Place · Level 3
Fish Oil · EPA / DHA Supplement
全球销售第一的补剂之一 · 机制清楚 · 但 RCT 证据近 5 年大翻转 · 食物胜过药丸的经典案例
Story path
Chapter 1
Three forms + freshness
Three forms + freshness
'Fish oil' isn't one thing — the forms on the market differ enormously, and the form directly drives absorption and safety.
① Natural triglyceride (TG / rTG, re-esterified) — the original form inside the fish
Highest absorption (~70%)Mid-to-high pricerTG (re-esterified) — purified through the EE step then re-attached to a TG backbone, giving high purity + high absorption
② Ethyl ester (EE) — the common output of industrial distillation
Glycerol on the TG is swapped for ethanolAbsorption 30-50% lower than TG (Dyerberg 2010)Works poorly on an empty stomach; needs a fatty mealMost cheap fish oils and REDUCE-IT's icosapent ethyl (Vascepa) are EE
③ Phospholipid-bound — krill oil
EPA / DHA covalently attached to phospholipids (PC)Theoretically the best absorption, but each capsule carries little EPA + DHAMost expensive; astaxanthin antioxidant is a byproduct
The point: TOTOX (oxidation level)
TOTOX = 2 × peroxide value (PV) + anisidine value (AV)TOTOX < 26 is the international standard (GOED + EU)Real-world testing: Albert 2015 NZ study showed ~83% of products on shelves failedOxidized fish oil isn't just ineffective — it's actively pro-oxidant, and produces nausea, fishy burps, and elevated liver enzymes
How to choose:
TOTOX < 10 is excellentLook for IFOS / GOED certificationRefrigerate after opening; use within 2-3 monthsStrong fishy taste or nausea → throw it out immediatelyForm preference: rTG > TG > EE
① Natural triglyceride (TG / rTG, re-esterified) — the original form inside the fish
Highest absorption (~70%)Mid-to-high pricerTG (re-esterified) — purified through the EE step then re-attached to a TG backbone, giving high purity + high absorption
② Ethyl ester (EE) — the common output of industrial distillation
Glycerol on the TG is swapped for ethanolAbsorption 30-50% lower than TG (Dyerberg 2010)Works poorly on an empty stomach; needs a fatty mealMost cheap fish oils and REDUCE-IT's icosapent ethyl (Vascepa) are EE
③ Phospholipid-bound — krill oil
EPA / DHA covalently attached to phospholipids (PC)Theoretically the best absorption, but each capsule carries little EPA + DHAMost expensive; astaxanthin antioxidant is a byproduct
The point: TOTOX (oxidation level)
TOTOX = 2 × peroxide value (PV) + anisidine value (AV)TOTOX < 26 is the international standard (GOED + EU)Real-world testing: Albert 2015 NZ study showed ~83% of products on shelves failedOxidized fish oil isn't just ineffective — it's actively pro-oxidant, and produces nausea, fishy burps, and elevated liver enzymes
How to choose:
TOTOX < 10 is excellentLook for IFOS / GOED certificationRefrigerate after opening; use within 2-3 monthsStrong fishy taste or nausea → throw it out immediatelyForm preference: rTG > TG > EE
Reading the label · actual dose
The most common label trap on a fish-oil bottle is the '1000 mg fish oil' claim:1000 mg of fish oil ≠ 1000 mg of EPA + DHAA typical '1000 mg fish oil' cap contains only ~180 mg EPA + 120 mg DHA = 300 mg ω-3To reach the AHA's 'EPA + DHA 1 g/day' target you need 3-4 caps'Concentrated high-potency' caps (60-80% omega-3) deliver 700-800 mg ω-3 per cap
Target dose reference (AHA / GOED):
General prevention (≥ 2 fatty-fish meals/week): no supplement neededCV secondary prevention / high TG: EPA + DHA 1-2 g/dayREDUCE-IT therapeutic (high TG + atherosclerotic cardiovascular disease: The plaque-clogged-artery family of disease — heart attack, stroke, peripheral artery disease.): icosapent ethyl 4 g/day (EPA-only)Pregnancy DHA: ≥ 200 mg/day (fetal brain development)UL: FDA sets the daily omega-3 ceiling at 3 g (from supplements); large doses extend bleeding time
Fish oil and anticoagulation boundaries:
Healthy adults at 1-2 g/day: no effect on INR / bleedingStopping 7 days before surgery is a conservative recommendationOn warfarin / antiplatelets: tell your physician; usually a minor issueReal bleeding impact appears only at chronic >4 g/day
Chapter 2
EPA vs DHA · different jobs
EPA vs DHA · different jobs
EPA and DHA differ by only two carbons and one double bond, but they do completely different jobs in the body.
