Charles Richet: Anaphylaxis, and What to Do When the Immune System Overreacts

Charles Richet — scientific infographic poster

In February 1902, in a laboratory in Paris, a dog named Neptune was given a tiny injection of a jellyfish-relative toxin he had already survived once before. He should have been protected. Within minutes he was dead. The physiologist standing over him, Charles Richet, had just discovered that the immune system — the thing that keeps you alive — can also kill you in under five minutes, and that it needs to have met the trigger before in order to do it.

He called it anaphylaxis. It won him the Nobel Prize in Physiology or Medicine in 1913. This page tells that story, and then does the more important job: it explains what anaphylaxis actually is, how to recognise it when the textbook signs are missing, and exactly what to do. If you are looking for the emergency information, go straight to section 7.

Table of Contents

  1. The Prize and the Man
  2. The Voyage That Started It
  3. The Unexpected Result
  4. Serum Sickness and the Clinical Context
  5. What Is Actually Happening
  6. Recognising Anaphylaxis
  7. Adrenaline, and Why It Is First
  8. Alpha-Gal Syndrome: The Rule-Breaker
  9. Testing and Diagnosis
  10. What Has Genuinely Changed
  11. Where Mainstream Medicine Agrees — and What Remains Debated
  12. Richet's Legacy and Its Shadow
  13. Key Research Papers
  14. Connections
  15. Featured Videos

1. The Prize and the Man

Charles Robert Richet (1850–1935) was a French physiologist, born in Paris to a family already inside medicine — his father, Alfred Richet, was a professor of surgery. Charles took the chair of physiology at the Faculty of Medicine in Paris in 1887 and held it for more than three decades. He was elected to the Academy of Medicine in 1898 and the Academy of Sciences in 1914. He worked on body-heat regulation, gastric secretion, muscle physiology, serum therapy against infection, and the physiology of respiration; he helped fund and design one of the earliest attempts at a heavier-than-air flying machine; he wrote plays and poetry under a pen name; and he was a lifelong, active pacifist who campaigned for international arbitration long before it was fashionable.

In 1913 he received the Nobel Prize in Physiology or Medicine "in recognition of his work on anaphylaxis." That work is genuinely first-rate. It is one of a small number of discoveries that did not just add a fact to medicine but reversed one of its assumptions, and the entire discipline of allergy grows out of it. Everything on this page from section 5 onward exists because of what Richet and his collaborator noticed in 1902.

The uncomfortable part, stated up front

This site does not launder records, so two things go here rather than in a footnote at the bottom.

Richet was a committed eugenicist. He was not a man of his time who made an offhand remark; he wrote a book arguing the case. La Sélection humaine (written in 1912, published in 1919) advocated preventing people he classified as defective from reproducing, and it contains explicitly racist ranking of human populations. He served as president of the French Eugenics Society. His views were wrong as science — the genetics he leaned on was crude even by the standards of his own lifetime — and the programme he advocated was, where it was implemented anywhere in the world, harmful to real people. Nothing about the quality of his physiology changes that, and nothing about his eugenics makes the physiology wrong either. Both are true at once, and a reader is entitled to know both.

Richet also devoted a large share of his later life to what he called "metapsychics" — the investigation of mediums, séances, telepathy and materialised spirits. He coined the word ectoplasm for the substance mediums supposedly exuded, coined cryptesthesia for a supposed sixth sense, served as president of the Society for Psychical Research in London in 1905, and published a book-length Traité de Métapsychique in 1922 treating the whole subject as an emerging science. He pursued it with exactly the seriousness he brought to physiology, and he did not abandon it when his star cases collapsed — most famously the 1905 Villa Carmen séances in Algiers, where the "materialised" spirit was reported by others to have been a local coachman draped in a sheet.

The lesson worth carrying forward is not that Richet was a fraud — he was not, and the anaphylaxis experiments replicate. It is that a Nobel Prize is awarded for a specific piece of work and confers no general authority on anything else its holder believes. That principle recurs across this section of the site; see Prizes That Aged Badly.

2. The Voyage That Started It

In the summer of 1901, Richet was a guest aboard the Princesse Alice II, the oceanographic research yacht of Prince Albert I of Monaco — a serious working oceanographer who had turned his fortune into a research programme and would later found the Oceanographic Museum in Monaco. Richet was joined by Paul Portier (1866–1962), a younger French physiologist.

