Do you need a vitamin or mineral supplement, and how can you tell?

Start with food. In a study that followed 30,899 US adults for about six years, getting enough vitamin A, vitamin K, magnesium, zinc and copper went with lower death rates only when the nutrients came from food, and taking supplements did not go with longer life¹. That study was observational, but it points the same way as the trials below.

What the multivitamin trials found for everyone else. The US Preventive Services Task Force (USPSTF) found too little evidence to say whether a multivitamin prevents heart disease or cancer, and it advises against beta-carotene and vitamin E supplements for that purpose². In the Physicians' Health Study II, 14,641 male doctors took a multivitamin or a placebo for a median of 11.2 years: heart attacks and strokes did not fall³, and total cancer was 8% lower⁴. In COSMOS, 21,442 older adults saw no effect on cancer, heart disease or death⁵, and among 390,124 adults followed for up to 27 years, multivitamin users did not live longer⁶. Our article on multivitamins goes further.

Test before you take. For several nutrients, a blood test can answer the question directly:

What you want to know	The test	How it is read
Am I low in vitamin D?	25-hydroxyvitamin D, written 25(OH)D	50 nmol/L (20 ng/mL) or more is enough for most people; under 30 nmol/L (12 ng/mL) raises the risk of deficiency; over 125 nmol/L (50 ng/mL) is linked to possible harm11
Am I low in B12?	Serum B12, plus methylmalonic acid (MMA) when B12 is borderline	B12 under 200 to 250 pg/mL is usually called low; MMA is the most sensitive marker and confirms a deficiency when B12 is between 150 and 399 pg/mL14
Are my iron stores low?	Ferritin, with a blood count	Ferritin measures stored iron, but inflammation can distort it, so it is read alongside other results20
Do I get enough omega-3?	Omega-3 index: EPA plus DHA in red blood cells	Reflects roughly the last four months of intake18; researchers who proposed it linked 8% or more with the lowest risk of heart death and 4% or less with the highest19

Two limits. Guideline panels do not advise testing everyone for vitamin D: the Endocrine Society suggests against routine 25(OH)D tests in healthy people, including those with obesity or darker skin, and the USPSTF found too little evidence to screen adults without symptoms^12,13. And some tests answer the question poorly: blood magnesium, for instance, says little about how much magnesium the body holds²⁶. Your clinician can order these tests, and several are sold by the direct-to-consumer lab services we review, which we judge on the lab behind them and who reads your results.

Who has a real need. The groups below have evidence behind them, not just marketing:

• Planning a pregnancy, or could become pregnant: folic acid, 400 to 800 micrograms a day, which the USPSTF gives its top grade⁷; iodine, 150 micrograms a day as potassium iodide on American Thyroid Association advice, which not every prenatal contains⁸; and enough choline, which most prenatal supplements barely provide⁹. See eating well during pregnancy.
• Breastfeeding: the same iodine advice applies⁸, and vegan mothers need a reliable B12 source, because their babies can become deficient very early¹⁴.
• Breastfed babies: 400 IU (10 micrograms) of vitamin D a day from soon after birth, until the baby drinks at least 1,000 mL (about a quart) of fortified formula or whole milk a day; breast milk alone does not supply enough^10,11.
• Over 50: many older people absorb the B12 in food poorly, so the National Academies advise getting it mainly from fortified foods or a supplement¹⁵. From 75, the Endocrine Society suggests vitamin D beyond the standard intake¹². See who needs B12.
• Vegans: B12 is not optional¹⁶; iodine can run short without dairy, seafood or eggs⁸; and an algae oil is the plant source of ready-made EPA and DHA, because the body converts little of the omega-3 in foods such as flaxseed^17,18. See vegan and vegetarian diets.
• Heavy periods, low ferritin or frequent blood donation: iron, after a test^20,21. See iron basics.
• Little sun: vitamin D intakes are set on the assumption of minimal sun, and people with little sun exposure or darker skin are more likely to run low¹¹. See vitamin D supplementation.
• Long-term metformin or acid-reducing medicine: B12 can fall, and so can magnesium with proton pump inhibitors^22,24,25. See which medicines deplete nutrients.

Decided you need something? Check how much is too much before you buy, read what to skip, then use the supplement schedule for when to take it.

Multivitamins in well-nourished adults

The Physicians' Health Study II randomized 14,641 male US physicians aged 50 or older (mean age 64.3) to a daily multivitamin or placebo. Over a median of 11.2 years there was no effect on major cardiovascular events (HR 1.01, 95% CI 0.91–1.10) or total mortality (HR 0.94, 95% CI 0.88–1.02)³. Total cancer was modestly lower (17.0 versus 18.3 cases per 1,000 person-years; HR 0.92, 95% CI 0.86–0.998), with no effect on prostate or colorectal cancer and no significant change in cancer deaths (HR 0.88, 95% CI 0.77–1.01)⁴.

