively unobtrusive.

• Another method is mental stress testing, involving monitoring biologicalresponses to standardized psychological or social stimuli. A wide range of mentalstress tests are employed, and experimental designs can be used. Limitations arethat the stimuli in mental stress testing are often divorced from everyday life andare brief, so only acute biological responses are recorded.


• Biological measures are frequently used in intervention studies to assess theimpact of lifestyle or cognitive-behavioural treatments on disease processes,although intervention methods have not been extensively used to test causalmodels.

• Some of the principle biological variables measured in psychosocial studiesinclude neuroendocrine function, cardiovascular activity, inflammatory processes,immune function and musculoskeletal activity.

• Care needs to be taken in the measurement and interpretation of biologicalmeasures, and in the design of the overall study.

References

Ader, R., Felten, D.L. and Cohen, N. (eds) (2001) Psychoneuroimmunology, 3rd edn, 2 vols. SanDiego, CA: Academic Press.

Antoni, M.H. (2003) Stress management effects on psychological, endocrinological,and immune functioning in men with HIV infection: empirical support for apsychoneuroimmunological model, Stress, 6: 173–88.

Baker, B., Helmers, K., O’Kelly, B. et al. (1999) Marital cohesion and ambulatory bloodpressure in early hypertension, American Journal of Hypertension, 12: 227–30.

Bartrop, R.W., Luckhurst, E., Lazarus, L., Kiloh, L.G. and Penny, R. (1977) Depressedlymphocyte function after bereavement, Lancet, 1: 834–6.

Berkman, L.F., Seeman, T.E., Albert, M. et al. (1993) High, usual and impaired functioning incommunity-dwelling older men and women: findings from the MacArthur FoundationResearch Network on Successful Aging, Journal of Clinical Epidemiology, 46: 1129–40.

Bigger, J.T., Fleiss, J.L., Rolnitzky, L.M. and Steinman, R.C. (1993) The ability of severalshort-term measures of RR variability to predict mortality after myocardial infarction,Circulation, 88: 927–34.

Biondi, M. and Picardi, A. (1999) Psychological stress and neuroendocrine function inhumans: the last two decades of research, Psychotherapy Psychosomatics, 68: 114–50.

Bjorntorp, P. (2001) Do stress reactions cause abdominal obesity and comorbidities?, ObesityReview, 2: 73–86.

Bongers, P.M., Kremer, A.M. and ter Laak, J. (2002) Are psychosocial factors risk factorsfor symptoms and signs of the shoulder, elbow, or hand/wrist?: a review of theepidemiological literature, American Journal of Industrial Medicine, 41: 315–42.

Brunner, E., Davey Smith, G., Marmot, M. et al. (1996) Childhood social circumstances andpsychosocial and behavioural factors as determinants of plasma fibrinogen, Lancet, 347:1008–13.

Brunner, E.J., Marmot, M.G., Nanchahal, K. et al. (1997) Social inequality in coronaryrisk: central obesity and the metabolic syndrome: evidence from the Whitehall II study,Diabetologia, 40: 1341–9.

Brunner, E.J., Hemingway, H., Walker, B.R. et al. (2002) Adrenocortical, autonomic, andinflammatory causes of the metabolic syndrome: nested case-control study, Circulation,106: 2659–65.

Brydon, L., Edwards, S., Mohamed-Ali, V. and Steptoe, A. (2004) Socioeconomic status andstress-induced increases in interleukin-6, Brain, Behavior and Immunity, 18: 281–90.

Burg, M.M., Jain, D., Soufer, R., Kerns, R.D. and Zaret, B.L. (1993) Role of behavioral andpsychological factors in mental stress-induced silent left ventricular dysfunction incoronary artery disease, Journal of the American College of Cardiology, 22: 440–8.

Buske-Kirschbaum, A., Jobst, S., Wustmans, A. et al. (1997) Attenuated free cortisol responseto psychosocial stress in children with atopic dermatitis, Psychosomatic Medicine, 59: 419–26.

