I could scarcely believe it when I came across this study. It almost seems too good to be true. But when I went through the methods and results, the children of mothers who took phosphatidylcholine supplements during pregnancy scored about 8 points higher on an IQ test at age four. And fewer of them screened positive for autism traits, compared with the placebo group [1].
Now, before we get ahead of ourselves, let's go through this methodically. Because while I am very excited, we do need to be robust in our analysis. If you're pregnant, or hoping to be, should you be taking choline, and if so, how much and in what form? And should the rest of us be taking it too? Let's find out.
Table of Contents
- Found in an Egg
- The Rat Studies
- Essential for Humans
- Choline in Pregnancy
- The Colorado Trial
- The Heart Scare
- More Trials
- Is TMAO Really the Problem?
- Does the Form Matter?
- The New Trial
- What They Found
- The Catches
- What About Adults?
- What I'd Tell a Patient Who's Pregnant
- Is It Too Good to Be True?
- References
Found in an Egg
The molecule in this new trial was first found in an egg. In 1850, a chemist in Paris, Theodore Gobley, pulled a fatty substance out of egg yolk. He called it lecithin, after the Greek word for egg yolk. Twelve years later, another chemist, Adolph Strecker, heated lecithin and got a new chemical out of it. He called it choline [2].
That lecithin is what we now call phosphatidylcholine. It's a fat with choline attached, and it's still the main form of choline in food. For about a century, though, choline was a curiosity. Nobody thought you needed to eat it.
So does it matter for the brain?
The Rat Studies
In the late 1980s, two researchers, Warren Meck and Christina Williams, gave rats extra choline around the time of birth. Those rats grew up with better memory as adults. And their later work suggested it protected them against memory decline as they aged. Extra choline early in life seemed to change the brain for the rest of the rats' lives [3].
But rats aren't people. And at the time, choline wasn't even officially a human nutrient. So do people need it at all?
Essential for Humans
In 1991, Steven Zeisel's group admitted healthy men to hospital, and half of them went without any choline. Within three weeks, their blood choline fell by about 30%. And a liver enzyme, a sign of liver damage, started to climb [4].
That same paper set out what choline does. Your body uses it to build the outer wall of every cell. It makes acetylcholine, one of the brain's chemical messengers, which is involved in memory and attention, and which nerves use to move muscles. And it supplies small chemical tags, called methyl groups, that help switch genes on and off [4].
And a growing baby is making new brain cells by the billion, each of which needs a new wall.
Studies like this one led to choline being officially recognised, in 1998, as an essential nutrient, with a recommended intake of 450 milligrams a day in pregnancy [5].
So what happens in pregnancy, when the baby needs it too?
Choline in Pregnancy
Kevin Klatt, a researcher at Cornell, and his colleagues later put it this way: pregnancy places a unique stress on choline metabolism [6].
Both mother and baby need choline at the same time. The baby is building a brain and a body, and every new cell needs choline for its wall. And the mother's supply gets stretched. Even in women eating the recommended amount, the substances choline makes, the ones that supply those methyl tags, fall during pregnancy [7].
And when Klatt's team gave pregnant women an extra 550 milligrams of choline a day, it didn't stay with the mother. One of choline's products went up in the placenta and in the baby's cord blood [6].
And most women don't even reach the basic target. About 90 to 95% of pregnant women eat less than the recommended 450 milligrams. The richest everyday source is eggs, at about 150 milligrams each [8].
That's all great in theory. So what happens when pregnant women actually take choline? Well, in 2012, an early trial of phosphatidylcholine in pregnancy came out of North Carolina. 140 women. And it found nothing. Their babies did no better on development, language or memory at 10 and 12 months. The researchers wondered whether a longer follow-up might show effects that appear later [9].
So was the rat story wrong for people?
The Colorado Trial
Not necessarily. A team in Colorado was looking for something different. Robert Freedman's group, at the University of Colorado, had spent years studying how the brain filters out a repeated sound. Play a baby two clicks, one straight after the other. A mature brain reacts strongly to the first click and less to the second. In schizophrenia, that filter often doesn't work well.
That filter develops around the time of birth, helped by a receptor in the brain that choline switches on. It's called the alpha-7 nicotinic receptor. So their question was simple: could extra choline in pregnancy help that filter develop [10]?
In early 2013 they published a trial of 100 pregnant women. The mothers took 6.3 grams of phosphatidylcholine a day, about 900 milligrams of choline, or twice the recommended amount. The babies also got a small dose for their first few months. At five weeks old, 76% of the choline babies reacted less to that second click, against 43% on placebo [10].
So the early signs were good. But that same spring, choline ran into a much bigger problem.
The Heart Scare
In April 2013, a study of about 4,000 heart patients linked a substance made from choline to heart attacks [11]. Gut bacteria turn choline into a compound called TMAO, and two years earlier the same Cleveland Clinic team had shown that TMAO sped up clogged arteries in mice [12]. Now they followed patients who'd had an angiogram, a scan of the heart's arteries. Those with the highest TMAO levels were much more likely to have a heart attack, a stroke or die over the next three years [11].