The prince made a suggestion that turned out to be one of the most productive pieces of scientific small talk on record. Bathers in the Mediterranean were being stung by the Portuguese man-of-war (Physalia physalis), and the prince wanted to know what the toxin actually was. Why not, he asked, study it?

Portier and Richet began on board, extracting material from Physalia tentacles and injecting it into laboratory animals. Back in Paris, Physalia was not available, so they switched to a local stand-in: the sea anemone (Actinia), whose stinging cells belong to the same phylum. They obtained a glycerine extract, found it lethal to dogs at sufficient dose, and named the active principle hypnotoxin for the drowsiness it produced.

Their stated goal was entirely conventional and, at the time, entirely reasonable: they wanted to find the dose that would immunise. Give an animal a small, survivable amount of a poison, let it mount a response, and it should tolerate a larger amount later. That was how antitoxin worked. That was how vaccination worked. It was the organising idea of the whole young field.

3. The Unexpected Result

Some of the dogs given a sublethal dose survived. Weeks later, those same dogs were given another dose — deliberately a small one, well below the amount they had already tolerated. The expectation was that nothing much would happen, or that the animals would prove more resistant than before.

Instead they collapsed. Vomiting, bloody diarrhoea, laboured breathing, loss of consciousness, and death — often within minutes. The dog Neptune, who had been perfectly healthy that morning and had survived a far larger dose three weeks earlier, died in under half an hour after a dose calculated to be harmless.

Richet's first response was the correct scientific one: he suspected he had made a mistake. He had not. The result held. And when he understood it, he understood that it was the exact inverse of what the experiment had been designed to demonstrate. The first exposure had not protected the animal. The first exposure was the thing that made the second one lethal.

Because phylaxis meant protection, he named the phenomenon aphylaxis — without protection — and then, reportedly for how it sounded, changed it to anaphylaxie. Portier and Richet presented the finding to the Société de Biologie in Paris on 15 February 1902, in a short French paper titled De l'action anaphylactique de certains venins.

Why this was a genuine shock

It is hard to feel the force of this now, because we grew up knowing that allergies exist. In 1902 nobody did. Immunity was understood as protective by definition. The word carried its meaning inside it: to be immune was to be exempt from harm. Twenty years of triumphant work — von Behring's diphtheria antitoxin, Ehrlich's side-chain theory of antibodies, Metchnikoff's phagocytes — had all pointed the same way. An immune response that had been raised against a substance was a good thing you wanted more of.

Richet showed that the machinery of immune memory could be the machinery of injury. The specificity was the damning part: an anaphylactic dog reacted to that toxin and not to others, and a naive dog given the same small dose was fine. It was not poisoning. It was recognition. The body had learned the substance and had learned it wrongly, and the learning was what killed.

Every autoimmune disease, every drug allergy, every peanut reaction, and every transplant rejection is downstream of the conceptual door Richet opened: immune specificity is a mechanism, not a moral position, and it can be pointed at the wrong thing.

Paul Portier, who did not get the prize

The 1913 Nobel went to Richet alone. Portier was a full partner in the discovery — he was on the yacht, he did the bench work alongside Richet, and his name is first on the 1902 paper. He was not included, and he never publicly complained about it; he went on to a distinguished career in marine and comparative physiology and lived until 1962. Richet himself credited Portier consistently, including in his Nobel lecture. The omission is one of the clearer single-name errors in the history of the prize, and this page names Portier as co-discoverer because that is what he was.

4. Serum Sickness and the Clinical Context

Richet's dogs did not land in an empty field. By 1902 doctors already had a problem they could not explain, and it came from medicine's greatest recent success.

Emil von Behring's diphtheria antitoxin — the work that won the very first Nobel Prize in Physiology or Medicine, in 1901 — was made by immunising horses and harvesting their serum. It saved an enormous number of children. But it meant injecting large volumes of horse protein into humans, sometimes repeatedly, and clinicians were seeing something odd: a first course usually went smoothly, while a later course in the same patient could produce fever, rash, joint pain, swollen glands and, occasionally, sudden collapse.

