Folic acid
is not folate.
Synthetic folic acid is added to food to help reduce spina bifida — a devastating condition, but one that affects a small minority of pregnancies. From December 2026 it will be in almost every loaf and biscuit sold in Britain, which means everyone will be eating it, every day. The original safety studies were done in rats. It was not until 2009 — by which point fortification was already global — that researchers showed human liver processes folic acid roughly fifty times more slowly than rats do, with a five-fold variation between individuals. The signals of harm that did emerge in later human trials — including an increased cancer risk — have never been adequately investigated. When did we consent to this?
From December 2026, you will not be able to opt out.
Synthetic folic acid will be mandatory in all non-wholemeal wheat flour sold in the United Kingdom, including imports. It will be present in bread, pasta, biscuits, cakes and pies — whether you bake at home or buy it. Most UK millers have already gone further, adding it to wholemeal and gluten-free products that the regulations do not require.
Only 9% of respondents to the government’s own consultation backed the policy we ended up with — and it became law with barely any parliamentary debate.
A public-health intervention that cannot be declined, that reaches every citizen regardless of their individual risk profile, deserves a higher evidence bar than this one has met.
Two molecules. One name.
Folate is the natural B-vitamin found in spinach, eggs, liver and beans. Folic acid is a different, oxidised molecule synthesised in a laboratory. Your body has to convert it — and does so slowly and variably.
Folate
- Found in dark leafy greens, beans, eggs, liver, citrus.
- Already in the reduced form the body can use immediately.
- Your body was built to receive this version directly — it crosses into the developing baby's brain through a dedicated receptor.
- Three billion years of evolution said yes to this molecule.
Folic acid
- Three billion years of evolution said no to this chemical. It was invented in 1943 for shelf life and industrial fortification.
- Must be reduced by the liver enzyme DHFR — a process humans perform at roughly 0.08% of the rate at which rats do it.
- Binds the folate receptor on the placenta around ten times more tightly than natural folate does — and in some studies, far more. Tighter is not the same as better. A receptor evolved to release its cargo cleanly may not work the way it should when gripping a foreign molecule.
- Unmetabolised folic acid is detectable in the blood of nearly every person in a fortified population — roughly 40% of older US adults still have it in their blood after an overnight fast.
On the most conservative reading of the only large randomised trial of folic acid in healthy pregnant women, there were nine foetal deaths in the supplemented arm for every neural-tube defect prevented. The trial's own principal investigator later asked, in The Lancet, whether folic acid works not by preventing malformation but by selectively killing malformed embryos. He called it terathanasia. Twenty-seven years later, the hypothesis remains untested.
Have the right questions even been asked?
None of what follows is proof. All of it is reason to ask questions that have not been asked.
The harm didn't stop at birth
The 9-to-1 ratio above describes pregnancies. The trial also followed the children who were born. In the 1998 follow-up paper, the authors reported that allergies and other conditions were broadly similar between the groups — except for middle-ear infections, which were significantly more common in the folic-acid group. Deaths among treated children were numerically higher than placebo at every follow-up visit. None reached significance on its own. The pattern was consistent. It was never followed up.
A precedent, and a question
Synthetic vitamin B₁₂ is documented to provoke autoantibodies against its own carrier protein in humans. If a lab-made B₁₂ can do this, could a lab-made B₉ do the same? Around 46% of children with autism carry autoantibodies against their own folate receptors, against roughly 5% in healthy children. Synthetic folic acid binds that receptor around ten times more tightly than the natural form — in some studies, far more — and releases poorly. Nobody has yet run the study that would close the loop between these facts. That is our hypothesis — and it deserves to be tested.
The mechanism points the same way
Unmetabolised folic acid circulates in the blood of almost every person in a fortified population — in roughly 40% of older US adults it persists even after an overnight fast. It engages a second folate receptor, FRβ, which sits on immune cells and is normally only activated during inflammation. Women with detectable unmetabolised folic acid in their blood have around 23% lower natural killer cell activity — the body's first-line defence against newly forming cancers. Chronic daily engagement of an immune checkpoint the body reserves for emergencies is an experiment no ethics committee could have approved.
