Supplements A–Z (2026): Evidence-Based Comparison Guides

By Erin Rose · Updated July 10, 2026 · Methodology · About Us

Before you browse, it's worth asking whether you need a supplement at all — see how to decide what you actually need, including the specific short list most adults should consider (and what to skip).

Check price →

• Algal Omega-3

  The vegan branch of the omega-3 family — algal-oil DHA is bioequivalent to fish-derived DHA on the biomarker level (Arterburn 2008 ran it head-to-head against cooked salmon) and matches fish oil's triglyceride-lowering at high doses. Two honest caveats: most algal oil is DHA-heavy or DHA-only (only 1 of 5 verified products discloses its EPA:DHA split), and no algal-oil product has run a hard cardiovascular-outcome trial.

• Alpha-GPC

  A real acetylcholine-precursor mechanism, real (disease-population, prescription-dose) dementia/stroke-recovery evidence, and a small positive sports-performance signal — weighed honestly against a 2021 observational study (12M+ people) linking use to higher stroke risk. Ranked by disclosed actual-mg, not label mg, exposing the industry-wide 50%-blend gap.

• Alpha-Lipoic Acid (ALA)

  Only the R-(+) enantiomer is the active form — and most "alpha-lipoic acid" is racemic (50/50), so a "600 mg" cap is really ~300 mg active R. But here is the counter-intuitive part the cost math reveals: priced per mg of actual active R, bulk racemic is the CHEAPEST source of R; premium R-ALA and stabilized sodium-R-lipoate cost several times more per mg of R — you pay for better absorption, not more active. The R-vs-racemic truth, the "R-ALA Complex" label trap, and the honest evidence (strong for IV diabetic neuropathy, weak long-term oral).

• Ashwagandha

  KSM-66 vs Sensoril vs generic, stress, sleep.

• Astaxanthin

  Real but narrow evidence: screen/eye fatigue, UV/skin tolerance, inconsistent exercise data, and modest lipid markers — plus why the "2 mg EFSA limit" you may have read about does not apply to the natural extract sold as a supplement.

• B-Complex

  A B-complex helps mostly when you are correcting a deficiency — outside that, the "energy" and heart-protection claims do not hold up: lowering homocysteine with B vitamins did not cut cardiovascular events in the big trials. Two things separate a good one: active/methylated forms (methylfolate, methylcobalamin, P5P) vs cheap folic acid and cyanocobalamin, and dose — the popular "B-50"/"B-100" products put B6 at 50-100 mg, right at the upper limit tied to nerve damage. What the evidence really shows, active vs cheap forms, and the megadose trap.

• BCAAs (Branched-Chain Amino Acids)

  BCAAs alone raise post-workout muscle protein synthesis about 22% (Jackman 2017) — but that's a submaximal, incomplete response next to whole protein or a full EAA blend, because building new muscle requires all 9 essential amino acids, not just the 3 in BCAAs (Wolfe 2017). For anyone already eating enough protein, standalone BCAA products are largely redundant. Real but modest soreness benefit; a genuine clinical niche in cirrhosis, walled off from the gym-supplement pitch. Ranked by cost per gram — bulk powder is far cheaper than capsules or flavored blends.

• Beetroot / Dietary Nitrate

  The active compound in beetroot is inorganic nitrate — converted to nitric oxide for blood pressure and endurance. Real evidence at ~6-13 mmol/day (372-800 mg nitrate), but the catch: nitrate content varies >200x between products and almost none disclose it, so "beetroot 1000 mg" tells you nothing about the dose that matters. The mmol-to-mg math, the cost per actual nitrate, and why the marketed milligrams are the wrong number.

• Berberine

  Dosing, dihydroberberine, blood sugar, safety.

• Beta-Alanine

  The carnosine-loading amino acid for high-intensity exercise — solid evidence for 1-4 minute efforts, but useless under 60 seconds or for pure strength. The 4-6 g/day loading dose, the cost per gram, and the label trap: most pre-workouts underdose it, and the tingle you feel is NOT proof of an effective dose.

• Beta-Carotene

  Two large RCTs (ATBC 1994, CARET 1996) found high-dose beta-carotene SUPPLEMENTS increased lung cancer risk and mortality in smokers, former smokers, and asbestos-exposed workers — CARET was stopped 21 months early. AREDS2 later dropped beta-carotene for the same reason. Dietary beta-carotene is a separate, generally favorable picture; the isolated pill is the problem, and most people don't need it at all.

