Cortisol secretory activity in older people in relation to positive and negative well-being

Summary

Secretion of the hormone cortisol, a physiological correlate of affect, has been studied mostly in relation to negative states, especially stress. By contrast, policy initiatives aimed at older populations now routinely emphasise well-being and a ‘positive ageing’ perspective.

In this study, we examined diurnal salivary cortisol profiles from 50 active seniors recruited into a wider community research project (mean age 74 years; 34 F/16 M). Participants’ wrist activity was continuously monitored by actimeters in their homes over a 48 h period. During this time two diurnal cycles of cortisol data were collected (8 samples per day); with actimeter data providing a compliance check in regard to timing of self-administered saliva collections. Prior to the trial, participants had completed the GHQ-30 which was scored separately to yield both positive and negative well-being scores which matched closely normative data from over 6000 cases in a large survey.

Our data suggest that positive and negative psychological well-being are quite strongly and inversely correlated. However, neither on their own was associated with basal levels of cortisol. Rather, for cortisol secretion in the 45-min period following awakening, but not during the rest of the day, we found a significant interaction between positive and negative well-being (p<0.024). Further analysis of this interaction showed that among participants low on negative well-being, higher positive well-being was significantly associated with lower cortisol; equally, among participants high on positive well-being, lower negative well-being was significantly associated with lower cortisol. Thus, a powerful synergy seemed to be operating in this early morning period such that cortisol secretion was 27% lower in participants with both higher-than-average positive well-being and lower-than-average negative well-being (comprising 34% of the sample). We conclude that cortisol secretion in the first 45 min following awakening is distinct from the rest of the day and most able to discriminate well-being states.

Introduction

It is well established that both laboratory-induced and naturalistic daily stressors activate the hypothalamic–pituitary–adrenal axis (HPA-axis) as evidenced by increased cortisol levels (e.g., Kirschbaum et al., 1993; Ockenfels et al., 1995; Smyth et al., 1998; Hanson et al., 2000; Peeters et al., 2003). However it is clear that individual differences known to contribute to general well-being (such as in self-esteem and affect) can modulate stress-induced elevations in cortisol (Smyth et al., 1998; Pruessner et al., 1999; Polk et al., 2005; Jacobs et al., 2007). The importance of subjective measures of well-being in relation to health outcomes is increasingly being recognised. However, to date relatively little has been published linking measures of well-being and cortisol status. Human well-being is a complex phenomenon and currently there is no single agreed formula for its measurement. In the past, measures of psychological well-being tended to focus upon the presence or absence of negative symptoms such as anxiety or low self-esteem. Recently, more recognition is being given to the importance of positive well-being including components such as autonomy, environmental mastery, personal growth, positive relations, purpose in life and self-acceptance (Ryff, 1989; Ryff and Singer, 1996). Interestingly, psychometric exploration of positive and negative well-being has tended to indicate that they can function somewhat independently and that positive well-being is not simply the absence of negative feelings and experience, or vice versa (Diener and Emmons, 1984; Watson et al., 1988; Huppert and Whittington, 2003).

An important psychometric approach to the measurement of well-being is seen in the work of Huppert and Whittington (2003) where they construct separate positive and negative well-being scales from the 30-item General Health Questionnaire (GHQ-30) (Goldberg, 1972). A particular benefit of using the GHQ-30 for measures of positive and negative well-being in this way derives from its extremely widespread and global use as self-report instrument in medical research. Using this approach it has been shown that although positive and negative well-being are quite strongly and significantly negatively correlated, there is also evidence of some degree of independence. Approximately, 40% of variance is shared, but just over a third of over 6000 participants in a large survey were either above average on both positive and negative well-being or below average on both. The importance of recognising this partial independence of scales is forcefully underlined by a previous report (Whittington and Huppert, 1998) that absence of positive well-being is more predictive of 7-year follow-up mortality than presence of negative well-being.

