If you are searching for the best supplements Peter Attia recommends, you have probably listened to enough of The Drive and read enough of Outlive to know the headline stack (EPA, magnesium, vitamin D, creatine, methylated B vitamins, the occasional sulforaphane and curcumin), and you want to know which of these have hard human-trial signal versus which are habitual.
Quick Answer: which of Peter Attia's supplements have the strongest evidence?
The 2 to 3 we would actually start with:
- EPA-dominant omega-3 fish oil, roughly 2 g of EPA+DHA per day with EPA at least 60% of the total. This is the most replicated cardiovascular-outcome signal in Attia's stack.
- Magnesium, 300 to 400 mg of elemental magnesium per day in a chelated form (glycinate, malate, or citrate). Background-population evidence for cardiometabolic risk reduction is solid and most adults under-consume it.
- Creatine monohydrate, 3 to 5 g per day. Originally a strength-training supplement, but the ISSN 2022 position stand documents real signal for muscle preservation and emerging cognitive endpoints in adults over 50.
Who should not start with these: anyone on warfarin or other anticoagulants needs to clear high-dose fish oil with the prescribing clinician. Anyone with stage 3 or worse chronic kidney disease should not start magnesium or creatine without nephrologist input. And nobody should swap a statin or PCSK9 inhibitor for any of the "ApoB-lowering" supplements listed lower in this article.
What to do FIRST: get the actual labs Attia builds the stack around. ApoB, Lp(a), fasting insulin, HbA1c, a 25-hydroxyvitamin D, an RBC magnesium if you can get one, and ferritin. The stack only makes sense against the biomarkers it is built to move. Without the labs you are guessing.
What "the Attia stack" actually is, in one paragraph
Peter Attia is a Stanford-trained surgeon turned longevity-focused physician whose practice and writing center on what he calls the "Four Horsemen" of late-life disease: atherosclerotic cardiovascular disease, type 2 diabetes and metabolic dysfunction, cancer, and neurodegeneration. His supplement framing follows from that mortality model. The stack he repeatedly discloses on The Drive and in Outlive is built around: lowering ApoB-containing lipoprotein burden (the cardiology lane he is most aggressive in), correcting common deficiencies that hurt cardiometabolic and bone health (vitamin D, magnesium, methylated B vitamins), supporting muscle and brain energetics with age (creatine, EPA/DHA), and a few adjuncts with thinner data that he treats experimentally rather than confidently (sulforaphane, curcumin, occasional ashwagandha and berberine). The conventional first line for his core target, elevated ApoB and atherosclerotic risk, is the 2018 AHA/ACC cholesterol guideline statin and PCSK9 inhibitor protocol. None of his supplement picks displace that; they sit on top of it.
Strongest evidence: the cardiometabolic and musculoskeletal core
EPA/DHA omega-3 fish oil
Why it helps. EPA and DHA are long-chain omega-3 fatty acids that incorporate into platelet, endothelial, and neuronal membranes. Mechanistically they lower fasting triglycerides through reduced hepatic VLDL secretion, modulate eicosanoid signaling toward less inflammatory mediators, and at very high doses (4 g) reduce major adverse cardiovascular events in secondary prevention.
What the trials show. The REDUCE-IT trial (Bhatt et al. 2019, n=8,179, median follow-up 4.9 years) used 4 g/day of icosapent ethyl, a prescription EPA ester, in statin-treated patients with elevated triglycerides and hit a 25% relative risk reduction in the composite primary endpoint. A broader meta-analysis (Bernasconi et al. 2021) of 40 RCTs found marine omega-3 supplementation associated with a 13% reduction in myocardial infarction risk. The benefit is dose-dependent and strongest with EPA-heavy preparations.
Dose used in trials. 2 g/day combined EPA+DHA at the lower-bound trial level for primary-prevention populations, escalating to 4 g/day icosapent ethyl in secondary-prevention disease.
Form to look for. EPA-dominant fish oil in triglyceride or re-esterified triglyceride form, third-party tested for oxidation (TOTOX) and heavy metals. The ConsumerLab fish oil review flagged several products as exceeding oxidation limits in recent rounds; potency on a label is not the same as quality in the bottle.