• Betaine (TMG)

  Two honest use cases, both with a catch: betaine reliably lowers homocysteine (well-replicated, dose-dependent) but that's a surrogate marker with no proven cardiovascular-outcome benefit, and 6g/day RAISES LDL cholesterol — a real trade-off. At ~2.5g/day for athletic performance, evidence is genuinely mixed and no trial showed a 1RM strength advantage. Not betaine HCl (a separate stomach-acid product). Ranked by cost per gram — bulk powder is 3-13x cheaper than capsules.

• Betaine HCl (Betaine Hydrochloride)

  Not betaine/TMG — betaine HCl is an acidic stomach-acid salt, usually sold with pepsin. The one genuinely solid finding (Yago et al., UCSF) is narrow and clinical: 1,500mg transiently re-acidifies the stomach and rescues absorption of a pH-dependent drug in volunteers with drug-induced hypochlorhydria — not proof it fixes bloating, reflux, or indigestion in healthy people, where no RCTs were found. Hurwitz 1997 (JAMA) found ~90% of elderly subjects still acidify normally, rebutting the "everyone gets low stomach acid with age" myth. The capsule-titration "warmth test" is unvalidated. Hard contraindication: do not use with ulcer/GERD history, NSAIDs, or PPI/H2-blocker acid reducers. Ranked by cost per capsule — plain vs. with-pepsin vs. with-gentian.

• Biotin

  Hair and nails evidence, dosing, lab-test interference.

• Boron

  Real-but-modest evidence: mineral/bone plausibility, a small short-term hormone shift (not replicated in athletes), and CFB-specific joint data — plus the boric-acid/borax scare that does not apply to supplement-dose boron.

• Boswellia

  Boswellia (Indian frankincense) has real, relatively fast-acting evidence for knee osteoarthritis — but only from extracts standardized to AKBA, the boswellic acid that actually inhibits inflammation. The catch: the trial-grade products disclose an AKBA % or a branded extract (5-Loxin, Aflapin, ApresFlex), while cheap ones print only "65% boswellic acids" (a generic spec) or a bare milligram — and lab testing has found commercial AKBA content ranging from 3.8% down to nearly zero. What the evidence shows, AKBA vs total boswellic acids, and cost by disclosure tier.

• Bromelain

  A pineapple-stem enzyme dosed in activity units (GDU/MCU/FIP), not just mg — identical 500mg labels here carry a 2.5x potency spread (1,200-3,000 GDU). Best-evidenced for post-surgical (dental) swelling, a meta-analyzed but split verdict; modest adjunct signal for sinusitis; osteoarthritis's one strong positive trial (Bolten 2015) tested an enzyme COMBO product, not bromelain alone.

• Calcium

  Citrate vs carbonate, absorption with/without stomach acid.

• Carnosine (L-Carnosine)

  Oral L-carnosine is largely broken down by serum carnosinase before it reaches muscle — for raising muscle carnosine, beta-alanine is the established, better-evidenced route (Harris 2006). Beyond muscle: modest, mixed glycemic-control evidence (one trial positive, one null) and real anti-glycation chemistry that is mechanism-only, not a proven human anti-aging outcome. Ranked by verified cost per day, with a combo-formula product flagged, not ranked as plain carnosine.

• Choline

  A genuine essential nutrient — 90% of Americans fall short of the Adequate Intake, and controlled human depletion trials prove a real deficiency syndrome (fatty liver, muscle damage). Pregnancy is the strongest thread (Caudill 2018 RCT); TMAO is a real, honestly-disclosed nuance, not a reason to avoid AI-range intake. Ranked by elemental choline mg, not bitartrate salt weight.

• Chondroitin

  Osteoarthritis evidence and how it pairs with glucosamine.

• Chromium

  Chromium is sold for blood sugar and weight, but the honest evidence is modest and mixed — meta-analyses in people with diabetes disagree, and it is not an established treatment for anyone. Three things the label gets wrong: picolinate is the most-studied form but no better absorbed than the others (all absorb ~1%); every product is 6-29x the ~35 mcg you actually need; and the identical picolinate dose sells for wildly different prices, which is brand, not formulation. What the evidence shows, the forms, and the cost.

• Citicoline (CDP-Choline)

  Two Cognizin-funded trials show a modest healthy-adult attention/memory signal — but the largest citicoline RCT ever run (ICTUS, n=2,298) was NULL for stroke recovery. Ranked by cost per mg; a Cognizin-branded pick, not a generic, turns out cheapest per mg.

• Citrus Bergamot

  Bergamot polyphenols (BPF) may lower LDL/triglycerides and raise HDL in short Italian trials — but the biggest effects come from studies with commercial or institutional ties to the extract, and the one independent systematic review found the evidence "quite limited." A real statin/CYP3A4 interaction caution applies. Ranked by disclosed polyphenol standardization, not just price.