In Vienna, the paediatricians Clemens von Pirquet and Béla Schick studied this systematically and published Die Serumkrankheit (Serum Sickness) in 1905. They saw what Richet had seen: the reaction depended on prior exposure, its timing shortened dramatically on re-exposure, and it was the patient's own response — not a contaminant in the serum — that caused it.

In 1906 von Pirquet coined the word that stuck. He needed a neutral term for "altered reactivity" that did not prejudge whether the alteration was protective or harmful, and from the Greek allos (other) and ergon (work) he built allergy. It was meant to cover immunity and hypersensitivity together. Usage narrowed it to the harmful half within a generation, which is why the word means what it means today.

So the field has a fairly precise birthday. Richet and Portier supplied the mechanism in 1902; von Pirquet and Schick supplied the clinical description in 1905 and the vocabulary in 1906. Serum sickness itself became much rarer once antitoxins moved to purified and then human or recombinant products — but it never vanished, and it turns up today after antivenoms, anti-thymocyte globulin, rabies immunoglobulin of equine origin, and certain monoclonal antibodies.

5. What Is Actually Happening

Here is the modern account in plain language. It is worth understanding, because it explains almost every practical rule in the sections that follow.

Step one: sensitisation (nothing visible happens)

The first meaningful encounter with the substance — a peanut protein, a bee venom component, a drug — is silent. Immune cells take a fragment of it, show it to helper T cells, and in susceptible people the response goes down a particular route (a "type 2" response) that instructs B cells to make an antibody class called IgE.

IgE is unusual. Most antibodies float in blood and lymph doing their work in solution. IgE hardly stays in circulation at all — it goes and parks on cells, binding tightly to high-affinity receptors on the surface of mast cells (which sit in skin, gut lining, airway walls and around blood vessels) and basophils (which circulate). Once bound, it can sit there for months.

The person now has loaded cells and no symptoms whatsoever. This is why a first exposure to a food is so often uneventful and the third or the thirtieth is not, and it is why "but she's eaten it before" is never a reason to doubt an allergic reaction. Prior uneventful exposure is a requirement, not a defence.

Step two: the trigger (everything happens at once)

On a later exposure, the allergen molecule bridges two adjacent IgE antibodies on the same mast cell. That physical cross-linking is the switch. The receptors cluster, a signalling cascade fires, and the cell empties its granules into the surrounding tissue in seconds.

Out come pre-formed mediators — histamine, tryptase, heparin, chymase — followed within minutes by newly manufactured ones: prostaglandin D2, the leukotrienes, and platelet-activating factor. Millions of cells across the whole body can do this simultaneously, which is what separates anaphylaxis from a local reaction such as a single hive.

Step three: what those mediators do to a body

Honest caveat: not every anaphylaxis is IgE-mediated

The IgE story above is the commonest route and the one worth learning first. It is not the only one, and pretending otherwise causes real reactions to be dismissed because "the allergy test was negative."

This is why the modern definition of anaphylaxis is clinical, not mechanistic. The old habit of calling non-IgE events "anaphylactoid" has been abandoned in current guidance, because it implied a lesser condition treated differently. It is not and it is not. If the clinical picture is anaphylaxis, it is anaphylaxis, and the treatment is the same regardless of the pathway that produced it.

6. Recognising Anaphylaxis

This is where anaphylaxis is won or lost, because the treatment is simple and the recognition is not. The single most useful thing on this page may be the sentence in the box below.

The pattern to hold in your head

Anaphylaxis is sudden (usually minutes to two hours after exposure), it is multi-system, and it is getting worse rather than better. Current international guidance recognises it in two situations:

  1. Skin or mucosal signs — widespread hives, flushing, itch, or swelling of lips, tongue or uvula — plus at least one of: difficulty breathing (wheeze, cough, stridor, throat tightness, hoarseness, blue lips), a drop in blood pressure or its consequences (faintness, collapse, confusion, incontinence), or severe gastrointestinal symptoms such as repeated vomiting and severe crampy abdominal pain.
  2. No skin signs at all, but sudden low blood pressure, bronchospasm or throat involvement after exposure to something that person is known or strongly suspected to react to. This second route exists precisely so that a reaction without a rash still counts.