Cacioppo, J.T., Tassinary, L.G. and Berntson, G. (eds) (2000) Handbook of Psychophysiology, 2ndedn. New York: Cambridge University Press.

Carney, R.M., Blumenthal, J.A., Stein, P.K. et al. (2001) Depression, heart rate variability, andacute myocardial infarction, Circulation, 104: 2024–8.


Carroll, D., Ring, C., Hunt, K., Ford, G. and Macintyre, S. (2003) Blood pressure reactions tostress and the prediction of future blood pressure: effects of sex, age, and socioeconomicposition, Psychosomatic Medicine, 65: 1058–64.

Charney, D.S. (2004) Psychobiological mechanisms of resilience and vulnerability:implications for successful adaptation to extreme stress, American Journal of Psychiatry, 161:195–216.

Chobanian, A.V., Bakris, G.L., Black, H.R. et al. (2003) The Seventh Report of the JointNational Committee on Prevention, Detection, Evaluation, and Treatment of High BloodPressure: the JNC 7 report, Journal of the American Medical Association, 289: 2560–72.

Cleeland, C.S., Bennett, G.J., Dantzer, R. et al. (2003) Are the symptoms of cancer and cancertreatment due to a shared biologic mechanism? A cytokine-immunologic model of cancersymptoms, Cancer, 97: 2919–25.

Clow, A., Thorn, L., Evans, P. and Hucklebridge, F. (2004) The awakening cortisol response:methodological issues and significance, Stress, 7: 29–37.

Cohen, S., Tyrrell, D.A.J. and Smith, A.P. (1991) Psychosocial stress and susceptibility to thecommon cold, New England Journal of Medicine, 325: 606–12.

Cohen, S., Doyle, W.J., Skoner, D.P., Rabin, B.S. and Gwaltney, J.M. (1997) Social ties andsusceptibility to the common cold, Journal of the American Medical Association, 277: 1940–4.

Da Costa, D., Dobkin, P.L., Pinard, L. et al. (1999) The role of stress in functional disabilityamong women with systemic lupus erythematosus: a prospective study, Arthritis CareResearch, 12: 112–19.

Danesh, J., Collins, R., Appleby, P. and Peto, R. (1998) Association of fibrinogen, C-reactiveprotein, albumin, or leukocyte count with coronary heart disease: meta-analyses ofprospective studies, Journal of the American Medical Association, 279: 1477–82.

Dantzer, R. (2001) Cytokine-induced sickness behavior: mechanisms and implications,Annals of the New York Academy of Sciences, 933: 222–34.

Dekker, J.M., Crow, R.S., Folsom, A.R. et al. (2000) Low heart rate variability in a 2-minuterhythm strip predicts risk of coronary heart disease and mortality from several causes: theARIC Study – atherosclerosis risk in communities, Circulation, 102: 1239–44.

Epel, E.S., McEwen, B., Seeman, T. et al. (2000) Stress and body shape: stress-induced cortisolsecretion is consistently greater among women with central fat, Psychosomatic Medicine, 62:623–32.

Esler, M., Jennings, G. and Lambert, G. (1989) Measurement of overall and cardiac norepin-ephrine release into plasma during cognitive challenge, Psychoneuroendocrinology, 14:477–81.

Evans, O. and Steptoe, A. (2001) Social support at work, heart rate, and cortisol: aself-monitoring study, Journal of Occupational Health Psychology, 6: 361–70.

Everson, S.A., Lynch, J.W., Chesney, M.A. et al. (1997) Interaction of workplace demands andcardiovascular reactivity in progression of carotid atherosclerosis: population-based study,British Medical Journal, 314: 553–8.

Fink, G. (ed.) (2000) Encyclopedia of Stress, 3 vols. San Diego, CA: Academic Press.Flor, H., Furst, M. and Birbaumer, N. (1999) Deficient discrimination of EMG levels and

overestimation of perceived tension in chronic pain patients, Applied Psychophysiology andBiofeedback, 24: 55–66.

