Allostasis in human evolution

McEwen and Wingfield (2003) discuss the concept of allostasis. I was unfamiliar with this concept myself until an interesting presentation by one of our graduate students the other day. It is distinguished from homeostasis in a way that makes for an interesting contrast that may have relevance to human evolution.

According to the paper (3):

Homeostasis is the stability of physiological systems that maintain life, used here to apply strictly to a limited number of systems such as pH, body temperature, glucose levels, and oxygen tension that are truly essential for life and are therefore maintained within a range optimal for the current life history stage.
Allostasis is achieving stability through change. This is a process that supports homeostasis, i.e., those physiological parameters essential for life defined above, as environments and/or life history stages change. This means that the "set-points" and other boundaries of control must also change. There are primary mediators of allostasis such as, but not confined to, hormones of the hypothalamo-pituitary-adrenal (HPA) axis, catecholamines, and cytokines. Allostasis also clarifies an inherent ambiguity in the term "homeostasis" and distinguishes between the systems that are essential for life ("homeostasis") and those that maintain the systems in balance ("allostasis") as environment and life history stage change.
We note, however, that another view of homeostasis is the operation of coordinated physiological processes that maintain most of the steady states of the organism. In this interpretation, homeostasis and allostasis might seem to mean almost the same thing. The reason they do not is that the notion of "steady state" is itself vague and does not distinguish between those systems essential for life and those that maintain them. It also does not differentiate changes in state to enable reproduction (and other life cycle processes) that are required for immediate survival.

The concept was presented to me in the context of the cost of social competition in primates. Competition--especially reproductive competition--requires that individuals increase their readiness state, their alertness, their rate of vocal communication, and many other behavioral correlates. Ultimately, this relatively excited allostatic state requires an increase in the rate of energy expenditure, mediated by hormonal interactions in the body. Within primate groups, such allostatic states may last for considerable periods of time. One example is the elevated cortisol levels found for male savanna baboons in subordinate ranks. These individuals suffer stress induced by competition from higher-ranked males, resulting in increases in the level of stress hormones, the maintenance of a high level of cardiovascular and mental readiness, and a lack of stress-mediating interactions with other individuals. Together, these behavioral components result in increased energy expenditure, a depression in immune response characterized by an increased rate of opportunistic infections, an increased chance of mortality. Considering the health effects and mortality associated with this allostatic state, we can conclude that the pattern of natural selection acting on energy expenditure and health in baboons would strongly reflect this allostatic state. Other allostatic states probably also have a strong component of selection associated with them, most notably that of the dominant males whose reproductive success depends immediately upon their ability to counter and resist threats from lower ranking males.

McEwen and Wingfield (2003) refer to the excess energetic expenditure associated with an allostatic state as an allostatic load. This represents the cost of maintaining the body in an altered state for a sustained period of time. Allostatic loads may be added to the basic requirements for survival as an important component of total energy expenditure. In the extreme, they lead to allostatic overload of one type or another. The most basic type of overload is a simple energetic shortfall, in which the body cannot acquire enough energy to meet its expenditure. A more complex type of overload accompanies sustained allostatic loads in individuals who can meet the energetic requirements. In these cases, the extension of altered hormonal and physiological states may lead to an increased chance of chronic or infectuous disease, psychological impacts, or other long-term health consequences. These consequences are the selective costs associated with the maintenance of elevated allostatic states; they are ultimately the reason why animals do not run at a higher energetic state all the time.

In the context of fossil hominids, I think this concept is important because it provides another piece of evidence that overall energy expenditure and other physiological variables cannot be predicted from body mass alone. Across much of human evolution, mortality risk across adulthood was significantly higher than in recent human populations. Humans lived in a broad range of environments, and at some times and places, they must have had relative resource abundance similar to, if not exceeding, that of the Upper Paleolithic and later Holocene people who had low adult mortality rates and elongated life histories. This is not to say that early humans would have matched the sustained population growth that commenced in the Late Pleistocene, but that at many times and places in the Early and Middle Pleistocene, such growth would have been probable for circumscribed times. Why did none of these people evidently live long and prosper to the extent seen later in the fossil record? A good answer (although not the only possible one) is that the real mortality risk came from the social factors. Absent direct violence (which did occur), the second most important social influence on health may have been allostatic loads of one kind or another.