EPA (20:5, eicosapentaenoic acid) — anti-inflammatory messenger:
Precursor → resolvins + protectins (endogenous anti-inflammatory mediators)Competes with AA (arachidonic acid, ω-6) for COX/LOX enzymes, lowering pro-inflammatory prostaglandinsHigh-concentration EPA RCTs show reductions in triglycerides + cardiovascular eventsREDUCE-IT (NEJM 2018) used pure EPA (icosapent ethyl) at 4 g/day → MACE ↓ 25%
DHA (22:6, docosahexaenoic acid) — membrane material:
Retina ~50% DHA (rod outer segment)Brain gray matter 15-20% DHACritical lipid for membrane fluidity and receptor functionDrives fetal brain development + infant visual maturation (see `fats-omega-3/membrane` Level-4 animation)
Can the two interconvert?
EPA → DHA: partially possible (full chain ALA → EPA → DPA → DHA)DHA → EPA: partially possible (retro-conversion)Both efficiencies are low — if you want one specifically, supplement that form directly
So:
Anti-inflammatory / TG-lowering / CV secondary prevention → prefer high-EPA formulasPregnancy / infants / brain-health focus → DHA can't be skippedBalanced supplements (~EPA:DHA = 1.5:1 to 2:1) suit general use'Pure EPA' formulas belong to REDUCE-IT-style indications, not everyday wellness
EPA (20:5, eicosapentaenoic acid) — anti-inflammatory messenger:
Precursor → resolvins + protectins (endogenous anti-inflammatory mediators)Competes with AA (arachidonic acid, ω-6) for COX/LOX enzymes, lowering pro-inflammatory prostaglandinsHigh-concentration EPA RCTs show reductions in triglycerides + cardiovascular eventsREDUCE-IT (NEJM 2018) used pure EPA (icosapent ethyl) at 4 g/day → MACE ↓ 25%
DHA (22:6, docosahexaenoic acid) — membrane material:
Retina ~50% DHA (rod outer segment)Brain gray matter 15-20% DHACritical lipid for membrane fluidity and receptor functionDrives fetal brain development + infant visual maturation (see `fats-omega-3/membrane` Level-4 animation)
Can the two interconvert?
EPA → DHA: partially possible (full chain ALA → EPA → DPA → DHA)DHA → EPA: partially possible (retro-conversion)Both efficiencies are low — if you want one specifically, supplement that form directly
So:
Anti-inflammatory / TG-lowering / CV secondary prevention → prefer high-EPA formulasPregnancy / infants / brain-health focus → DHA can't be skippedBalanced supplements (~EPA:DHA = 1.5:1 to 2:1) suit general use'Pure EPA' formulas belong to REDUCE-IT-style indications, not everyday wellness
RCT pivot · weak signal in general population
The fish-oil RCT story is one of the most dramatic pivots in modern nutrition medicine.2002-2010 (the optimistic era): GISSI-Prevenzione, JELIS, and other early RCTs showed reduced recurrence in post-MI patients taking fish oil.
2018-2020 (major recalibration): large, rigorous RCTs overturned the general-population benefit.
