Analysis based on dosage and long follow up all points one way
Dr Clare Craig
Why cancer is the question
Folic acid forms the building blocks of DNA and drives cell division. Early embryos have cells dividing at enormous speed and they need a ready supply. Anything that helps cell division presents a hypothetical risk for people with cancer and also for the initiation of cancer in healthy people. Consequently, studies have been carried out to try and measure where that hypothetical concern is present in reality.
Measuring the harm
The strongest test for harm is a randomised controlled trial, in which people are assigned by the toss of a coin to receive either the treatment or a placebo and are then followed up to see what happens. In 2013 researchers pooled every large folic acid trial, around fifty thousand people, into one analysis in the Lancet. They reported no significant effect on cancer. That result became the basis of the official safety verdict.
A year earlier, a separate more careful review split the same trials by dose given. It found an increase in cancer in the trials using lower doses, close to the amount fortification delivers but none in the trials using much higher doses.
That headline, a six per cent increase reported as no significant effect, is an average across every dose the trials used. Split the trials into three dose groups, on both the published figures and the patient-level data behind the 2013 analysis, and the average comes apart.
| Folic acid dose | Patient-level data (13 trials) Increased risk | Sample size (N) Relative RiskConfidence intervals | Increased risk – published pool (10 trials) | Sample size (N) Relative RiskConfidence intervalspool (10 trials) |
|---|---|---|---|---|
| Below 1 mg (0.4–0.8) — the fortification range | 23% *(statistically significant) | N 10,279 (4 trials); RR 1.23 (1.03–1.46) | 18%*(statistically significant) | N 7,776 (2 trials); RR 1.18 (1.02–1.37) |
| 1 mg | 30% | N 1,713 (2 trials); RR 1.30 (0.76–2.24) | 29% | N 1,693 (2 trials); RR 1.29 (0.83–1.99) |
| Above 1 mg (2–40) | 2% | N 37,629 (7 trials); RR 1.02 (0.95–1.09) | 3% | N 28,764 (5 trials); RR 1.03 (0.96–1.11) |
Table 1: The two columns are independent analyses of the same body of trials, one built from individual patients’ records and the other from the summary results each trial published. They include slightly different trials and measure follow-up over slightly different periods, so the figures are close rather than identical. RR is the risk ratio: 1.00 means no difference between the folic acid and placebo groups, and 1.23 means cancer was 23 per cent more common with folic acid. Both find the increase at fortification-level doses, and neither finds one above 1 mg.
The increase is seen at the exact range that fortifying flour will deliver. That is the reverse of how a drug normally behaves, since a larger dose would usually carry more risk, not less. The result below 1 milligram is the same even if the single polyp-prevention trial that critics single out is excluded and it is seen in ordinary adults rather than just in high-risk groups.
Prostate cancer in particular is a concern. In one trial, men given 1 milligram of folic acid a day were diagnosed with prostate cancer at 2.6 times the rate of those on a placebo, a ten-year risk of 9.7 per cent against 3.3 per cent. The trial had been designed to study bowel polyps rather than the prostate. Sixteen years later there have still been no further trials to measure this properly. Instead, researchers re-examined the data from the other folic acid trials already run. Pooled together, six trials covering around 25,000 men showed a 24 per cent increase in prostate cancer among the men given folic acid (relative risk 1.24, 95 per cent confidence interval 1.03 to 1.49). A reanalysis at a case level claimed there was no signal but they specifically excluded dosages within the fortification range. Folate obtained from food carried no such risk, which points to the synthetic form rather than the vitamin as it occurs in the diet.
What fortification delivers
Under the new regulations, flour must carry 250 micrograms of folic acid per 100 grams. This works out at about 188 micrograms once processing losses are allowed for. Government modelling puts the extra intake for an average woman of childbearing age at roughly 90 micrograms a day, which on its own looks modest. The implication would be that a woman’s wheat intake each day would not exceed either two thin slices of bread or a quarter of a pizza. However, this is clearly diet dependent. It is easy for a carbohydrate heavy diet to quickly breach safe levels as shown in the table.
Folic acid on a heavy carbohydrate day
| Food | Assumption | Folic acid (µg) |
|---|---|---|
| Breakfast cereal (Special K) | 100 g, existing voluntary fortification | 334 |
| Biscuits | ~40 g flour | 75 |
| 2 slices of toast | 60 g flour | 113 |
| Sausage roll | Pastry made with 25 g flour | 47 |
| Cake | Slice made with ~50 g flour | 94 |
| Supermarket thick crust pizza | Base made with ~250 g flour | 470 |
| From food, no supplement | 1,133 | |
| Multivitamin | Supplement | 400 |
| Total | 1,533 |
Wheat foods use 188 µg of folic acid per 100 g of flour, the amount left after processing and baking losses; the regulations add 250 µg to raw flour and about a quarter is lost before it is eaten. The cereal figure is folic acid already added under existing voluntary fortification and is eaten as sold, so no cooking loss applies. The multivitamin is a supplement. Food alone, with no supplement, already exceeds the 1,000 µg (1 mg) upper limit.
Women planning a pregnancy are still advised to take a 400 microgram supplement, many breakfast cereals are already fortified, and heavy wheat eaters will take in more than the average. The government estimate that between 225,000 and 400,000 people will, once their annual intake is considered, have on average exceeded the upper limit every day. Many more will have exceeded it on some days. The government also admits that even before fortification 72,000 primary school children already exceed the limit on folic acid intake largely from breakfast cereals. They did not estimate what proportion of children would breach the upper limit after fortification and there are no studies to say whether the “upper limit” is safe anyway in children. It is just a guess.
Why the signal is more likely understated
Every feature of the cancer trials limited detection of any potential safety signal.
| Feature of the trials | What it does to the result |
|---|---|
| Most ran only about 5 years | Misses cancers that take longer to appear; the 2013 authors said their work could not address later cancers or later effects in younger people. |
| Two Norwegian trials, followed for cancer for about six and a half years, well after treatment ended. | Found significantly more cancer and more deaths, including deaths from cancer, in a country that does not fortify its flour. (The study was in heart patients). |
| Many run in the US and Canada where background fortification dampens any signal | Where the placebo group ate folic acid too the trial was comparing folic acid with slightly more folic acid, which blurs any difference |
| Conducted in patients who were already ill | Heart, kidney and bowel-growth patients, not the healthy adults, children and non-pregnant people that fortification will reach |
Given the inadequacy of the current evidence perhaps the government should run one before spending money on an intervention that would make safety studies hard to carry out and expose the whole population to any potential risk.
What it means
The claim that the existing studies show the policy to be safe is false. They do not. Once dosage is considered there is a detectable increase in cancer at the exact dose that would result from fortification. This pattern was found independently by two teams, even in trials too short to capture a delayed effect. At fortification doses the harm is clearly seen with longer follow-up.
The burden of proof is the point. Fortification is not a pill that a person chooses to take. It is added to a staple food and eaten, unavoidably and for a lifetime, by a whole population, most of whom gain nothing from it. For an exposure that broad and that involuntary, the standard should be a positive demonstration of benefit for all as well as safety. The safety studies have been inadequate but even they, at the dose fortification will use, point towards harm.
