Eustress vs. distress: how to tell whether stress is helping or harming you

In 1936, Hans Selye was giving rats shots of a crude extract from ovaries, hoping to isolate a new hormone. The animals sickened the same odd way each time. Adrenal glands swelled, thymus and lymph tissue shrank, and ulcers appeared in the lining of the gut. Then he ran his controls: extract of kidney and spleen produced the same three things. So did cold, surgery, forced exercise, and low doses of drugs with no chemistry in common. Whatever he did to the animals, if it was harsh enough, the body answered the same way.1

The sameness is what made his name. Selye called the pattern the general adaptation syndrome. For four decades he argued that stress is the body's broad answer to any demand at all, including demands a person wants.2 Two words attached themselves to his work much later. Eustress means demand that builds you up. Distress means demand that wears you down.

What became of that idea is close to the opposite: health writing took the finding and built a filing system out of it. Eustress in one bin: the new job, the wedding, the cold shower. Distress in the other: the deadline, the divorce, the difficult parent.

That costs people something real: it sends them looking in the wrong place at the moment they most need the answer.

Nothing about an event settles whether it'll build you. Not how hard it is, not whether it feels good, not whether you chose it. Three other things settle that, and all three sit on your side of the meeting. The first is the size of the demand next to what you can absorb now. The second is whether recovery follows. The third is whether it connects to something you find worth doing.

The same hour of hard training builds you in March and takes you apart in October, with no change in the training. So the question most people bring to stress has no answer. Is this good for me or bad for me? That asks about the wrong object. The answerable version is narrower. Can my body, right now, turn this demand into a gain?

The dose is the whole story

The strongest evidence against the two-kinds model comes from the study of poisons, where toxicologists have logged a strange shape for decades. Some agents do harm at a high dose and measurable good at a low one. The name for it's hormesis. Calabrese and Baldwin gave the effect a working definition. The good part of the curve is small, often a few tens of percent above the control group at its best. It sits in a dose range that's narrow next to the range where the same agent harms.3

Mattson carried the same rule into human biology, where the examples are training, eating less, and plant compounds. A demand the body can handle switches on repair systems inside the cell. What comes out the far side is better stocked than what went in.4 That's what training is, and it's also what overtraining is. The difference isn't the thing you did.

A small dose helps and a large dose harms, which means the useful zone is thin. It sits below a line, past which the same input starts subtracting. Reading hormesis as a licence for hardship gets the shape backwards.

An honest version of this argument has to concede something. Not everything works by dose inside the range a human being is likely to meet. Sustained violence, going without food or care, the loss of a child: there's no dose of these that leaves a person better off. It would be ugly to suggest otherwise. There's a huge middle ground where nearly everyone lives: work, training, cold, fasting, conflict, ambition, duty. That ground doesn't sort into kinds. It sorts by dose, by the state you're in, and by what follows.

The bill arrives late, and separately

McEwen and Stellar gave the cost a name in 1993: allostatic load, the running cost of adapting, which adds up over time.5 Allostasis is the ordinary work of adjusting to conditions that won't hold still, not a fault in the system.

McEwen's 1998 paper makes the argument clear, because it refuses to cast the chemicals as villains. Cortisol and adrenaline protect you: they free up fuel, sharpen attention, and tune the immune system on the clock of an emergency. The same molecules harm tissue under a different pattern in time, and he set those patterns out plainly. Demands arrive with no space between them. The response fails to settle down to a challenge already met many times. Or it keeps running long after the challenge has ended.6

There's a fourth pattern in that paper that almost nobody quotes. The running cost also grows when a response is too weak. One part does too little, other parts cover for it, and the cover carries a price of its own. Hold onto that; it matters later.

The brain does the judging, and it's reshaped by its own answer. Under long load you can measure change in the hippocampus, the amygdala and the front of the brain. Much of that reshaping looks reversible rather than fixed, which is an important caveat and a reason to refuse the language of permanent damage.8 Sapolsky's shorthand is still the best available: machinery built for a three-minute emergency, left running for three years.7

The timing problem is why the popular model survives. The cost of adapting isn't billed at the point of purchase. The body covers it first, quietly and well, by borrowing from jobs that can wait. Scores built from hormone, immune, metabolic and heart markers predict later physical and mental outcomes across many groups.9 So the build-up is real, but it shows on a research timescale, not a Tuesday one. Symptoms tend to mark the end of the covering phase, not the start of the load. That's why so many people describe an onset that came from nowhere and mean it honestly.

Recovery is the variable

If the same demand can go either way, what decides is what happens afterwards. A demand followed by enough recovery is a training signal. Remove the recovery and the same demand becomes slow wear, with nothing at the time to tell it from progress.

Culture files recovery as the reward at the end of effort. In the body it's the phase where the gain gets locked in. Skip it and you've paid the full price of the challenge and collected none of the return. Over enough repeats, that's how careful, hard-working people make themselves less able.

