13.6 — Allergy, Autoimmunity and Immunodeficiency

A system powerful enough to destroy any invader, and specific enough to recognise almost any molecule, can fail in exactly three ways: it can attack something harmless, it can attack you, or it can fail to attack anything.

All three are common. Between them they account for a substantial share of chronic disease in wealthy countries, and — encouragingly — the treatments for all three have improved more in the last twenty years than in the preceding fifty.

Allergy

An immune response to something harmless.

Around 20 to 30 percent of people in industrialised countries have at least one allergic condition, and the rate has risen substantially over the last half-century.

The mechanism

Sensitisation happens first, and produces no symptoms at all.

On first exposure, the allergen is processed and presented, a Th2 response is generated, and B cells class-switch to IgE (Chapter 13.3). That IgE binds to mast cells throughout the body and sits there, waiting.

So the first exposure is silent. The person is now loaded.

On re-exposure, the allergen cross-links two adjacent IgE molecules on a mast cell. That cross-linking is the trigger — one molecule binding is not enough, which is a genuine safety mechanism.

The mast cell degranulates within seconds, releasing histamine and other mediators.

Immediate phase (minutes) — vasodilation, leakiness, smooth muscle contraction, mucus, and itch. Sneezing, wheeze, hives, swelling.

Late phase (4 to 12 hours) — eosinophils and other cells arrive, producing more sustained inflammation. This is why an asthma attack can recur hours after apparently settling, and why someone with anaphylaxis is observed for several hours rather than sent home once they look better.

The four types of hypersensitivity

A classification worth knowing because it separates conditions that look similar and are treated differently.

Type I — immediate, IgE-mediated. Minutes. Anaphylaxis, hay fever, allergic asthma, food allergy.

Type II — antibody against cell surfaces. Hours to days. Haemolytic transfusion reactions (Chapter 13.7), some drug-induced blood disorders, Graves' disease and myasthenia gravis.

Type III — immune complexes. Antibody and antigen form complexes that deposit in tissues, particularly in small vessels, joints and kidneys, triggering complement and inflammation. Serum sickness, lupus.

Type IV — delayed, T cell-mediated. 48 to 72 hours, and no antibody involved at all. Contact dermatitis from nickel or latex, the tuberculin skin test, and coeliac disease.

The timing is the practical discriminator. A reaction in minutes is IgE-mediated; a reaction two days later is T cell-mediated, and treating the second with antihistamines and adrenaline achieves nothing.

Food allergy

Affects around 3 to 6 percent of children and 1 to 2 percent of adults.

The commonest triggers are milk, egg, peanut, tree nuts, soy, wheat, fish and shellfish. In children, milk and egg allergies are usually outgrown; peanut, tree nut and shellfish allergies usually are not.

And it must be distinguished from food intolerance, because they are confused constantly and the implications differ enormously.

	Allergy	Intolerance
Mechanism	Immune	Enzyme or chemical
Onset	Minutes	Hours
Dose	Trace amounts can kill	Dose-dependent
Danger	Can be fatal	Unpleasant, not fatal
Example	Peanut	Lactose

Lactose intolerance is not an allergy (Chapter 1.2). Coeliac disease is not an allergy either — it is autoimmune (Chapter 9.3).

And the advice on preventing peanut allergy has completely reversed, which is worth knowing because the old advice is still repeated.

Guidance from around 2000 recommended avoiding peanuts in infancy in high-risk children. Peanut allergy rates roughly tripled over the following decade.

The LEAP trial, published in 2015, randomised high-risk infants to early peanut introduction or avoidance. Early introduction reduced peanut allergy at age five by around 80 percent.

The guidance now is the opposite of what it was: introduce allergenic foods early, from around four to six months, including in high-risk infants — with medical supervision where there is severe eczema or existing egg allergy.

It is a good example of a plausible intervention causing harm, and of a trial reversing it.

Treatment

Avoidance, where practical.

Antihistamines — blocking H1 receptors. Second-generation agents do not cross the blood–brain barrier and are therefore non-sedating (Chapter 11.2).

Nasal steroids — the most effective treatment for allergic rhinitis, and consistently underused because people expect an immediate effect. They take several days to work and should be taken regularly rather than as needed.

Adrenaline for anaphylaxis (Chapter 23.5).