So now there were two questions. Does choline really help the developing brain? And is it safe for your heart?
More Trials
In 2016, the Colorado children were checked again, at just over three. The gene for that receptor has also been linked to ADHD and autism, so the team asked parents about attention problems and social withdrawal [1].
The parents reported fewer attention problems and less social withdrawal in the choline group. But by then the trial was down to 49 children [13].
Then in 2018, a team at Cornell reported on a very carefully run trial. Back in 2009 and 2010, pregnant women in their third trimester had all their food provided by the researchers, so the team knew exactly how much choline each woman got. Half got 480 milligrams a day, about the recommended amount. The other half got 930. The babies in the higher group processed what they saw faster [14]. And when the children were followed to age seven, the higher-choline group was better at keeping their attention on a task [15].
It was tiny, just 26 women, and it was partly funded by egg and beef industry groups. But it was a direct test of about double the recommended amount, from a different group, and it pointed the same way. So what about that heart scare?
Is TMAO Really the Problem?
In 2019, researchers tested it with genetics. Some people are born with genes that raise their TMAO. If TMAO caused heart disease, those people should have more of it. They didn't. If anything, it ran the other way: diabetes and kidney disease push TMAO up. So TMAO may be a marker of people who are already unwell, rather than the cause [16].
Then the mouse results stopped lining up. In 2021, another team fed mice prone to clogged arteries extra phosphatidylcholine. Their TMAO doubled. But they ended up with less plaque in their arteries, not more [17].
And as far as I can find, nobody has yet run a trial that lowers TMAO in people and then counts heart attacks. So the link with heart disease is an association. It hasn't been shown to be the cause.
But if you're still wary of TMAO, there's a simpler question. Does the form of choline matter?
Does the Form Matter?
Choline comes in several forms. Phosphatidylcholine, the form in food, is only 14% choline by weight, which is why the women in the new trial needed ten large capsules a day. Supplements often use choline salts like choline bitartrate, which pack more choline into a much smaller tablet. And two forms are sold for the brain: citicoline, also called CDP-choline, and alpha-GPC.
- Choline by weight, calculated from molecular weight: phosphatidylcholine about 14%, choline bitartrate about 41%, choline chloride about 75%, citicoline (CDP-choline) about 21%, alpha-GPC about 40%.
In 2021, the Cleveland Clinic group tested forms for four weeks. Choline bitartrate pushed TMAO up. Phosphatidylcholine supplements didn't, and neither did four eggs a day [18].
A small German trial found alpha-GPC raised it too [19].
So out of an abundance of caution, the new trial selected phosphatidylcholine. So as of last year, 2025, how did the choline evidence for neurodevelopmental outcomes stack up, before this new 2026 study was released?
The New Trial
A 2025 review of the four human trials so far concluded the evidence wasn't strong enough to say either way [20].
Which is why this new, much larger 2026 trial is so critical. Here's what the new study did. Between 2017 and 2021, about 200 pregnant women at the University of Colorado's clinics were randomly assigned at 16 weeks of pregnancy. Half got 9 grams of phosphatidylcholine a day. That's about 1,300 milligrams of choline, nearly three times the recommended amount. The other half got identical-looking placebo capsules, filled with corn oil. They took them until the baby was born [1].
Both groups also took prenatal vitamins and were told to eat choline-rich foods. So this tests a supplement on top of good standard care.
They chose phosphatidylcholine rather than a choline salt for two reasons. It keeps blood choline up for longer. And it mostly escapes the gut bacteria that make TMAO.
The trial was paid for by the US National Institutes of Health and two foundations, not a supplement company. One author has advised, unpaid, on a company's genetic tests for how the body handles choline [1].
They followed the children to age four. The main things they measured were attention and social behaviour, using a parent questionnaire. And at four, the children also sat a formal IQ test. So what did they find?
What They Found
The children in the choline group averaged 106 on the IQ test. The placebo group averaged 98. Because they were testing three outcomes, the researchers had set themselves a stricter bar for calling a result real. This one cleared it [1].
The uncertainty around that number is wide, anywhere from about 2 points to about 14. For comparison, when very premature babies were given DHA after birth, in a trial of 480 children, the gain at age five was about 3.5 points. So 8 points would be more than double that [21].
Social behaviour improved too. Parents score problems like poor eye contact, or little interest in other people. In the placebo group, 24% of children scored 2 or more on social withdrawal. In the choline group, it was 4% [1].
The same questionnaire has an autism scale, and those withdrawal questions make up a big part of it. Eleven children scored in the range that would prompt a referral for an autism assessment, and also had language difficulties. Ten were in the placebo group. One was in the choline group [1].
And the click test fitted the theory again. At 12 weeks old, 70% of the choline babies reacted less to that second click, against 48% on placebo. But attention problems, one of the two main things they measured, didn't clearly differ [1].
