The folate-deficient women nobody could find
Dr Clare Craig
For decades researchers hunted for the women who ought to have proved the folate theory: women whose pregnancies were affected because they were deficient in folate. They could not find them.
Low folate inside a cell can disrupt development. It seemed to follow that affected pregnancies must occur in women who did not have enough folate, and that giving them more would prevent the defect. But the deficient group could not be identified. Most women carrying an affected pregnancy had plentiful folate. Every attempt to measure it failed to find the deficient cohort. Then came the study that every policy to fortify has relied on since.
The data for the study came from Kirke and colleagues who took blood from over 56,000 pregnant women in Ireland and followed what happened to their pregnancies. Folate was actually measured in 84 affected pregnancies and 266 controls. The median was higher in the healthy pregnancies. But they could not show that the defects were confined to women who were deficient.
The missing link arrived two years later, and it did not come from new women or new measurements. Daly took the same blood results, divided the women into bands by folate level and compared the risk in each. What had changed was the analysis.
How one study became the missing link
It showed that women in the lowest band of red cell folate had around eight times the risk of those in the highest, with the risk falling steadily in between. It never reached zero. Even women at the very top of the range still had affected pregnancies, which is awkward for an explanation that the cause is inadequate folate.
Under other circumstances such a finding might have been dismissed with the words “correlation is not causation”. The study could not distinguish the effect of folate itself from the other characteristics that go with having a high folate concentration. Instead, the idea of a deficient group had been dispelled but the idea that interventions could help lived on.
Folate stores depend on absorption, on the enzymes that process folate, on common genetic variants and on vitamin B12. Folate reaching the embryo depends on many more steps than that. Diet is just one of several variables.
So knowing that there is a relationship between folate stores and risk tells us nothing on its own about how much diet could reduce that risk.
The question is not whether folate matters. It plainly does. The question is whether a lack of it in the mother’s diet, as opposed to her blood, has any effect on risk. The evidence compares groups of women, not quantities of folate in the diet.
Women with higher folate were not identical women who happened to have eaten more of it. They differed in vitamin B12, in absorption, in metabolism, in genetics, in diet and probably in many other respects. The same team had already shown B12 to be an independent risk factor for these defects in its own right and it also governs how much folate registers in the blood in the first place. A woman high in one tends to be high in the other, so the lower risk in the top band could be due either to their high folate or B12 levels. A single measurement taken once in pregnancy cannot tell them apart.
What fortification proposes is to try to move an entire population across an invisible threshold by adding a drug to bread. If the low risk in that group was due to the many variables that made the low risk women different, not just their folate stores, then raising everybody’s blood folate chemically will not reproduce the desired result.
A threshold drawn after the event
A lot hangs on this single study. Twenty years after Daly published, a commentary on the World Health Organization’s folate guideline described the primary data on deficiency as scant. It referred to this paper as the only case-control data available linking red cell folate to these defects. Despite plenty of blood samples being taken from pregnant women globally, the study has never been replicated.
Thirty years on there is still little consensus on whether the curve, a normal distribution, has any kind of meaningful threshold below which women could be said to be at risk.
Despite this the WHO uses a threshold based on this study. They recommend that red cell folate be kept above 906 nanomoles per litre. Nine hundred and six was the line drawn to differentiate the group at the extreme of low risk. The cut off is not objective. The same data yielded a threshold nearer 1,300 when the curve was fitted.
A later model of Chinese data landed close to the Irish figure and has been presented ever since as confirmation, which is not the same as repeating the study and finding the same thing again. Besides, deciding that only a tiny fraction of the population were representative of a normal healthy state is a bold move from the WHO.
The WHO themselves admitted that although this was “a strong recommendation” it relied on “low/very low quality evidence”. The guideline states that the data came from one case-control study with small numbers and that because low folate cannot explain every case the threshold cannot predict any individual woman’s risk. They then pushed for global introduction of the drug folic acid into everyone’s food as part of their “Sustainable Development Goals”.
The embryo is not fed through a pipe
At four weeks, as the risk period ends, the embryo is about four millimetres long and weighs around two milligrams, suspended in a fluid-filled sac perhaps a centimetre or two across. At the folate concentration of ordinary adult tissue its entire folate content would be about half a nanogram. Allow a hundredfold more for tissue dividing this fast and it is still only some tens of nanograms.
A woman carries fifteen to thirty milligrams of folate in her body, about half of it in her liver. Her stores exceed the embryo’s entire folate content hundreds of thousands of times over. Even a single day’s eating covers it thousands of times.
Folate is recycled rather than consumed, so what moves through the body matters as much as what sits in the stores. A warehouse full of something does not guarantee that it reaches where it is needed. The problem may not be the size of the warehouse. It may be the delivery system.
The mother’s blood is not, in any case, the right thing to be measuring. The neural tube closes by about the end of the fourth week after conception. Maternal blood does not flow through the placenta until around the tenth to twelfth week. For the whole of the period that matters her blood is the source of what the embryo receives, but it does not reach the embryo directly.
