Supplements¶

The supplement market is mostly theater. Tier 1: fix structural deficits (D3, omega-3, magnesium, B12, iodine). Tier 2: creatine and caffeine have robust evidence for performance. Everything else requires a tested deficiency or specific clinical reason.

🌿 budding tended 2026-05-21 research health supplements nutrition performance

flowchart LR
  def[structural deficit] --> t1[Tier 1: always fix]
  perf[performance goal] --> t2[Tier 2: real evidence]
  diag[diagnosed deficiency] --> t3[Tier 3: conditional]
  hype[everything else] --> t4[Tier 4: theater]
  t1 --> stack[your stack]
  t2 --> stack

L0 — TL;DR (≤5 lines)¶

The supplement industry sells certainty on soft evidence. Most products have marginal or zero effect in a well-fed adult; the few that work, work in specific contexts. Tier 1 — vitamin D3, omega-3 EPA/DHA, magnesium, B12 (if vegan/50+), iodine (if off iodized salt) — are structural deficits produced by the built environment, not compliance failures; fix all five before anything else. Tier 2 — creatine monohydrate for strength and power, caffeine for performance and alertness — have robust trial data and a known mechanism. Everything else requires a tested deficiency or specific clinical reason. Default: one good varied diet plus Tier 1 coverage beats any nootropic stack.

L1 — Overview¶

Core question¶

Which supplements have strong enough evidence to justify regular use in a well-fed, active adult, and how do you distinguish signal from the surrounding noise?

Why it matters¶

The supplement industry generates ~$200B/year globally on evidence that is, on average, much weaker than pharmaceutical-grade claims.
Several Tier 1 gaps are structurally inevitable at modern latitudes and with modern diets — not fixable by "eating well" because the deficit is in the food system and built environment, not the plate (L-1917).
Tier 2 supplements (creatine above all) have a stronger evidence base per-compound than many pharmaceuticals — and are systematically ignored by the same people spending heavily on unproven stacks.
Supplements interact with medications and each other; "natural" ≠ safe. The dose-response can invert (vitamin A toxicity; selenium toxicity above 400 µg/day; B6 neuropathy from overdose).
For a training, brain-priority profile, the correct stack is short and cheap.

Mermaid map (L1)¶

flowchart LR
  food[food-first] --> gap{structural gap?}
  gap -- yes --> t1[Tier 1 fixes]
  gap -- no --> tested{tested deficiency?}
  tested -- yes --> t3[Tier 3 conditional]
  tested -- no --> goal{clear performance goal?}
  goal -- strength/power --> creatine[creatine 3-5g/d]
  goal -- alertness/endurance --> caff[caffeine ~3mg/kg]
  goal -- no specific --> stop[stop here]
  t1 --> d3[D3 1500 IU + K2]
  t1 --> o3[omega-3 2g EPA+DHA]
  t1 --> mg[Mg glycinate 400mg]
  t1 --> b12[B12 if vegan/50+]
  t1 --> iod[iodine if non-iodized salt]

Skeleton sub-claims¶

Tier 1 gaps are environmental, not personal. Vitamin D: sunscreen + indoor work
winter latitude. Omega-3: modern omega-6:3 ratio ~15:1 vs. ancestral 1–4:1. Magnesium: soil depletion + ultraprocessed food. These persist in healthy eaters.
Creatine is the most evidence-backed performance supplement. 700+ trials, known mechanism (phosphocreatine recycling), consistent 8–20% strength output increase, cognitive upside in sleep-deprived and older adults. Cheap and safe at 3–5 g/day indefinitely.
Caffeine works but has a ceiling and a sleep cost. 3 mg/kg is the well-studied dose for endurance and strength output. Chronic use blunts the effect; the sleep cost of late caffeine is larger than the daytime gain for most people.
Most adaptogens, nootropics, and longevity supplements are Tier 4. Ashwagandha has marginal cortisol evidence in stressed populations; NMN/NR has no human lifespan data; most herbal stacks have effect sizes too small to detect without N>1000 trials.
Blood testing gates the conditional tier. Iron, folate, zinc, vitamin A, iodine vary enough person-to-person that supplementing blind risks overshoot. Test, then treat.

L2 — Deep dive¶

1. Tier 1 — structural deficit fixes¶

Gaps that recur in well-fed populations because the food system and built environment create them, not individual failure (L-1917).

