Stress and illness: The decades-long search for a link
Share on Linkedin
(Copyright: Thinkstock)
Common sense tells us that stress makes us prone to illness, but proving any connection has been virtually impossible. Until now, perhaps.

A bereaved man suffers a heart attack. An unemployed graduate is plagued by eczema. A divorced woman develops high blood pressure. Are these situations connected… or mere coincidences?

Of all the influences on our health and well-being, chronic stress is among the most ubiquitous. And because the misery of stress is often experienced as much in the body as the mind – tiredness, headaches, tense muscles and the like – common sense tells us that psychological stress leaves us prone to physical illness.

While most people outside medicine happily accept this view, it has been viewed with a much greater degree of scepticism within the field. And that’s partly on account of the difficulty in proving exactly how stress might work, not to mention understanding why some people succumb to it but others do not. There are plenty of ideas and intuitive hunches, but rather less incontestable evidence.

But a recent finding from a team led by Sheldon Cohen at Carnegie Mellon University in Pittsburgh could finally have revealed a link, offering perhaps the best evidence so far of how stress operates at the biological level. What Cohen thinks stress is doing – a surprise, perhaps, to many of us – is undermining the body’s capacity to deal with inflammation.

It’s not the only possible stress mechanism; nor is it likely to be the whole truth. But the work represents an important step in understanding the mechanism of stress. It not only adds plausibility to claims about its role in promoting ill health, it’s a mechanism that fellow researchers find plausible and convincing.

Seeking proof

Cohen has been working on stress for 30 years, and despite a fair amount of epidemiological evidence accumulated during this time, he and other professionals have been cautious about accepting a link without firm evidence about a possible mechanism. In a 2007 review of psychological stress published in the Journal of the American Medical Association, Cohen said: “Despite widespread public belief that psychological stress leads to disease, the biomedical community remains [sceptical] of this conclusion.” But speaking now, little more than half a decade later, he takes a rather more positive view. There’s been a shift, he says. A plausible mechanism to explain the effects stress will reinforce it.

The issue is how you design an experiment that proves any ill-effects are caused by stress. Up until now, the real hard evidence has had to come from animal studies. “With humans you can bring them to a lab and expose them to stress, and you can get elevations in heart rate and blood pressure,” he says. Physiological changes, in other words, that often play a role in disease. But, for ethical reasons what you can’t do is give human subjects an illness and see if stress makes it worse.

Well, not quite. An upper respiratory tract infection like a cold is an illness, but it is mild enough to make it ethically acceptable to break the rule, and do human experiments. 

Cohen began the research that propelled him to the forefront of this field back in the early 1990s. His approach, originally developed at Britain’s (now defunct) Common Cold Research Unit, relied on giving measured doses of cold virus to volunteers and checking to see if they developed symptoms. Before infecting his subjects with the virus Cohen rated their stress levels using standard questionnaires. The higher his subjects’ stress scores, he found, the greater the likelihood they would catch a cold.

Further studies showed that the principal enduring and stressful life events that rendered people susceptible to the cold virus were problems with family and friends, overwork, and unemployment. The longer these events lasted, the greater the susceptibility.

British stress researcher Andrew Steptoe, a professor of psychology at University College, London, describes these experiments as probably the most elegant of all the studies of stress and infection done to date. “What you don’t usually know under everyday circumstances is exactly when people were exposed to an infectious organism,” he says. “There are bugs floating around all the time.” Thanks to those experiments, stress levels could be linked precisely to the time of infection.

Prime suspect

While this work confirmed the association between stress and illness, it left Cohen little wiser about the mediating influence. Part of the explanation is quite likely behavioural. Stressed people smoke and drink more. They sleep badly and often take less exercise. All these things have detrimental consequences on your health. But Cohen worked on the assumption that as well as lifestyle issues there are also specific biochemical pathways linking stress and health.

One of the key molecular players in these pathways is also among the most familiar of the body’s signalling chemicals: cortisol, a steroid commonly referred to as a “stress hormone”. Produced by the adrenal gland in response to stress, the original view of cortisol was straightforward: more stress prompts your body to make more cortisol, and the higher the level of cortisol in your circulation, the worse the outlook for your health.

However, many studies have undermined this simple idea, says Professor Phil Evans, a psychologist at the University of Westminster with a long established interest in stress and cortisol. “Generally, levels of cortisol in naturalistic studies [i.e. those carried out in the real world] do not predict health outcomes strongly, or with any great consistency,” he says.

So if cortisol is involved, but not in the simplistic sense of “the more, the worse”, what is the nature of the link? Cohen’s view is that what matters more than the level of circulating cortisol is the body’s response to it. Cortisol molecules exert their effects on the body’s cells via a set of specific receptor sites: the glucocorticoid receptors, to give them their full name. When a cortisol molecule attaches itself to a receptor it triggers a chain of chemical events within the cell. Stress, says Cohen, changes the sensitivity of these receptors; they become resistant to the activating effects of cortisol.

One of cortisol’s key roles in the body is the suppression of inflammation. This is why its synthetic equivalent, hydrocortisone, is used in treating a range of inflammatory conditions, from eczema to ulcerative colitis. It’s Cohen’s contention that if the glucocorticoid receptors on the cells of the immune system fail to respond as they should to the presence of cortisol –“glucocorticoid resistance”, as it’s known – the body’s arrangements for keeping inflammation in check break down.

