MSG Doesn't Cause Headaches. In 130 People Convinced It Did, the Reactions Didn't Survive a Retest

The Letter That Launched a Panic

In the spring of 1968, a Maryland physician named Robert Ho Man Kwok wrote a short letter to the New England Journal of Medicine. After meals at Chinese restaurants, he felt numbness at the back of his neck spreading to his arms and back, with general weakness and a racing heart. The symptoms began 15 to 20 minutes after the first dish and faded after about two hours, and Kwok offered several possible culprits: the soy sauce, the cooking wine, the salt, and monosodium glutamate, the flavor enhancer known as MSG. He closed by asking colleagues to look into it.

What followed was not a careful investigation: newspapers picked up the letter, gave the symptoms a name, Chinese Restaurant Syndrome, and spent the next fifty years warning readers about a seasoning, while restaurants began advertising "No MSG" the way they would advertise a peanut-free kitchen. The name itself did quiet damage: a panel convened by the Federation of American Societies for Experimental Biology later called the term pejorative, since the symptoms had nothing specifically to do with Chinese food.

MSG is the sodium salt of glutamate, an amino acid your body makes and that shows up naturally in tomatoes, Parmesan cheese, mushrooms, and soy sauce, so the question was never whether glutamate is foreign to the human diet. It was whether the small amount sprinkled on a meal could make people sick, and for decades the answer came from small, flawed challenges rather than a definitive trial.

130 People, Four Rounds, No Peeking

In 2000, a team led by Raif Geha at Harvard Medical School ran the trial the question deserved. They recruited 130 volunteers who said MSG made them ill, across several medical centers. The design was a multicenter, double-blind, placebo-controlled crossover study with four protocols, the largest of its kind at the time, according to the authors, and in the first protocol each volunteer came in on two separate days and drank either 5 grams of MSG or a placebo, with no food, and without knowing which was which, with a reaction counted as two or more of ten listed symptoms within two hours.

Five grams is an enormous dose by kitchen standards: a typical serving of food with added MSG contains less than half a gram, and the average adult gets about 0.55 grams of added MSG in an entire day, so the volunteers swallowed roughly nine days' worth at once, on an empty stomach. Responders went on to further rounds: challenged again without food in two more protocols, then given MSG with food in a fourth.

What Happened on the First Round, and What Happened Next

The first round gave MSG's critics exactly what they wanted: fifty of the 130 volunteers, 38.5 percent, reacted to MSG only, seventeen, 13.1 percent, reacted to the placebo only, and nineteen more reacted to both, a gap that was statistically significant, so if the study had ended there, the headline would have written itself.

It didn't end there. That is the point of a multiple-challenge design: one reaction proves little. When the responders were challenged again, the pattern fell apart. The reactions did not reproduce. Only two of the 130 reacted to MSG in every round of the no-food retesting, and even those two failed to react consistently when the MSG arrived with food, while in the final protocol MSG served with food produced no more reactions than placebo and nobody suffered a persistent or serious effect.

Look at the placebo numbers again, because they carry the finding. Thirteen percent of people who were certain MSG made them sick reacted to a drink that contained no MSG at all. Nearly one in seven. The trial measured something real. It was not a reaction to a seasoning. It was expectation, plus the ordinary background rate of headaches and flushing in people watching themselves closely for two hours. Remember the dose: 5 grams, nine times a day's worth of added MSG, swallowed at once on an empty stomach, and still nothing reproducible happened.

The Strongest Counterargument

The strongest case against this result starts with the first round. The difference between MSG and placebo in protocol A was statistically significant, 38.5 percent against 13.1 percent, and it is possible to read the retests the other way. A variable trigger produces variable results, and retesting a shrinking group of responders loses statistical power with each round. On that reading, the study did not prove there is no effect; it proved that the effect, if it exists, is inconsistent. Inconsistency is not the same as absence.

FASEB's expert panel reached a similar middle position in 1995. After reviewing the evidence for the FDA, the panel concluded that a subgroup of otherwise healthy people may develop a short-lived symptom complex, headache, flushing, tingling, palpitations, within an hour of taking 3 grams or more of MSG without food. Geha's own paper agreed that large doses given without food may elicit more symptoms than a placebo in people who believe they react. The panel also flagged a second group, people with severe, poorly controlled asthma, who might experience bronchospasm, a question that has never been settled and that this trial did not test.

A third objection concerns the challenge itself. Five grams of MSG in a drink on an empty stomach is not a meal, and an exaggerated dose can miss a real effect that depends on how food is actually eaten. None of this rescues the headline version of the scare, but it does narrow the claim: the trial is strong evidence that MSG in food, at ordinary amounts, does not reliably cause symptoms, even in people who are convinced it does. It is weaker evidence about large doses on an empty stomach, and it says nothing about severe asthmatics.

What We Didn’t Prove

This was one trial, published in 2000, and no independent team has repeated its exact four-protocol design. The volunteers were self-selected believers, not a random sample of the public, so the study speaks to people who think they react, not to everyone, and it tested a single large dose in a drink, not the small amounts in a lifetime of meals. It tested neither children nor people with severe asthma.

What You Can Do

Stop paying a premium for "No MSG" labels. The same glutamate is already in your tomatoes, Parmesan, and mushrooms, and your body makes it too. If restaurant meals give you headaches, look at the likelier suspects first: a salty meal, alcohol, dehydration, and a short night of sleep. If you want a real test, ask a friend to cook two identical meals, one with a pinch of MSG and one without, and see whether you can tell them apart without knowing which is which. If you have severe asthma and suspect a food trigger, raise it with your doctor: the asthma question is the one piece of this story that science has not closed.

Sources

1. Geha, R. S., Beiser, A., Ren, C., Patterson, R., Greenberger, P. A., Grammer, L. C., Ditto, A. M., Harris, K. E., Shaughnessy, M. A., Yarnold, P. R., Corren, J., & Saxon, A. (2000). Multicenter, double-blind, placebo-controlled, multiple-challenge evaluation of reported reactions to monosodium glutamate. Journal of Allergy and Clinical Immunology, 106(5), 973-980. doi:10.1067/mai.2000.110794
2. Geha, R. S., et al. (2000). Review of alleged reaction to monosodium glutamate and outcome of a multicenter double-blind placebo-controlled study. Journal of Nutrition, 130(4S Suppl), 1058S-1062S. doi:10.1093/jn/130.4.1058S
3. Kwok, R. H. M. (1968). Chinese-restaurant syndrome. New England Journal of Medicine, 278(14), 796. doi:10.1056/NEJM196804042781419
4. Raiten, D. J., Talbot, J. M., & Fisher, K. D. (1995). Executive summary from the report: Analysis of adverse reactions to monosodium glutamate (MSG). Journal of Nutrition, 125(11), 2891S-2906S. doi:10.1093/jn/125.11.2891S
5. Tarasoff, L., & Kelly, M. F. (1993). Monosodium L-glutamate: A double-blind study and review. Food and Chemical Toxicology, 31(12), 1019-1035. doi:10.1016/0278-6915(93)90012-N
6. U.S. Food and Drug Administration. (2012). Questions and Answers on Monosodium Glutamate (MSG). fda.gov
7. Yang, W. H., Drouin, M. A., Herbert, M., Mao, Y., & Karsh, J. (1997). The monosodium glutamate symptom complex: Assessment in a double-blind, placebo-controlled, randomized study. Journal of Allergy and Clinical Immunology, 99(6), 757-762. doi:10.1016/S0091-6749(97)80008-5