VITAL trial (NEJM 2019, N = 25,871, 5.3 yr):
Healthy adults on EPA + DHA 1 g/day vs placeboPrimary CV / cancer endpoints: no significant benefitSubgroups: MI events ↓ 28% (P=0.13, borderline)Black subgroup MI ↓ 77% (striking but needs replication)
ASCEND trial (NEJM 2018, N = 15,480, 7.4 yr):
Diabetic adults with no CV history on ω-3 1 g/daySerious vascular events: no significant benefit (RR 0.97, P=0.55)
STRENGTH trial (JAMA 2020, N = 13,078):
High-CV-risk patients on carboxylic-acid-form EPA + DHA 4 g/dayStopped early: no benefit, plus atrial fibrillation up 70%Stark contrast with REDUCE-IT's pure-EPA 4 g
REDUCE-IT vs STRENGTH controversy:
Same dose (4 g), similar high-CV-risk populationREDUCE-IT positive (pure EPA), STRENGTH negative (EPA + DHA)Three explanations: EPA-without-DHA is the key driver (mechanism hypothesis); REDUCE-IT's mineral-oil placebo was harmful (FDA investigated, partially supported); the truth lies betweenThis is why icosapent ethyl was FDA-approved while other fish-oil formulations were not
Current consensus:
Healthy population using fish oil for CV prevention: weak evidenceHigh TG (> 500 mg/dL): first-line indicationVery-high CV risk + high TG: icosapent ethyl (Vascepa) by prescription≥ 2 fatty-fish meals/week (salmon / sardines / mackerel): stronger evidence than any supplement (DGAC + AHA)
Chapter 3
Food vs pill
Food vs pill
'Eat fish or take fish oil' is a question with a data-driven answer.
Typical fatty-fish ω-3 content (per 100 g cooked):
Salmon (Atlantic farmed): 1.8-2.5 g EPA + DHASardines (oil-packed): 1.4-1.5 gMackerel: 1.0-2.7 g (large seasonal variation)Cod: 0.2-0.3 g (lean fish, low)Tuna (deep-sea canned): 0.2-0.5 g (water-packed loses some)Trout: 0.8-1.5 g
For comparison: one 'standard' fish-oil cap is ~0.3 g ω-3, so two 100 g salmon meals/week ≈ 20 caps/week = 3 caps/day.
But 'eating fish' is more than ω-3:
High-quality protein (~20-25 g / 100 g)Vitamin D (fatty fish is one of the few natural D-rich sources)Iodine (marine fish)Selenium (marine fish + shellfish)Vitamin B12Satiety, substituting for red and processed meat
Multiple epidemiological and RCT meta-analyses show that consuming fatty fish vs taking ω-3 capsules gives stronger cardiovascular protection. This is the food matrix effect: many bioactives in a whole food acting in synergy, not a single isolated compound.
Practical:
2-3 fatty-fish meals/week (≥ 150 g each): most people get adequate ω-3 from food, no supplement neededDon't eat fish / vegan: consider algal oil for DHA, 200-500 mg/dayPregnancy: pick low-mercury species (salmon / sardine / trout); limit large predatory fish (shark / swordfish / bigeye tuna)Liver disease or on anticoagulants: discuss with a clinician
Typical fatty-fish ω-3 content (per 100 g cooked):
Salmon (Atlantic farmed): 1.8-2.5 g EPA + DHASardines (oil-packed): 1.4-1.5 gMackerel: 1.0-2.7 g (large seasonal variation)Cod: 0.2-0.3 g (lean fish, low)Tuna (deep-sea canned): 0.2-0.5 g (water-packed loses some)Trout: 0.8-1.5 g
For comparison: one 'standard' fish-oil cap is ~0.3 g ω-3, so two 100 g salmon meals/week ≈ 20 caps/week = 3 caps/day.
But 'eating fish' is more than ω-3:
High-quality protein (~20-25 g / 100 g)Vitamin D (fatty fish is one of the few natural D-rich sources)Iodine (marine fish)Selenium (marine fish + shellfish)Vitamin B12Satiety, substituting for red and processed meat
Multiple epidemiological and RCT meta-analyses show that consuming fatty fish vs taking ω-3 capsules gives stronger cardiovascular protection. This is the food matrix effect: many bioactives in a whole food acting in synergy, not a single isolated compound.
