Case studies in neural tube defect causation

Dr Clare Craig

Key points

• Neural tube defects present in regional epidemics.
• In Birmingham the rate nearly doubled, almost halved, then doubled again within twenty years.
• In Cameron County, Texas, a cluster followed a mould toxin in the corn crop. The toxin disables the folate receptor.
• In mice the toxin causes the defects and folic acid can mitigate that harm.
• In Shanxi province in China which has the highest rates in the world, pollutants from coal burning were found at higher concentrations in the placentas of affected pregnancies.
• In 1972 a serious attempt was made to identify an environmental cause in Britain. Work began and it was partly supported by one animal study but was then dropped.
• Folic acid can prevent defects caused by poisons. Folic acid treating a toxin induced condition is a drug effect not the replacement of a dietary deficiency with a nutrient.
• Britain is adding folic acid to flour, knowing groups that are at higher risk of toxin exposure have a higher rate, yet the exposures are not being measured.

There is a firm belief that it is folate deficiency that leads to neural tube defects. This is despite a lack of evidence of deficiency and the established effects of famine. Because of this belief other questions do not get asked about causation.

The risk of neural tube defects is not confined to a deficient group and, despite much effort, variation in incidence between countries could never be put down to differences in diet. Furthermore, higher incidence epidemics arrive in waves and cluster in particular places. Here are four case studies to examine examples of those clusters.

Birmingham, England

In 1966, Ian Leck reported the neural tube defect epidemics in Birmingham in the Lancet. The incidence was 2.0 per 1,000 births in 1936 and 1937. In 1940 it rose by 40 percent to 2.8 per 1,000. By 1948, the rate had fallen to 1.54 but it peaked again in 1956 at 2.8.

Elwood documented the same pattern elsewhere, comparing Dublin with the Boston and Providence lying-in hospitals across the years from 1900 to 1965 and finding epidemics that peaked in 1935 to 1939 and again in 1960 to 1965. The Dublin wave resembled the epidemic of neurological malformations in the north-east United States, which was at its highest between 1930 and 1934. Rochester, New York recorded a rise at much the same years as Boston and Providence. Rates rose and fell across whole cities on a timescale of years.

Looking closer at Birmingham where the rate nearly doubled, almost halved, then doubled again inside twenty years. Diet really does not fit as an explanation, especially when the 1948 low of 1.54 per 1,000 came during rationing. There was no dramatic change in diet and no intermittent preventative intervention. Genetics did not change. At a time with no screening and abortion, diagnostic rates at birth would have stayed steady too. What caused these periods of high rates?

Cameron County, Texas

In the spring of 1991, six anencephalic babies – where the brain has not formed – were born in six weeks at one hospital in Brownsville, Texas. Brownsville is in Cameron County and the incidence there for all neural tube defects was 1.47 per 1,000 births from 1986 to 1989. In 1990 and 1991 it was 2.71. The incidence for anencephaly alone rose from 0.96 to 1.97 per 1,000. Surveillance was then extended along the whole frontier and found that the raised rate was not confined to Brownsville. It ran across all fourteen Texas counties on the border with Mexico. The mothers were 93 percent Hispanic.

It is worth looking at another event shortly before. In the autumn of 1989, horses across the US began dying of equine leukoencephalomalacia. This is a description of liquefaction of the white matter of the brain. More than forty clusters were reported in badly hit Texas in under two months compared to a more usual one to five. Consequently, the corn feed for livestock was stopped. People carried on eating it.

It was later found that the contaminant was fumonisin, which is a toxin produced by a mould of maize. It affects brain development by disabling folate receptor alpha, the door that admits natural folates into brain cells, by destroying the fats that hold the receptor in the correct position. It was shown in 1997 that a 40 percent reduction in those fats almost completely blocked folate transport through that receptor which a developing brain depends on. The result of exposure to this toxin is brain cells which are starved of folate in the midst of plenty elsewhere in the body.

A retrospective study five to ten years later found that women who ate a moderate amount of tortilla in the first trimester had more than double the risk, an adjusted odds ratio of 2.4. There was a dose response relationship. There is a blood marker of fumonisin exposure and comparing exposure in six categories with resulting risk resulted in a high exposure group having a 4.4 fold increased risk. The twelve women with the highest exposure did not have affected births but the authors concluded that this was due to fetal death above a threshold of damage.

