Supplements

The honest meta-conclusion of supplement research is sobering: after hundreds of thousands of person-years of randomized data, no supplement on the market has been shown to extend life or prevent disease in healthy middle-aged adults at anything like the magnitude of basic lifestyle interventions — Mediterranean-pattern eating, regular exercise, decent sleep, and not smoking dominate every supplement effect by an order of magnitude.

What remains is a more useful question: which supplements have enough human evidence to be worth buying as gap-fillers, and which cause harm at common doses? Five clear that bar for a generally healthy 35–55-year-old who exercises moderately:

1. Vitamin D3 — strong for correcting deficiency; weak for outcomes in the already-replete
2. Omega-3 (EPA and DHA, the two long-chain marine fats) — modest cardiovascular benefit, reliable for triglycerides, and one of the few supplements with a randomized signal on epigenetic-age clocks; matters most if you don't eat fish
3. Magnesium — strong for repletion (widespread inadequacy) and blood pressure; small and low-certainty for sleep
4. Vitamin B12 — strong in specific groups (vegetarians, metformin users, older adults)
5. Creatine monohydrate — strong for muscle, moderate for cognition under metabolic stress (not in well-rested healthy young adults) and for bone alongside resistance training

Everything else is optional, gap-specific, or speculative — including the heavily marketed "longevity molecules" like nicotinamide mononucleotide (NMN), resveratrol, urolithin A, spermidine, and lithium orotate, which rest on animal data, surrogate biomarkers, or industry-funded short-term trials. This page carries the conclusions and the dosing; the evidence behind each one lives in the linked deep dive.

Test before you supplement

For anything fat-soluble, mineral, or chronically dosed, test first — supplementing blind is how people end up with hypercalcemia from megadose vitamin D, copper depletion from chronic zinc, or nerve damage from "energy" formulas loaded with B6.

One caveat on the most popular test: both the US Preventive Services Task Force (USPSTF) and the Endocrine Society advise against routine 25-hydroxyvitamin D screening in healthy adults, because correcting an asymptomatic low reading hasn't changed hard outcomes in trials. Testing earns its keep when you want to dose deliberately rather than guess — not as a population screen.

The general midlife panel lives in Midlife labs. These add to it when you're deciding what to supplement:

• 25-hydroxyvitamin D — the single most informative supplement-related lab, with the caveat above
• Vitamin B12 with methylmalonic acid (MMA) or holotranscobalamin (the fraction of B12 actually available to cells) — serum B12 alone is insensitive
• Ferritin and transferrin saturation — especially premenopausal women (low iron) and men over 50 (occasionally high)
• Red blood cell magnesium — more sensitive than serum, which the body holds constant by draining stores
• Omega-3 index — the percentage of EPA + DHA in red blood cell membranes; target 8–12%
• Homocysteine — a downstream readout of B12, folate, and B6 status
• Thyroid-stimulating hormone (TSH) in context — 4–6 mIU/L in an older adult is most likely physiological. See Thyroid management.

1. Vitamin D3

Strong for correcting deficiency and for bone health in deficient adults. Weak for outcomes in the already-replete.

The largest trial, VITAL, gave 25,871 US adults 2000 IU/day for an average of 5.3 years and found no effect on cancer, cardiovascular events, or all-cause mortality — a 1% difference well inside the range expected from chance.^[1] The extra-skeletal signals are softer than the headlines: cancer mortality about 12% lower with daily but not intermittent dosing,^[2] new autoimmune disease 22% lower,^[3] and a post-hoc slowing of methylation ageing clocks in DO-HEALTH — a trial that missed all six of its pre-specified primary endpoints.^[4]

• Target 75–125 nmol/L (30–50 ng/mL). Below that, 1000–2000 IU/day usually suffices; up to 4000 IU/day for severe deficiency.
• Take with a fat-containing meal. Daily beats weekly or monthly megadosing, which is at best useless and at worst harmful.
• Above 4000 IU/day without monitoring, hypercalcemia is rare but real.
• Above about 35°N, winter ultraviolet-B is too weak to make any vitamin D from roughly October to March, and a good summer cannot cover it — 25-hydroxyvitamin D has a half-life of weeks, not months.

→ Vitamin D — dose-response, who benefits, the telomere and epigenetic-clock signals, and the guideline split over how high to aim.

2. EPA/DHA Omega-3

Weak-to-moderate for cardiovascular outcomes, best in low-intake populations. Moderate and reliable for triglycerides. Weak but real for slowed biological aging. Caution: a dose-dependent atrial-fibrillation excess.

