Why Zebras Don't Get Ulcers: The Acclaimed Guide to Stress, Stress-Related Diseases, and Coping

The summary

A zebra sprinting from a lion and a person lying awake over a mortgage are running the same physiology. That is the argument. The mammalian stress response is a superb three-minute emergency system, and Sapolsky’s claim is that humans switch it on for weeks over things that cannot be outrun, then get sick from the response rather than the threat. We live well enough and long enough, he writes, “to generate all sorts of stressful events purely in our heads.” His example: no hippo worries whether Social Security will outlast it.

Sapolsky is a Stanford neurobiologist who started following wild baboons in Kenya’s Masai Mara in 1978 and kept at it for roughly thirty years, drawing blood from animals whose main occupation is making one another anxious. That fieldwork is the book’s spine; the argument is endocrinology.

The response is what damages you

Under threat, adrenaline and glucocorticoids mobilise stored energy, drive it to muscle, raise heart rate and blood pressure, and postpone everything that can wait — digestion, growth, tissue repair, reproduction, immune surveillance. That triage is exactly right when the emergency lasts minutes. In his own review work, Sapolsky sorted glucocorticoid actions into four jobs: permitting other systems to run at baseline, amplifying the response in a crisis, suppressing it so it does not overshoot, and preparing you for the next one. Nothing on that list was designed to run for a decade.

Immunity shows the shape of it. Stress lasting minutes to hours strengthens innate and adaptive immune responses, largely by moving immune cells where they will be needed. Stress lasting months does the reverse: dysregulated signalling, chronic inflammation, fewer protective cells. Same hormones; duration flips the sign.

One caveat on the book’s most-quoted brain claim. That chronic stress kills hippocampal neurons was Sapolsky’s own hypothesis — he titled a 1996 review of the evidence “the current state of confusion” — and the field has since softened it. What holds up best is the reversible part. Sustained high cortisol retracts hippocampal dendrites and shrinks hippocampal volume — and when twenty-two people with Cushing’s disease were scanned before and after surgery lowered their cortisol, that volume grew back by 3.2 percent on average, up to 10 percent in some. Outright neuron death has proved much harder to pin down.

Stress does not cause diseases, it tips odds

The title is doing some work. Ulcers turned out to be largely infectious: Marshall and Warren took the 2005 Nobel for showing that Helicobacter pylori causes more than 90 percent of duodenal ulcers and up to 80 percent of gastric ones.

The bacterium does not finish the story, though, and the book says so. Roughly half of all humans carry it and only 10 to 15 percent ever develop an ulcer. In a Danish cohort of 3,379 adults tracked for eleven years, people in the top third of a life-stress index were about twice as likely to get a confirmed ulcer as those in the bottom third — and the effect was the same whether or not they carried the bug. That is the shape of nearly every claim here. Sapolsky has said he is drawn to stress precisely because its impact is distributed: it rarely causes anything outright, but makes a great many things somewhat worse. A smaller claim than the genre usually makes, and far more defensible.

Rank matters less than control

The obvious guess about hierarchy is that the bottom is the bad place to be. In a stable baboon troop that holds — subordinate males carry higher resting cortisol while dominants enjoy predictability and control. Destabilise the hierarchy and the advantage disappears: through a period of instability in Sapolsky’s troop, the high-ranking males lost that low-cortisol profile and carried elevated basal cortisol instead. Reviewing the primate literature, Sapolsky concluded the rank-health relationship depends on how a given society is organised.

Two things cut across rank. Personality is one: the low-cortisol dominants were the males best at telling a genuinely threatening rival from a neutral one and, when it was the former, most likely to strike first — dominants without that knack ran cortisol as high as subordinates. Affiliation is the other — among high-ranking males, more grooming and friendly contact with females and infants goes with lower baseline cortisol and a blunted cortisol stress response.

Humans show the same gradient with better paperwork. Whitehall II followed 10,314 British civil servants and found health worsening step by step down the employment grades. What failed to explain it is the interesting part: in the original Whitehall study, standard coronary risk factors accounted for only about a quarter of the mortality gradient. In Whitehall II, low control at work made the largest single contribution to the coronary gradient — ahead of those risk factors, and ahead of height.

Four things change how hard it lands

Jay Weiss’s rats made the point before anyone measured a civil servant. Give two rats identical shocks and the one that cannot predict them ulcerates far worse; give one a lever that ends the shock and it fares better than the yoked twin taking the same current. Identical stressor, different disease.

Sapolsky compresses the coping literature into four questions. Do you have control? Can you see it coming? Do you have an outlet for the frustration? Do you have people? Lose all four and glucocorticoid output climbs and stays there. He is blunt that social isolation is a health risk factor in its own right.

These are levers, not cures. Weiss’s warned rats ulcerated less than unwarned ones, not less than unshocked ones: predictability makes a bad thing more survivable, not good. And because for a creature like us a stressor includes merely anticipating one, a sense of control over what you cannot influence becomes its own trap.

The bottom line

Most of what kills people in wealthy countries accumulates slowly, and running an acute emergency system chronically is one of the accelerants — not the cause of any single disease, but a thumb on the scale for many. Most of the gain comes not from removing stressors but from changing the control, predictability, outlets and connection around them.

Read it if you want the physiology instead of the pep talk, from a scientist funnier than his subject deserves and scrupulous about what the evidence will carry.