Frankenhaeuser, M., Lundberg, U., Fredrikson, M. et al. (1989) Stress on and off the job asrelated to sex and occupational stress in white-collar workers, Journal of OrganizationalBehavior, 10: 321–46.

Glaser, R., Kiecolt-Glaser, J.K., Marucha, P.T. et al. (1999) Stress-related changes inproinflammatory cytokine production in wounds, Archives of General Psychiatry, 56:450–6.

Goodyer, I.M., Park, R.J., Netherton, C.M. and Herbert, J. (2001) Possible role of cortisol anddehydroepiandrosterone in human development and psychopathology, British Journal ofPsychiatry, 179: 243–9.

Grassi, G. and Esler, M. (1999) How to assess sympathetic activity in humans, Journal ofHypertension, 17: 719–34.

Hankinson, S.E., Willett, W.C., Colditz, G.A. et al. (1998) Circulating concentrations ofinsulin-like growth factor-I and risk of breast cancer, Lancet, 351: 1393–6.

Heim, C., Ehlert, U. and Hellhammer, D.H. (2000a) The potential role of hypocortisolism inthe pathophysiology of stress-related bodily disorders, Psychoneuroendocrinology, 25: 1–35.


Heim, C., Newport, D.J., Heit, S. et al. (2000b) Pituitary-adrenal and autonomic responses tostress in women after sexual and physical abuse in childhood, Journal of the American MedicalAssociation, 284: 592–7.

Hemingway, H., Malik, M. and Marmot, M. (2001) Social and psychosocial influences onsudden cardiac death, ventricular arrhythmia and cardiac autonomic function, EuropeanHeart Journal, 22: 1082–1101.

Hemingway, H., Shipley, M., Mullen, M.J. et al. (2003) Social and psychosocial influences oninflammatory markers and vascular function in civil servants (the Whitehall II study),American Journal of Cardiololgy, 92: 984–7.

Herbert, J. (1995) The age of dehydroepiandrosterone, Lancet, 345: 1193–4.Hjemdahl, P. (1993) Plasma catecholamines: analytical challenges and physiological

limitations, Baillieres Clinical Endocrinology Metabolism, 7: 307–53.Huppert, F.A. and Van Niekerk, J.K. (2001) Dehydroepiandrosterone (DHEA) supplementa-

tion for cognitive function, Cochrane Database Systematic Reviews: CD000304.Jarrett, D.B., Greenhouse, J.B., Thompson, S.B. et al. (1984) Effect of nocturnal intravenous

cannulation upon sleep-EEG measures, Biological Psychiatry, 19: 1537–50.Jorgensen, R.S., Johnson, B.T., Schreer, G.E. and Kolodziej, M.E. (1996) Elevated

blood pressure and personality: a meta-analytic review, Psychological Bulletin, 120:293–320.

Juul, A., Scheike, T., Davidsen, M., Gyllenborg, J. and Jorgensen, T. (2002) Low serum insulin-like growth factor I is associated with increased risk of ischemic heart disease: a popula-tion-based case-control study, Circulation, 106: 939–44.

Kamarck, T.W., Debski, T.T. and Manuck, S.B. (2000) Enhancing the laboratory-to-lifegeneralizability of cardiovascular reactivity using multiple occasions of measurement,Psychophysiology, 37: 533–42.

Kaplan, J.R., Pettersson, K., Manuck, S.B. and Olsson, G. (1991) Role of sympathoadrenalmedullary activation in the initiation and progression of atherosclerosis, Circulation, 84:VI23–32.

Kario, K., Pickering, T.G., Umeda, Y. et al. (2003) Morning surge in blood pressure as apredictor of silent and clinical cerebrovascular disease in elderly hypertensives: a prospect-ive study, Circulation, 107: 1401–6.