The trap: skin signs are absent in a meaningful minority

Most people — including many non-specialist clinicians — carry a mental image of anaphylaxis that is essentially "covered in hives and struggling to breathe." Hives are present in the large majority of cases, but in roughly one in ten to one in five they are absent, and their absence is associated with delayed recognition and delayed adrenaline.

Reactions that present as sudden collapse with no rash, or as severe abdominal pain and vomiting with no rash, are exactly the ones that get labelled a faint, a panic attack, food poisoning, or a vasovagal episode. If the timing fits an exposure and more than one body system is involved, treat it as anaphylaxis. You are not required to see hives.

Gastrointestinal symptoms are a legitimate presenting feature

Repetitive vomiting and severe cramping abdominal pain after an exposure count as an organ system in their own right — this was made explicit in the 2020 guidance rather than being treated as an incidental detail. In children with food-triggered reactions, vomiting is often the first objective sign. In insect-sting and drug reactions, severe gut symptoms may be the only warning before cardiovascular collapse.

Biphasic reactions: it can come back hours later

In a minority of people the reaction settles, the person feels genuinely better — and then it returns hours later without any further exposure. A meta-analysis pooling twelve studies of nearly 2,900 adults found biphasic reactions in about one in twenty, and calculated that six hours of observation after the initial reaction resolves rules out recurrence in more than 97% of cases; extending observation to 8–12 hours adds a little more.

Two practical consequences: go to hospital even if the adrenaline worked, and if you were treated and sent home, do not spend the next several hours alone. Second-phase reactions can be as severe as the first.

The common triggers

Some people carry an amplified baseline that makes any trigger worse: mastocytosis, hereditary alpha tryptasemia, and poorly controlled asthma. Asthma in particular is the strongest single predictor of a fatal food reaction, and getting asthma properly controlled is one of the most useful things a food-allergic person can do about their allergy.

7. Adrenaline, and Why It Is First

Read this section before you need it. It is written to be acted on.

The rule

Intramuscular adrenaline (epinephrine) into the outer thigh is the first-line treatment for anaphylaxis. Give it immediately. Nothing else comes first, and nothing else substitutes for it.

Adrenaline is the only treatment that acts on every part of the problem at once. It tightens leaking blood vessels and raises blood pressure, it relaxes the airway muscle, it reduces swelling in the throat, it supports the heart, and it dampens further mediator release from mast cells. It works within minutes. In a person having anaphylaxis, given intramuscularly at the correct dose, it is a safe drug — the risk of withholding it is far greater than the risk of giving it.

How to give it

Carry two

Two auto-injectors, together, wherever you go. Not one at home and one at school. Not one in the car. The reasons are unglamorous and all real: a device can misfire, a dose can be given into the wrong place, a first dose can be inadequate for the severity of the reaction, and the ambulance can be more than fifteen minutes away. A device in a drawer at home has never helped anyone.

Check the expiry date and set a reminder before it lapses. Look through the viewing window: the liquid should be clear and colourless. If it is brown, cloudy, or has solid material in it, replace it — it has degraded. Do not leave devices in a hot car or a cold one; heat in particular destroys adrenaline. And learn your own device by handling a trainer, because they are not all operated the same way — the familiar "blue to the sky, orange to the thigh" mnemonic belongs to one brand and does not describe the others.

Antihistamines and steroids are not substitutes

This needs to be unambiguous, because the mistake is common and it is lethal.

Both classes have a place after adrenaline, as add-ons for symptom relief. Neither has a place instead of it.

Delay is the factor most consistently associated with death

The clearest evidence comes from a UK register of fatal anaphylaxis. In that series, the median time from onset to respiratory or cardiac arrest was 5 minutes for drug-triggered reactions, 15 minutes for insect venom, and 30 minutes for food. Adrenaline was used at some point in 62% of the fatal cases — but it was given before arrest in only 14%. In other words, in most deaths adrenaline was eventually administered; it was administered too late to matter.

That same study carries an honest counterweight worth stating: at least three deaths in the series were attributed to adrenaline overdose. Those were intravenous administrations in a clinical setting, not intramuscular auto-injector doses. It is a reason for hospitals to be careful with the IV route in a patient who is not in arrest. It is not a reason for anyone to hesitate over an auto-injector in the thigh.

