To review the benefits and harms of vitamin and mineral supplementation in healthy adults to prevent cardiovascular disease (CVD) and cancer.
MEDLINE, PubMed (publisher-supplied records only), the Cochrane Central Register of Controlled Trials (CENTRAL), the Cochrane Database of Systematic Reviews, and Embase, between January 2013 and February 1, 2022. Additionally, we evaluated all studies included in the prior USPSTF review for inclusion in the current review.
We reviewed 17,459 unique citations and 379 full-text articles against a priori inclusion criteria. We included English-language randomized clinical trials (RCTs) of vitamin or mineral use among persons without CVD or cancer reporting all-cause mortality, CVD, cancer, or adverse outcomes as well as observational cohort studies examining serious harms of supplement use. Critical appraisal was completed independently by two investigators. Data were extracted from studies by one reviewer and checked by a second.
We conducted quantitative pooling when at least three studies of the same supplement reported the same outcome. Because most outcomes occurred in less than 10 percent of the study sample, we typically used methods appropriate for rare events, including a fixed effects Mantel-Haenszel model or a random effects restricted maximum likelihood model using Peto odds ratio. Stratified or subgroup analyses and meta-regression were used to explore effect modification for trials of vitamin D, which had the largest body of evidence.
A total of 84 studies (n=739,803) were included. Pooled effects indicated that multivitamin use may be associated with a reduced risk of incidence of any cancer (OR, 0.93 [95 % CI, 0.87 to 0.99]; 4 RCTs; n=48,859; I=0%; range in absolute risk difference among adequately powered trials, −0.2% to −1.2%) and lung cancer (OR, 0.75 [95% CI, 0.58 to 0.95]; 2 RCTs, n=36,052; I=30%; absolute risk difference, 0.2%). However, the evidence for multivitamins had important limitations, including only three adequately powered trials, one with a median of only 3.6 years of multivitamin use and another that was limited to antioxidants. Pooling of studies of vitamin D with or without calcium showed no association with all-cause mortality (OR, 0.96 [95% CI, 0.91 to 1.02]; 27 RCTs [n=117,082]; I=0%), CVD (e.g., composite CVD events: OR, 1.00 [0.95 to 1.05]; 7 RCTs [n=74,925]; I=0%), or cancer outcomes (e.g. any cancer incidence: OR, 0.98 [0.92 to 1.03]; 19 RCTs [n=86,899]; I=0%). Beta-carotene, with or without vitamin A, was associated with an increased risk of cardiovascular mortality (OR 1.10 [95% CI, 1.02 to 1.19]; 5 RCTs [n=94,506]; I=0%; range in absolute risk difference −0.8% to 0.8%), and lung cancer (OR 1.20 [95% CI, 1.01 to 1.42]; 4 RCTs [n=94,830]; I=38.8%; range in absolute risk difference −0.1% to 0.6%). In addition, we found less robust evidence that folic acid was associated with an increased risk of cancer incidence. We found clear evidence that vitamin E (with or without vitamin C or selenium) offers no benefit for all-cause mortality (OR, 1.02 [95% CI, 0.97 to 1.07]; 9 RCTs [n=107,772]; I=0%), CVD events (OR, 0.96 [95% CI, 0.90 to 1.04]; 4 RCTs [n=62,136]; I=0%), and cancer (OR, 1.02 [95% CI, 0.98 to 1.08]; 5 RCTs [n=76,777]; I=0%), and more equivocal evidence that multivitamins (antioxidant-focused or broad spectrum), vitamin A (without beta-carotene), vitamin C, calcium (without vitamin D), and selenium also had no impact on all-cause mortality, CVD, and cancer. There was also weak evidence that supplements increased the risk of some other serious harms, such as hip fracture (vitamin A), hemorrhagic stroke (vitamin E), and kidney stones (vitamin C, calcium). Several supplements were associated with an increased risk of some minor and reversible adverse outcomes, such as skin yellowing (beta-carotene) and gastrointestinal symptoms (calcium).
Some studies lacked full outcome ascertainment or had insufficient followup or power for the main review outcomes; varied background interventions (primarily due to factorial designs) may cloud supplement effects; people of color were minimally represented. Most supplements had too few studies to explore effect modification.
Vitamin and mineral supplementation provides little to no benefit in preventing cancer, CVD, and death, with the exception of a possible small benefit for cancer incidence with multivitamin use, where lung cancer showed the largest benefit. Beta-carotene is associated with increased risk of lung cancer and other harmful outcomes in persons at high risk of lung cancer. Data were absent or insufficient to draw conclusions for any of the B vitamins, iron, zinc, or magnesium.