Karlamangla, A.S., Singer, B.H., McEwen, B.S., Rowe, J.W. and Seeman, T.E. (2002)Allostatic load as a predictor of functional decline: MacArthur studies of successful aging,Journal of Clinical Epidemiology, 55: 696–710.

Kiecolt-Glaser, J.K. and Newton, T.L. (2001) Marriage and health: his and hers, PsychologyBulletin, 127: 472–503.

Kiecolt-Glaser, J.K., McGuire, L., Robles, T.F. and Glaser, R. (2002) Emotions, morbidity, andmortality: new perspectives from psychoneuroimmunology, Annual Review of Psychology,53: 83–107.

Kiecolt-Glaser, J.K., Bane, C., Glaser, R. and Malarkey, W.B. (2003a) Love, marriage, anddivorce: newlyweds’ stress hormones foreshadow relationship changes, Journal of Consultingand Clinical Psychology, 71: 176–88.

Kiecolt-Glaser, J.K., Preacher, K.J., MacCallum, R.C. et al. (2003b) Chronic stress andage-related increases in the proinflammatory cytokine IL-6, Proceedings of the NationalAcademy of Sciences of the USA, 100: 9090–5.

Kirschbaum, C. and Hellhammer, D.H. (2000) Salivary cortisol, in G. Fink (ed.) Encyclopediaof Stress, vol. 3. San Diego, CA: Academic Press.

Klumb, P.L. and Lampert, T. (2004) Women, work, and well-being 1950–2000: a review andmethodological critique, Social Science and Medicine, 58: 1007–24.

Komesaroff, P.A., Esler, M.D. and Sudhir, K. (1999) Estrogen supplementation attenuatesglucocorticoid and catecholamine responses to mental stress in perimenopausal women,Journal of Clinical Endocrinolgy Metabolism, 84: 606–10.

Kristenson, M., Orth-Gomer, K., Kucinskiene, Z. et al. (1998) Attenuated cortisol responseto a standardized stress test in Lithuanian versus Swedish men: the LiVicordia study,International Journal of Behavioral Medicine, 5: 17–30.

Kunz-Ebrecht, S.R., Kirschbaum, C., Marmot, M. and Steptoe, A. (2004a) Differences incortisol awakening response on work days and weekends in women and men from theWhitehall II cohort, Psychoneuroendocrinology, 29: 516–28.

Kunz-Ebrecht, S.R., Kirschbaum, C. and Steptoe, A. (2004b) Work stress, socioeconomic


status, and neuroendocrine activation over the working day, Social Science and Medicine, 58:1523–30.

Lamberts, S.W., van den Beld, A.W. and van der Lely, A.J. (1997) The endocrinology of aging,Science, 278: 419–24.

Light, K.C., Girdler, S.S., Sherwood, A. et al. (1999) High stress responsivity predicts laterblood pressure only in combination with positive family history and high life stress,Hypertension, 33: 1458–64.

Lundberg, U., Dohns, I.E., Melin, B. et al. (1999) Psychophysiological stress responses, muscletension, and neck and shoulder pain among supermarket cashiers, Journal of OccupationalHealth Psychology, 4: 245–55.

Lupien, S.J. and Lepage, M. (2001) Stress, memory, and the hippocampus: can’t live with it,can’t live without it, Behavioural Brain Research, 127: 137–58.

Lupien, S.J., Nair, N.P., Briere, S. et al. (1999) Increased cortisol levels and impaired cognitionin human aging: implication for depression and dementia in later life, Reviews in Neuro-science, 10: 117–39.

Lupien, S.J., King, S., Meaney, M.J. and McEwen, B.S. (2000) Child’s stress hormone levelscorrelate with mother’s socioeconomic status and depressive state, Biological Psychiatry, 48:976–80.

Mazat, L., Lafont, S., Berr, C. et al. (2001) Prospective measurements of dehydroepiandroster-one sulphate in a cohort of elderly subjects: relationship to gender, subjective health,smoking habits, and 10-year mortality, Proceedings of the National Academy of Science, USA,98: 8145–50.

