The search for the reason was abandoned halfway through
Dr Clare Craig
Britain and Ireland once carried the highest recorded rates of neural tube defect in the world. They no longer do. The rates rose for 60 yrs, peaked twice and then collapsed and neither the rise nor the fall has been explained.
If we have not explained causation then we cannot pretend that all neural tube defects are caused by folate deficiency. Before mass medicating the entire population surely we should look for answers as to why these islands developed extraordinary rates in the first place, why epidemics occurred in waves, and why current levels are far more aligned with levels across Europe.
The epidemic that disappeared
In three Dublin maternity hospitals between 1900 and 1965, a review of 476,459 births showed neural tube defects were commoner in the 1960s than at the start of the century. The Irish diet improved over those decades but the rate climbed anyway. Through the war years 1942 to 1945, when food was scarcer, rates fell.
There were two epidemic peaks of anencephaly in Dublin, one in 1935 to 1939 and one in 1960 to 1965. The earlier one resembled an epidemic of neurological malformations in the north-eastern United States that had peaked in 1930 to 1934. Something other than diet was changing on both sides of the Atlantic. Was there an environmental exposure?
After these two peaks the record shows a fourfold collapse between 1980 and 1994, from 4.69 per 1,000 births to 1.16 per 1,000. Public health advice to pregnant women to take folic acid only started at the tail of that decline. Since 1994 the Irish rate has sat flat at about 1 per 1,000 for thirty years.
In England a similar decline was at least in part due to abortion. Ultrasound techniques were improving and becoming more readily available, leading to earlier diagnosis and increasing terminations. The rate of affected pregnancies and affected births disconnected in England in that time. In Ireland, termination was against the law throughout, and the researchers examining the issue stated that the change was real and not accounted for by prenatal screening. It was not just births, the number of affected pregnancies fell by three quarters in a single generation.
The usual explanations for the fall do not fit. The fall across the British Isles was not accompanied by any substantial increase in dietary folate consumption.
In Europe the rates were between 0.6 and 1.3 per 1,000 throughout this period. This is a baseline that Ireland and Britain have since adopted. Whatever was causing the excess in Ireland and Britain eased in the 1980s and had gone by mid-1990s.
Searching for clues
At the time there was a real drive to find the answer to that question. In the early 1970s the cause of anencephaly was an open research problem and the busiest people working on it were British and Irish. Knox tested dietary intakes in 1972. Field and Kerr proposed potato blight in the Lancet in 1973 and Kinlen and Hewitt tested it in Scotland within months. Fedrick reported an association with maternal tea drinking in 1974. Leck drew those epidemiological clues together the same year.
Then in December 1976 Smithells, Sheppard and Schorah published their findings on vitamin levels in early pregnancy. Six mothers in the study went on to deliver a baby with a neural tube defect, and their first-trimester red cell folate and white cell vitamin C were lower than those of controls. The first ideas about supplementing for prevention were articulated in 1980.
At that point the question changed. Before 1976 it had been “what exposure causes this?”. After, the question was whether vitamins prevent it. These are not mutually exclusive questions but the earlier lines of inquiry disappear from the literature at that point and were not resumed. There was never evidence that something other than an environmental exposure was causing the epidemics and the search for it was left unfinished. Since then there is fragmentary evidence that helps flesh out some possibilities.
Shared pathways
The functioning of the folate pathway in the developing embryo is clearly key to neural tube defect causation. Broadly, folates are needed for two purposes: to build the blocks that make DNA and RNA and to carry methyl groups that help switch genes on and off. It is therefore fundamental to many of the processes taking place in a cell. Methionine is essential to the second process as it is a link in the chain taking the methyl group from folate and enabling the methylation involved in gene switching.
Ethionine, a drug that displaces methionine, results in a failure of methylation and produces neural tube defects in the developing mouse embryo. Other methionine cycle inhibitor drugs prevent closure in mouse embryos more generally. Arsenic is one. It acts through the same depletion because consuming methyl groups is how the body disposes of arsenic. In 2026 a group led from Peking University reported that mercuric chloride, given to mice across the days of neural tube formation, disrupted placental function and embryonic methionine metabolism and produced the defects.
Three chemically unrelated agents, methionine cycle inhibitor drugs, arsenic and mercury, all produce the defects and all act by disturbing methionine metabolism. Based on this evidence it is not actually folate that is the critical determinant it is gene methylation. Folate depletion is a fourth way that the methylation process can fail.
There has never been evidence of a specific group with folate deficiency who bore all the risk. In fact there is only one study that has shown a relationship in humans and even those with the highest folate levels were still exposed to some risk of neural tube defects. What is seen is a correlation whereby women with lower folate levels have increasing risk. Factors, like heavy metal exposure, result in draining of the folate resource while the body disposes of the toxin. Heavy metal exposure would be one example of a third factor that was the underlying cause that would itself result in a relationship whereby lower folate correlated both with more heavy metal exposure and higher risk of neural tube defects.
What do we know about the role of heavy metals?
Although the inquiry was abandoned we can piece together parts of the puzzle as to the role of heavy metals in neural tube defect causation.
The epidemiology of metals and neural tube defects was systematically reviewed in 2023. Thirty studies were examined and, unfortunately, a risk of bias from confounding and selection was judged high in most of them. Based on this evidence, which is the best we have, the verdict was support for an association with prenatal mercury and with manganese but with the evidence on arsenic, cadmium and lead inadequate either to establish an effect or to rule one out.
A Peking University group measured metals in placental tissue from 36 anencephaly cases, 44 spina bifida cases and 50 controls in Shanxi province. Median placental mercury was 2.25 nanograms per gram in cases compared to 1.16 in controls. The odds ratio for having a measure above the median was 8.80, with confidence intervals between 3.80 and 20.36. There was a dose response relationship with risk. The same group measured methylmercury separately in the same placentas and found an adjusted odds ratio of 3.64, interval 1.66 to 7.99.
