What should you eat and avoid during pregnancy?

Nutrients worth getting right

• Folic acid. The neural tube, which becomes the brain and spinal cord, forms very early, often before a pregnancy is confirmed. The US Preventive Services Task Force therefore advises everyone who could become pregnant to take 0.4–0.8 mg daily, with high certainty of substantial benefit¹. Food folate helps too: see folate and methylation.
• Choline. In a tightly controlled feeding trial, babies of women who ate about 930 mg a day in late pregnancy processed information faster than those on about 480 mg⁸. Eggs, meat and fish are the main sources; see choline.
• Iodine. The fetal brain depends on thyroid hormone, which is built from iodine. UK children of mothers with low early-pregnancy iodine were likelier to score in the bottom quarter for verbal IQ⁹, though supplement trials are too small to settle the question¹⁰. See iodine.
• Omega-3 (DHA) and fish. US guidance is 2–3 weekly servings from the lower-mercury list, such as salmon, sardines, shrimp or cod². Across 70 trials, omega-3 supplements reduced preterm birth but slightly increased post-term pregnancy¹¹. See EPA and DHA.
• Iron and vitamin D. The right amount depends on blood tests, skin tone and season, so let your midwife or doctor guide you. Iron plus folic acid is the long-standing standard supplement in many countries⁷.

Things to limit or avoid

• Alcohol. No safe level has been proven. A pooled analysis of light drinking (up to 32 g a week, a little over two US standard drinks) found slightly higher odds of a small baby, and the authors suggested abstaining as a precaution³.
• Listeria. Pregnant women are roughly 18 times likelier to catch this infection, which spreads through ready-to-eat foods, deli meats and soft cheeses. It is often a mild fever for the mother but can be dangerous for the baby⁴.
• High-mercury fish. Avoid shark, swordfish, king mackerel, marlin, orange roughy, bigeye tuna and Gulf of Mexico tilefish².
• Caffeine. Under 200 mg a day does not appear to be a major factor in miscarriage or preterm birth⁶.
• Vitamin A. Keep preformed vitamin A from supplements below about 10,000 IU a day⁵.

Do prenatal vitamins matter? In a Cochrane review of 20 trials (about 142,000 women), multiple-micronutrient pills with iron and folic acid produced fewer low-birthweight babies than iron and folic acid alone, but 19 trials were in low- and middle-income countries⁷. A standard prenatal vitamin is a low-risk safety net; it does not replace food, and more is not better.

Folic acid: the best-supported advice here

In 2023 the US Preventive Services Task Force reaffirmed its top-grade recommendation that everyone planning or able to become pregnant take 0.4–0.8 mg of folic acid daily, citing high certainty of substantial net benefit against neural tube defects, which affect an estimated 3,000 US pregnancies a year¹. Few other items here have comparable certainty.

Choline: promising, but one tiny trial

Caudill and colleagues randomized 26 women entering the third trimester to 480 or 930 mg of choline a day until delivery, in a double-blind feeding study⁸. Of 24 infants tested at 4, 7, 10 and 13 months, those in the 930 mg group had faster average reaction times. Limits: 13 women per group, a lab measure of processing speed rather than school outcomes, and one site. It is a reason to eat eggs and fish, not proof that supplements raise IQ.

Iodine: a cohort signal and thin trial data

Bath and colleagues measured urinary iodine in first-trimester samples from 1,040 women in the ALSPAC cohort, then tested their children at ages 8 and 9⁹. After adjusting for 21 confounders, children of women with an iodine-to-creatinine ratio under 150 µg/g were likelier to score in the lowest quartile for verbal IQ (odds ratio 1.58, 95% CI 1.09–2.30) and reading accuracy (1.69), with scores worsening as iodine fell. A single spot urine sample is a crude marker, and observational data cannot show that supplements fix the gap. The Cochrane review of iodine trials (11 trials, over 2,700 women) found less postpartum hyperthyroidism (RR 0.32, low-quality evidence) but no clear effect on preterm birth or birthweight, and judged the data insufficient to assess routine supplementation¹⁰.

Omega-3 and fish: trials for capsules, cohorts for food

Middleton's Cochrane update pooled 70 randomized trials (19,927 women). Omega-3 supplements lowered preterm birth before 37 weeks (13.4% versus 11.9%; RR 0.89, 95% CI 0.81–0.97) and before 34 weeks (4.6% versus 2.7%; RR 0.58, 95% CI 0.44–0.77), both high-quality evidence. The trade-off was more pregnancies running past 42 weeks (1.6% versus 2.6%), and effects on children's cognition were mostly absent or low-quality¹¹. For whole fish, Hibbeln's ALSPAC analysis of 11,875 women found graded odds of a child scoring in the lowest verbal IQ quartile: 1.48 (95% CI 1.16–1.90) for no seafood and 1.09 (0.92–1.29) for under 340 g a week, versus more than 340 g (12 oz), trend p = 0.004. The authors argued that advice to limit fish could be harmful¹². It is observational, but it fits FDA and EPA advice to eat fish chosen by mercury level².

Alcohol, caffeine and Listeria

Mamluk's meta-analysis of 24 cohort and two quasi-experimental studies compared light drinking (up to 32 g a week) with none. Odds of a small-for-gestational-age baby were slightly higher (OR 1.08, 95% CI 1.02–1.14), while preterm birth (OR 1.10, 95% CI 0.95–1.28) was compatible with no effect. The authors favoured abstinence as a precaution, noting sparse evidence³. ACOG's opinion says under 200 mg of caffeine a day does not appear to be a major contributing factor in miscarriage or preterm birth, the link with growth restriction is undetermined, and no final conclusion is possible about high intake and miscarriage⁶. Madjunkov's review puts listeriosis risk in pregnancy at about 18 times the general population's; fetal listeriosis carries 25–35% mortality depending on gestational age, and neonatal infection about 20% case fatality, although maternal illness is often only a mild fever⁴.

Vitamin A and prenatal vitamins

Rothman's cohort screened 22,748 pregnant women; 339 babies had birth defects, 121 in tissue arising from the cranial neural crest. Those defects were 3.5 times as common (95% CI 1.7–7.3) with more than 15,000 IU a day of preformed vitamin A from food and supplements than with 5,000 IU or less. For supplements alone above 10,000 IU a day the ratio was 4.8 (95% CI 2.2–10.5), with an apparent threshold near 10,000 IU⁵. It was observational, with self-reported diet. Keats's Cochrane review (20 trials, 141,849 women) found multiple-micronutrient pills reduced low birthweight versus iron and folic acid (RR 0.88, 95% CI 0.85–0.91; high-quality evidence) and small-for-gestational-age births (RR 0.92), with a borderline effect on preterm birth (RR 0.95, 95% CI 0.90–1.01) and none on perinatal mortality⁷. The single UK placebo-controlled trial showed no clear change in birth outcomes but less third-trimester anaemia (RR 0.66).

Where the evidence disagrees or is missing

Cohorts link low iodine and low fish intake to poorer child cognition, yet the omega-3 trials found little effect on cognition; fish may matter for other nutrients such as iodine and choline, or confounding may explain part of the cohort signal^9,11,12. Choline rests on one 26-woman trial, iodine trials are underpowered, and alcohol and caffeine thresholds come from observational data where a firm line cannot be drawn.

Bottom line (confidence 4 of 5)

Folic acid has high-certainty support and omega-3 for preterm birth has high-quality trials; the Listeria, mercury and vitamin A cautions guard against rare but serious harms. Specific caffeine, choline and iodine amounts rest on weaker evidence, so follow your midwife's or obstetrician's advice where it differs from this page.