What else to do, in order

  1. Give adrenaline. First. Before anything below this line.
  2. Call emergency services (911, 999, 112, or your local number). Say the word "anaphylaxis." Call even if the person improves — because of biphasic reactions, and because the effect of a single dose can wear off before the reaction has finished.
  3. Position them. Lie the person flat with their legs raised. If breathing is easier sitting up, let them sit — but keep the legs out straight and do not let them stand. If they are vomiting or unconscious, put them on their side in the recovery position. Pregnant patients should lie on the left side.
  4. Do not let them stand up or walk. This matters more than it sounds. Sudden upright posture in someone whose circulating volume has collapsed can stop the heart from filling; deaths have occurred at the moment a patient stood up or was walked to a car. Keep them down and let the ambulance come to them.
  5. Second dose at 5–15 minutes if there is no improvement, or if symptoms return.
  6. Stay with them. Anaphylaxis can deteriorate quickly and the person may not be able to tell you.
  7. Go to hospital and be observed. Guidance generally supports at least six hours of observation after symptoms resolve, longer for severe reactions, for people on beta-blockers, for those with asthma, and for anyone who needed more than one dose.

Afterwards

Get a replacement prescription before you leave, get a written anaphylaxis action plan, and get a referral to an allergist — identifying the trigger is what prevents the next episode. Consider medical ID jewellery. Teach the people around you, including at school or work, how and when to use the device; the person having the reaction is often too unwell to do it themselves.

One recent development is worth knowing about: a needle-free nasal adrenaline spray was approved in the United States in 2024 for adults and children of about 30 kg and over, with a lower-dose version subsequently approved for smaller children. It is an option for people who cannot or will not carry a needle device, and it does not change any of the rules above about acting fast and calling for help.

8. Alpha-Gal Syndrome: The Rule-Breaker

Everything in section 6 teaches you that allergic reactions are fast. Alpha-gal syndrome breaks that rule, which is why it goes undiagnosed for years, and why it deserves its own section here. This site has detailed coverage — see Alpha-Gal Syndrome and its deep-dive articles — and what follows is the short version.

What it is

Galactose-alpha-1,3-galactose ("alpha-gal") is a sugar molecule found on the cells of essentially all non-primate mammals. Humans, apes and Old World monkeys lost the enzyme that makes it, which is why we can safely treat it as foreign. Ordinarily we ignore it in food. In alpha-gal syndrome, a person develops IgE antibodies against that sugar and then reacts to mammalian meat — beef, pork, lamb, venison, rabbit — and sometimes to dairy, gelatin, and mammal-derived medical products.

The tick

The sensitising event is a tick bite. In the United States it is chiefly the lone star tick (Amblyomma americanum), whose range has expanded well beyond the historic southeast; other species are implicated in Australia, Europe, Asia, Africa and South America. Something in tick saliva presents alpha-gal to the immune system in a way that drives an IgE response, and people often report intense, prolonged itching at the bite site. Repeated bites raise antibody levels; avoiding further bites lets them fall.

The delay, and why it hides the diagnosis

Reactions typically begin three to six hours after eating, not within minutes. The leading explanation is that alpha-gal arrives attached to fats and must be digested, packaged into lipid particles and delivered to the circulation before it can reach the sensitised mast cells — and that takes hours.

The consequence is that nobody connects the dots. Someone eats a steak at seven, wakes at midnight covered in hives or vomiting, and blames the last thing they ate or drank, or the bedroom, or stress. Reactions are frequently nocturnal. Many patients are carried for years with a label of chronic idiopathic urticaria, irritable bowel syndrome, or idiopathic anaphylaxis. Gastrointestinal-predominant presentations — abdominal pain, diarrhoea, nausea — with no skin involvement at all are common, and those are the hardest to catch.

How it was found

The discovery is a good example of medicine noticing a geographic anomaly. Around 2007, oncologists observed that severe first-dose reactions to the cancer antibody cetuximab were far more common in parts of the southeastern United States than elsewhere. Investigators showed in 2008 that affected patients had pre-existing IgE against alpha-gal, which is present on the cetuximab molecule because of how it is manufactured. Within a year the same research group had connected that antibody to the puzzle of delayed red-meat reactions in the same region, and the syndrome had a name.