McEwen, B. (2002) Estrogen actions throughout the brain, Recent Progress in HormoneResearch, 57: 357–84.

McEwen, B.S. (1998) Protective and damaging effects of stress mediators, New England Journalof Medicine, 338: 171–9.

McEwen, B.S. and Wingfield, J.C. (2003) The concept of allostasis in biology andbiomedicine, Hormones and Behavior, 43: 2–15.

McEwen, B.S., Biron, C.A., Brunson, K.W. et al. (1997) The role of adrenocorticoids asmodulators of immune function in health and disease: neural, endocrine and immuneinteractions, Brain Research Review, 23: 79–133.

Meaney, M.J. (2001) Maternal care, gene expression, and the transmission of individualdifferences in stress reactivity across generations, Annual Review of Neuroscience, 24:1161–92.

Miller, G.E., Freedland, K.E., Carney, R.M., Stetler, C.A. and Banks, W.A. (2003) Pathwayslinking depression, adiposity, and inflammatory markers in healthy young adults, Brain,Behavior and Immunity, 17: 276–85.

Penninx, B.W., Kritchevsky, S.B., Yaffe, K. et al. (2003) Inflammatory markers and depressedmood in older persons: results from the Health, Aging and Body Composition study,Biological Psychiatry, 54: 566–72.

Pickering, T. (2002) Future developments in ambulatory blood pressure monitoring andself-blood pressure monitoring in clinical practice, Blood Pressure Monitor, 7: 21–5.

Pickering, T.G., James, G.D., Boddie, C. et al. (1988) How common is white coat hyperten-sion? Journal of the American Medical Association, 259: 225–8.

Pollard, T.M., Ungpakorn, G., Harrison, G.A. and Parkes, K.R. (1996) Epinephrine andcortisol responses to work: a test of the models of Frankenhaeuser and Karasek, Annals ofBehavioral Medicine, 18: 229–37.

Pruessner, M., Hellhammer, D.H., Pruessner, J.C. and Lupien, S.J. (2003) Self-reporteddepressive symptoms and stress levels in healthy young men: associations with the cortisolresponse to awakening, Psychosomatic Medicine, 65: 92–9.

Raison, C.L. and Miller, A.H. (2003) When not enough is too much: the role of insufficientglucocorticoid signaling in the pathophysiology of stress-related disorders, American Journalof Psychiatry, 160: 1554–65.

Ridker, P.M., Rifai, N., Stampfer, M.J. and Hennekens, C.H. (2000) Plasma concentration ofinterleukin–6 and the risk of future myocardial infarction among apparently healthy men,Circulation, 101: 1767–72.

Rissen, D., Melin, B., Sandsjo, L., Dohns, I. and Lundberg, U. (2000) Surface EMG andpsychophysiological stress reactions in women during repetitive work, European Journal ofApplied Physiology, 83: 215–22.


Ritz, T. and Steptoe, A. (2000) Emotion and pulmonary function in asthma: reactivity in thefield and relationship with laboratory induction of emotion, Psychosomatic Medicine, 62:808–15.

Sabban, E.L. and Kvetnansky, R. (2001) Stress-triggered activation of gene expression incatecholaminergic systems: dynamics of transcriptional events, Trends in Neurosciences, 24:91–8.

Sandhu, M.S., Heald, A.H., Gibson, J.M. et al. (2002) Circulating concentrations of insulin-like growth factor-I and development of glucose intolerance: a prospective observationalstudy, Lancet, 359: 1740–5.

Sapolsky, R.M., Romero, L.M. and Munck, A.U. (2000) How do glucocorticoids influencestress responses? Integrating permissive, suppressive, stimulatory, and preparative actions,Endocrine Reviews, 21: 55–89.

Sauter, S.L. and Moon, S.D. (eds) (1996) Beyond Biomechanics: Psychosocial Factors and Musculo-skeletal Disorders in Office Work. New York: Taylor & Francis.

