The goal of this review was to examine the evidence for age-related changes in HPA axis functioning in humans, specifically, to evaluate the hypothesis that, with age, the HPA axis becomes less resilient in responding to stimulation. Evidence from rodent studies suggests that with age the HPA axis becomes less sensitive to negative feedback signal of elevations in GCs, resulting in a pattern of prolonged activation in response to stimulation. Available data for humans do not permit a final conclusion regarding the possibility that there is a similar age-related change in patterns of HPA response to stimulation. While studies generally find that the pattern of initial HPA response is well preserved with age in humans, few studies have gathered data on longer term HPA 'recovery' or down-regulation poststimulation. Thus, the hypothesis that with age there is a prolongation of HPA response to stimulation has not been fully tested. Our own position is that while age may indeed be associated with an increased prevalence of prolonged patterns of HPA response to challenge, aging per se is unlikely to be uniformly associated with such changes in HPA resiliency. Older age groups exhibit substantial heterogeneity in their relative burdens of morbidity and disability indicating that increases in morbidity and disability are neither uniform nor inevitable consequences of aging (1-3). Such heterogeneity would suggest that aging may not be associated with uniform changes in HPA function either. Rather, aging can be hypothesized to be associated with heterogeneous patterns of change in HPA function with some individuals experiencing substantial change while others maintain patterns of HPA function that are more like those of younger individuals. Thus, while age may indeed by associated with an increase in the likelihood of a decline in HPA resiliency, such declines would not be a uniform feature of the aging process. We hypothesize that the risk of decline in HPA resiliency with age is related to life-long, cumulative exposures to GCs. Though we also acknowledge that other factors such as genetic endowment, initial number of hippocampal CSRs, diet, etc. would be expected to influence this process as well. Individuals with greater net GC exposure would be at increased risk for decline, with older age being the time when such declines in HPA resiliency become manifest. Some older individuals would thus develop a pattern of prolonged HPA response while others would maintain better HPA resiliency. Figure 4 illustrates some of the patterns of HPA resiliency that might be observed in response to a challenge. Pattern A depicts a pattern of good resiliency, a pattern that would be most common at younger ages where maximal resiliency is the norm. Pattern B is one associated with some minimal to moderate decline in HPA resiliency (e.g. perhaps representing that seen at older ages among those with lesser exposures to GCs). Pattern C represents that seen among older individuals experiencing substantial declines in HPA resiliency. Data from a pilot study of ours looking at patterns of ACTH and cortisol response to a driving simulation test (T. E. Seeman, L. F. Berkman, B. Gulanski, R. J. Robbins, S. Greenspan, and J. W. Rowe, submitted) as well as those of Greenspan et al. (163) illustrate these differential patterns and their associations with both age and health. Our data illustrate patterns A and B, showing a possible effect of age rather than health status on HPA resiliency. In contrast to earlier studies using pharmacological challenges, we used a driving simulation as a more 'realistic' type of challenge. The driving simulation, with its attendant requirements for responses to varied and sometimes dangerous situations, represented a challenge that all of our subjects actually deal with in their daily lives (i.e. driving a car). Preliminary data revealed that older 'healthy' subjects (aged 70-79) had more prolonged cortisol responses, with levels remaining elevated above baseline after 2 h (e.g. pattern B). By contrast, younger individuals (aged 30-39) showed a rapid return to baseline before the end of the 2-h follow-up period (e.g. pattern A). The contrast between patterns B and C can be seen in the Greenspan et al. study (163) in which older subjects with diabetes or high blood pressure were found to have a greater response to their oCRF challenge when compared with older 'healthy' subjects. These data suggest that there is heterogeneity within older age groups in terms of patterns of HPA response to challenge.