There are relatively few studies which have examined the importance of positive in addition to negative well-being in relation to cortisol secretion. Seeman and McEwen (1996) reported that HPA axis reactivity was lower in participants with more supportive social relationships. Moreover, among women with breast cancer, mean daytime salivary cortisol levels were significantly lower in those whose perceived quality of social support was higher (Turner-Cobb et al., 2000). More recently, trait optimism in a sample of students was shown to be inversely related to salivary cortisol levels, particularly in a 40-min period immediately after awakening. By contrast, trait pessimism was not related to cortisol levels (Lai et al., 2005). Similarly significant associations between aspects of eudaimonic positive well-being (e.g., personal growth and purpose in life) and lower cortisol secretion have been reported, with the effect being most marked for morning cortisol levels and in those over 75 years of age. In the same study, negative well-being (e.g., anxiety, anger) was not associated with cortisol secretion across the day (Ryff et al., 2006). These results also support the partial ‘independence view’ of positive and negative well-being, at least in relation to associations with cortisol secretion. Consistent with these findings, state happiness, assessed from aggregated momentary experience samples of happiness over the working day, was inversely related to total cortisol secretion over the day (from 08:00 to 22:00 h) an effect independent of psychological distress (Steptoe et al., 2005). A follow-up study from the same group using a similar method of momentary assessment of state happiness found a comparable inverse association between cortisol and happiness (again, independent of psychological distress) but this time the effects were limited to the first 60 min after awakening—the awakening cortisol response (ACR) (Steptoe et al., 2007). The authors claimed that these results supported the notion that positive well-being was directly related to health-related biological processes.

The ACR is a discrete aspect of the cortisol circadian cycle. In healthy adults salivary free cortisol concentrations increase by 50–160% in the first 30 min immediately after awakening (see Clow et al., 2004 for a review). Although the physiological role of the ACR has not been clearly defined, evidence suggests that it is under a distinct regulatory influence, different from the rest of the diurnal cortisol secretory cycle and sensitive to a range of psychosocial variables (Clow et al., 2004). Indeed, this is illustrated by the findings of Steptoe et al. (2007) cited above, where state happiness was found to be associated with reduced cortisol secretion in the first 60 min post-awakening but not over the rest of the diurnal cycle. Other studies (Steptoe and Wardle, 2005; Ryff et al., 2006) have not examined cortisol immediately after awakening (the ACR) but more generally point to morning, rather than evening, cortisol secretion as most sensitive to individual differences in well-being. Examination of the relationship between well-being and cortisol secretion therefore necessitates careful examination of time-of-day effects and warrants separate analysis of both aspects of the diurnal cycle: the ACR and the diurnal decline over the rest of the day.

Because positive and negative well-being, while clearly related to each other, do not always relate to other variables in the same way, we believe that more detailed examination of interactive as well as main effects of the positive and negative dimensions of well-being are warranted. In the light of the role of cortisol in a range of health related outcomes and policy initiatives aimed at the promotion of ‘positive ageing’, we chose to explore these questions in a sample of community dwelling older people. This study population was also warranted as it has been shown that well-being and cortisol secretion were most clearly associated in those over 75 years of age (Ryff et al., 2006). This may partly reflect greater potential variation in cortisol measures as a result of the aging process. There is evidence that the HPA-axis may become more dysfunctional as people age, with generally higher levels of cortisol frequently being observed and changes in normal circadian rhythms (Sapolsky, 1999). However, such changes vary considerably amongst older adults, especially in circadian patterning (see, for example, Ice et al., 2004) which argues the importance of examining for psychosocial correlates of such variation. Thus, in this study we set out to explore possible relationships between diurnal cortisol secretion (the ACR and the diurnal decline) and aspects of well-being in a group of high functioning home dwelling senior citizens.

It was our general guiding hypothesis that positive and negative well-being would not simply represent opposite ends of a bi-polar continuum and this may be evidenced by more complex associations with measures of cortisol secretion across the day. We specifically wished to extend our analytic strategy by examining interactive as well as simple effects of positive and negative well-being. It was also our hypothesis that well-being would be more closely related to cortisol secretion in the morning (the ACR) than cortisol secretion across the rest of the day.