Supplement	Why structural	Dose	Notes
Vitamin D3	Indoor work + sunscreen + latitude >40°N + winter. Dietary sources are few. Melanin reduces synthesis.	1000–2000 IU/day (Oct–Apr ≥40°N); 600–800 IU year-round if limited sun	Target: 25(OH)D 40–60 ng/mL. Take with fat. Toxicity begins at chronic >10,000 IU. Pair with K2 if supplementing.
Omega-3 EPA+DHA	Western omega-6:3 ratio ~15:1; ALA (flax) → EPA conversion ~5%; few eat fatty fish 2×/wk	1–2 g combined EPA+DHA/day	TG reduction measurable at 2–4 g. Cardiovascular, brain, anti-inflammatory. Algae-based DHA is vegan-complete.
Magnesium glycinate	Soil depletion reduces food content; ultraprocessed food is near-zero; absorption competes with calcium	300–400 mg/day, glycinate or malate form	Threonate for brain-targeted delivery (plausible BBB advantage; small human data). Oxide is poorly absorbed. Night timing improves sleep.
B12 (cobalamin)	Only animal sources (or fortified); absorption via intrinsic factor decreases after 50; vegans: zero dietary source	500–1000 µg sublingual/day	Sublingual bypasses intrinsic factor decline. Methylcobalamin preferred for MTHFR variant carriers; marginal difference otherwise.
Iodine	"Natural" and sea salts are not iodized; soil iodine is regional	150 µg/day; 220 µg if pregnant	Seaweed is variable (often far above RDA). Excess (>1100 µg) can trigger thyroiditis in susceptible individuals.

Shortcut stack: D3 1500 IU + K2 MK-7 100–200 µg + omega-3 2 g EPA/DHA + magnesium glycinate 400 mg covers the three biggest structural gaps. Monthly cost: ~€10–15 at any reasonable brand.

Why K2 alongside D3: D3 raises calcium absorption; K2 (MK-7 form) routes calcium to bone and arteries rather than soft tissue. Evidence base is mechanistically solid; RCT confirmation at scale is still accumulating. Low risk, plausible upside.

2. Tier 2 — performance adjuncts (robust evidence)¶

Creatine monohydrate¶

The best-evidenced performance supplement across five decades and 700+ randomized trials.

Mechanism: creatine phosphate is the rapid-replenishment currency for explosive ATP recycling in fast-twitch muscle. Supplementing saturates the muscle phosphocreatine pool, extending the duration and peak power of anaerobic bursts.

Outcome	Effect size	Notes
Strength / power output	+8–20% peak	Clearest on compound movements and sprint-type efforts
Lean mass (with training)	+1–2 kg over 4 weeks	Acute: intracellular water. Real hypertrophy follows over months.
Cognitive (general healthy adults)	Small (<5%)	Larger in vegetarians and elderly (lower baseline phosphocreatine)
Cognitive (sleep-deprived)	Moderate	5 g maintains cognitive performance across ~24h sleep loss
Inter-set recovery	Faster	More phosphocreatine available; repeatable effort over more sets

Protocol: - No loading required. 3–5 g/day indefinitely. Loading (20 g/day × 5 days) saturates in 5 days vs. ~3 weeks; same long-term result. - Creatine monohydrate is the correct form. Kre-Alkalyn, HCl, ethyl ester — no demonstrated advantage at higher price. - Take with water; timing and food context are irrelevant past the first week. - ~30% are "non-responders" — those with already-high baseline phosphocreatine via high meat intake. Safe to trial 4 weeks and assess. - Safe at 3–5 g/day indefinitely for healthy kidneys.

Caffeine¶

The world's most-used psychoactive compound and one of the best-characterized performance supplements.

Outcome	Dose	Notes
Aerobic endurance	3 mg/kg (210 mg for 70 kg)	Consistent 3–7% improvement in time trials
Strength	3 mg/kg	Smaller effect; most reliable on compound lower-body movements
Alertness / reaction time	1–3 mg/kg	Dose-dependent; acute; tolerance limits chronic effect
Fat oxidation	Small acute increase	Effect size too small to be meaningful for body composition

Ceiling and tolerance: daily use down-regulates adenosine receptors, blunting the effect. Standard fix: 10–14 day washout every 6–8 weeks. Withdrawal peaks at 24–48 h (headache, fatigue); resolves by day 4–5.

Sleep cost (the critical constraint): half-life is 5–6 hours. 400 mg at 2 pm measurably fragments sleep architecture at 10 pm. The performance cost of the lost deep sleep exceeds the daytime gain for most people. Hard cut-off: no caffeine after 1 pm if bedtime is 22:00–23:00.

Protein powder (conditional Tier 2)¶

Protein powder is food, not a supplement. It belongs in Tier 2 only when whole-food intake can't hit the training target (1.6–2.2 g/kg/day for strength).