Schneider, H.J., Pagotto, U. and Stalla, G.K. (2003) Central effects of the somatotropic system,European Journal of Endocrinology, 149: 377–92.

Schneiderman, N., Weiss, S.M. and Kaufman, P.G. (eds) (1989) Handbook of Research Methodsin Cardiovascular Behavioral Medicine. New York: Plenum.

Schwartz, J.E. and Stone, A.A. (1998) Strategies for analyzing ecological momentary assess-ment data, Health Psychology, 17: 6–16.

Schwartz, J.E., Warren, K. and Pickering, T.G. (1994) Mood, location and physical position aspredictors of ambulatory blood pressure and heart rate: application of a multi-level randomeffects model, Annals of Behavioral Medicine, 16: 210–20.

Seeman, T.E., McEwen, B.S., Rowe, J.W. and Singer, B.H. (2001) Allostatic load as a markerof cumulative biological risk: MacArthur studies of successful aging, Proceedings of theNational Academy of Sicence, USA, 98: 4770–5.

Sluiter, J.K., Frings-Dresen, M.H., Meijman, T.F. and van der Beek, A.J. (2000) Reactivity andrecovery from different types of work measured by catecholamines and cortisol: a system-atic literature overview, Occupational and Environmental Medicine, 57: 298–315.

Steenland, K., Fine, L., Belkic, K. et al. (2000) Research findings linking workplace factors toCVD outcomes, Occupational Medicine, 15: 7–68.

Steptoe, A. (1997) Behavior and blood pressure: implications for hypertension, in A.Zanchetti and G. Mancia (eds) Handbook of Hypertension – Pathophysiology of Hypertension.Amsterdam: Elsevier Science.

Steptoe, A. (1998) Psychophysiological bases of disease, in M. Johnston and D. Johnston (eds)Comprehensive Clinical Psychology Volume 8: Health Psychology. New York: Elsevier Science.

Steptoe, A. and Ayers, S. (in press) Stress, health and illness, in S. Sutton, A. Baum and M.Johnston (eds) Sage Handbook of Health Psychology. London: Sage.

Steptoe, A. and Willemsen, G. (2002) Psychophysiological responsivity in coronary heartdisease, in S.A. Stansfeld and M.G. Marmot (eds) Stress and the Heart. London: BMJ Books.

Steptoe, A. and Willemsen, G. (in press) The influence of low job control on ambulatoryblood pressure and perceived stress over the working day in men and women from theWhitehall II cohort, Journal of Hypertension.

Steptoe, A., Feldman, P.M., Kunz, S. et al. (2002) Stress responsivity and socioeconomic status:a mechanism for increased cardiovascular disease risk?, European Heart Journal, 23: 1757–63.

Steptoe, A., Kunz-Ebrecht, S., Owen, N. et al. (2003a) Influence of socioeconomic status andjob control on plasma fibrinogen responses to acute mental stress, Psychosomatic Medicine,65: 137–44.

Steptoe, A., Kunz-Ebrecht, S., Rumley, A. and Lowe, G.D. (2003b) Prolonged elevationsin haemostatic and rheological responses following psychological stress in lowsocioeconomic status men and women, Thrombosis and Haemostasis, 89: 83–90.

Steptoe, A., Kunz-Ebrecht, S., Owen, N. et al. (2003c) Socioeconomic status and stress-related biological responses over the working day, Psychosomatic Medicine, 65: 461–70.

Steptoe, A., Owen, N., Kunz-Ebrecht, S. and Brydon, L. (2004) Loneliness and neuroendo-crine, cardiovascular, and inflammatory stress responses in middle-aged men and women,Psychoneuroendocrinology, 29: 593–611.

Stone, A.A., Schwartz, J.E., Smyth, J. et al. (2001) Individual differences in the diurnal cycle ofsalivary free cortisol: a replication of flattened cycles for some individuals, Psychoneuroendo-crinology, 26: 295–306.