The headline trial, REDUCE-IT, gave 8,179 statin-treated adults with high triglycerides purified EPA (icosapent ethyl) at 4 g/day and cut major cardiovascular events by 25% — partly confounded by a mineral-oil placebo that raised low-density lipoprotein (LDL) cholesterol and inflammatory markers in the control arm.^[5] STRENGTH, a matched trial at the same 4 g/day dose using mixed EPA+DHA, found no benefit at all,^[6] and neither did the much lower-dose omega-3 arm of VITAL, whose benefit in low-fish-intake participants was a secondary finding rather than a headline.^[7]

Beyond events, DO-HEALTH produced the first randomized signal that 1 g/day slows several DNA-methylation ageing clocks. Read it with its caveats: a post-hoc analysis of 777 of the trial's 2,157 participants, all over 70, in a trial that missed all six pre-specified primaries; the clocks are surrogates, not hard outcomes; and vitamin D and exercise moved them too, with the three clearly stacking on only one clock.^[8] It remains the most interesting readout in the category — and it came at an ordinary dose.

• Two or more servings of fatty fish a week makes supplementation unnecessary.
• If you don't eat fish, 1 g/day of combined EPA+DHA is the sensible default — it lifts the omega-3 index into target range, matches the DO-HEALTH dose, and sits at the bottom of the arrhythmia signal below. Going to 2 g/day is a deliberate trade, not a free upgrade. Algae oil is non-inferior for vegans; re-esterified triglyceride absorbs best, ethyl ester is fine with food.
• Treat 4 g/day as prescription territory. The trials show a dose-dependent atrial-fibrillation excess: about 25% higher overall, roughly 50% above 1 g/day, and a smaller but real 12% even at 1 g/day or less.^[9] Higher dietary omega-3 tracks the opposite way, but that doesn't license high supplemental doses.
• Refrigerate after opening; look for an International Fish Oil Standards (IFOS) or Labdoor mark. Of 171 North American products surveyed, roughly half exceeded standard oxidation thresholds, and rancid oil is pro-inflammatory.^[10]

→ Omega-3 — the REDUCE-IT versus STRENGTH puzzle, the omega-3 index, the atrial-fibrillation paradox, and the EPA:DHA ratio that matters for mood.

3. Magnesium

Strong for repletion and for modest blood-pressure reduction, concentrated in people with something to correct. Weak and low-certainty for sleep. Preliminary for cognition.

About half of adults don't get enough from food.^[11] Pooling 34 trials at a median ~368 mg/day, supplemental magnesium lowered blood pressure about 2/2 mmHg overall.^[12] A 2025 update across 38 trials sharpened it: the effect concentrates in people already treated for hypertension — about 7.7 mmHg systolic — and in the deficient, with no significant effect in those already normotensive.^[13] It behaves as a repletion agent, not a drug.

Worth knowing: drinking water counts. Over 16 years in 26,733 postmenopausal Swedish women, those whose tap water carried about 10 mg/L of magnesium had 31% lower risk of ischemic stroke than those at about 5 mg/L — and only the magnesium, not the calcium, carried it.^[14] Unremineralised reverse-osmosis filtration works against this. See Water.

• 200–400 mg/day elemental — though the top of that range sits above both regulatory ceilings: the EU limit for supplemental magnesium is 250 mg/day and the US limit 350 mg/day, so the upper end is not "more is better" territory. Food-based magnesium has no limit.
• Glycinate or threonate in the evening for sleep; citrate any time. Split doses above 200 mg.
• Be cautious in chronic kidney disease; separate by 2 hours from quinolones, tetracyclines, and bisphosphonates.

→ Magnesium — the full form table, the blood-pressure and diabetes evidence, testing, and drug interactions.

4. Vitamin B12

Strong for repletion in at-risk groups; weak as a routine longevity supplement in healthy non-vegetarians. Caution on B6: chronic intake above the European upper limit of 12 mg/day can cause a reversible peripheral neuropathy (tingling and numbness in the hands and feet) — a reason to avoid loaded "energy" B-complexes.

Three groups should supplement: vegetarians and vegans (plant foods contain essentially none); people on metformin or chronic acid-suppressing drugs, where absorption is impaired with documented depletion over years;^[15] and adults past about 60, where age-related atrophic gastritis impairs the stomach acid and intrinsic factor B12 needs — roughly 6% of adults under 60 are deficient or marginally depleted, against about 20% over 60.

• Cyanocobalamin and methylcobalamin work equally well for most people.
• 25–250 µg/day for prevention; 1000 µg/day for documented deficiency.
• Don't trust serum B12 alone. Confirm with MMA above ~270 nmol/L or holotranscobalamin below 35 pmol/L.

