Stress Does Cause Depression. A Third of the Link Is Something Else.

That stress affects mental health is not in dispute. What is worth knowing is how much of the link is stress doing the causing, and what happened to the explanation everyone was taught about why it affects some people more than others.

Both questions have unusually clean answers, and one of them is a reversal.

The month of a stressful event

The best-designed study of this used a twin registry, which lets you do something a survey cannot.

Researchers tracked female twins over a year — 24,648 person-months, containing 316 onsets of major depression — and recorded fifteen classes of stressful life event, each rated for how threatening it was and for how far it might have arisen from the person’s own behaviour.

Here is why that design matters. Identical twins share their genes. If a stressful event still predicts depression when you compare one identical twin with the other, the effect cannot be explained by the genes they hold in common — so what remains is closer to the causal part.

The authors’ own summary is the sentence to keep: stressful life events have a substantial causal relationship with the onset of depression, and about a third of the association is not causal, because people predisposed to depression select themselves into higher-risk situations.

Both halves of that are important. The raw association overstates what stress does. It does not overstate it to zero.

The explanation that collapsed

If stress causes depression in some people and not others, the obvious question is what distinguishes them. For about fifteen years there was a famous answer: a variant in a serotonin transporter gene that made carriers vulnerable to stress. It appeared in textbooks, popular science and a great deal of health journalism.

It did not survive.

A 2009 meta-analysis in JAMA examined the interaction between that gene, stressful life events and depression risk and did not find the effect. Then, in 2019, a much larger study settled it.

Researchers identified the 18 candidate genes that had been studied ten or more times for depression, and tested them in samples ranging from 62,138 to 443,264 people, with the analyses preregistered in advance.

They found no clear evidence for any candidate gene association with depression, and none for any gene-by-environment moderator effect. As a set, the depression candidate genes were no more associated with depression than genes never proposed as candidates.

Their conclusion is unusually direct for a journal: the early hypotheses were incorrect, and the large number of associations in that literature are likely to be false positives.

What that leaves standing, and what it does not

It is worth being precise about what collapsed, because it is easy to over-read in both directions.

Depression is heritable — that is not in question and the twin work above depends on it. What failed was the specific claim that a handful of named genes, identifiable by a cheek swab, moderate how much damage stress does to you. Those studies were run on samples orders of magnitude smaller than the ones that overturned them, which is the whole story of how they happened.

So the honest position on individual vulnerability is that nobody can currently tell you where you sit. There is no test, the gene that was supposed to provide one does not, and confident claims about your personal stress-sensitivity are not coming from this evidence.

The physiological half, which is on firmer ground

Alongside all this runs a separate and better-behaved literature about what sustained stress does to the body.

The framework is allostatic load: the mediators that help you handle an acute demand — the same hormonal and cardiovascular responses that are genuinely protective in the short term — impose a cost when they are activated too often, for too long, or fail to shut off when the demand ends.

That framing is more useful than “stress is bad” because it locates the problem in the pattern rather than the presence. A stress response that arrives, does its job, and switches off is functioning. One that does not switch off is the thing being described.

1. Take the month seriously. The elevated risk in the study is concentrated in the month of a stressful event. That is a specific window in which to be less demanding of yourself, not a permanent condition.
2. Notice the selection effect in your own life. A third of the association runs through people in difficulty ending up in more difficult circumstances. That is not a character flaw; it is a documented pattern, and it is the part that compounds.
3. Distrust anything that tells you your genes make you fragile. Eighteen candidate genes, samples up to 443,264, preregistered, nothing. Any product or article claiming to assess your genetic stress vulnerability is selling a hypothesis that has been tested and failed.
4. Aim at the switch-off, not the switch-on. The allostatic framework is about a response that does not end. Sleep, recovery time and the end of the working day are interventions on the shutting-off, which is where the load accumulates.
5. Do not treat “most people are fine” as reassurance. Most people exposed to a severe event do not develop depression. That is a fact about populations and it tells you nothing about whether you are all right.

Three of those five are about narrowing the claim rather than acting on it, which is what an evidence base looks like when its most confident-sounding part has just been withdrawn.

When this is not the right article

Everything above is population-level evidence about the relationship between life events and the onset of depression. It is not a diagnostic tool, it cannot tell you what is happening to you, and it is not a substitute for care.

If low mood, anxiety or exhaustion has persisted for weeks, is affecting how you sleep, work or relate to people, or if you are having thoughts of harming yourself, that is a matter for a doctor or a crisis service rather than for an article about odds ratios. In the United States that service is the 988 Suicide & Crisis Lifeline — call or text 988, or chat at chat.988lifeline.org. It is free, it answers at any hour, and it takes calls from people worried about someone else as readily as from people worried about themselves. The evidence here describes a pattern across thousands of people. It has nothing useful to say about one.

The parts of this that are genuinely hard

The collapse of the candidate-gene literature is a good outcome for science and an uncomfortable one for anybody who found it explanatory. “I am wired to take this badly” was a usable framing, and losing it does not replace it with anything as satisfying.

And the selection finding is easy to read cruelly. That people predisposed to depression end up in higher-risk circumstances is an observation about how disadvantage and illness compound, not an allocation of blame. Read the other way round it is one of the strongest arguments there is for treating circumstances rather than character.

The bottom line

Stressful life events do cause depression, with an odds ratio of 5.64 in the month one occurs — and about a third of that association is not stress causing anything, but predisposition producing exposure. The co-twin comparison puts the causal figure nearer 3.58.

What is gone is the explanation for who. Eighteen genes, hundreds of thousands of people, preregistered analyses, and no evidence for any of it. The honest summary is that stress matters, that it matters less than the headline number, and that nobody can currently tell you how much it matters to you.

Keep exploring

Continue reading

Deeper into what you just read.

FamilyHow to Help Someone Who’s Depressed When Nothing You Try WorksYourselfResilience Is the Common Outcome, and It Is Not Toughness

Explore more

Adjacent ground, if this was useful.

HealthNothing You Eat at Ten Fixes the Coffee You Drank at ThreeFamilyThe Part of Parental Burnout That Nobody Says Out Loud