• CLA (Conjugated Linoleic Acid)

  The classic animal-vs-human-gap supplement: dramatic fat loss in rodents/livestock, but two human meta-analyses find only a small, plateauing effect (~0.09 kg/week) — and several RCTs found it worsens insulin resistance and lowers HDL in the exact population most likely to buy it, plus two rare hepatotoxicity case reports.

• Cod Liver Oil

  The real safety constraint is the vitamin A load, not the omega-3 — some disclosed servings already use 8-25% of the adult UL before a multivitamin is added. Cod-liver-oil-specific evidence is thin: one real RA RCT (NSAID-sparing, not disease-modifying) plus pediatric trials that combined it with a multivitamin. Only 2 of 5 tracked products disclose vitamin A to a number.

• Collagen

  Types, skin and joint evidence, dosing.

• Colostrum

  The one spec that decides quality is IgG — and most brands, including ARMRA, hide it. Disclosed IgG %, cost per gram of IgG, and an honest read on what the evidence supports.

• Copper

  The honest headline is a warning, not a benefit: most iron-replete adults already exceed the 900mcg RDA from food, and the real reason anyone lands here is offsetting long-term high-dose zinc, which competitively blocks copper absorption and can cause sideroblastic anemia and neutropeny. Wilson's disease is a hard supplementation contraindication. "Copper reverses gray hair" and oral GHK-Cu peptide anti-aging drops are marketing overreach with no human RCT support.

• CoQ10

  Ubiquinol vs ubiquinone, absorption, statins.

• Cordyceps

  Most US "cordyceps" is mycelium-on-grain — mycelium dried on a rice or oat substrate, largely starch, low in the actives. Premium products are fruiting body / Cordyceps militaris that disclose a real beta-glucan % (the validated active) and cordycepin; the rest hide behind a big "milligram" number or a "polysaccharide %" that just counts grain starch. You can even spot-check it at home with an iodine drop. The fruiting-body-vs-mycelium truth, the beta-glucan test, and the honest (modest) evidence.

• Cranberry

  The compound studied for UTI prevention is the A-type proanthocyanidin (PAC), and the studied dose is ~36 mg PAC/day. But only a handful of products disclose PAC in milligrams; the rest sell a meaningless "25,000 mg equivalent" number that says nothing about the compound that matters. The 2023 Cochrane reversal (a real reduction in recurrent-UTI risk for women), the 36 mg PAC dose, and cost per day at the dose the trials actually used — for prevention, not treatment.

• Creatine

  Monohydrate vs HCl, dosing, what actually works.

• Curcumin

  Bioavailability problem, enhanced forms, joint pain.

• D-Mannose

  The mechanism (blocking E. coli bladder adhesion) is real, and early trials looked promising — but the largest, best-designed trial ever run (MERIT, 2024, 598 women, JAMA Intern Med) found no benefit for recurrent-UTI prevention. Cost per 2 g studied dose, the capsule under-dosing trap, and the honest evidence arc from hype to null.

• Digestive Enzymes

  Use-by-use, not one verdict: lactase for lactose intolerance is strong (Montalto 2005, Ojetti 2010 RCTs); alpha-galactosidase (Beano-type) for bean/legume/FODMAP gas is real, RCT-backed, but its one IBS trial (Hillila 2016) was null; pancreatic enzyme replacement therapy is strong evidence but a prescription medicine for diagnosed EPI, not an OTC bloating fix; broad-spectrum "digestive enzyme complex" blends for general dyspepsia in healthy people are thin, small-trial evidence, one 100% manufacturer-authored (DigeZyme/Sami-Sabinsa).

• Elderberry

  The honest version: elderberry has real but small and mixed evidence for shortening cold/flu symptoms — several small (mostly industry-funded) trials found a benefit, but the single most rigorous independent trial found none. On the shelf, the marker actives are anthocyanins, and only one product in our set discloses an anthocyanin % — the rest sell a bare milligram, an "equivalent" number, or a sugar-forward gummy. What the evidence really shows, the anthocyanin-disclosure problem, and the raw-berry and autoimmune safety notes.

• Electrolytes

  Sodium/potassium/magnesium ratios, sugar-free picks, keto and sweat loss.