Inflammation in this context refers not to the sudden burst of activity you get following acute infection or injury – a protective attempt to deal with some form of harm to the body and to initiate the healing process – but to a low level elevation of chemical processes that are there all the time. “We know that stress, acutely and chronically, increases inflammation,” says Steptoe. If Cohen is right, we now know how. Stress reduces the sensitivity of the receptors (in effect chemical switches) which are supposed to control the level of inflammation in the body.

Detecting resistance

This possibility has actually been under consideration for some time. Cohen himself published some of the early work on glucocorticoid resistance. One study looked at parents of children with cancer: an extremely stressed group. “Our argument was that when you look at people who are under chronic, extreme stress you should find an increase in glucocorticoid resistance. And in comparison with matched control parents that’s exactly what we did find.” But this work didn’t set out to demonstrate any subsequent link between glucocorticoid resistance and illness in those parents.

Back to the common cold. Cohen realised that data from two of his previous studies could be used to test this glucocorticoid resistance theory. In one of these studies he’d assessed 276 volunteer subjects for major stressful life events, exposed them to common cold viruses, and then checked over the next five days for signs and symptoms of a cold. Although he didn’t have any direct measurements of the sensitivity of their receptors, he did have details of the proportions of the various types of immune cells present in their blood. Other research has shown that this acts as a surrogate measure of glucocorticoid resistance.

The experiment revealed that prolonged stress is correlated not only with the likelihood of developing a cold, but also with higher levels of glucocorticoid resistance. In other words, the people whose receptors responded inadequately to the cortisol in their blood were the ones who got the colds.

In a second study, Cohen had made direct measurements of glucocorticoid resistance in a group of 79 subjects before exposure to cold viruses. His hope was that it would allow him to predict which of his subjects would catch a cold and which wouldn’t. In fact the group wasn’t large enough to make a statistically significant prediction about actual colds; but he was able to predict which subjects showed a pattern of biological activity of the kind associated with the symptoms of a cold.

The findings of both studies point in the same direction. What counts is not the amount of cortisol circulating in the body, but how much our cells react to it.

Health predictor

Colds may be unpleasant but, in the scheme of things, they’re hardly a big deal. So what of more serious disease? The real importance of the work, says Cohen, lies in its relevance to the many other conditions in which inflammation is a factor. “It’s a model that can be applied to many diseases. Cardiovascular disease, asthma, autoimmune disease, diabetes…The regulation of inflammation plays a big role in the progression of all of them.”

Steptoe agrees. “People are already interested in inflammation in relation to chronic diseases such as diabetes and coronary heart disease,” he says. And Phil Evans too joins the chorus of acceptance. Cohen’s conclusions, he says, are “potentially generalisable to a range of other inflammatory illnesses”.

At University College London, Steptoe works with the epidemiologist Professor Michael Marmot, who’s spent many years studying peoples’ health in relation to their socioeconomic class and their status at work. Some of the variation that Marmot has identified is clearly attributable to differences in peoples’ material circumstances and behaviour. But when it comes to the health effects of their status in a hierarchy, the forces at work are more subtle. 

One possibility is that people fare less well when they find themselves unable to determine their own actions and make their own choices. The crossover with stress is obvious. Professor Steptoe is investigating the cortisol biochemistry associated with their predicament, especially as it affects their risk of heart disease.

Definitive proof

Cohen readily admits that proving his hypothesis will be difficult, at least for now. One of the classic methods of studying a biological system in animal models is to disturb it in some way – stimulate it or inhibit it, for example – and see what happens. “There’s no way at present we could directly manipulate glucocorticoid resistance without upsetting all kinds of other systems as well,” he says. While animal work showing that chronic stress affects glucocorticoid resistance is persuasive, it is not definitive. “And we can’t do this in humans,” he adds. “We can’t randomly assign people to chronic stressful and non-stressful conditions.”

Steptoe agrees. We’re reliant on the differing stress levels that people themselves report, he says. “And in an observational study in humans there’s always the possibility that some other factor is involved which you can’t avoid. For example, it could be that a certain type of person chooses a job that puts them under higher work stress, so any changes aren’t necessarily to do with the stress per se, but with the type of person who’s attracted to that kind of occupation.”

In the meantime, Cohen wants to know if there is evidence that other forms of ill health can be linked to stress-induced resistance in the glucocorticoid receptors. He talks of finding opportunities “to examine the role of glucocorticoid resistance in predicting the onset and progression of other inflammatory diseases, probably in epidemiological studies.” And he’s not alone. He knows, for example, of another group of researchers planning a study of the role of glucocorticoid resistance in asthma.

There is always the possibility that the definitive experiment will always remain out of reach, in which case stress researchers might find themselves in the position of doctors trying to demonstrate a causal link between lung cancer and smoking. Unable to do the definitive experiment they have had to rely on an accumulation of observational studies, animal experiments, and studies of the physiological mechanisms most likely involved.

After thirty years of exploration, Cohen says he remains undaunted by the prospect that definitive experiment might elude him, and he laughs off any suggestion of frustration. “When you work in chronic stress,” he says, “you don’t expect to be able to do experiments like that.”

To comment on this or anything else you have seen on Future, head over to our Facebook page or message us on Twitter.

Around the BBC