If you're also taking the omega-3 fat DHA in pregnancy, there's a bonus. In 2022, a small trial from Kevin Klatt and the Cornell group found that extra choline raised DHA in the mothers' blood, probably because the liver uses choline's methyl tags to package DHA and send it out. Their main measure wasn't clearly different, the effect didn't show up clearly in the babies' cord blood, and the trial was part-funded by Balchem, a company that makes choline [7].
So is this too good to be true?
The Catches
There are three things that make me want to see it repeated.
First, half the children are missing from the IQ result. About 200 women started, and 103 children sat the test at four. Families with more education were much more likely to stay, and a mother's education also affects how children score. The authors say the result held when they used statistical modelling to fill in the missing children, but that appendix isn't published yet [1].
Second, the autism result. The trial didn't diagnose autism. It's a parent questionnaire flagging children for possible assessment. The authors call it exploratory, and it rests on just 11 children [1].
Third, both positive phosphatidylcholine trials come from the same Colorado group. The one from a different group, back in 2012, found nothing, though it only followed babies to 12 months [9]. This new trial is bigger than any before it. But it still needs someone else to repeat it.
So where does that leave us?
What About Adults?
First, the rest of us. Adults need choline too. Remember those men in 1991, whose livers started to struggle within three weeks without it. And most adults fall short of the recommended amount, which is 550 milligrams a day for men and 425 for women [8].
But does extra choline help an adult brain? In Finland, men who ate the most phosphatidylcholine had a 28% lower risk of dementia over about 22 years, though total choline intake wasn't linked to dementia at all [22].
That's an observational study, though. People who eat more eggs and fish differ in other ways. And the trials in adults are small. One trial of citicoline found some improvement on memory tests in older adults, but it was run by the company that makes it [23].
So for adults, I think choline is probably under-consumed. A supplement to fill that gap is a reasonable option. But I don't condone megadosing. Aim for around the recommended amount, not several times it.
What I'd Tell a Patient Who's Pregnant
Pregnancy is different. Demand goes up, and most of the human trials point towards a benefit. The downside looks small. In the Colorado trials, mothers taking far more than the recommended amount had no more side effects than placebo [1]. And a trial in South Africa gave 2 grams of choline a day, over four times the recommended amount, to pregnant women who were drinking heavily. Their babies did better on memory tests, and the only side effect in the mothers was a small increase in nausea and indigestion [24].
So here's the maths. In pregnancy, you want at least 450 milligrams a day. A typical diet gives about 300, so most women are around 150 short. A large egg has about 150, so that's an extra egg every single day, on top of everything else you eat. In practice most women don't manage that, which is why 90 to 95% fall short.
So for patients who are pregnant or planning to be, I encourage eggs, and using a supplement to make sure you're reaching that 450 milligrams is probably a reasonable option. For some women, that might mean a full 450 milligram supplement. For others, it might be 150. Either way, you'd stay far below the upper limit of 3,500 milligrams a day. And the form that's been studied in pregnancy is phosphatidylcholine, so it's probably best to stick with that.
Why not more, like the new trial? The same researchers found their earlier, lower dose gave almost the same results for social withdrawal, and they now suggest a 900 milligram supplement [1]. I'm being more cautious than they are.
Check your prenatal vitamin first, because many contain little or no choline [8].
Phosphatidylcholine has the most trial data and didn't raise TMAO, though 450 milligrams of choline that way is three or four of those large capsules.
And talk to your midwife or obstetrician before you start.
Is It Too Good to Be True?
So is it too good to be true? I don't think so. I think it's an exciting result that needs repeating. But you don't need to wait for the repeat to make sure you're getting enough choline in pregnancy. Eat eggs, and consider a supplement to make sure you're reaching 450 milligrams a day.
References
1. https://doi.org/10.1016/j.ajog.2026.10.017
2. https://doi.org/10.1159/000343120
3. https://doi.org/10.1016/s0149-7634(03)00069-1
4. https://pubmed.ncbi.nlm.nih.gov/2010061/
5. https://doi.org/10.3945/ajcn.111.022772
6. https://doi.org/10.1096/fj.202101401RR
7. https://doi.org/10.1093/ajcn/nqac147
8. https://ods.od.nih.gov/factsheets/Choline-HealthProfessional/
9. https://doi.org/10.3945/ajcn.112.037184
10. https://doi.org/10.1176/appi.ajp.2012.12070940
11. https://doi.org/10.1056/NEJMoa1109400
12. https://doi.org/10.1038/nature09922
13. https://doi.org/10.1176/appi.ajp.2015.15091188
14. https://doi.org/10.1096/fj.201700692RR
15. https://doi.org/10.1096/fj.202101217R
16. https://doi.org/10.2337/db19-0153
17. https://doi.org/10.1016/j.jnutbio.2021.108617
18. https://doi.org/10.1016/j.amjmed.2021.03.016
19. https://doi.org/10.1007/s00394-021-02637-6
20. https://doi.org/10.3390/nu17050886
21. https://doi.org/10.1056/NEJMoa2206868
22. https://doi.org/10.1093/ajcn/nqz148