Dose-in-trial vs dose-people-buy. A standard 1,000 mg fish oil softgel from a generic store brand typically contains only 180 mg EPA + 120 mg DHA. To match the 2 g/day Attia targets, a reader would have to take roughly seven of those softgels. Most "fish oil did nothing" complaints trace to underdosing, not to fish oil failing.
Skip if you are on warfarin or another anticoagulant without prescriber clearance, or if you have a known fish allergy (algal EPA exists but few brands offer high-EPA algal products).
Actionable takeaway: Read the back of the label and count combined EPA+DHA milligrams, not "fish oil" milligrams. Match the trial doses or do not expect the trial outcomes.
Magnesium (multiple forms)
Why it helps. Magnesium is a cofactor in over 300 enzymatic reactions, including ATP binding, neuromuscular transmission, and glucose handling. Most US adults under-consume relative to RDA per the NIH ODS magnesium fact sheet. Mechanistically, low intake associates with insulin resistance, higher blood pressure, and worse sleep architecture, all of which fall squarely inside Attia's healthspan targets.
What the trials show. A 2016 meta-analysis (Veronese et al.) of 40 prospective studies linked higher dietary magnesium with a 10% lower risk of coronary heart disease, 12% lower risk of stroke, and 26% lower risk of type 2 diabetes per 100 mg/day increment. Smaller RCTs of magnesium supplementation show modest reductions in systolic blood pressure (about 2 mmHg) and improvements in fasting glucose in insulin-resistant adults.
Dose used in trials. 300 to 500 mg/day of elemental magnesium. Attia commonly mixes forms (glycinate at night for sleep, malate or citrate during the day) to hit roughly this range without GI side effects.
Form to look for. Magnesium glycinate, malate, citrate, or threonate. Magnesium oxide is poorly absorbed and mostly useful as a laxative. Stack-splitting across two or three forms reduces the GI load at any one dose.
Skip if you have stage 3 or worse chronic kidney disease, or are on potassium-sparing diuretics or digoxin without clearance.
Creatine monohydrate
Why it helps. Creatine raises intramuscular phosphocreatine stores, the rate-limiting substrate for ATP regeneration during short, high-intensity work. The newer story for an Attia-style readership is in the brain. Creatine supplementation raises brain phosphocreatine on 31P-MRS imaging, supports mitochondrial energetics in neurons, and appears to buffer cognitive performance under sleep deprivation and acute hypoxia. In aging muscle the picture is cleaner: creatine reliably preserves lean mass and strength in older adults paired with resistance training. Mechanism is solid in both compartments; the brain trials are still smaller than the muscle trials.
What the trials show. The 2022 ISSN position stand on creatine (Forbes et al.) summarized over 500 studies and concluded that creatine is "the most effective ergogenic supplement available," with consistent effects on muscle mass, strength, and recovery, and emerging effects on cognition in older adults and vegetarians.
Dose used in trials. 3 to 5 g/day of creatine monohydrate. Loading at 20 g/day for 5 days is not necessary and is mostly a faster-onset choice for athletes.
Form to look for. Plain creatine monohydrate, micronized or not, third-party tested. Branded forms (HCl, ethyl ester, buffered) do not outperform monohydrate in head-to-head trials and cost more.
Skip if you have stage 3 or worse CKD without nephrologist clearance. The "creatine harms healthy kidneys" claim is not supported by the ISSN review but caution remains warranted in pre-existing disease.
Actionable takeaway: Take 3 to 5 g of plain creatine monohydrate daily, every day, with or without food. Loading is optional, branded forms are unnecessary, and skipping rest days does nothing useful.
Vitamin D3
Why it helps. Vitamin D modulates calcium absorption, bone remodeling, and immune signaling. The bone and falls evidence in older adults with low baseline 25-hydroxyvitamin D is reasonably strong; the cardiometabolic and cancer evidence is more mixed.