Form	Profile	Best use
Whey concentrate	Complete, fast-absorbing, 70–80% protein	Post-workout; budget option
Whey isolate	Very low lactose, faster still	Lactose sensitivity
Casein	Slow-release (6–8h)	Pre-sleep muscle protein synthesis
Pea isolate	DIAAS 0.73; incomplete alone	Vegan; combine with rice protein for leucine adequacy
Soy isolate	DIAAS 0.91; complete profile	Vegan, full amino acid coverage

The "anabolic window" is largely myth — total daily protein is the lever. Timing within 2 hours of training is fine, not mandatory.

3. Tier 3 — conditional (only with tested need)¶

Real mechanisms; enough individual variation that blind supplementation risks overshoot or wasted spend. Test before treating.

Supplement	Test first	If deficient
Iron	Serum ferritin + hemoglobin	18–25 mg/day bis-glycinate form (gentlest GI). Vitamin C 60 mg at same meal doubles non-heme absorption. Do not supplement without confirmed deficiency — iron excess accelerates oxidative stress.
Zinc	Serum zinc (morning, fasted)	15–25 mg/day picolinate or gluconate. UL is 40 mg/day; zinc competes with copper — add 1–2 mg copper alongside if >20 mg zinc.
B9 (folate)	Homocysteine or MTHFR panel	400 µg/day dietary folate equivalent. MTHFR C677T carriers: prefer 5-MTHF (methylfolate) over folic acid.
CoQ10 (ubiquinol)	Statin use; fatigue out of proportion to sleep/training	100–200 mg ubiquinol/day. Statins deplete endogenous CoQ10; mechanism is sound; trial evidence in non-statin users is mixed.
Melatonin	Jet lag; circadian shift; shift work	0.5–1 mg, 30–60 min before target bedtime. Most OTC doses (5–10 mg) are 5–20× above effective. Lower is the correct dose.

4. Tier 4 — theater¶

Widely sold, widely taken, effect sizes undetectable or inconsistently replicated in healthy well-fed adults.

Compound	What's claimed	What evidence actually shows
Ashwagandha (KSM-66)	Cortisol reduction, stress	3–5 small trials: ~10–15% cortisol reduction in high-stress populations. No clear effect in well-rested healthy adults. Hepatotoxicity cases at high dose, prolonged use.
NMN / NR	Longevity, NAD+ repletion	Blood NAD+ rises — no demonstrated downstream benefit in healthy humans. Animal data only for longevity. Multi-year RCTs underway as of 2025.
BCAA	Muscle protein synthesis	Redundant if total protein >1.6 g/kg/day. MPS requires all nine essential amino acids; BCAAs alone are incomplete. Whey is better and cheaper.
Resveratrol	Cardiovascular, longevity (sirtuin)	Oral bioavailability <1% due to rapid metabolism. Sirtuin findings have not translated to human trial benefit.
Collagen peptides	Skin, joints, tendons	Glycine + proline load is the mechanism — adequate protein from food supplies this. Some marginal joint-pain benefit; placebo-robust. Not a protein replacement (incomplete profile).
Detox / cleanse	Liver / kidney clearing	The liver and kidneys already do this. No clinical signal in any controlled cleanse trial.
Pre-workout stacks	Pump, energy, performance	Performance comes from caffeine (Tier 2) and creatine (Tier 2). Beta-alanine tingling, sub-clinical citrulline, proprietary blends — packaging.
Testosterone boosters	Natural T increase	Effect sizes negligible vs. sleep, training load, and dietary fat — the real T stack.
Multivitamin	Micronutrient insurance	Better than individual Tier 4 singletons; covers few Tier 1 gaps adequately (D usually 400 IU — under-dosed; omega-3 absent; magnesium as oxide — poorly absorbed).

5. Decision tree for a well-fed active adult¶

1. Protein ≥ 1.6 g/kg from food? No → fix food first.
2. Five Tier 1 gaps covered? No → start there, nothing else.
3. Train for strength or power? Yes → add creatine 3–5g/day.
4. Use caffeine? → time it strategically; cycle; protect sleep.
5. Blood panel: ferritin, 25(OH)D, zinc, TSH, B12, folate → treat deficiencies.
6. Still want more? → Read original trials, not company summaries.

6. Minimal personal stack (active, daily gym, brain-priority, lean)¶

For a 68 kg / 1.78 m adult, daily training, 15k steps, 8h sleep, no alcohol:

Time	Supplement	Dose	Why
Morning with breakfast	D3 + K2	D3 1500 IU + K2 MK-7 100 µg	Structural; fat-soluble — take with food
Morning with breakfast	Omega-3	2 g EPA+DHA	Structural; brain, heart, anti-inflammatory
Anytime (water)	Creatine monohydrate	5 g	Strength output; cognitive resilience under fatigue
Evening meal or before bed	Magnesium glycinate	400 mg	Structural; sleep quality; muscle recovery
If vegan or 50+	B12 sublingual	500 µg	Structural; not needed if regularly eating meat/dairy

Monthly cost at reasonable brands: ~€15–25. Outperforms any €100/month stack for a well-nourished, active, non-deficient adult.