Mice experiments have shown fumonisin can produce neural tube defects. Folic acid reduces the proportion affected from 79 percent to 50 percent. Replacing the fats that hold the receptor in place reduced it right down to 5 percent. When folate deficient mice were exposed to fumonisin the result was unexpected. The toxin caused neural tube defects in 91 percent of mice on a normal diet compared to 36 percent in the deficient mice. Fetal death only accounted for half of that difference.

Shanxi, China

Shanxi in rural northern China has the unfortunate title of the place with the highest neural tube defect rates ever recorded.

Researchers have thought about unusual exposures. Coal fires are used for cooking but also heating bedrooms through the winter. Women with any exposure to indoor coal combustion had around a 60 percent higher risk of an affected pregnancy. In the county with the highest rate in the province a cluster sat within six kilometres of the coal mines in almost every year from 1998 to 2005, and incidence tracked mine production. What was in the coal? 

Incomplete coal combustion produces polycyclic aromatic hydrocarbons which were found in placentas. The levels were significantly higher in affected pregnancies than in controls. Higher levels were also present in maternal blood and even in cord blood, where polycyclic aromatic hydrocarbons were bound to DNA.

All coal contains arsenic in pyrites, or fool’s gold and in the organic matter. Western coal is processed and washed to remove it. In China, villagers dig their own coal from local pits and burn it raw in unvented stoves. Across 5,314 samples the arithmetic mean for Chinese coals is 6.97 milligrams per kilogram, and the same authors list southern Shanxi among the areas where arsenic in coal is slightly enriched.

Shanxi is one of China’s recognised areas of arsenic poisoning from well water. Around half of the pump wells exceeded the safe limit. A woman there is exposed through water but also through the unprocessed coal. The body processes arsenic by adding a methyl group to it so it can be disposed of. Folate provides the methyl groups for gene switching and depleting this resource could impact gene switching in the same way low folate levels do. 

When placental arsenic was measured in women from northern China, the women with a level above the median were 63 percent more likely to have an affected pregnancy. The same risk was not evident in women voluntarily taking folic acid but this result is bound to show confounding with poorer, less educated women being at higher risk and not taking folic acid.

A folic acid intervention resulted in a large reduction of neural tube defects in Shanxi and this evidence is held up as evidence for flour fortification. A folate-responsive condition is not the same as a folate-deficient causation. Also, no one checked whether the folic acid resulted in miscarriage and pregnancy loss such that the rate fell because development failed altogether, as was seen in the only trial of low dose supplements.

The search for the toxin in Britain

The epidemic nature of neural tube defect outbreaks led to a concerted effort to find an environmental cause in Britain. In 1972 James Renwick noted a relationship between the geography, seasonality and timing of potato blight and neural tube defect incidence. He suspected a compound produced by potatoes to defend themselves from the fungus.

The hypothesis was tested by feeding blighted potato concentrate to pregnant rats and marmosets. The rats did not have birth defects but four of eleven marmoset fetuses had gross cranial defects. Next a trial of high risk women who had had an affected pregnancy was carried out with one group asked to avoid potatoes. The trial was too small to show an effect but there was no reduction in recurrence. If potatoes were the only cause, this trial might have detected it, but if potatoes accounted for as much as half of the cases, a trial this small would have failed to detect the effect about 92 percent of the time.

The question has never been properly answered. The search for an environmental cause in Britain ended in the mid-1970s, while the rates were falling steeply.

What do these cases have in common?

None of these investigations found inadequate dietary folate was the cause. In each situation there was an external agent: a mycotoxin that disabled the folate receptor, coal pollutants in the placenta and a plant toxin that produced cranial defects in primates. The list of accepted non-dietary causes that have a contributing role has kept growing since: the epilepsy drug valproate, arsenic, maternal hyperthermia, pesticides, maternal diabetes.

Despite this growing body of evidence the consensus remains that dietary folate deficiency, something that has never been properly demonstrated, is the sole underlying cause of neural tube defects. The result is that other causative agents no longer get investigated and exposure continues.