What the trials show. The D2d trial (Pittas et al. 2019, n=2,423) tested 4,000 IU/day vitamin D3 in adults at high risk for type 2 diabetes and did not hit the primary endpoint, although a pre-specified subgroup with lower baseline 25-OH-D showed benefit. VITAL (Manson et al. 2019) found no overall reduction in invasive cancer or cardiovascular events from 2,000 IU/day in vitamin D-replete adults, but did show signals in deficient subgroups. The honest read: vitamin D3 corrects deficiency reliably; benefits accrue mostly to deficient people.
Dose used in trials. 2,000 to 5,000 IU/day in adults, targeted to a serum 25-OH-D level in the 40 to 60 ng/mL range per the NIH ODS vitamin D fact sheet and Attia's own framing.
Form to look for. Vitamin D3 (cholecalciferol), ideally co-formulated with vitamin K2 (MK-7) when calcium intake is high. Test serum 25-OH-D at baseline and again at 3 to 6 months; the dose-response is highly individual.
Skip if your baseline 25-OH-D is already in the upper-normal range. Supplementing past that does not add benefit in current trials and can edge into hypercalcemic risk at very high doses.
Moderate evidence: methylated B vitamins, sulforaphane, curcumin
Methylated B-complex (with methylfolate and methylcobalamin)
Worth considering if your homocysteine is elevated or you carry MTHFR variants that reduce conversion of folic acid to active 5-methyltetrahydrofolate. Methylfolate at 1 to 15 mg/day plus methylcobalamin (B12) at 1,000 to 5,000 mcg/day reliably lowers homocysteine, which Attia tracks on labs. The cardiovascular-event benefit of homocysteine lowering itself has been disappointing in large RCTs (HOPE-2, VISP) at modest doses, but the methylation-pathway support is still useful in patients with documented poor methylation. The real question is not whether methylated B vitamins lower a number; it is whether the number you are lowering is yours to lower. B-complex is a "fix what your labs show" intervention, not a population-wide prescription.
Sulforaphane (broccoli sprout extract)
Sulforaphane is a Rhonda Patrick favorite that Attia has discussed and used. It activates the Nrf2 antioxidant response and has been shown in animal models and small human studies to support hepatic detox enzyme expression. A 2022 trial (Chandrasekhar et al.) showed reductions in oxidized LDL and inflammatory markers in healthy adults supplementing standardized broccoli sprout extract for 12 weeks. Mechanistically clean, human trials small. Treat as moderate, not strong.
Theracurmin curcumin
Standard turmeric extract is poorly absorbed; Theracurmin is a colloidal submicron form with substantially better serum levels. A 2017 trial (Panahi et al.) of bioavailable curcumin in metabolic syndrome patients improved lipid and inflammatory markers over 12 weeks. Mechanism is anti-inflammatory through NF-kB modulation. Effect sizes are modest and most trials are short. Useful adjunct, not a foundation.
Popular but evidence-thin in Attia's context: AG1, berberine, bergamot
AG1 (Athletic Greens) has appeared in Attia's stack briefly and many listeners read this as endorsement. A supplement can look impressive on a label and still miss the basics: the AG1 label is a multivitamin plus a small dose of various greens and adaptogens, and the evidence base is the underlying ingredient stack, not the proprietary blend itself. You can replicate roughly what AG1 delivers with a quality multi plus a separate magnesium and EPA product at lower cost. "Famous-podcaster-takes-it" should never be the only reason to buy.
Berberine is a plant alkaloid Attia has discussed for metabolic markers. RCTs show fasting glucose and lipid improvements in metabolic syndrome populations at 500 mg two or three times daily. Mechanism touches AMPK activation, similar enough to metformin that some have called it "nature's metformin." Honest framing: if you have impaired glucose tolerance, the actual standard of care is metformin, a generic drug with decades of safety and outcome data, not berberine.
Bergamot polyphenols show a small LDL- and ApoB-lowering signal in RCTs (Mollazadeh et al. 2019 review), typically 500 to 1,000 mg/day, roughly 15% LDL reduction at best. A statin outperforms bergamot by 2 to 3 times on the same endpoint. Position bergamot as adjunct, not replacement, consistent with Attia's own framing.