• Essential Amino Acids (EAA)

  The honest "complete" answer to BCAAs — free-form EAAs stimulate muscle protein synthesis more than an equivalent amount of intact protein, gram for gram, but whey is already ~45-50% EAA by weight, and two trials found small EAA doses matched much larger whey doses for MPS. Largely redundant for anyone eating adequate protein; the most defensible niche is older adults, and even there the evidence is split. Ranked by cost per gram — bulk powder is far cheaper than premium capsules.

• Fiber

  Psyllium vs inulin vs methylcellulose vs wheat dextrin — why the type matters more than the gram count, who each one helps (and harms), and the cost per gram of soluble fiber.

• Fisetin

  Senolytic mouse-lifespan data (Yousefzadeh 2018) is real and strong — but the one completed human efficacy trial (ROPE, knee osteoarthritis, using the actual mouse-derived dose) found NO benefit vs placebo. Retail sells 100-500mg/day continuously; the tested protocol is ~1,400mg/day pulsed.

• Flaxseed Oil

  The plant-ALA branch of the omega-3 family — flaxseed oil reliably raises plasma/erythrocyte EPA (Harper 2006: +60% on 3g ALA/day), but the ALA→DHA conversion pathway runs under 0.5% (Plourde & Cunnane 2007) and DHA essentially never rises, even at high trial doses. A 2022 scoping review found high-dose flaxseed oil produced no omega-3-index increase while algal oil raised it in every study. Real, if modest, uses exist (BP, small-dense LDL) and ALA carries a real observational — not RCT — CVD-risk association. Oxidation/rancidity is a first-class quality issue for this liquid PUFA.

• GABA

  Whether oral GABA meaningfully crosses the blood-brain barrier is genuinely disputed (Boonstra 2015) — the mechanism itself is unresolved, not just "the trials are small." The stress/sleep trial base is small, short, and mostly industry-funded (Pharma Foods International, maker of PharmaGABA); most positive trials used ~100mg of the branded fermented form, while most retail capsules are synthetic GABA at 500-750mg, 5-7.5x the tested dose.

• Garlic

  Garlic modestly lowers blood pressure in people with hypertension — but most garlic pills may deliver almost none of the active. Allicin isn't in the clove; it forms from alliin when garlic is crushed, and the enzyme that makes it is destroyed by stomach acid, so 83% of even enteric-coated brands were found to release under 15% of their labeled allicin. Two routes actually work: enteric-coated allicin, and aged garlic extract (the form with the real BP trial data). Why "1000 mg garlic" tells you nothing, which products deliver, and the honest blood-pressure evidence.

• Ginkgo Biloba

  The honest headline first: the strongest, most independent evidence is negative — the Cochrane review calls the benefit "inconsistent and unreliable," and two large trials show ginkgo does NOT prevent dementia. The narrow positive signal is EGb 761 (the specific trial extract, 240 mg/day) for treating — not preventing — dementia. So which product matters: only Nature's Way Ginkgold is that trial extract; and only a couple of brands disclose their ginkgolic-acid level (the allergenic contaminant capped at <5 ppm). What the evidence really shows, EGb 761 vs generic, and the bleeding-risk caution.

• Glucomannan

  Konjac glucomannan, honestly evidenced: the EU authorized a 3g/day weight-loss claim, but the best RCT and meta-analysis found no significant weight loss. LDL/cholesterol is the better-supported effect — and the water/choking safety rule is load-bearing, not boilerplate. See /fiber/ and /psyllium/ for how it compares to other fibers.

• Glucosamine

  Osteoarthritis evidence, sulfate vs HCl, dosing.

• Glutathione

  Oral glutathione raising body stores is real (Richie 2015, dose-dependent RCT) — but "proven skin treatment" is small/short evidence, and "form X absorbs better" is largely unverified in humans. IV glutathione is never endorsed: no RCT efficacy data and a Philippine FDA safety warning.

• Glycine

  One identical molecule that costs up to 12x more from some brands. The 3 g sleep dose (honestly graded), cost per gram, and the glycine you already get from magnesium glycinate and collagen.

• Grape Seed Extract

  Three independent meta-analyses (Feringa 2011, Zhang 2016, Foshati 2022) converge on a real but small blood-pressure-lowering effect — bigger in younger, obese, or metabolic-syndrome people — but a well-designed single RCT (Ras 2013, 300mg/day) found no significant effect. A modest risk-marker mover, not a hypertension drug. Ranked by disclosed OPC/proanthocyanidin standardization, with a mild antiplatelet-mechanism caution for anyone on blood thinners.