Immunotherapy — and this is the one that changes the disease rather than the symptoms.

Repeated exposure to gradually increasing doses of the allergen shifts the response from IgE-dominated to a tolerant one, generating blocking IgG antibodies and regulatory T cells.

It genuinely works, particularly for insect venom, grass and tree pollen, and dust mite. Venom immunotherapy reduces the risk of a systemic reaction to a subsequent sting from around 60 percent to under 5 percent, which is one of the largest treatment effects in allergy.

And oral immunotherapy for peanut allergy is now available. It does not cure the allergy; it raises the threshold so that an accidental exposure is unlikely to be dangerous, which changes daily life substantially for a family.

Anti-IgE antibodies (omalizumab) bind free IgE and prevent it attaching to mast cells, and are used in severe asthma and chronic urticaria.

Why allergy is increasing

The hygiene hypothesis, proposed by David Strachan in 1989, noted that hay fever was less common in children with more older siblings, and suggested that early infections protect.

It has been substantially refined. The current framing — the "old friends" hypothesis — is that the immune system develops normally through exposure to the microbes it co-evolved with, particularly gut bacteria and helminths, rather than through childhood infections as such.

Supporting observations: children raised on traditional farms have substantially lower allergy rates; the Amish, with high farm exposure, have far lower asthma than the Hutterites, who are genetically similar and use industrialised farming; antibiotic use in infancy is associated with increased allergy; and caesarean delivery, which alters initial gut colonisation, is associated with a modest increase.

What it does not support is avoiding hygiene. Handwashing prevents infections and does not cause allergy. The relevant exposure is microbial diversity in the environment, not dirt or disease.

The honest position is that this is a well-supported framework with an incompletely established mechanism, and that no intervention based on it has yet been shown to prevent allergy in a trial.

Autoimmunity

The immune system attacking your own tissue — the failure of the tolerance mechanisms in Chapter 13.2.

Around 5 to 10 percent of people are affected, and roughly 80 percent of them are women.

Why women

The female predominance is one of the most striking and least explained facts in immunology. Ratios reach 9 to 1 in lupus and Sjögren's syndrome.

Several contributing factors are established rather than one explanation.

Sex hormones — oestrogen generally enhances immune responses; testosterone suppresses them. Which fits the fact that women mount stronger responses to infection and vaccination, and also explains part of the autoimmune risk: a more vigorous system has more opportunity to err.

X chromosome genes. A disproportionate number of immune genes are on the X chromosome, and X-inactivation (Chapter 2.4) is incomplete — some genes escape it, so women express higher levels. And people with Klinefelter syndrome (XXY) have lupus risk approaching that of women, which points strongly at gene dosage rather than hormones alone.

Microchimerism — fetal cells persist in the mother for decades after pregnancy, and may contribute.

Why it happens at all

Genetic susceptibility — particularly HLA type (Chapter 13.2), plus many other variants each contributing a little.

Environmental triggers, and the best-established ones are:

Molecular mimicry — a microbial protein resembles a human one closely enough that the response cross-reacts.

The clearest example is rheumatic fever. Antibodies against streptococcal throat infection cross-react with heart valve tissue (Chapter 7.2), and the result is permanent valve damage. It is entirely preventable by treating the sore throat with penicillin, which is why that unglamorous intervention matters so much where the disease is still common.

And the strongest recently established example is Epstein–Barr virus and multiple sclerosis. A 2022 study following ten million US military personnel found that EBV infection increased the risk of multiple sclerosis 32-fold, and that MS essentially never occurred in people who had never had EBV. It is now considered a necessary though not sufficient cause, and it has made an EBV vaccine a serious research priority.

Smoking is a strong risk factor for rheumatoid arthritis, particularly in genetically susceptible people. Which makes it one of the few modifiable risk factors for an autoimmune disease.

Vitamin D deficiency shows consistent associations, though causation is not established.

The pattern of disease

Organ-specific — one target. Type 1 diabetes, Hashimoto's and Graves' thyroiditis, pernicious anaemia, coeliac disease, multiple sclerosis, myasthenia gravis, vitiligo.

Systemic — many targets. Lupus, rheumatoid arthritis, Sjögren's syndrome, systemic sclerosis, vasculitis.