Practical points

The site's deep-dive coverage includes an anaphylaxis emergency plan and a dedicated testing and diagnosis article.

9. Testing and Diagnosis

Allergy testing is enormously useful when it is used to answer a question raised by a history, and actively harmful when it is used as a fishing expedition. The distinction is the single most important thing in this section.

Sensitisation is not allergy

A skin-prick test puts a drop of allergen extract on the skin and pricks through it; a wheal means IgE against that allergen is present on skin mast cells. A specific IgE blood test measures the same antibody in serum. Both answer one question: is there IgE against this substance? Neither answers the question you care about: will this person react if they eat it?

Plenty of people have detectable IgE to foods they eat happily every week. A positive test in someone with no history of reacting to that food often means nothing at all. Acting on it — removing the food "just in case" — is not a neutral choice: unnecessary avoidance is uncomfortable, nutritionally costly in children, and, as section 10 explains, can actually cause the allergy it was meant to prevent.

What these tests are good at is ruling things out. A negative test in someone with a doubtful history is reassuring. And higher values do shift the probability of a real reaction upward — but the number does not predict how severe a reaction would be. A modest result can precede a severe reaction and a very high one can accompany mild symptoms.

Tests that are sharper

Tryptase during an episode

Tryptase is a mast-cell enzyme released during degranulation, and measuring it can confirm after the fact that mast cells fired. Two rules make it useful:

Crucially, a normal tryptase does not exclude anaphylaxis. It is frequently normal in food-triggered reactions, and it is most often elevated in drug- and venom-triggered ones. Anaphylaxis remains a clinical diagnosis. Separately, a persistently raised baseline tryptase is a clue to mastocytosis or hereditary alpha tryptasemia, both of which raise the risk of severe reactions.

See an allergist

A proper assessment is a detailed history first, then targeted testing to confirm or refute what the history suggested, then a challenge if there is doubt. That sequence produces answers. The reverse sequence — a broad panel followed by an attempt to explain the results — produces long avoidance lists and no answers. An allergist also handles the things that genuinely change outcomes: venom immunotherapy, drug allergy de-labelling (most people labelled "penicillin allergic" are not, and the label leads to worse antibiotics), action plans, and the immunotherapy options in the next section.

A plain warning about commercial "food intolerance" and IgG panels

Direct-to-consumer tests that measure IgG or IgG4 antibodies against dozens or hundreds of foods are heavily marketed, often through clinics, gyms, wellness practitioners and online kits, and are typically sold as identifying "food sensitivities" or "food intolerances." They do not diagnose allergy, and they do not diagnose intolerance either.

The reason is not subtle. Food-specific IgG and IgG4 are a normal physiological response to eating food. Their presence indicates exposure and, if anything, tolerance — rising IgG4 is one of the markers that accompanies successful immunotherapy. The European Academy of Allergy and Clinical Immunology issued a formal Task Force statement in 2008 recommending against the use of IgG4 food testing as a diagnostic tool, and major allergy societies internationally have since reiterated the point.

What these panels reliably produce is a long list of foods the person eats often, followed by an elimination diet built on nothing. That costs money, restricts nutrition, generates anxiety around eating, and — the most serious harm — can delay the diagnosis of a real condition, whether that is a genuine IgE-mediated allergy, coeliac disease, inflammatory bowel disease, or alpha-gal syndrome. The same applies to hair analysis, applied kinesiology, cytotoxic testing, electrodermal or "bioresonance" devices, and pulse testing: none has demonstrated validity for diagnosing food allergy. See also Food Intolerance, which covers the real, non-allergic mechanisms that these tests are usually being bought to explain.

10. What Has Genuinely Changed

For most of the century after Richet, allergy management consisted of one instruction: avoid the food, and carry adrenaline in case you fail. That is no longer the whole of it. Three changes since 2015 are real, and one of them is an outright reversal of previous advice.

Early introduction: the advice used to be wrong

For years, parents of infants at high risk of food allergy were told to delay introducing peanut — US paediatric guidance in 2000 suggested waiting until age three. The reasoning was intuitive and completely untested. During the same period, peanut allergy rates rose sharply.