• HMB (β-Hydroxy β-Methylbutyrate)

  HMB's famous "7+ lbs of muscle in 12 weeks" trials (Wilson 2014, Lowery 2016) were funded by and co-authored with Metabolic Technologies Inc., the company that patented and sells HMB — and independent researchers with no MTI tie found trivial-to-null effects in the same trained-athlete population (Rowlands 2009, Sanchez-Martinez 2018). The more defensible signal: untrained lifters, older adults, and bed rest, where HMB modestly slows muscle LOSS, not builds new muscle. Ranked by cost per 3g/day dose — cheap Ca-HMB powder beats the pricier HMB-FA form on price, with no proven outcome advantage for FA.

• Hyaluronic Acid (Oral)

  Oral HA capsules are a separate product from HA dermal fillers and topical serums — oral evidence never transfers. Small, replicated skin-hydration trials (mostly Kewpie-funded, the Hyabest maker) at 120mg/day; thinner, dose-inconsistent joint evidence. Ranked by trial-matched ingredient vs generic, then cost per mg.

• Inulin

  The best-evidenced prebiotic fiber — and the one most likely to bloat you. A fermentable chicory-root fiber that feeds Bifidobacteria (not a bulking/viscous fiber like psyllium or glucomannan), with a real BENEO ingredient-manufacturer COI on its flagship meta-analysis and a genuine gas/bloating downside for IBS. See /fiber/ for how it compares to other fiber types.

• Iodine

  The U-shaped curve: Teng 2006's 5-year China cohort found BOTH mild deficiency AND excess iodine raised thyroid-dysfunction and autoimmune-thyroiditis rates — more is not automatically better. Pregnancy's the highest-stakes deficiency thread; Lugol's/Iodoral megadose 'protocols' (11-45x the UL) are flagged, not recommended.

• Iron

  Ferrous sulfate vs bisglycinate, low ferritin, tolerability.

• Krill Oil

  The phospholipid "better absorbed" edge over fish oil is real but small and inconsistent (positive, non-significant, and null results across three trials) — the bigger issue is density: krill delivers only 108-240mg EPA+DHA per gram of oil vs 300-500mg+/g in concentrated fish oil, so you often pay more for less omega-3. Half the products checked don't even disclose EPA+DHA mg.

• L-Arginine

  Oral L-arginine has poor, variable bioavailability (~68%, 51-87% range) due to first-pass arginase metabolism — a head-to-head trial found L-citrulline raises blood arginine MORE effectively. Blood pressure is arginine's strongest evidence (~5/3 mmHg, Dong 2011); a landmark trial found it raised mortality when taken after a heart attack.

• L-Carnitine

  It comes in four forms with different jobs — but they're all sold as generic "energy and fat burning." L-carnitine L-tartrate (LCLT) has the exercise-recovery evidence; acetyl-L-carnitine (ALCAR) crosses into the brain for cognition and nerve pain; propionyl-L-carnitine (PLC) is the circulation form for claudication; plain L-carnitine is the base. Buy the wrong form and you paid for the wrong evidence. Which form for which goal, the cost per gram, and the honest read on the TMAO heart question.

• L-Citrulline

  The nitric-oxide / pre-workout amino acid, judged honestly: modest blood-pressure and small performance evidence — and the label trap that "6,000 mg citrulline malate" is not 6,000 mg of citrulline. The clinical dose vs label dose, cost per actual gram, and why independent testing found most "2:1" products under-deliver.

• L-Glutamine

  The honest version: L-glutamine is a cheap commodity amino acid your body already makes in abundance, and the "seals leaky gut / heals your gut" marketing isn't supported — controlled trials show no effect on gut permeability at normal doses, none in athletes, and high doses were actually linked to higher mortality in critical illness. Its one FDA-approved use is reducing sickle-cell crises. So the real questions are just: cheapest gram (it's all the same molecule), powder vs capsule, and whether you need a sport certification.

• L-Lysine

  Cold-sore prevention rests on a genuinely mixed, mostly 1978-1987 trial record — 3 positive RCTs vs 3 null RCTs at similar doses — and Cochrane 2015 couldn't confirm it works, unlike oral antivirals. Ranked by cost per gram at the ~1-3g/day trial-relevant dose, not sticker price.

• L-Theanine

  Calm-focus evidence, dosing, and stacking with caffeine.

• L-Tryptophan

  The serotonin/melatonin precursor amino acid: a real but modest, dose-dependent sleep effect at ~1 g, the 1989 EMS contamination history explained honestly, and the serotonin-syndrome interaction with SSRIs/MAOIs.