7. Timing, cycling, interactions¶

Pair / rule	Why
D3 + K2 together	D3 raises calcium absorption; K2 routes it to bone, not arteries. Separate intake is suboptimal.
Magnesium at night	Sleep improvement signal; also GI-gentler than morning on empty stomach.
Iron + vitamin C; away from tea/coffee	C triples non-heme iron absorption. Tannins in tea/coffee block it. Separate by 2h.
High-dose zinc + copper	>20 mg zinc long-term depletes copper. Add 1–2 mg copper alongside.
Caffeine cut-off	No caffeine after 1 pm if bedtime is 22:00–23:00 (half-life 5–6h).
Creatine + hydration	Creatine draws water into muscle. Hydration becomes more important, not less.
Fat-soluble vitamins (A, D, E, K)	Take with a fat-containing meal; absorption is markedly better.

8. Safety and upper limits¶

Supplement	UL	Above UL
Vitamin D	4000 IU/day (likely conservative)	>10,000 IU chronic: hypercalcemia, kidney damage
Vitamin A (retinol)	3000 µg RAE/day	Hepatotoxicity; teratogenicity in pregnancy
Iron	45 mg/day (non-therapeutic)	GI damage, organ damage; dangerous in haemochromatosis
Zinc	40 mg/day	Copper depletion, nausea, immune suppression
Selenium	400 µg/day	Hair loss, nail brittleness, neuropathy
Iodine	1100 µg/day	Thyroiditis, hypothyroidism in susceptible individuals
Vitamin B6	100 mg/day	Peripheral neuropathy (overdose mimics deficiency)
Magnesium (oral)	No hard UL	Osmotic diarrhea; kidneys excrete excess; not systemic
Caffeine	~400 mg/day (healthy adults)	Arrhythmia, anxiety, sleep fragmentation, dependence

9. How to read a supplement study¶

The supplement literature is structurally noisy: - Industry funding is the norm; industry-funded effect sizes are systematically inflated. - Publication bias — negative trials are underreported. - Short durations (4–8 weeks) miss chronic effects. - Surrogate endpoints (blood markers) are not outcomes (function at 80, all-cause mortality). - Most subjects are young males in university towns — not a representative target population.

Five questions for any supplement claim: 1. Who funded it? 2. Is the endpoint a surrogate or an outcome? 3. Is the study population comparable to you? 4. Is there a plausible, independently testable mechanism? 5. Is it replicated in ≥3 independent trials?

One study is a hypothesis. Three independent replications is a weak signal. Ten RCTs with consistent effect sizes is the entry point for strong evidence — creatine clears this bar; most supplements don't.

Open questions¶

Optimal 25(OH)D target: 40–60 vs. 60–80 ng/mL? Trials above the deficiency floor (20 ng/mL) are mixed on outcomes. Upper range may confer additional immune and cardiovascular benefit, or may just raise calcium absorption past the knee.
Magnesium threonate (L-threonate) for cognition: BBB penetration advantage over glycinate is mechanistically plausible; human cognition trials are small (N<100).
Creatine in sleep deprivation: the cognitive preservation signal is promising but underpowered. A well-powered trial would resolve this for variable-sleep use cases.
NMN / NR long-term: 5-year human RCTs reportedly underway as of 2025. If outcomes (not just blood NAD+) are positive, would move from Tier 4 to Tier 3/2.
Ashwagandha hepatotoxicity threshold: case reports exist; dose-duration threshold unknown. Ongoing pharmacovigilance.

References¶

Holick, M. F. (2011). Vitamin D deficiency. New England Journal of Medicine.
Institute of Medicine. (2010). Dietary Reference Intakes for calcium and vitamin D.
Rawson, E. S., & Venezia, A. C. (2011). Use of creatine in the elderly. Current Aging Science.
Lanhers, C., et al. (2017). Creatine supplementation and lower-limb strength: systematic review and meta-analysis. Sports Medicine.
Grgic, J., et al. (2020). Wake up and smell the coffee: caffeine supplementation and exercise performance. British Journal of Sports Medicine.
Walker, M. (2017). Why We Sleep — caffeine half-life and sleep architecture.
L-1917: Six structural micronutrient gaps are systemic deficits (swarm, health, Sharpe 7).
Examine.com — independent supplement database; evidence tiering.
ConsumerLab.com — third-party purity testing.

Inspiration sources¶

L-1917 (S573): structural deficit framing — gaps are systemic, not personal.
HEALTH-AS-INFRASTRUCTURE: "most supplement effort is theatre" as the entry signal.
FOOD-AS-FUEL §3: six predictable micronutrient gaps in well-fed populations.