Evidence that the stress response itself isn't the culprit comes from an odd direction. In a small trial, healthy young men trained in breathing, cold and meditation raised their own adrenaline at will. Their early immune response to an injected toxin was weaker than in men who hadn't trained.10 One technique, one short protocol, a small sample, short-term markers only: it settles nothing about long-term disease. What it shows is narrower. A large surge you start yourself, held inside a window and followed by a return, is one kind of event. A low hum that never resolves is another, and the shape of the episode is what matters, not which nerves it uses.

The half of the curve nobody writes about

Stress writing has a systematic blind spot: it watches one end of the dose curve and treats the other end as safety. Too much is the problem, and less is always the fix.

The exercise research is where that breaks most visibly. Booth and colleagues argued that not moving is a primary cause of chronic disease in its own right. It's an exposure with results, not an empty space where an exposure could have been.11 Stillness holds a position on the curve rather than marking where it begins, and carries costs of its own.

This next step goes further than the evidence strictly allows, and it should be labelled that way. Recall McEwen's fourth pattern: a response that's too weak counts as load.6 He was describing a narrow case in the body: a hormone axis gone flat, with other systems covering for it. He wasn't describing a person whose life asks too little of them. Reading across from one to the other is our extension of his framework, not a finding of his.

Here's our reading, stated plainly. Too little demand, held long enough, carries a cost of its own kind rather than acting as neutral rest. The likely reason is a loss of the capacity that regular challenge maintains. That would be why telling an under-stretched person to rest more so often changes nothing, and sometimes makes things worse. The reading fits the physiology, but it hasn't been shown to be one single effect. Nobody has ruled out the plainer story: too much and too little may be unrelated processes that only look alike from outside.

So "reduce your stress" is a specific instruction wearing the clothes of general advice. Given to someone at the other end of the curve, it's precisely wrong.

What the instruments can and can't tell you

A device would be convenient, and the market behaves as though it has one. It doesn't.

Heart rate variability is the most marketed stand-in, and at the group level it's useful. It tracks the part of the nervous system you don't run on purpose, and predicts real outcomes. But the measures are many and can't be swapped for one another, and normal ranges differ hugely between people and across ages.12 The methods research is blunter still. Breathing rate, posture, time of day, recent movement, coffee and alcohol all move the number. The rules for holding those steady are strict enough that most home devices miss them.13

A morning reading is a real signal wrapped in a lot of noise. Read as a trend across weeks, under steady conditions, it can tell you something. Read as a verdict on yesterday, it's close to a coin toss with a decimal point.

Allostatic load scores have the same problem in another form. They were built as research tools for studying whole populations, and which markers go in, and how they're scored, differ between groups.9 No cut-off on such a score is a personal diagnosis, and none was ever meant to be.

So the honest position is this. No test has been proven to separate demand that builds you from demand that wears you down, in one person, in real time. Anyone selling one is selling something else. What remains is a set of questions, asked over time, by a person willing to be wrong about the answers.

Better questions than "is this good for me?"

Does the demand connect to anything you find meaningful? Antonovsky placed meaning at the centre of a person's ability to meet demands. Not optimism, but a durable sense that what's being asked of you is worth the engagement.14 The caveat is substantial. This research mostly asks people to describe themselves at one point in time. It's hard to separate meaning as a cause of strength from meaning as a sign of resources the person already had. It may work the other way round. Still, asked of yourself about one demand, it sorts well.

Does your energy come back inside the window it used to? This is the most useful of the four: it needs no device and uses your own history as the norm. A hard session that used to cost you a day and now costs you four has told you something exact. What it hasn't told you is that the session got harder. The same test works beyond training, for a difficult talk or a week of travel. The readable thing is the gap afterwards.

Is your response the right size for the trigger? When a minor annoyance produces a reaction far bigger than itself, the reaction is rarely about the annoyance. Under a growing load the margins narrow, and things that used to pass without comment start producing one. That mismatch is one of the earlier readable signs that the covering is running thin.

Is it adding up across weeks? Almost nothing here's visible at the scale of a day. Almost all of it's obvious at the scale of two months. One bad night means nothing. Six weeks of shortening sleep is trend data. So is a resting heart rate that keeps creeping up. Trend is the only resolution at which allostatic load is legible to the person carrying it, and it's the one nobody watches.

None of these produces a score, and running through them in order defeats the point. They're what remains answerable once the question with no answer has been set aside.

The cost of the reframe

The reframe has a cost, and it should be stated rather than smoothed over. If there's no category of good stressor, there's no list to consult, no routine that stays correct, and no answer that holds. The judgement has to be made against a moving quantity: what you can absorb this month, not what you could last year. And it expires quietly, without notice.

What you get in exchange is that the failures stop being moral. Say the training that built you in spring starts taking you apart in autumn. That's not a stretch in which you became worse at trying. It's a stretch in which the arithmetic changed underneath a routine that stayed the same. Handed that, most people try harder at the thing subtracting from them, because trying harder is what the moral version recommends. The arithmetic version asks for something duller and more difficult. Work out where on the curve you're standing before deciding how much more to add. The curve keeps moving. So the working out isn't something you do once.