The cognitive case is narrower than it sounds: VITACOG, the trial that anchors it, enrolled adults with mild cognitive impairment and high homocysteine, and slowed brain atrophy only in those who already had adequate omega-3. The cardiovascular case is closed the other way — the large trials that successfully lowered homocysteine did not reduce heart attacks, strokes, or cardiovascular death.

→ B vitamins — who needs them, the form debates, the B6 ceiling, and why NAD⁺ boosters don't belong here.

5. Creatine Monohydrate

Strong for muscle and strength; moderate for cognition under metabolic stress and in older adults, and for slowed bone loss alongside resistance training. Caution: not disease-modifying for motor neurodegeneration — large trials in Parkinson's, Huntington's, and amyotrophic lateral sclerosis were null or stopped for futility.

Its safety record is unusually deep: a 2025 review across 685 trials and 12,839 participants, at doses up to 30 g/day for up to 14 years, found no clinically significant excess of serious adverse events versus placebo.^[16] The case that it is closer to a conditionally essential nutrient than a sports supplement rests on intake: about 70% of adults over 65 get under 1 g/day from food, and the body makes only another 1–2 g/day.^[17]

• 3–5 g/day of monohydrate, whenever convenient. No loading needed. Skip the HCl, nitrate, ethyl ester and "buffered" forms — 2–5× the price, no advantage.
• Look for Creapure branding or USP / NSF / Informed Sport verification.
• Mention it before routine labs: serum creatinine runs slightly high (a benign byproduct, not kidney damage — the kidney's actual filtration rate is unaffected).
• For brain effects, roughly 10 g/day for weeks — the blood-brain barrier needs more than the muscle dose.
• For bone it only works paired with training: in postmenopausal women on creatine plus supervised resistance training, femoral-neck bone density fell 1.2% over a year versus 3.9% on placebo — slowed loss, not a gain.^[18]

→ Creatine — the muscle meta-analyses, the brain and sleep-deprivation work, the honest limits of the cognition literature, and the female-specific picture.

A practical timing template

Timing matters less than consistency — all five work by saturating body stores rather than producing an acute spike.

Sleep and anxiety supplements

Weak-to-moderate for the best two, weak or preliminary for the rest, caution on indefinite nightly melatonin.

Most of this category delivers modest help with few side effects, only a couple have real evidence, and none replace cognitive behavioural therapy for insomnia (CBT-I) for chronic insomnia. The best risk-benefit belongs to magnesium glycinate (200–400 mg) and L-theanine (200 mg, weak-to-moderate evidence, excellent safety). Magnesium's sleep effect is genuine but small, and worth stating in absolute terms: in the best recent trial, 155 adults with poor sleep took 250 mg/day of elemental magnesium as bisglycinate, and about one extra person in seven reached a clinically meaningful improvement who would not have on placebo.^[19]

Low-dose melatonin (0.3–1 mg) earns a narrow place — jet lag, delayed sleep-phase syndrome, insomnia in older adults whose own melatonin has fallen — but the 2017 American Academy of Sleep Medicine guideline gave a conditional recommendation against it for general insomnia,^[20] and a 2025 observational cohort linked a year or more of use to higher heart-failure and mortality rates — confounded, not proof of cause, but enough to favour short-term strategic use over an indefinite nightly habit.^[21]

Ashwagandha, glycine, valerian, and cannabidiol (CBD) sit at the weak-to-preliminary end, with ashwagandha carrying rare liver-injury reports and CBD interfering with the liver enzymes that clear most prescription drugs. None of these produces clinically significant addiction, tolerance, or withdrawal at studied doses — the key advantage over prescription hypnotics.

→ Sleep and anxiety supplements — the tier ranking, the addiction question for each, and the case against chronic over-the-counter antihistamines above age 50.

Geroprotectors: rich mechanism, thin human translation

Weak / preliminary across the category. The gap between mechanistic promise and clinical translation has widened over the past five years.

The closest thing to a true gerotherapeutic is rapamycin, which extends mouse lifespan 10–25% across replicated trials. The human data are thinner than the enthusiasm: trials of drugs that damp the same nutrient-sensing pathway rapamycin acts on improved immune function and reduced respiratory infections in older adults,^[22] but the first randomized trial of intermittent low-dose rapamycin in healthy adults — PEARL, 48 weeks — missed its primary endpoint of visceral fat, and its positives were secondary and subgroup-confined.^[23] There is still no human trial with hard longevity endpoints.