Other Chinese studies found similar results based on cord blood, placenta or maternal serum at delivery. The main criticism of these findings is that they are measuring eight months after the event. That is true but the signal just keeps reappearing.
There is a small amount of research outside China with one substantial study in 2006 looking at 184 mothers of affected pregnancies and 225 controls. The study was in Texan counties along the Mexican border and tested Mexican American women for blood lead and urinary arsenic, cadmium and mercury. Median levels did not differ between cases and controls. However, among the women in the highest income group, cases were nine times as likely as controls to have raised urinary mercury, with a confidence interval running from 1.4 to 57. It was only this one subgroup in this small study and there was clearly a need for more research to answer this question.
What about places with high mercury exposure? In gold mining districts of Tanzania raised maternal mercury was associated with stillbirth with an odds ratio of 2.49 i.e. more than double the risk. For congenital anomalies the odds ratio was 2.24. Arsenic exposure was also related to miscarriage, stillbirth and preterm birth. In the copper and cobalt mining city of Lubumbashi in the Democratic Republic of Congo, babies with visible birth defects had more manganese in the fetal side of the placenta. The strongest single factor was the father working in mining, at 5.5. Neither study reported neural tube defects separately, and mining communities carry arsenic, lead and poverty alongside the mercury.
Amalgam
Dental amalgam is roughly half metallic mercury. Cord blood mercury rises with the number of maternal amalgam fillings and falls with the years since the last was placed. At autopsy mercury was measured in fetuses and infants under 1 year old and was found in the kidneys, the liver of fetuses and the cerebral cortex of infants all in proportion to the number of the mother’s fillings.
Hot drinks can cause mercury to be released from fillings. Warming a mouth rinse from 35 to 45 degrees raised mercury evaporation 1.7-fold, and chewing tripled exhaled mercury vapour. Hot drinks are therefore a plausible repeated stimulus. The amount released at higher temperatures has not been measured. The inverse correlation with filling age suggests exposure spikes depending on timing of the filling or its removal and any disturbance.
There is counter-evidence from a Norwegian cohort of 69,474 pregnancies which found no association between the number of amalgam-filled teeth and malformations overall. They did not look specifically at neural tube defects nor the timing of the filings.
The women with the heaviest occupational exposure were the dental nurses who mixed amalgam by hand. Their pregnancies have been studied, and no increase in spina bifida was found. However the study was only large enough to expect four or five cases. Had the rate been more than twice as high the study could have shown this but anything less than a doubling would not have been measurable this way.
Dental nurses have fillings in their own mouths like everyone else, so what is being compared is the extra mercury breathed at work. The control group were potentially exposed from their own fillings. Swedish data showed raised neonatal mortality in sons of dental nurses in the 1960s, falling steadily across the following two decades as amalgam use declined.
Tea drinking
A 1974 study compared 464 mothers of anencephalic babies with 1,785 controls and found that the cases drank significantly more tea. The whole difference came from the fact that cases were under-represented among women who drank no tea or only one or two cups. Women exposed to more than three cups had the full risk with no higher risk from further exposure. Part of the finding was that there was a a strong social class gradient.
The assumption was that the causative mechanism was caffeine exposure. Investigations into caffeine showed that it was not capable of being the toxin and so the tea link was forgotten. A 2022 meta-analysis reported an odds ratio of 1.37 with an interval from 0.96 to 1.95. That means that on current evidence tea either is not relevant to risk or perhaps doubles it. Other studies of tea drinking have found no signal or even a suggestion that it was protective.
As well as risk from mercury fillings, manganese, which is a known risk factor, is concentrated in tea. Coffee is hot and caffeinated and does not accumulate manganese. If caffeine were the agent, coffee should behave like tea but this question has not been examined. It may still be possible to examine this question in a population where mothers still have mercury fillings.
What about the fathers?
It is natural to assume the exposure risk must relate to the mother but that may be short sighted. In Lubumbashi, in the Democratic Republic of Congo, the strongest single factor was the father working in mining. Sperm DNA is methylated and that methylation determines which genes will switch on and when. Male mice fed a folate-deficient diet throughout life fathered litters with around thirty percent more birth defects. Their sperm showed altered methylation across genes critical for development. A separate group found that continuing the folate deficiency into the second generation tripled the rate of new mutations in the embryos. Mice lacking the MTHFR gene which is needed for folate dependent methylation also showed damage which was detectable still in the following generation.
The social class gradient in neural tube defects has been clear in literature for fifty years and was determined based on the father’s occupation. Exposure of the father to heavy metals at work is a variable that should not be ignored.
The high rates in Ireland and Britain
Ireland and Britain were the world’s heaviest tea drinkers. From the 1970s dental caries started to fall and subsequent composite fillings displaced amalgam. The cohorts reaching childbearing age from the 1980s carried less mercury in their mouths each year.
In the 1930s, during the epidemic peak, lead arsenate spraying and organomercury seed dressings were in wide use on both sides of the Atlantic. The body consumes methyl groups to dispose of both.
There is no proof that heavy metals were a factor in these epidemic peaks. There is enough suggestive evidence to say that the hypothesis should be investigated. Heavy metals are certainly a candidate as one causative factor. The fact that there were epidemic peaks a generation apart but coinciding in two countries suggests environmental exposures should be examined more closely. Neural tube defects as a whole are almost certainly caused by more than one factor. Some are genetic, some metabolic, some environmental. Folate deficiency is one factor among them and is not of relevance in well nourished countries. Treating folate as if it was the one and only cause was never a plausible account of this disease.