The LEAP trial (Learning Early About Peanut Allergy), published in the New England Journal of Medicine in 2015, tested it properly. It randomised 640 infants aged 4 to 11 months who already had severe eczema, egg allergy or both — that is, the highest-risk group, the ones the old advice was aimed at — to either consume peanut regularly or avoid it entirely until age five.

Among the children who started with a negative peanut skin test, peanut allergy at age five was 13.7% in the avoidance group and 1.9% in the consumption group. Among those who already had a small positive skin test, it was 35.3% versus 10.6%. Regular early eating did not cause peanut allergy. It prevented most of it. A follow-up study showed the protection persisted after a further twelve months of avoidance, so it was not merely temporary desensitisation.

Guidelines were rewritten. Current advice in the US, UK, Australia and elsewhere is to introduce peanut and egg-containing foods in infancy, around four to six months alongside other solids, with a specialist assessment first for infants with severe eczema or existing food allergy. Say it plainly: the old guidance to avoid early was wrong, it was based on plausibility rather than evidence, and it was reversed. It is one of the clearest self-corrections in modern paediatrics, and it is a fair reminder that "be cautious" is itself an intervention with consequences.

Oral immunotherapy: real, and honestly limited

Oral immunotherapy (OIT) means eating tiny, precisely measured, gradually increasing amounts of the allergen under medical supervision, then continuing a maintenance dose indefinitely. For peanut, the PALISADE trial published in 2018 tested a standardised peanut protein product in nearly 500 participants. Among the 4-to-17-year-olds who completed treatment, 67.2% could tolerate at least 600 mg of peanut protein at the exit challenge, versus 4.0% on placebo. The product was approved in the United States in 2020 for children and adolescents, and later extended to younger children.

The caveats are as important as the result, and a good clinician will lead with them:

Omalizumab: protection without eating the allergen

Omalizumab is an injected antibody that mops up circulating IgE and, indirectly, strips IgE receptors from mast cells. It has been used in asthma and chronic hives for years. The OUtMATCH trial, published in the New England Journal of Medicine in 2024, tested it in people — mostly children — allergic to peanut plus at least two other foods. After 16–20 weeks of injections, 67% of those on omalizumab could consume at least 600 mg of peanut protein without dose-limiting symptoms, compared with 7% on placebo, with similar results for the other foods tested.

It was approved in the United States in 2024 as an adjunct to food avoidance, to reduce reactions from accidental exposure in people aged one and over with IgE-mediated food allergy. The wording is deliberate and matters: it is not permission to eat the food. Avoidance continues, adrenaline continues to be carried, the injections continue every two to four weeks, and about a third of participants did not reach the protective threshold. It is expensive. But for someone with multiple severe food allergies, where OIT to each food separately is impractical, it is the first genuinely new option in a very long time.

A smaller change worth knowing

The addition of sesame to the US list of major allergens requiring declaration produced an unintended consequence: rather than segregate production lines, some manufacturers added sesame flour to products that had not contained it and declared it on the label, which is legal and which reduced the number of safe products for sesame-allergic people. It is a good illustration that labelling law shapes behaviour in ways nobody drafted, and a reason to keep reading labels on products you have bought safely for years.

11. Where Mainstream Medicine Agrees — and What Remains Debated

Broad agreement

Genuinely unsettled

12. Richet's Legacy and Its Shadow

Richet gave medicine something it could not have reasoned its way to. The idea that the immune system can be the agent of injury was not a refinement of existing theory; it contradicted it. From that single reversal come the modern understanding of allergy, of autoimmunity, of transplant rejection, and of drug hypersensitivity. Every auto-injector in every school bag traces back through a straight line to a dog in a Paris laboratory in February 1902 — and to Paul Portier, who should have shared the prize.

And Richet spent decades of his working life on ideas that were wrong, and one set of them was harmful.

The eugenics is the serious one. La Sélection humaine is not a stray paragraph; it is a sustained argument for state control of who may reproduce, with a racial hierarchy built into it, written by a man whose scientific authority gave the argument weight it had not earned. He led the French Eugenics Society. Whatever he intended, the movement he lent his name to supplied intellectual cover for forced sterilisation programmes across several countries and, in its most extreme form, for far worse. The fact that he was simultaneously a pacifist who campaigned against war does not resolve the contradiction; it is simply part of the record, and it should make anyone cautious about assuming that decency in one domain travels to another.