COSMOS randomized 21,442 US adults (women aged 65 or older, men aged 60 or older) to a multivitamin-multimineral or placebo. Over a median of 3.6 years, invasive cancer (HR 0.97, 95% CI 0.86–1.09), cardiovascular events (HR 0.98, 95% CI 0.86–1.12) and all-cause mortality (HR 0.93, 95% CI 0.81–1.08) did not differ. Lung cancer was lower (HR 0.62, 95% CI 0.42–0.92), a secondary finding⁵.

Because the trials had limited follow-up, a 2024 cohort study pooled 390,124 adults without chronic disease from three US cohorts, with up to 27 years of follow-up and 164,762 deaths. Daily multivitamin use was not linked to lower mortality (HR 1.04, 95% CI 1.02–1.07 in the first half of follow-up; HR 1.04, 95% CI 0.99–1.08 in the second)⁶. Users were more often women and college educated and less often smokers, a reminder that people who take supplements differ from those who do not.

The USPSTF weighed this evidence in 2022. It issued an I statement (insufficient evidence) for multivitamins and for single or paired nutrients other than beta-carotene and vitamin E, and a D recommendation (against) for those two, for preventing cardiovascular disease or cancer. At the time, 52% of US adults reported using at least one supplement and 31% a multivitamin-mineral².

The food comparison comes from 30,899 adults in the National Health and Nutrition Examination Survey (NHANES), followed for a median of 6.1 years (3,613 deaths). Ever-use of supplements was not associated with mortality. Adequate intakes of vitamin A, vitamin K, magnesium, zinc and copper were associated with lower all-cause or cardiovascular mortality, but only when they came from food¹. Calcium from supplements at 1,000 mg a day or more was associated with more cancer deaths (rate ratio 1.53, 95% CI 1.04–2.25)¹. Residual confounding and recall bias are possible in a study like this. All of these studies enrolled mostly well-nourished US adults; none tested people with a known deficiency.

What guideline panels say about testing

The 2024 Endocrine Society guideline found no trial evidence to support routine 25(OH)D screening in the general population, including people with obesity or a dark complexion, and no clear evidence for an optimal blood level, so it suggests against routine testing¹². The same panel suggests vitamin D beyond the dietary reference intake for ages 1 to 18, adults over 75, pregnancy and high-risk prediabetes, and against it for healthy adults under 75; it calls the best doses unclear¹². The USPSTF noted that no single blood level defines deficiency and gave screening of adults without symptoms an I statement¹³.

The cut-offs in the table come from the Food and Nutrition Board for vitamin D¹¹, and from laboratory practice for B12 and MMA, where an MMA above 0.271 micromol/L suggests deficiency¹⁴. Ferritin reflects iron stores but can be affected by inflammation, and hemoglobin is neither sensitive nor specific, so clinicians often combine measures²⁰. Red-cell fatty acids reflect about 120 days of intake, while plasma levels swing with the last meal¹⁸. The omega-3 index was proposed by its developers as a heart-risk marker: in their analysis an index of 8% or more went with the greatest protection and 4% or less with the least¹⁹.

Pregnancy and infancy

The USPSTF concluded with high certainty that folic acid has a substantial net benefit and recommends 0.4 to 0.8 mg (400 to 800 micrograms) a day for anyone planning or able to become pregnant (grade A); about 3,000 US pregnancies a year are affected by neural tube defects⁷. For iodine, the American Thyroid Association advises 150 micrograms a day as potassium iodide for anyone planning a pregnancy, pregnant or breastfeeding, and the American Academy of Pediatrics advises at least 150 micrograms plus iodized salt; of 59 best-selling prenatal supplements in 2016 to 2017, only 34 contained iodine⁸. About 90% to 95% of pregnant women take in less choline than the adequate intake of 450 mg a day, and prenatal supplements typically contain little or none⁹.

Iron needs rise in pregnancy (the RDA is 27 mg a day)²⁰. Yet the USPSTF found the evidence insufficient to judge either screening or routine iron supplementation in pregnant people without symptoms, although iron deficiency affects about 18% of pregnancies and rises from 6.9% to 28.4% between the first and third trimesters²⁷. That is a question for your prenatal care team rather than a reason to add iron on your own. Preformed vitamin A needs a ceiling. Among 22,748 pregnant women, those taking more than 10,000 IU a day from supplements had 4.8 times the rate of cranial-neural-crest birth defects seen with 5,000 IU or less (95% CI 2.2–10.5), mostly when exposure came before the seventh week; the authors estimated that about 1 in 57 exposed babies had a malformation caused by the supplement²⁸. That was a single observational cohort; its apparent threshold, near 10,000 IU a day of supplemental vitamin A, is the same as the adult upper limit in how much is too much.