Among nutraceuticals, urolithin A has the best industry-funded data — its ATLAS trial found about 12% leg-strength gain and 10% higher peak oxygen uptake in 88 overweight middle-aged adults, but missed its primary endpoint of peak power.^[24] Spermidine has a striking observational signal — 39% lower all-cause mortality in the top tertile of dietary intake in the Bruneck cohort^[25] — but its one randomized supplement trial, SmartAge, in 100 older adults over 12 months, was negative on its primary cognitive endpoint.^[26]

Below that it gets worse. NMN and nicotinamide riboside raise blood levels of nicotinamide adenine dinucleotide (NAD⁺), the cofactor every cell uses to run its energy metabolism, but produce small and inconsistent functional signals — NMN improved muscle insulin sensitivity in postmenopausal prediabetic women,^[27] nicotinamide riboside at 1000 mg gave a small systolic blood-pressure drop in stage-1 hypertensives,^[28] and at 2000 mg was completely negative in obese men.^[29] A 2024 paper raised the further concern that they may raise end-products of that pathway which the body cannot recycle, and which track with cardiovascular events in cohort data.^[30] Resveratrol at supplement doses has effectively no human evidence, and lithium orotate has no randomized trial in healthy adults at all.

→ Geroprotectors — each agent in detail, and why industry funding distorts this literature.

The rest of the shelf

Curcumin — Moderate for joints, Caution on the liver. The defensible uses are mild-to-moderate osteoarthritis pain, where a well-formulated extract performs about as well as a non-steroidal anti-inflammatory drug with fewer gut side effects,^[31] and recovery from heavy exercise, where it blunts soreness without improving performance.^[32] The "reverses your epigenetic age" claim comes from a trial of a whole diet, not curcumin. The catch: the high-absorption capsules engineered for systemic effect are a documented cause of unpredictable, dose-independent liver injury, with onset at 1–4 months, with a strong genetic risk marker (HLA-B*35:01).^[33] Keep the scale in view — serious but rare, roughly 1 in 10,000 to 1 in 100,000 exposed.^[34] Take it in time-limited courses, not as a daily anti-aging molecule. → Curcumin

Glucosamine — Moderate as association, Weak as cause. In UK Biobank, habitual users had about 15% lower all-cause and roughly 18% lower cardiovascular mortality,^[35] with a companion analysis putting cardiovascular death about 22% lower.^[36] But it is observational and almost exactly the size healthy-user bias can manufacture alone; no trial has tested it. Its original use is contested and form-dependent — the publicly funded GAIT trial and an independent network meta-analysis were both negative, while the positive trials nearly all used one prescription formulation.^[37][38] Cheap and safe, but it interacts with warfarin. A defensible reason to take it is knee osteoarthritis — not life extension. → Glucosamine

Calcium — food-first, and never as a bolus. Dietary calcium is cardioprotective: over 10 years, the MESA cohort found the highest dietary intake had the lowest risk of new arterial calcification.^[39] Large isolated supplement doses do the opposite — a single large pill of 1,000 mg or so spikes serum calcium for hours and over years tracks with a probable rise in heart attacks. The evidence is contested (a 2020 umbrella review was null), but the asymmetry is clear: supplements offer no cardiovascular upside and a plausible downside food doesn't carry. The one hard-outcome benefit, fewer fractures, is confined to the deficient and institutionalised elderly; across more than 51,000 community-dwelling adults over 50, calcium, vitamin D, or both showed no effect on fractures in any subgroup.^[40] Target 1,000–1,200 mg/day from food (don't count spinach — oxalate binds nearly all of it), supplement only the gap, never more than ~500 mg elemental at once, always with a meal, and prefer calcium citrate or microcrystalline hydroxyapatite. Pair it with vitamin D3, adequate magnesium, and — on mechanism rather than hard outcomes, see below — vitamin K2. → Calcium

Vitamin K2 — Moderate on surrogates, unproven on outcomes. Vitamin K2, in its menaquinone-7 (MK-7) form, decides where absorbed calcium goes, activating the proteins that pull it into bone and keep it out of artery walls. Two trials are positive, both on surrogate endpoints: 180 µg/day over three years improved arterial stiffness in postmenopausal women,^[41] and a 2026 trial slowed coronary calcium progression on imaging.^[42] An equal weight is null — a larger, higher-dose Danish trial found no effect on calcification despite moving the biomarker,^[43] and in kidney disease the biomarker moves while outcomes don't. The positive trials all used product from the dominant commercial manufacturer. A cheap, safe pairing with D3 — not the longevity molecule the aisle implies. It directly opposes warfarin, though not the newer direct oral anticoagulants. → Vitamin K2