Practical:
2-3 fatty-fish meals/week (≥ 150 g each): most people get adequate ω-3 from food, no supplement neededDon't eat fish / vegan: consider algal oil for DHA, 200-500 mg/dayPregnancy: pick low-mercury species (salmon / sardine / trout); limit large predatory fish (shark / swordfish / bigeye tuna)Liver disease or on anticoagulants: discuss with a clinician
Mercury & microplastics
Real risk considerations when eating fish:① Methylmercury
Biomagnifies — long-lived, large predatory fish carry the highest mercuryHigh risk: shark, swordfish, marlin, bigeye tuna, tilefish, Pacific king mackerelLow risk: salmon, sardines, mackerel (Atlantic, NOT king mackerel), trout, shrimp, scallops, tilapiaEPA-DHA vs mercury net: NEJM 2002 + Mozaffarian 2006 analyses show benefit > risk for most adults (pregnant women, children under 5, and women of childbearing age need particular caution)
② POPs (persistent organic pollutants — PCB / dioxins)
Concentrate in farmed fish (feed-driven accumulation)Modern certified fish oil is distilled and largely free of PCB / mercury'Purity' labels: IFOS 5-star certification is meaningful
③ Microplastics + PFAS
Microplastics and perfluoroalkyl substances have been detected in most fish (and in water + salt — not exclusive to fish)Health impact is still under study, but not a reason to stop eating fish
④ Fish-oil oxidation (covered in the earlier scene): a far more common clinical 'supplement quality' issue than mercury or PCB.
Pregnancy / lactation recommendations (FDA + EPA 2017):
2-3 servings (~ 8-12 oz / 230-340 g) of low-mercury fish per weekAvoid: shark, swordfish, marlin, tilefish, bigeye tuna, Pacific king mackerelLimit: yellowfin tuna (1 serving/week), canned white tuna'Eat no fish at all' is actively NOT recommended by FDA + EPA — DHA is too critical for the fetus
Summary: 2-3 low-mercury fatty-fish meals per week + no supplement is the strongest current evidence-based strategy for the average adult, and this clarity has only sharpened after the recent wave of RCT reversals.
Chapter 4
Plant ω-3 (ALA) · incomplete sub
Plant ω-3 (ALA) · incomplete sub
Vegetarian ω-3 substitution is incomplete — the most commonly oversimplified topic in nutrition science.
ALA (α-linolenic acid, 18:3 ω-3) sources:
Flaxseed oil: 53-57% ALAChia seeds: 18% ALAWalnuts: 9% ALAHemp seeds: 22% ALASoybean / canola oil: 7-10% ALA
Key point: endogenous ALA → EPA → DHA conversion is very low
ALA → EPA: 5-8% (women slightly higher, estrogen-related)EPA → DHA: 0.5-5%Total ALA → DHA: 0.1-4%
This means a fully fish-free diet relying only on flaxseed oil produces noticeably low blood EPA / DHA — but it doesn't mean 'fully deficient'; ALA itself is an essential fatty acid. Long-term vegetarians' blood EPA / DHA runs 20-50% lower than fish-eaters', but CV event rates don't track that proportionally, possibly because the overall diet pattern is healthier.
Algal oil is the true vegan source of DHA / EPA:
DHA / EPA extracted from microalgae (the original ω-3 source that fish eat) is biochemically equivalent to fish oilTypical product: 200-500 mg DHA per capsuleCost is higher (about 3-5× fish oil)Strongly recommended for vegan women in pregnancy and lactation
ω-6 / ω-3 ratio:
Modern Western diet 15-20 : 1Estimated evolutionary ratio 1-4 : 1'High ω-6 is the source of inflammation' is a popular claim but the actual evidence is weak: RCTs replacing saturated fat with ω-6 (linoleic acid) consistently show lower LDL and lower CV eventsThe point is not 'lower ω-6' — it's 'absolutely increase ω-3'Don't agonize over ω-6; focus on total ω-3 intake instead
ALA (α-linolenic acid, 18:3 ω-3) sources:
Flaxseed oil: 53-57% ALAChia seeds: 18% ALAWalnuts: 9% ALAHemp seeds: 22% ALASoybean / canola oil: 7-10% ALA
Key point: endogenous ALA → EPA → DHA conversion is very low
ALA → EPA: 5-8% (women slightly higher, estrogen-related)EPA → DHA: 0.5-5%Total ALA → DHA: 0.1-4%
This means a fully fish-free diet relying only on flaxseed oil produces noticeably low blood EPA / DHA — but it doesn't mean 'fully deficient'; ALA itself is an essential fatty acid. Long-term vegetarians' blood EPA / DHA runs 20-50% lower than fish-eaters', but CV event rates don't track that proportionally, possibly because the overall diet pattern is healthier.
