Since writing this (in 2012) I've read some other things on the topic. Mindsight by Daniel Siegal is worth a read (he gave a talk at Google).
Graduate school can be pretty stressful. Between classes, conferences, research or teaching assistantships, publication pressure, the looming job-hunt, and having to see your undergraduate class-mates get married and take "real" jobs with a salary far in excess of your own pathetic stipend, the psychological stress can be pretty serious. After having read Robert Sapolsky's (neuroendocrinologist at Stanford) Why Zebras Don't Get Ulcers, it is apparent that this type of chronic stress-response is very, very bad for your health in a number of ways. I will go over a few of the ways in which repeated activation of the stress response leads or contributes to health problems, and then some coping mechanisms that he suggests (adapted for grad school life).
How does psychological stress translate into a physiological effect? There are, of course, a myriad of ways this occurs, many of which are the subjects of ongoing research in a number of subfields within physiology. Sapolsky covers how stress affects stroke, heart attack, dwarfism, sex and reproduction, immunity and vulnerability to disease, the perception of pain, memory, sleep, aging and death, depression and socio-economic status (or rather the perception of it). I am just going to cover those that I think are relevant to graduate students, as most of us are too young to have the sorts of problems chronically stressed professors will have. To understand the effects of stress on any one of these areas, first we need a general understanding of the physiology of the stress response.
The stress-response was defined by Walter Cannon as the well-known "fight or flight" response. During a stress-response the body rapidly mobilizes stored energy, and puts it in a readily usable form (glucose). Unnecessary tasks that don't weigh heavily on the immediate outcome of the stressor are shut down (digestion, growth, tissue repair, etc.). Heart rate increases, pumping all the nutrients that have just flooded into your bloodstream to the muscles that are presumably going to carry you to safety. If the stress is psychological, your muscles aren't going to carry you to safety, and the principal hormones that trigger the stress-response remain in the bloodstream, especially if the stress-response is being chronically activated.
The hormonal response works basically as follows (and I am surely glossing over a fair amount of important detail). The hypothalamus projects CRH (corticotropin releasing hormone) which gets the pituitary gland to release ACTH (corticotropin) which triggers the release of glucocorticoids (steroid hormones, synthetic versions you may be familiar with: hydrocortisone, prednisone, etc.). Glucocorticoids have a range of different effects, depending on what is receiving them. The release of glucocorticoids occurs more or less simultaneously with epinephrine, norepinepherine, glucagon, endorphins, enkephalins, and vasopressin.
Chronic activation of the stress-response is well known to have a variety of deleterious effects on the functioning of the circulatory system. As noted, the stress-response prepares you for intense physical activity such as running from a lion or something. This can raise your blood pressure and lead to hypertension. Capillaries and other small blood-vessels respond to increases in blood-pressure (from a faster and harder beating heart) by building extra muscle around them to better control blood flow. This in turn results in more rigid blood-vessels, which results in higher blood-pressure… etc. Increased pressure can damage artery branch points, which results in an inflammatory response. Additionally, blood is already more vicious during a stress-response, as epinephrine increases clotting factor in circulating platelets. So along with a range of other factors, this makes you more likely to have atherosclerosis, which in turn, increases the risk of heart disease and stroke.
There are three ways in which stress can negatively affect the body that I think particularly impact graduate students. The first is the effects on digestion. Stress can modulate functional gastrointestinal disorders such as irritable bowel syndrome, among others. Stress can increase the risk of "acquiring" this functional disorder, and make pre-existing cases worse. This can be because the colon is too contractile, or because contractions lose their directionality. Additionally, the stress-response decreases sensitivity to cutaneous (skin) pain, while increasing sensitivity to visceral pain (internal organs). Apparently, the hyper-contractility of the colon is typically not observed during sleep, meaning that the functional disorder may be due in large part to stress. Of course the pain induced by IBS itself can be a stressor, leading to another viscous cycle.