• L-Tyrosine

  A real but narrow tool: tyrosine improves cognition under acute stress (cold, sleep loss, heavy load) but does nothing at rest — it is not a daily focus pill. And the dose is the story: trials used ~2 g up to 7-13 g of plain L-tyrosine, so a 500 mg capsule is a fraction of a studied dose and powders cost a fraction per gram. Plus the NALT trap: N-acetyl-L-tyrosine is marketed as "more bioavailable" but is poorly converted to usable tyrosine. What works, the dose gap, and why the fancy form is worse.

• Lactoferrin

  Two of the most-cited pregnancy-anemia trials behind it were retracted in 2023 — and the two dedicated COVID-19 RCTs are clean nulls. What the surviving iron, immune, and gut evidence actually shows.

• Lion's Mane

  What the human evidence actually shows, and how to tell a real fruiting-body extract from grain filler by its beta-glucan.

• Lutein & Zeaxanthin

  AREDS2 is genuinely strong evidence (HR 0.82-0.91) — but only in people who already have intermediate/advanced AMD, not a prevention claim for healthy eyes and not the "blue light" screen-strain story most marketing borrows it for. The 10mg:2mg AREDS2 ratio vs what is actually on shelves, and the honest CREST read on meso-zeaxanthin "triple carotenoid" formulas.

• Maca

  Maca has one effect with repeat trial support: a modest lift in self-reported sexual desire, seen at 1,500-3,000 mg/day. It does this WITHOUT changing testosterone — the 2002 trial measured hormones specifically to check, and found nothing. So the shelf full of maca "testosterone boosters" is selling you a mechanism the research already ruled out. The trials are small and the systematic review calls the evidence limited. On the shelf, the catch is arithmetic: a 750 mg capsule means the studied dose is four capsules a day, so a "180 servings" bottle is a 45-day supply. Cost per day at the dose that was actually studied, what the colours mean, and where the evidence is thinner than the marketing.

• Magnesium

  Forms compared, sleep/anxiety/cramps picks, dosage, deficiency signs.

• Manganese

  Another warning-first essential mineral: ordinary diets already supply several times the tiny 1.8-2.3mg AI, so standalone supplementation is rarely necessary. The load-bearing story is excess — manganism, a Parkinsonism-like neurotoxic syndrome from occupational, drinking-water, and impaired-excretion (liver disease/long-term parenteral nutrition) exposure. Most retail standalone products are dosed at 3-20x the AI, and several exceed the 11mg UL outright.

• MCT Oil

  One real, reproducible effect (C8-driven ketone/BHB elevation) oversold into three more — modest weight loss, a null coconut-oil thermogenesis claim, and a hype-zone cognition signal. The GI-tolerance titration ramp, cost per gram of C8 (the ketogenic driver most labels hide), and why coconut oil is not the same product.

• Melatonin

  How much to take (less is often more), timing for sleep vs jet lag, and the label-accuracy problem.

• Methylfolate

  Vs folic acid, MTHFR, pregnancy dosing.

• Milk Thistle (Silymarin)

  The active is silymarin — and the milligram number on the front of the bottle usually tells you nothing about it. Quality products disclose an 80% silymarin standardization (so you can compute the actual active); cheap ones headline "1000 mg" or "3000 mg / 4:1 / full-spectrum" with no silymarin figure, and independent testing found silymarin content varying ~2,800% between brands. The evidence is real but modest (lower ALT/AST; no confirmed mortality benefit), the standardization gap is the moat, and IV silibinin (the real mushroom-poisoning antidote) is not the same as an oral capsule.

• MSM

  MSM (methylsulfonylmethane) is a commodity — the identical molecule in every bottle — with modest, mixed evidence for knee osteoarthritis (physical function improves; pain often does not reach significance). So the only thing worth comparing is cost per gram, which ranges widely, and powder is far cheaper than capsules. The one quality signal is OptiMSM (a branded, distilled MSM) vs generic — and joint blends bury a tiny MSM dose at a much higher effective price. The honest evidence, the cost-per-gram math, and OptiMSM vs generic.

• Multivitamins

  Who actually needs one, what to look for, best picks.

• Myo-Inositol

  PCOS evidence, the 40:1 ratio, and dosing.

• NAC

  Liver/glutathione and mental health, dosing.

• Nicotinamide Riboside (NR)

  NAD+ reliably rises in all 8 trials, but five independent, non-industry RCTs each testing a different organ system found the functional/clinical endpoint null or underpowered. NR's regulatory story is NMN's opposite: never excluded, never delisted. Only two products cleared the ASIN gate — a near-duopoly, disclosed honestly.