During the weeks when the neural tube closes, maternal blood does not wash directly over the embryo. What feeds it instead is the lining of the womb. Folate reaches the embryo along a chain of handovers:
maternal blood → uterine gland → secretions around the pregnancy → yolk sac → embryo
The uterine glands remain active into the tenth week and discharge their secretions where the early placenta can take them up. Fluid collecting in the cavity around the embryo acts as a reservoir from which the yolk sac draws. Folate receptor alpha is responsible for folate uptake at each handover, on the surface of the early placenta, on the glandular epithelium of the womb and on the yolk sac membrane.
What the embryo sees is not the same as what is in the mother’s blood. The only attempt at measuring the difference used bloods drawn between six and twelve weeks to compare to folate in the fluid around the embryo. It found levels were significantly lower in the sac than in the mother’s blood. Later in pregnancy folate is concentrated for the fetus and the levels run higher than in maternal blood.
More folate in the blood does not mean more reaches the embryo
When researchers finally measured both compartments in the same pregnancies, the amount in the mother’s blood did not predict the amount around the embryo. Forty-two women had folate assayed in their blood and in the sac fluid of the same pregnancy, matched pair by pair, at six to twelve weeks. There was no significant correlation between the two.
The same study contains a natural experiment on the point. Smokers in the group had markedly less folate in their blood than non-smokers:
| Maternal serum | Sac fluid | Estimated receptor occupancy | |
| Non-smokers | 13.9 nmol/L | 10.5 nmol/L | 78% |
| Smokers | 7.6 nmol/L | 7.5 nmol/L | 71% |
Maternal serum folate in the smokers was forty-five per cent lower. Their sac fluid was only twenty-nine per cent lower, and the gradient between blood and cavity, clearly present in the non-smokers, had disappeared altogether. Although the direction was similar the magnitude of the difference is massive. Admittedly this is a small observational comparison and more research is needed.
Taking this data we can figure out what it means for the folate receptor binding on the uterine gland cells itself. Folate receptor alpha binds the circulating form of folate tightly, with a dissociation constant of about three nanomoles per litre. Every sac concentration ever measured is well above this level. Even the lowest concentration would be enough, on a simple binding model, to occupy at least seventy per cent of the receptors.
Those are calculated estimates rather than measurements, taken from cultured cells rather than human uterine glands. What they suggest is that the system may already be operating on the flatter part of its response curve, where adding more produces progressively less. Whether that limits what actually crosses has never been measured.
The delivery system
Binding to uterine gland cells is still not the same as delivery to the embryo. Folate receptor alpha only takes folate in. Getting it out of the glandular cell and into the cavity is a different operation altogether, done either by pumps that cost the cell energy or by packaging it into the gland’s secretion. Neither is driven by how much folate is in the blood.
How much arrives must also depend on how many receptors the cell has made, how fast they cycle and what it does with the folate once inside. In cell experiments those things respond to the cell’s own folate status. None of it has been measured in a human uterine gland while a neural tube was closing.
How much folate receptor a cell puts on its surface is regulated by the same signalling that reads the cell’s nutrient status. Cells with plenty of folate make fewer receptors. This is a system with feedback rather than a pipe with pressure at one end. Its answer to more folate arriving is to need less apparatus for collecting it.
Having said that, the only other direct human study found the opposite. That earlier work, published in 1993, used a different assay and found folate concentrated inside the sac. It concluded that both folate and vitamin B12 are “transferred to the coelomic cavity against a concentration gradient”, the coelomic cavity being that fluid-filled space around the embryo. On that reading, the sac acts as a store held ready for the yolk sac.
The 2007 study, the one whose figures are used above, opened by observing that “there is currently no information on maternofetal folate transfer during the first trimester”. Those two papers are the entire direct evidence on folate concentrations inside the early human gestational sac, and they do not agree with each other. It has barely been measured.
What the high-dose trial actually proved
The key human evidence was the Medical Research Council trial. It showed that a very high pharmaceutical dose could reduce recurrence in women already known to be at high risk. It did not show that ordinary women were folate deficient, nor that the much smaller dose delivered through fortification works by the same route.
Four milligrams a day is ten times the standard supplement dose and roughly twenty times what fortification is expected to add to an ordinary diet. It cut recurrence by seventy-two per cent, six affected pregnancies in the treated arms against twenty-one in those not given it.
The only study at lower doses shared the pregnancy loss data and shows nine losses for every neural tube defect prevented. Without proper outcome data in the high dose trials it is impossible to know whether the effect was a benefit or just the illusion of one thanks to a survivorship bias.
In any case, even if a high dose could overcome a transport problem by overwhelming it, it is not proof that the original problem was dietary deficiency.
After decades of research, there is still no clearly defined population of women in whom folate-deficiency is responsible for neural tube defects. The principal evidence instead shows an association across a continuum, based largely on one small case-control dataset that has never been directly replicated. We know that a very high dose of folic acid can reduce recurrence in selected high-risk women. We do not know that this proves ordinary pregnancies are harmed by dietary folate deficiency or that adding a much smaller dose to everyone’s bread reproduces the same effect.
If deficiency is truly the cause, famine should provide the clearest natural experiment. That is the evidence examined next.
