Multivitamins prevent disease

In large randomized trials of older US adults, a daily multivitamin did not lower rates of heart attack, stroke or death, and its effect on cancer was small and uneven between trials. Pooling four trials, the evidence review for the US Preventive Services Task Force found 7 percent lower odds of any cancer (odds ratio 0.93), which in absolute terms ranged from 0.2 to 1.2 fewer people with cancer per 100 across the trials large enough to measure it. The longest trial, in male doctors, found a modest drop in total cancer; the largest and most recent, COSMOS, found none.

The trials enrolled mostly older US adults with no known nutrient deficiency, the two largest tested a single common product, and all measured whether new cancer, heart disease or death occurred. None of these tested a multivitamin as treatment for an existing disease. Two separate trials of a high-dose product in people who had already had a heart attack found no significant reduction in a combined count of deaths, heart attacks, strokes, heart procedures and hospital stays for unstable angina. The Task Force judged the evidence insufficient to weigh the benefits and harms of multivitamins for preventing heart disease or cancer, and recommended against beta carotene and vitamin E supplements for that purpose. In trials of high-dose beta carotene, smokers and people exposed to asbestos developed more lung cancer, and the NIH Office of Dietary Supplements applies that caution to multivitamins providing large amounts of beta carotene or vitamin A; it also notes that vitamin K, found in many multivitamins, decreases the effectiveness of warfarin. Pregnancy sits outside this evidence, and the Task Force separately recommends a daily folic acid supplement for anyone planning or capable of pregnancy.

The Physicians' Health Study II randomly assigned 14,641 male US doctors aged 50 and over to Centrum Silver or a placebo for a median of 11.2 years. Total cancer was 17.0 per 1,000 person-years with the multivitamin and 18.3 with placebo (hazard ratio 0.92, with a confidence interval reaching 0.998). Prostate cancer, colorectal cancer and cancer deaths did not differ significantly (for cancer deaths, hazard ratio 0.88, interval 0.77 to 1.01). In the same trial, major cardiovascular events were unchanged (hazard ratio 1.01), as were heart attack, stroke and cardiovascular death, and death from any cause did not differ significantly (hazard ratio 0.94). Men with a past cancer had fewer new cancers on the multivitamin (hazard ratio 0.73), a difference that was not significantly larger than in men without one.

COSMOS assigned 21,442 US adults, women aged 65 and over and men aged 60 and over, to the same product or a placebo for a median of 3.6 years. Invasive cancer occurred in 518 people on the multivitamin and 535 on placebo (hazard ratio 0.97, not significant). Lung cancer, a secondary outcome, was lower with the multivitamin (hazard ratio 0.62), while breast and colorectal cancer did not differ significantly. Cardiovascular events (hazard ratio 0.98) and deaths (hazard ratio 0.93) did not differ significantly. Across nine trials, the Task Force review found no significant effect on death from any cause (odds ratio 0.94) and no benefit for cardiovascular disease. Both large trials were funded by the National Institutes of Health, with study pills and packaging donated by manufacturers who, according to the papers, had no role in design, analysis or reporting.

In people who already had heart disease, the TACT2 trial assigned 1,000 adults in the US and Canada with diabetes and a past heart attack to six caplets a day of a 28-ingredient high-dose multivitamin and mineral or a placebo and followed them for a median of four years. Death, heart attack, stroke, a procedure to reopen heart arteries or a hospital stay for unstable angina occurred in 175 of 500 people in each group (hazard ratio 0.99, interval 0.80 to 1.22). Heart attack, stroke or cardiovascular death occurred in 105 people on the product and 81 on placebo (hazard ratio 1.30, interval 0.97 to 1.73), a difference that was not statistically significant and that the authors call indeterminate, and deaths from any cause did not differ (hazard ratio 1.02). The paper reports that the same product in an earlier trial of 1,708 people after a heart attack also showed no significant reduction (hazard ratio 0.89, interval 0.75 to 1.07). TACT2 was funded by the National Institutes of Health. Outside trials, a 2024 analysis of three US cohorts followed 390,124 adults free of cancer and chronic disease for up to 27 years. Daily multivitamin users had no lower risk of death; their risk was 4 percent higher in the first half of follow-up (hazard ratio 1.04, interval 1.02 to 1.07) and similar in the second half (1.04, interval 0.99 to 1.08). Daily users were more often women and college educated and less often smokers, and an observational design cannot fully separate the effect of the product from differences between the people who take it.

Reported harms in the multivitamin trials were few. The Physicians' Health Study II recorded slightly more skin rash (29.0 versus 27.3 percent) and nosebleeds (21.6 versus 19.8 percent), which COSMOS did not repeat. The Task Force statement does not apply to children, people who are chronically ill or hospitalized, or people with a known nutritional deficiency. The NIH Office of Dietary Supplements notes that no two trials have used the same combination and amounts of nutrients, so a result for one product does not carry over directly to another.