This is not exactly news, since many paleoanthropologists have been investigating allostatic load for a long time (although usually under other names). For example, the energetic cost of pregnancy and lactation is perhaps the largest allostatic load faced by any humans as a nonpathological result of their biology. However, much in the study of social stresses in ancient humans has yielded to the wake of the observation that ethnographic hunter-gatherers are relatively egalitarian in social structure. This would appear on the surface to rule out the idea that there were major differences between people in their social status. But a relatively egalitarian structure only means that status differences were more fluid, not that they were less significant. A rigid pecking order can be distinguished easily on the basis of binary interactions between individuals. In contrast, fluid status distinctions depend on broader parts of the social network to resolve them. Such a system may result in a decreased total level of stress on lower-ranking individuals, but it need not do so. In early humans, there is no strong evidence that interpersonal relations fit the ethnographic hunter-gatherer model, but even if they did that is not sufficient to exclude the idea that significant allostatic loads occurred as a result of social interactions.

I might even offer a hypothesis that allostatic loads were driving factors behind the life history changes associated with the Late Pleistocene. At present, we know that major behavioral changes (at least in the circummediterranean region) included an increase in faunal dietary breadth, an increase in sedentism, growing population sizes and densities, and increased technological variability. All of these factors may reflect more basic changes in human interactions. In particular, they may correlate with a growing degree of social stratification and cultural differentiation. Together, these forces might have sharply increased the degree of competition for resources within social groups, while circumscribing to a greater extent the possibility of individual dispersal to other groups. In other words, the stress on lower ranking people could be expected to have increased while the opportunity to escape by moving would have decreased. If true, then allostatic load may have become an increasingly important factor at this time late in human evolution.

This hypothesis is potentially of great interest for the establishment of social systems predating the innovation of agricultural subsistence. A growing, less mobile, population would increase the opportunity for epidemic disease. The imposition of greater allostatic load might create a pool of individuals at greater risk for disease due to immune depression. Greater population pressure would have restricted both group fissioning and opportunities for dispersal. As a result, lower ranking individuals would be forced (to a greater extent than previously) to cooperate with people of higher status--possibly against their interest. Thus, people would have been forced away from egalitarian social structures into a more hierarchical organization. All of these forces are potentially synergistic, drawing an energetic load from lower-ranking individuals that would inhibit them from rising in the social hierarchy.

Were there energetic and disease factors contributing to the establishment of social hierarchies? It is clear that the very hierarchical nature of early civilizations must ultimately have arisen from simpler hierarchies in transitional societies of collectors, pastoralists, and early agriculturalists. Likewise, these societies represent a transformation of small-scale, egalitarian hunter-gatherer bands. But these transformations have usually been considered to result from purely social changes as a result of the network effects of accumulating wealth and increasing population size. Both of these factors were undoubtedly contributors.

Yet the idea that an accumulating health differential may also have emerged early in the prehistory of growing population size is appealing. Certainly people--especially men--would not tolerate the reproductive advantage of a few high-ranking individuals at their expense. By the origins of state-level societies, they simply had no choice: the accumulation of wealth and social influence of the highest-ranking kings, princes, priests, and other elite were strongly entrenched. But what enabled the initiation of hierarchicalization in earlier, smaller societies? People had no innate quest for wealth, since material wealth was entirely absent as we know it in all but the latest Pleistocene humans. Wealth was only a partial proxy for status, and was substituted for it only after accumulation of material goods became possible. But health was always a primary factor differentiating people. Mate choice in Pleistocene humans was likely highly influenced by health, with people strongly attuned to the health of potential partners. If status differences in growing populations led to a health differential, the innate ability to judge health in potential mates would tend to reinforce those status differences. In effect, the realized heritability of status (including this nongenetic effect) would increase; people would tend to inherit their status more and more from their parents. An initial chance hierarchy in status would snowball into one that was reinforced by biology.

Even so, small revolutions of the social order would have been common; just as large revolutions occurred later in human history. Ultimately people do not tolerate overarching power by a dominant elite, and they resist to the extent that social circumstances and material inequity will allow. This aspect of human nature must be the primary reason for the continuation of egalitarian social structure in hunter-gatherer groups. We can speculate that short-term hierarchies may have emerged in incipient form countless times during the Pleistocene evolution of humans. Resource shortfalls, group extinctions, and social fissioning would reverse this trend whenever it emerged. But the unique circumstances of the latest Pleistocene suspended many of the corrective mechanisms of small-scale hunter-gatherer bands. Human biology stepped into a new phase, from which only extreme technological changes would eventually begin to reclaim it.


McEwen BS and Wingfield JC. 2003. The concept of allostasis in biology and biomedicine. Hormones and Behavior 43:2-15.