• NMN (Nicotinamide Mononucleotide)

  Blood NAD+ reliably rises and a walking-speed/muscle-function signal replicates across several small trials — but a 2024 meta-analysis of 8 RCTs found no glucose/lipid benefit, and the FDA excluded then reinstated NMN as a legal supplement (2022-2025). Ranked by purity/third-party testing, not just price.

• Omega-3

  EPA vs DHA, triglyceride vs ethyl ester, best fish oil.

• Panax Ginseng

  The active compounds are ginsenosides — found ONLY in true Panax ginseng (Asian) and Panax quinquefolius (American). "Siberian ginseng" (Eleuthero) is not a ginseng and has ZERO ginsenosides. Most labels list "ginseng root mg" with no ginsenoside % (content varies 15-36x vs label), and evidence is real-but-modest: an ED benefit tempered by a low-certainty Cochrane review, memory (not overall cognition), and a small glucose effect. The species trap, the standardization gap, and what the evidence actually shows.

• Phosphatidylserine (PS)

  The FDA's 2003 "qualified health claim" is real but the weakest tier FDA issues — and the trials behind it used bovine-cortex PS, discontinued for decades. Every product sold today is soy or sunflower, a different substance; the direct soy replication of the classic 300mg dose (Jorissen 2001) was null. Ranked by disclosed source (soy/sunflower/undisclosed) and branded vs generic raw material, then cost per mg.

• Pine Bark Extract (Pycnogenol)

  Nearly all human pine bark evidence is on one patented, industry-funded ingredient — Pycnogenol® — and the one fully independent review (Cochrane 2020, 27 RCTs) found "no definitive conclusions... possible" for any condition. Three blood-pressure meta-analyses of overlapping trials disagree (null, modest-positive, positive-but-not-in-well-designed-trials). A generic pine bark extract has never been tested head-to-head against Pycnogenol, so equivalence is unproven — ranked by genuine-vs-generic split, not price alone.

• Potassium

  Why OTC pills are capped at 99 mg (2% of what you need), who should never supplement it (kidney disease, blood-pressure meds), and how to actually raise your intake.

• PQQ (Pyrroloquinoline Quinone)

  Small cognitive/attention/cerebral-blood-flow signals cluster around 20mg BioPQQ trials — but nearly every positive trial traces back to Mitsubishi Gas Chemical (BioPQQ's maker) or a Cepham-affiliated author, and the mitochondrial-biogenesis mechanism is mouse/cell-culture: in the one human trial that measured it in muscle (Hwang 2020), the biomarker rose with ZERO functional benefit. Ranked by BioPQQ-branded vs generic disodium, then cost per mg.

• Probiotics

  Strain-specific evidence, condition matching.

• Protein

  Whey vs plant, dosing, protein for GLP-1 users and weight loss.

• Psyllium

  The single best-evidenced fiber supplement, deep-dived: FDA-authorized cholesterol claim, glycemic control in type 2 diabetes, constipation, and the honest husk-vs-powder-vs-capsule buying math. See /fiber/ for how psyllium compares to other fiber types.

• Quercetin

  Three independent meta-analyses find a small, real blood-pressure reduction that only clears significance at 500mg/day plain quercetin — but plain and phytosome (Quercefit) forms deliver very different absorption per mg (Riva 2019: ~20x), so a milligram isn't a milligram across forms.

• Rhodiola

  The adaptogen with real-but-modest evidence — dose-dependent endurance/VO2max gains above ~600 mg/day and honest, mixed anti-fatigue data — plus a hard buying problem: rosavins are the R. rosea species fingerprint, yet many products disclose only salidroside (or nothing), and surveys have found 20-40% of R. rosea products mislabeled or the wrong species. The studied dose, the rosavin standardization moat, and the evidence graded honestly.

• Saffron

  The mood spice with unusually real evidence — small trials find saffron extract comparable to SSRIs for mild-to-moderate depression (studied for, not a treatment) — and a hard buying problem: saffron is one of the most adulterated commodities on earth, and most products never disclose their crocin/safranal potency. Dose, the standardization moat, and the honest depression evidence.

• SAMe

  Depression evidence, dosing, and stability caveats.

• Saw Palmetto

  The honest headline first: the strongest, most independent placebo-controlled trials (STEP, CAMUS) and the Cochrane review found saw palmetto no better than placebo for prostate/urinary (BPH) symptoms — even at 2-3x the dose — and it does not lower PSA. The one nuance is that standardized extracts (the ~85-95% fatty-acid, 320 mg/day form the trials used, and the European "Permixon" extract) are a different, better-characterized product than the cheap non-standardized "berry powder" most brands sell. What the evidence really shows, standardized vs berry powder, and the honest read on hair loss.