What to look for when buying
Use the Attia-style criteria the same way you would for any lab-driven protocol:
- Form matters more than brand. EPA percentage, not "fish oil" milligrams. Magnesium glycinate or malate, not oxide. Methylfolate, not folic acid. Creatine monohydrate, not exotic salts.
- Third-party verification. USP Verified, NSF Certified for Sport, ConsumerLab Approved, or a brand-published Certificate of Analysis batch report.
- Dose to the trial. Most disappointing supplement experiences trace to underdosing, not to the molecule failing.
- Track the lab. If a supplement is meant to move a biomarker (ApoB, homocysteine, 25-OH-D, fasting glucose, hsCRP), retest at 8 to 12 weeks. If the marker has not moved, the intervention has not worked for you regardless of how it worked for someone else.
When supplements are not enough
The Attia framing is explicit: supplements do not lower lifetime ApoB exposure the way statins and PCSK9 inhibitors do. They do not replace metformin in a patient with impaired glucose tolerance. They do not substitute for resistance training, Zone 2 cardio, sleep, and protein intake. If your ApoB is over 100 mg/dL, your fasting insulin is elevated, your HbA1c is above 6.0%, or you have a family history of premature ASCVD, the conversation belongs with a cardiologist and an endocrinologist, not on a podcast. Supplements come after that conversation, not before it.
FAQ
Does Peter Attia take NAD+ or NMN? He has discussed both on The Drive and has not endorsed them as core stack items. Human RCT evidence on NMN and NR for hard healthspan endpoints is still thin; mechanism is interesting and most of the data lives in cell and rodent models.
Does he take Rapamycin? Rapamycin is a prescription mTOR inhibitor, not a supplement, and Attia has discussed off-label use under physician supervision. It is outside the scope of this article and outside what we cover in the supplement cluster.
Why does Attia stack magnesium and glycine for sleep? Glycine is an inhibitory neurotransmitter at the brainstem that promotes slow-wave sleep at 3 g pre-bed, and magnesium glycinate provides both the magnesium dose and additional glycine. The combination has small RCT support for sleep onset and quality.
Is creatine safe for women and older adults? Yes. The ISSN 2022 position stand reviewed evidence across populations including post-menopausal women and adults over 65 and supported safety and efficacy.
If you are stacking a few supplements for this, StackMyMed (our companion app) tracks what you actually take, schedules the best time for each one, and flags any combinations worth a second look.
Conclusion: the bottom line on best supplements Peter Attia recommends
The honest summary is shorter than the stack itself. EPA-dominant fish oil at trial dose, magnesium in a chelated form, vitamin D titrated to a target serum level, and creatine monohydrate at 3 to 5 g/day are the four supplements in Attia's public protocol with the strongest human-trial backing for the cardiometabolic and musculoskeletal endpoints he targets. Methylated B vitamins are valuable when the labs justify them. Sulforaphane and curcumin sit at moderate evidence, useful adjuncts without being foundations. ApoB-lowering supplements like bergamot are real but small compared to first-line cardiology. The unifying theme: Attia treats supplements as biomarker-driven adjuncts to a lifestyle and pharmacology protocol, not as the protocol itself.
Next steps:
- Get the labs first (ApoB, Lp(a), fasting insulin, HbA1c, 25-OH-D, ferritin, homocysteine). Build the stack against the values.
- Compare with the sibling roundup at best supplements Huberman recommends for where the two stacks overlap and diverge.
- For the broader picture, read our best supplements for longevity in 2026 overview, and see our supplement review methodology and the Maria Rodriguez author page for our standards.
This article is for informational purposes and not medical advice. Supplements can interact with medications and health conditions. Consult a licensed physician before starting any supplement, particularly if you are pregnant, nursing, taking prescription medications, or managing a chronic condition.
Reviewed by Maria Rodriguez, MS Nutrition Science, focused on cognitive and mood biochemistry.
Recommended Products
As an Amazon Associate, UsefulVitamins.com earns from qualifying purchases at no extra cost to you. We only recommend products supported by published research or third-party testing.