Algal oil is the true vegan source of DHA / EPA:
DHA / EPA extracted from microalgae (the original ω-3 source that fish eat) is biochemically equivalent to fish oilTypical product: 200-500 mg DHA per capsuleCost is higher (about 3-5× fish oil)Strongly recommended for vegan women in pregnancy and lactation
ω-6 / ω-3 ratio:
Modern Western diet 15-20 : 1Estimated evolutionary ratio 1-4 : 1'High ω-6 is the source of inflammation' is a popular claim but the actual evidence is weak: RCTs replacing saturated fat with ω-6 (linoleic acid) consistently show lower LDL and lower CV eventsThe point is not 'lower ω-6' — it's 'absolutely increase ω-3'Don't agonize over ω-6; focus on total ω-3 intake instead
What ALA itself does
ALA itself (even without conversion to EPA / DHA) is an essential fatty acid — the body can't synthesize it. Its roles:1. Energy substrate: like other fatty acids, β-oxidized for energy
2. A small fraction joins membrane phospholipids: contributing to fluidity
3. Cardiovascular evidence (independent of EPA/DHA): the Sacks AHA 2017 review shows a positive (though weak) correlation between high ALA intake and lower CV events; PREDIMED-Plus + several cohorts show 1-2 g/day ALA correlates with ~10% lower all-cause mortality; mechanism partially via anti-arrhythmic effects
4. Modest inflammation marker improvement (limited evidence): high-ALA diet produces small drops in C-reactive protein: A liver protein that rises with inflammation — a common blood marker for 'is the body inflamed'. / interleukin-6: A pro-inflammatory signal molecule (cytokine) released by immune cells during inflammation.
Practical recommendations (vegetarian or no fish):
1-2 tablespoons of ground flaxseed daily (ground, not whole — whole seeds pass through largely intact) ≈ 2-3 g ALA1 handful of walnuts (28 g) ≈ 2.5 g ALA1-2 tablespoons of chia seeds ≈ 5 g ALAAdd algal DHA 200-500 mg/day (especially in pregnancy, lactation, and children under 5)Don't cook with flaxseed oil — PUFAs oxidize rapidly under heat
Common wrong claims to flag:
'Flaxseed oil equals fish oil': wrong — conversion is insufficient'Pure vegan needs no supplements': B12, DHA, and often iron / zinc all deserve attention'ω-3:ω-6 must be tuned to 1:1 for health': no evidence supports this extreme
Chapter 5
Other indications · tiered evidence
Other indications · tiered evidence
Beyond cardiovascular disease, fish oil / EPA + DHA has other evidence-based indications, with widely varying evidence strength.