The metapsychics is different in kind — a distinguished scientist spending his later decades chasing ectoplasm and materialised spirits is a story about credulity, not cruelty. But it makes the same point in a gentler way. Richet applied real methodological seriousness to séances and still could not see through them, because he wanted the phenomena to be real and because a laboratory reflex is not proof against a conjuror. Expertise does not transfer. Rigour in one field does not immunise you in another.

This is, in the end, the reason to keep both halves of the man on one page. Richet's anaphylaxis work is not credible because Richet said it — it is credible because Portier and Richet reported an experiment that others repeated and extended, and because the mechanism they stumbled into turned out to explain a mountain of clinical observation they never made. Strip his name off it entirely and the science stands. Attach his name to his eugenics and it still falls, because the evidence was never there. That is how it is supposed to work, and it is the most useful thing this page can leave you with besides section 7.

13. Key Research Papers

  1. Portier P, Richet C. De l'action anaphylactique de certains venins. C R Soc Biol (Paris) 1902;54:170-172. The original report; French, pre-dating the modern indexes and not carried in PubMed. Search PubMed for coverage of this paper
  2. Cohen SG, Zelaya-Quesada M. Portier, Richet, and the discovery of anaphylaxis: a centennial. J Allergy Clin Immunol 2002;110(2):331-6
  3. Sampson HA, Muñoz-Furlong A, Campbell RL, et al. Second symposium on the definition and management of anaphylaxis: summary report. J Allergy Clin Immunol 2006;117(2):391-7
  4. Shaker MS, Wallace DV, Golden DBK, et al. Anaphylaxis — a 2020 practice parameter update, systematic review, and GRADE analysis. J Allergy Clin Immunol 2020;145(4):1082-1123
  5. Cardona V, Ansotegui IJ, Ebisawa M, et al. World Allergy Organization anaphylaxis guidance 2020. World Allergy Organ J 2020;13(10):100472
  6. Pumphrey RS. Lessons for management of anaphylaxis from a study of fatal reactions. Clin Exp Allergy 2000;30(8):1144-50
  7. Kim TH, Yoon SH, Hong H, Kang HR, Cho SH, Lee SY. Duration of observation for detecting a biphasic reaction in anaphylaxis: a meta-analysis. Int Arch Allergy Immunol 2019;179(1):31-36
  8. Chung CH, Mirakhur B, Chan E, et al. Cetuximab-induced anaphylaxis and IgE specific for galactose-alpha-1,3-galactose. N Engl J Med 2008;358(11):1109-17
  9. Commins SP, Satinover SM, Hosen J, et al. Delayed anaphylaxis, angioedema, or urticaria after consumption of red meat in patients with IgE antibodies specific for galactose-alpha-1,3-galactose. J Allergy Clin Immunol 2009;123(2):426-33
  10. Platts-Mills TAE, Li RC, Keshavarz B, Smith AR, Wilson JM. Diagnosis and management of patients with the alpha-gal syndrome. J Allergy Clin Immunol Pract 2020;8(1):15-23.e1
  11. Du Toit G, Roberts G, Sayre PH, et al. Randomized trial of peanut consumption in infants at risk for peanut allergy (LEAP). N Engl J Med 2015;372(9):803-13
  12. Vickery BP, Vereda A, Casale TB, et al; PALISADE Group of Clinical Investigators. AR101 oral immunotherapy for peanut allergy. N Engl J Med 2018;379(21):1991-2001
  13. Wood RA, Togias A, Sicherer SH, et al. Omalizumab for the treatment of multiple food allergies. N Engl J Med 2024;390(10):889-899
  14. Stapel SO, Asero R, Ballmer-Weber BK, et al. Testing for IgG4 against foods is not recommended as a diagnostic tool: EAACI Task Force Report. Allergy 2008;63(7):793-6

Live PubMed Searches

  1. Anaphylaxis epinephrine management
  2. Alpha-gal syndrome red meat allergy
  3. Peanut oral immunotherapy trial
  4. Early peanut introduction (LEAP)
  5. Biphasic anaphylaxis incidence

Connections

Back to top