Taylor, S.E., Klein, L.C., Lewis, B.P. et al. (2000) Biobehavioral responses to stress in females:tend-and-befriend, not fight-or-flight, Psychological Review, 107: 411–29.

Tuomilehto, J., Lindstrom, J., Eriksson, J.G. et al. (2001) Prevention of type 2 diabetes mellitusby changes in lifestyle among subjects with impaired glucose tolerance, New EnglandJournal of Medicine, 344: 1343–50.

Uchino, B.N., Cacioppo, J.T. and Kiecolt-Glaser, J.K. (1996) The relationship between socialsupport and physiological processes: a review with emphasis on underlying mechanismsand implications for health, Psychological Bulletin, 119: 488–531.

Van Cauter, E., Leproult, R. and Plat, L. (2000) Age-related changes in slow wave sleep andREM sleep and relationship with growth hormone and cortisol levels in healthy men,Journal of the American Medical Association, 284: 861–8.

van Eck, M., Berkhof, H., Nicolson, N. and Sulon, J. (1996) The effects of perceived stress,traits, mood states, and stressful daily events on salivary cortisol, Psychosomatic Medicine, 58:447–58.

Vedhara, K., Cox, N.K.M., Wilcock, G.K. et al. (1999a) Chronic stress in elderly carers ofdementia patients and antibody response to influenza vaccination, Lancet, 353: 627–31.

Vedhara, K., Fox, J.D. and Wang, E.C. (1999b) The measurement of stress-related immunedysfunction in psychoneuroimmunology, Neuroscience and Biobehavioral Reviews, 23:699–715.

Wamala, S.P., Murray, M.A., Horsten, M. et al. (1999) Socioeconomic status and determin-ants of hemostatic function in healthy women, Arteriosclerosis, Thrombosis and VascularBiology, 19: 485–92.

Watson, D. and Pennebaker, J.W. (1989) Health complaints, stress, and distress: exploring thecentral role of negative affectivity, Psychological Review, 96: 234–54.

Weiner, H. (1992) Perturbing the Organism: The Biology of Stressful Experience. Chicago:University of Chicago Press.

Wittrock, D.A. (1997) The comparison of individuals with tension-type headache andheadache-free controls on frontal EMG levels: a meta-analysis, Headache, 37: 424–32.

Wolf, O.T. (2003) HPA axis and memory, Best Practice in Research Clinical Endocrinology ofMetabolism, 17: 287–99.

Wolf, O.T. and Kirschbaum, C. (1999) Actions of dehydroepiandrosterone and its sulfate inthe central nervous system: effects on cognition and emotion in animals and humans, BrainResearch Reviews, 30: 264–88.

Yan, L.L., Liu, K., Matthews, K.A. et al. (2003) Psychosocial factors and risk of hypertension:the Coronary Artery Risk Development in Young Adults (CARDIA) study, Journal of theAmerican Medical Association, 290: 2138–48.

Yehuda, R. (2002) Current status of cortisol findings in post-traumatic stress disorder,Psychiatric Clinics of North America, 25: 341–68.

Young, A.H., Gallagher, P. and Porter, R.J. (2002) Elevation of the cortisol-dehydroepiandrosterone ratio in drug-free depressed patients, American Journal ofPsychiatry, 159: 1237–9.

Young, E.A., Tolman, R., Witkowski, K. and Kaplan, G. (2004) Salivary cortisol andposttraumatic stress disorder in a low-income community sample of women, BiologicalPsychiatry, 55: 621–6.

Zautra, A.J., Hamilton, N.A., Potter, P. and Smith, B. (1999) Field research on the relationshipbetween stress and disease activity in rheumatoid arthritis, Annals of the New York Academyof Science, 876: 397–412.

Zorrilla, E.P., Luborsky, L., McKay, J.R. et al. (2001) The relationship of depression andstressors to immunological assays: a meta-analytic review, Brain, Behavior and Immunity, 15:199–226.


tools of psychosocial biology in health care research 20

Documents