CoQ10 — narrow and real, oversold everywhere else. The one unambiguous indication is primary CoQ10 deficiency, a rare genetic disorder where it is genuinely disease-modifying. Beyond that the heart-failure signal is real but narrow — a 2024 meta-analysis of 33 trials put all-cause mortality about a third lower (relative risk 0.64), resting heavily on one pivotal trial, at low-to-moderate certainty, and it hasn't changed guidelines.^[44] The other mortality signal is KiSel-10, in community-dwelling Swedes aged 70–88 given CoQ10 with selenium — a cohort that turned out to have low average selenium, though selenium status was not an entry criterion. Both hard-endpoint trials carry industry sponsorship. In the general population there is no mortality benefit at all, and in obese adults roughly 45% higher all-cause mortality among users.^[45] The reconciling principle: CoQ10 only helps if the body can keep it in its active reduced form, which fails under the oxidative load of metabolic dysfunction. Migraine prophylaxis has a modest positive signal. Take it with fat; every hard-endpoint trial used plain ubiquinone, so don't pay for branded delivery systems. → CoQ10

Supplements to avoid (or use only with documented deficiency)

Caution — documented harm at doses people commonly take. The pattern is consistent: each was once enthusiastically recommended on observational data or mechanism, then large randomised trials revealed harm.

• Vitamin E ≥400 IU/day. A 2005 meta-analysis found about 39 extra deaths per 10,000 people at high doses;^[46] HOPE-TOO showed more heart-failure hospitalisations;^[47] SELECT reported 17% more prostate cancer in men.^[48] Stay under about 30 IU supplemental; nuts, seeds and olive oil are sufficient.
• Beta-carotene in current or former smokers. ATBC showed 18% more lung cancer in male smokers,^[49] and CARET was stopped early for a 28% increase.^[50] The US Preventive Services Task Force issued a Grade D recommendation against routine vitamin E or beta-carotene for cardiovascular or cancer prevention.^[51]
• Preformed vitamin A above 3000 µg/day (measured as retinol activity equivalents, RAE). The Nurses' Health Study found 48% more hip fractures.^[52] Teratogenic in pregnancy at high doses.
• High-dose calcium supplements above about 1000 mg/day in non-deficient adults — a 2010 meta-analysis reported 31% more heart attacks,^[53] and the Women's Health Initiative calcium-plus-vitamin-D arm showed 17% more kidney stones.^[54]
• High-dose antioxidants around training. The burst of reactive oxygen species a hard session produces is the signal that drives adaptation. In a controlled trial, 1,000 mg vitamin C plus 400 IU vitamin E daily abolished the exercise-induced gain in insulin sensitivity.^[55] Also avoid vitamin C above 2,000 mg/day and selenium above 200 µg/day in the already-replete.
• Iron without documented deficiency. Overload accelerates oxidative damage and is implicated in cardiovascular and neurodegenerative disease. Adult men and post-menopausal women shouldn't supplement unless ferritin is documented low.

The class-level result is the honest summary: pooling 78 randomized trials in almost 300,000 people, antioxidant supplements did not prevent death, and in the low-bias trials beta-carotene, vitamin E, and higher-dose vitamin A were associated with slightly higher mortality.^[56]

→ Supplements to avoid — each agent with full citations, threshold doses, and the case for testing first.

Quality assurance

The industry is poorly regulated; independent testing routinely finds 10–30% of products underdosed, contaminated with heavy metals, or mislabelled.^[57] Look for a USP Verified, NSF Certified for Sport, or Informed Sport mark.

• Fish oil — third-party tested for oxidation (the total oxidation, or TOTOX, value) and refrigerated after opening.
• Melatonin is the worst offender on the shelf: across 31 commercial US products, actual content ranged from 83% below to 478% above label, with lot-to-lot variation within a single product reaching 465%.^[58] Prolonged-release prescription melatonin (Circadin 2 mg in the EU) is far more reliable.
• NMN products from major online retailers have repeatedly failed label-claim assays.
• Cheap multivitamins often use the least bioavailable forms at doses that aren't ideal.

Optional add-ons

Situational rather than universal — the case for each is narrower than the core five.

What's overrated

The whole emerging-longevity category, covered above: mechanism without translation, and the gap has widened rather than closed.

Meanwhile the foundational story — vitamin D in deficient populations, omega-3 for cardiovascular nuance and now epigenetic ageing, magnesium for blood pressure, creatine for muscle and for cognition in the groups where it actually shows an effect — has become steadily more practical to act on.

There is no case for taking everything at once. Use supplements as targeted gap-fillers where biomarkers, diet, age, or sex create a specific rationale. Prescription pharmacology — including glucagon-like peptide-1 (GLP-1) receptor agonists and testosterone therapy — lives under Clinical Care, not here.

For the full ordered action list drawn from every pillar, see the healthspan long list.