In the only folic acid trial to follow its participants for a full decade, men taking folic acid were three times as likely to be diagnosed with prostate cancer as those taking placebo — 9.7% versus 3.3%. Some systematic reviews of shorter trials have agreed. Others have not. No randomised trial has ever run long enough to find the cancers that take fifteen or twenty years to emerge. Methotrexate — the original chemotherapy drug — works by blocking folate. The NHS's own guidance is that people with cancer should not take folic acid. From December 2026, they will.
A capacity–load–trigger framework.
"Autism" may be an umbrella — several distinct conditions with similar symptoms, grouped under a single name. This framework is about one of them: the subset where children carry autoantibodies against their own folate receptors, a plausibly autoimmune condition with a plausibly environmental trigger. We propose how it arises, and the role synthetic folic acid may play in driving it. The framework has three components.
Capacity
Intrinsic susceptibility. Genetic variants in MTHFR, DHFR and FOLR1 reduce the body's ability to process folic acid and deliver usable folate to the developing brain. Some populations are more affected than others.
Load
Exogenous pressures. Synthetic folic acid intake from supplements, fortification and infant formula. Drugs like valproate and paracetamol that deplete one-carbon metabolism. Bovine milk protein with folate-receptor cross-reactivity.
Trigger
Events that break immune tolerance. Infection, fever, or metabolic stress in a predisposed host. The result is folate-receptor auto-antibodies that block folate delivery to the foetal and infant brain — a chronic, self-sustaining deficit.
The framework reconciles the long-standing paradox that both folate deficiency and excess synthetic folic acid are associated with higher autism risk: in each case, the net effect is the same — compromised delivery of reduced folate to the developing brain during the windows when it is most needed.
Four questions. None has been properly asked.
What follows is not a comprehensive list. It is the four research priorities we believe most urgently need funding, using existing tools and existing samples where possible. Each one is answerable. None is being pursued at scale.
-
Did folate-receptor autoantibodies rise in step with fortification?
Stored sera and newborn dried blood spots from the pre- and post-fortification eras are held in biobanks in the United States, Canada, Ireland and elsewhere. An interrupted time-series analysis using those archived samples could tell us whether folate-receptor autoantibody prevalence changed as fortification was introduced. The samples already exist. The assays already exist. This study could be running inside six months. Why isn't it?
-
In which other conditions do these autoantibodies turn up?
Around 30% of women with unexplained infertility carry the same folate-receptor autoantibody. Beyond that, a near-vacuum. Asthma, eczema, inflammatory bowel disease, polycystic ovary syndrome, prostate cancer — conditions with high folate demand or documented folic acid signals — have barely been screened.
-
What are folate receptors beta and gamma doing under chronic exposure — and would we even know if they were being attacked?
Most of the autoantibody evidence so far concerns one of three folate receptors — alpha. The other two, beta and gamma, are barely studied. Synthetic folic acid binds them too. We have assays for autoantibodies against alpha; we do not have reliable ones for beta or gamma. Could they be developed? Until they are, we are measuring with the lights half-off.
-
Does substituting natural folate reverse the signals?
The cleanest test of the whole framework is a head-to-head randomised trial: synthetic folic acid against natural reduced folate (L-5-methyltetrahydrofolate) — in infant formula, in pregnancy supplementation, and in the wider population. Do folate-receptor autoantibodies fall? Does natural killer cell activity recover? Do developmental outcomes differ? A primary biomarker endpoint could be reached in three to four years. The catch is that in a country where synthetic folic acid is already in the food supply, the control arm is contaminated — every participant receives background fortification regardless of what is in their supplement. A clean trial needs either an unfortified population (increasingly rare) or a comparison across countries with different fortification policies. This is difficult. That is not a reason to postpone it.
Primary sources.
This is not a position we ask you to take on trust. Read the work. Check the citations. Challenge the arguments. If the evidence is what we say it is, the case speaks for itself.