And they cluster. A person with one autoimmune condition has a substantially raised risk of another, and a family history of any autoimmune disease raises the risk of all of them. This is why finding coeliac disease prompts a thyroid check, and why type 1 diabetes prompts screening for both.

Treatment, and where it has improved dramatically

The historical approach was broad immunosuppression — steroids, methotrexate, azathioprine. Effective and indiscriminate, with the infection and cancer risks that implies.

The modern approach targets specific molecules, and the results have transformed several diseases.

TNF inhibitors — infliximab, adalimumab. Rheumatoid arthritis, Crohn's disease, psoriasis, ankylosing spondylitis.

And the change in rheumatoid arthritis is worth stating. The deformed hands shown in older textbooks have become uncommon in countries with access to these drugs, because early aggressive treatment prevents the joint destruction rather than managing its consequences (Chapter 5.9).

B cell depletion — rituximab, for rheumatoid arthritis, vasculitis and multiple sclerosis.

IL-17 and IL-23 blockers — for psoriasis, with response rates that were unimaginable a generation ago.

JAK inhibitors — oral drugs blocking a shared intracellular signalling pathway.

And the direction of travel is toward restoring tolerance rather than suppressing the system — regulatory T cell therapies, and antigen-specific approaches. Early results in type 1 diabetes with an antibody that delays onset in high-risk relatives are the first success of that approach.

Immunodeficiency

Failure to respond adequately.

Primary — inherited

Over 450 identified conditions, individually rare and collectively affecting perhaps 1 in 1,200 people.

And the pattern of infection points to which arm has failed, which is genuinely useful diagnostically:

Antibody deficiency → recurrent bacterial infections of the sinuses, ears and lungs, starting after about six months when maternal IgG has waned.

T cell deficiency → viral, fungal and opportunistic infections, and severe reactions to live vaccines.

Phagocyte defects → skin and deep abscesses, and specific organisms (Chapter 13.1).

Complement deficiency → recurrent Neisseria infections, and nothing much else.

Severe combined immunodeficiency (SCID) — the most severe, with essentially no adaptive immunity.

Untreated, it is fatal in the first year. It is the condition of the "bubble boy", David Vetter, who lived from 1971 to 1984 in a sterile isolator.

And it is now curable. Bone marrow transplant, if performed before infections have taken hold, cures over 90 percent.

Which is exactly why newborn screening for SCID matters, and it is a beautiful piece of testing: the same dried blood spot used for metabolic screening is tested for a DNA by-product produced when T cells rearrange their receptor genes (Chapter 13.2). No T cells means none of that by-product. A single test on an existing sample identifies a condition that is curable before symptoms and fatal after.

Gene therapy has also cured SCID — and its history includes the early cases of leukaemia caused by vector insertion (Chapter 2.9), which drove the safety improvements now in use.

Common variable immunodeficiency is the commonest primary deficiency presenting in adults. Recurrent chest and sinus infections, and it is frequently diagnosed years late because each infection is treated individually without anyone stepping back. Regular immunoglobulin replacement transforms it.

The general warning signs of a primary immunodeficiency: unusually frequent or severe infections, infections with unusual organisms, infections needing intravenous antibiotics repeatedly, failure to thrive in a child, and a family history.

Secondary — acquired, and far commoner

HIV — Chapter 17.9.

Medication — chemotherapy, steroids, biologics, transplant immunosuppression. This is now the largest category by a wide margin, and it is a consequence of successful treatment elsewhere.

Malnutrition — the leading cause of immunodeficiency worldwide, and protein deficiency particularly impairs cell-mediated immunity.

Diabetes, kidney failure, liver disease, cancer, and splenectomy (Chapter 7.8).

Age — thymic involution and reduced vaccine responses (Chapter 12.7).

And the practical points for anyone immunosuppressed are specific and worth stating:

Vaccination matters more, not less — with the exception that live vaccines are contraindicated.

Fever is treated urgently rather than watched.

Food and travel precautions matter.

And skin surveillance matters, because reduced immune surveillance substantially increases skin cancer risk (Chapter 10.6).

What the next page fixes

The immune system's ability to distinguish self from non-self is what makes it useful — and it is also what makes giving someone else's blood or organ so difficult. Chapter 13.7 covers blood groups and transfusion, where getting the immunology wrong kills within minutes.