For infants, the American Academy of Pediatrics raised its advice to at least 400 IU of vitamin D a day for all infants, including those exclusively breastfed, starting soon after birth¹⁰. Human milk provides only about 25 to 78 IU per liter, and the advice lasts until a baby drinks at least 1,000 mL a day of vitamin D-fortified formula or whole milk¹¹.

B12: age, diet and medicines

The National Academies estimated that 10% to 30% of older people may be unable to absorb the B12 naturally present in food, while absorption of the crystalline B12 used in fortified foods and supplements does not appear to fall with age; hence the advice that people over 50 meet their needs mainly from those sources¹⁵. Atrophic gastritis, which cuts the acid and intrinsic factor needed to absorb food B12, affects 2% of the general population but 8% to 9% of adults aged 65 and older¹⁴.

A review of 18 studies that measured deficiency with MMA or holotranscobalamin found it in 62% of pregnant vegetarian women, 25% to almost 86% of children, 21% to 41% of adolescents and 11% to 90% of older adults, with higher rates in vegans and in people vegetarian since birth¹⁶. Fortified foods and supplements substantially reduce the risk¹⁴.

In a randomized trial in the Netherlands, 390 people with type 2 diabetes on insulin took metformin 850 mg three times a day or placebo for 4.3 years; B12 ended 19% lower than with placebo, and the absolute risk of deficiency was 7.2 percentage points higher, a number needed to harm of 13.8²³. In the Diabetes Prevention Program Outcomes Study, low B12 was more common with metformin at 5 years (4.3% versus 2.3%), low or borderline B12 at 5 and 13 years (19.1% versus 9.5%, and 20.3% versus 15.6%), and each year of use raised the odds (OR 1.13, 95% CI 1.06–1.20)²². Both teams suggested regular B12 checks. In a case-control study in Kaiser Permanente Northern California, 25,956 people with B12 deficiency were compared with 184,199 without: two or more years of proton pump inhibitors (OR 1.65, 95% CI 1.58–1.73) or H2 blockers (OR 1.25, 95% CI 1.17–1.34) were associated with deficiency²⁴. For magnesium, a meta-analysis of nine observational studies (109,798 patients) found a pooled risk ratio of 1.43 (95% CI 1.08–1.88) with proton pump inhibitors²⁵. The FDA advises clinicians to consider checking magnesium before and periodically during long-term use; in a quarter of the cases it reviewed, supplements did not correct the low level and the drug had to be stopped²⁶.

Iron: when it helps and when it harms

A trial in France randomized 198 menstruating women aged 18 to 53 with unexplained fatigue, ferritin under 50 µg/L and normal hemoglobin to 80 mg of elemental iron a day or placebo for 12 weeks. Fatigue scores fell by 47.7% with iron and 28.8% with placebo (difference −18.9%, 95% CI −34.5 to −3.2), with no change in quality of life, depression or anxiety, and ferritin rose 11.4 µg/L more than with placebo²¹. The authors suggest checking blood markers after six weeks²¹. Iron is not harmless when stores are full: about 1 in 10 white people carry the most common hemochromatosis gene variant and about 4.4 per 1,000 have two copies, and liver specialists advise people with hemochromatosis to avoid iron and vitamin C supplements²⁰. People on vegetarian diets need about 1.8 times as much iron, because plant iron is absorbed less well²⁰.

Omega-3 for vegans

Among 165 long-term vegans, the mean omega-3 index was 3.7%, similar to omnivores who get very little EPA and DHA¹⁷. In 46 with an index below 4%, an algae-derived supplement giving 254 mg of EPA plus DHA a day for 4 months raised the index from 3.1% to 4.8%¹⁷. That study measured blood levels, not health outcomes. The body converts less than 15% of plant omega-3 (ALA) into EPA and DHA¹⁸.

Bottom line (confidence 4 of 5)

For well-nourished adults, long randomized trials found little benefit from a daily multivitamin, and a very large cohort points the same way. Real needs in pregnancy, infancy, older age, vegan diets, low iron stores and long-term metformin or acid-reducer use rest on task-force reviews, randomized trials and deficiency surveys. Testing helps when there is a reason to suspect a gap; for vitamin D, guideline panels advise against routine testing of healthy people. None of this replaces advice from your own clinician or pharmacist, especially if you are pregnant, take prescription medicines or have a chronic condition.