Strong evidence (A grade):
High triglycerides (TG > 500 mg/dL): EPA + DHA 2-4 g/day lowers TG 20-50%Icosapent ethyl (Vascepa) 4 g + statin in high TG + atherosclerotic cardiovascular disease: The plaque-clogged-artery family of disease — heart attack, stroke, peripheral artery disease. patients: MACE ↓ 25% (REDUCE-IT)Pregnancy DHA ≥ 200 mg/day: fetal brain + visual development, ACOG recommended
Moderate evidence (B grade):
Rheumatoid arthritis (RA): EPA + DHA 3-4 g/day reduces morning stiffness and NSAID use (Goldberg 2007 meta)Major depression as adjunct: Liao 2019 meta shows EPA-dominant formulas (> 60% EPA) at ≥ 1 g produce improvement; DHA alone doesn't workDry eye: DREAM trial 2018 showed no significant benefit (earlier small studies had been optimistic)
Disputed / weak evidence:
Cognitive function / Alzheimer's prevention: most RCTs negative (Yurko-Mauro 2010 DHA-alone showed some, but ADCS-MeM 2010 + VITAL 2019 negative)Pediatric ADHD: Bloch 2011 meta shows a small effect (SMD ~0.31), not first-lineMacular degeneration: AREDS2 showed no additional benefit from added ω-3Exercise recovery: limited evidence for reducing muscle sorenessAllergy / asthma: early maternal-infant supplementation may slightly reduce offspring allergy risk
No / negative evidence:
Cancer prevention: VITAL negativeDiabetes prevention: ASCEND negative'Make you smarter' / 'anti-aging': marketing claims, no RCT support
So the real role of fish oil:
Adjunct for specific patients (high TG / RA / specific MDD subtypes)ω-3 source for those who don't eat fishPregnancy DHA supplementNot a universal 'wellness for healthy people' supplement
Strong evidence (A grade):
High triglycerides (TG > 500 mg/dL): EPA + DHA 2-4 g/day lowers TG 20-50%Icosapent ethyl (Vascepa) 4 g + statin in high TG + atherosclerotic cardiovascular disease: The plaque-clogged-artery family of disease — heart attack, stroke, peripheral artery disease. patients: MACE ↓ 25% (REDUCE-IT)Pregnancy DHA ≥ 200 mg/day: fetal brain + visual development, ACOG recommended
Moderate evidence (B grade):
Rheumatoid arthritis (RA): EPA + DHA 3-4 g/day reduces morning stiffness and NSAID use (Goldberg 2007 meta)Major depression as adjunct: Liao 2019 meta shows EPA-dominant formulas (> 60% EPA) at ≥ 1 g produce improvement; DHA alone doesn't workDry eye: DREAM trial 2018 showed no significant benefit (earlier small studies had been optimistic)
Disputed / weak evidence:
Cognitive function / Alzheimer's prevention: most RCTs negative (Yurko-Mauro 2010 DHA-alone showed some, but ADCS-MeM 2010 + VITAL 2019 negative)Pediatric ADHD: Bloch 2011 meta shows a small effect (SMD ~0.31), not first-lineMacular degeneration: AREDS2 showed no additional benefit from added ω-3Exercise recovery: limited evidence for reducing muscle sorenessAllergy / asthma: early maternal-infant supplementation may slightly reduce offspring allergy risk
No / negative evidence:
Cancer prevention: VITAL negativeDiabetes prevention: ASCEND negative'Make you smarter' / 'anti-aging': marketing claims, no RCT support
So the real role of fish oil:
Adjunct for specific patients (high TG / RA / specific MDD subtypes)ω-3 source for those who don't eat fishPregnancy DHA supplementNot a universal 'wellness for healthy people' supplement
4 questions before buying
If you've decided to supplement fish oil, these 4 questions avoid ~80% of bad products.1. What specific problem are you trying to solve? (indication)
'Cardiovascular wellness' without a specific diagnosis: eat fish first, not supplements; if you really want to, EPA + DHA 1 g/day is enoughHigh TG > 500: 2-4 g/dayPregnancy: DHA-dominant 200-500 mg/dayRA / arthritis: 3-4 g/day EPA + DHADepression adjunct: EPA-dominant formula (> 60% EPA), > 1 g/dayCan't answer? Don't buy
2. How much actual EPA + DHA is in this bottle?
Read the nutrition panel, not the headline on the front'1000 mg fish oil' ≠ '1000 mg ω-3' (see scene 1)Target: at least 500 mg EPA + DHA per capsule for a 'concentrated' product
3. Form (TG / rTG / EE)?
rTG or TG outperforms EE (30-50% absorption difference)EE form must be taken with a fatty mealVascepa (icosapent ethyl) is EE but pure EPA at 4 g — prescription use, not wellness
4. Oxidation (TOTOX) + certification?
IFOS 5-star / GOED certificationTOTOX < 10 is excellentFishy taste / nausea / burps: already oxidized — return itRefrigeration + brown glass bottle > clear plasticUse within 2-3 months after opening
3 red flags — put it down immediately:
'Deep-sea fish oil' with no EPA / DHA milligram breakdown on the label: deceptive'X dollars for 100 caps' suspiciously cheap: usually EE form with high oxidation'Beauty / anti-aging / makes-you-smarter' marketing: outside the evidence base