• Sea Moss

  What the evidence really shows (almost no human trials), the iodine/thyroid risk nobody labels, heavy metals, and how to buy a tested one instead of a 16-in-1 blend.

• Selenium

  Thyroid (Hashimoto’s) evidence, dosing, and the narrow safe range.

• Shilajit

  What the thin human evidence really shows, and the moat almost no one covers: heavy metals (including thallium), the fulvic-acid % confusion, and how to buy a tested one.

• Spermidine

  Kiechl's dietary-intake mortality association vs SmartAge's null primary memory endpoint (n=100, RCT) — autophagy is a mouse/yeast mechanism, not human proof. Ranked by disclosed actual spermidine mg, not wheat-germ-extract weight.

• Spirulina

  Cultivated Arthrospira with real-but-modest evidence for cholesterol, blood pressure and fasting glucose — and a widely misunderstood safety story. The microcystin scare belongs to WILD "blue-green algae" (AFA), not true spirulina, which tests clean; the real quality axis is cultivation source + heavy-metal testing, which most labels hide behind a vague "third-party tested." What the evidence shows, the AFA vs Arthrospira distinction, and how to tell a transparent source from marketing.

• Taurine

  The energy-drink amino acid, judged honestly: caffeine (not taurine) drives the "energy" buzz, but taurine has real, separately-earned wins for blood pressure and exercise. Dose by goal, the cost-per-gram (one identical molecule, ~8x price spread), and what the 2023 longevity paper actually showed.

• Tocotrienols

  The other vitamin E — structurally distinct from tocopherols. The classic cholesterol claim is small, 1990s, industry-linked, and doesn't hold up in later independent trials; the more promising recent signal is fatty liver (NAFLD), still limited to one research group. Annatto (tocopherol-free) vs palm (full-spectrum) is a real, evidence-grounded sourcing distinction, not just marketing.

• Tongkat Ali

  Testosterone evidence, dosing, tongkat vs fadogia.

• Urolithin A

  A replicated muscle-endurance/strength signal across 7 human RCTs — but 5 of 7 are funded by the company that sells the ingredient (Zhao 2024 is the lone independent trial). The trial-grade product (Timeline Mitopure, 500mg free UA) isn't reliably buyable on Amazon; both live Amazon options are compromised (under-dosed or the wrong molecule).

• Vitamin A

  Preformed vitamin A (retinol/retinyl esters) is a proven human teratogen at high doses in pregnancy (Rothman 1995, ~10,000 IU/day threshold — the same number as the adult UL) and, in chronic excess, causes liver and bone toxicity. Most people are already replete from food and don't need a supplement; provitamin-A beta-carotene is a different, non-teratogenic substance covered separately.

• Vitamin B12

  Methylcobalamin vs cyanocobalamin, energy, deficiency risk.

• Vitamin B6

  Vitamin B6 is the rare water-soluble vitamin with a real toxicity ceiling: chronic high doses cause sensory nerve damage, so the tolerable upper limit is 100 mg/day — yet many products sell exactly 100 mg, thousands of percent over the ~1.7 mg you need. The other axis is pyridoxine HCl vs P5P (the active coenzyme form, marketed as superior on thin evidence). Where B6 actually helps (PMS, pregnancy nausea), the dose that is risky, and how to buy a sane one.

• Vitamin C

  Forms, the absorption ceiling, immune evidence.

• Vitamin D

  D2 vs D3, dosing (1000–5000 IU), K2 pairing, deficiency.

• Vitamin E

  Vitamin E is a "more is not better" vitamin: the biggest, most independent trials found no heart or cancer prevention and signals of harm at the 400 IU dose most products sell — a heart-failure signal, increased prostate cancer, and higher all-cause mortality at high doses. The one proven use is a narrow liver condition (NASH) at high dose. On the label, natural d-alpha is ~twice as active as synthetic dl-alpha, and gamma/tocotrienol forms are a different product entirely. What the evidence really shows, natural vs synthetic, and why the dose matters.

• Vitamin K

  K1 vs K2, how much you actually need (90/120 mcg AI), and the one warning that matters: vitamin K directly opposes warfarin. Routes into /vitamin-k2/ for MK-7-specific depth.

• Vitamin K2

  MK-7 vs MK-4, how much to take with vitamin D, and the calcium-routing (bone vs artery) evidence.

• Zinc

  Forms, copper balance, immunity and colds, upper limits.

How we compare supplements

We pull product label data from the NIH Dietary Supplement Label Database, clinical evidence from PubMed systematic reviews, and third-party certifications from USP and NSF. Read our methodology and editorial standards.