A second negative effect relevant to graduate students is the effect of the stress-response on memory. For obvious reasons, memory is very important, especially when you are preparing for your comprehensive/prelim exams, have a presentation to give to your colleagues, or any number of other things. It is well known that short-term stressors of mild to moderate severity enhance cognition (and certain types of memory), while major or lengthy stressors have the opposite effect. Sapolsky differentiates between short-term and long-term, and explicit and implicit (which includes procedural) memory. Explicit memory concerns facts and events, as well as your ability to retrieve them. Stress seems to have the greatest effect on the functioning of the hippocampus, which is involved with turning short-term memories into long-term ones, memory consolidation, and a number of other things. High amounts of synthetic glucocorticoids or abnormally high amounts secreted by your own body (such as in Cushing's syndrome) are well known to result in explicit memory problems (cushingoid dementia for one). Elevated glucocorticoid levels have been shown to create this response through three pathways. First, they seem to result in the degradation of neuronal networks, decreasing our ability to recall information, or increasing the number of associate cues required to recall the information. Second, the birth of new neurons (neurogenesis) is inhibited. Third, hippocampal neurons can become endangered. While initially the stress-response increases the amount of glucose available to the brain, these resources decline rather precipitously in an extended stress response, which can result in increased vulnerability to neuronal death if there is another neuron insult going on at the same time (such as a stroke, seizure, etc.) In short, chronic stress decreases your ability to recall facts, your ability to store new facts, and, if coupled with an additional neural insult, can result in neuronal death.
Sleep is important for a variety of reasons. There seem to be several stages of sleep: shallow sleep, deep-sleep (slow-wave sleep), and rapid-eye movement sleep (REM sleep). Generally, people cycle through these different stages about every 90 minutes or so. Different stages of sleep are unsurprisingly associated with different patterns of activation in the brain. Slow-wave sleep, for example, seems to be associated with the consolidation and retrieval of memories. The brain's store of glycogen is also renewed during slow-wave sleep. Some research has shown that explicit memories in particular are consolidated during slow-wave sleep; patterns of activation in the hippocampus of rats while creating explicit memories are repeated during slow-wave sleep.
Where does stress factor into this? Sleep-deprivation has an obvious effect. Slow-wave sleep is hypothesized to result in the release of an unknown but important corticotropin inhibiting factor (CIH) which would inhibit ACTH, which is the pituitary's hormone that signals the release of glucocorticoids. Thus if you are sleep-deprived, your basal level of glucocorticoids is likely to be elevated. One study of sleep deprivation has also shown that working memory (short-term) is severely impaired by sleep-deprivation. CRH seems to be the most important part of the stress-response as it relates to the inhibition of sleep. CRH not only deprives you of sleep (because your sympathetic nervous system is activated and your heart-rate is elevated) but also the quality of your sleep (time spent in slow-wave sleep is decreased). So in short, low-quality sleep or little sleep results in decreased memory consolidation and decreased functional working memory, bad news.
So, lots of bad news. To add to the heap, it may be the case that grad students who seem particularly in control, are extremely productive, have rigorous schedules that they rarely deviate from (i.e. those who live highly structured lifestyles) may actually have rather elevated basal levels of glucocorticoids, reflecting an underlying persistent anxiety (which presumably has to do with encountering deviations or disruptions to this scheduling). Sapolsky offers a number of recommendations based of the literature from experimental psychology, which I have adapted somewhat to our specific situation.
Control
In graduate school and in academia, we are faced with lots of stressors, over which we have varying degrees of control. Our articles and conference proposals get rejected, we are not selected for scholarships or grants, our advisors email us back 50% of the time (if we are lucky), our other friends move on with their lives while we remain "in school." These are all things that we have only some control over. Gary King noted that he could "wallpaper his house with rejection letters." So how do we avoid being stressed by these things? We have to recognize the degree of control we do possess, while not overestimating it. This requires that we do not stress over things that are in the past (as I am quite guilty of), and realize that there is rarely one perfect solution to some huge stressor, but rather small footholds of control that allow us to manage the larger stressor.
Social Support
There often is a lot of resistance to this one, especially amongst men. We often deal with our problems (well try to anyhow) internally, resulting in a chronically activated stress-response. Or our way of dealing with is by meting it out to others. Since most graduate students are not the physically aggressive types, this comes out in more verbal degradation in seminars. Underlying stress has been the cause of many a condescending, caustic remark in class, I suspect. Social support can help mitigate this. That does not mean going out and drinking a whole bunch. Having lots of friends does not necessarily mean that you have high levels of social support either. Quality matters more. Having the ability to "cry on someones shoulder" so to speak, is the important part. Having a significant other or a close friend in a similar situation is most important (assuming they satisfy this requirement of course).
Therapy
There is still, unfortunately a stigma against psychotherapy, especially amongst men. This is unfortunate as it can be highly effective, especially for those of us that qualify as having "type-A" personalities. In the absence of the previous two factors, this can be especially helpful. Fortunately those of us still in graduate school typically have easy access to a therapist through the university health center. Especially for those of that don't naturally deal well with stress, or during particularly stressful periods, therapy can be a great way to deal with stress.
So essentially we should follow this dictum (apparently heard during a Quaker meeting)
In the face of strong winds, let me be a blade of grass.
In the face of strong walls, let me be a gale of wind.