Supplements

Supplements are bought without a prescription or a diagnosis, and usually without a source. This pillar applies the same rule to a bottle of capsules as to a prescription: what did the trials measure, who paid for them, and what did the relevant guideline conclude. Each guide below names the trial or guideline behind every figure, and the vitamin D, B12 and magnesium reference ledgers print the lab ranges with their sources.

Multivitamins: the USPSTF (2022) gives multivitamins for preventing heart disease or cancer an I statement, insufficient evidence, and recommends against beta-carotene and vitamin E with a D grade because the harms outweigh the benefits. COSMOS, the largest recent trial at 21,442 participants, missed its primary cancer endpoint, its cardiovascular endpoint and all-cause mortality, and it enrolled women 65 and older and men 60 and older, so it says little about adults 35 to 50. Do multivitamins do anything? has the trial table.

Vitamin D: the baseline intake from NASEM is 600 IU a day for adults 19 to 70 and 800 IU over 70, from food, fortification or a supplement. The 2024 Endocrine Society guideline supports supplementing beyond that in adults 75 and older, in pregnancy and in high-risk prediabetes, suggests against it in healthy adults under 75 to lower disease risk, and suggests against routine 25(OH)D blood testing in every population it considered. Vitamin D supplements: who actually needs them works through the guideline group by group.

Magnesium: the often-quoted claim that 48 percent of Americans are deficient is an intake shortfall against a requirement estimate, not a measured deficiency rate, and serum magnesium holds under 1 percent of the body's supply. Cochrane rates the evidence against magnesium for older-adult leg cramps high certainty, and the migraine recommendation still printed on federal pages comes from a guideline the American Academy of Neurology retired in 2015. Does magnesium help you sleep? covers the sleep trials specifically.

Collagen and NAD+ precursors: the 2025 meta-analysis of collagen for skin that stratified by funding found no effect in non-industry-funded studies and none in high-quality studies in any category, while the knee-pain effect is real but small, 5.6 to 6.3 mm on a 100 mm scale against a clinically meaningful threshold of 18 mm. NMN and NR reliably raise blood NAD+, and the best 2025 meta-analysis found no effect on muscle mass, grip strength, gait speed or chair-stand time in adults over 60. Does collagen actually work? and NMN and NR report who funded what.

Ask three questions of any bottle. Does a guideline recommend it for someone like you, and with what grade? Did the trial that is quoted measure the outcome you care about, or a blood level that stands in for it? Who paid? Raising a blood marker is the finding that replicates across this whole category, from NAD+ to vitamin D, and it is rarely the finding that matters. The guides on this pillar were written to answer those three questions for multivitamins, vitamin D, magnesium, collagen, and NMN and NR.