Omega-3 as EPA and DHA
The strongest single lifestyle lever on triglycerides in our catalogue and at the same time the textbook example of an offset: the same amount that lowers triglycerides often raises LDL with DHA-containing preparations.
Dose in the studies
In the underlying studies: The cited studies used two to four grams of EPA and DHA combined per day. What counts is the amount of EPA and DHA, not the amount of fish oil or the number of capsules.. That is the reported study figure, not a recommendation to take it.
A supplement is not a lifestyle recommendation. Speak with your doctor.
Plant-based alpha-linolenic acid from linseed or rapeseed oil is not the same thing: conversion to EPA and DHA is limited, and this batch has no defensible percentages. There is no effect row on uric acid or ferritin for the supplement. Fish as food is named as purine-rich; milligrams are in the food search. Meta-analyses describe excess atrial fibrillation from about one gram per day. Time to the full triglyceride effect is not stated in the opened abstracts.
Effects in detail
Each row is a directed edge from our database, with range, latency, evidence grade and the condition under which it applies at all.
| Marker | Effect | Latency | Mode of action | Outcome | Evidence | Condition | Sources |
|---|---|---|---|---|---|---|---|
| Apolipoprotein B | helps. Lowers Direction known, size not established Meta-analyses consistently show a small LDL-C rise with EPA+DHA (Eslick +0.06 mmol/L; AbuMweis +0.15 mmol/L) that is attributable to DHA (Wei & Jacobson); the AHA advisory confirms the rise at 4 g/day in very high TG and its absence with EPA-only products. This is the classic omega-3 'offset': TG down, LDL-C slightly up. Trial proxy: ldl_c. Pulling the other way, the same AHA advisory states that in the largest 4 g/day trials non-HDL-C and ApoB were modestly decreased, so total atherogenic particle number fell — with no pooled figure in the abstract. Both readings stand; we carry the quantified direction and record the counter-reading here rather than showing two contradictory edges. | not specified | chronic | unclear | Evidence grade A | applies regardless of the starting value | |
| Diastolic blood pressure | helps. Lowers -0.99 mmHg confidence interval of the estimate, not the range one person can expect Seventy randomised trials give a pooled systolic reduction of 1.52 mmHg (95 % CI -2.25 to -0.79) and diastolic of 0.99 mmHg with EPA plus DHA supplementation. This adds a second, modest benefit to the batch-1 omega-3 lever, which was built on triglycerides. | not specified | chronic | unclear | Evidence grade A | applies regardless of the starting value | |
| Systolic blood pressure | helps. Lowers -1.52 mmHg confidence interval of the estimate, not the range one person can expect Seventy randomised trials give a pooled systolic reduction of 1.52 mmHg (95 % CI -2.25 to -0.79) and diastolic of 0.99 mmHg with EPA plus DHA supplementation. This adds a second, modest benefit to the batch-1 omega-3 lever, which was built on triglycerides. | not specified | chronic | unclear | Evidence grade A | applies regardless of the starting value | |
| Triglycerides | helps. Lowers Direction known, size not established the absolute bounds are two pooled estimates, -0.34 and -0.37 mmol/L, from two meta-analyses at a mean 3.25 g/day; neither is an interval. The percentage pair is a different thing again: 15 % is derived at a baseline of 200 mg/dL and 30 % is the advisory's figure at 4 g/day in very high triglycerides The effect grows with baseline and dose. About −30 mg/dL at a mean 3.25 g/day; at least 30 percent at 4 g/day in very high triglycerides. REDUCE-IT (pure EPA) cut events, STRENGTH (EPA+DHA) did not. | not specified | postprandial and chronic | unclear | Evidence grade A | TG at or above 150 mg/dL |
What this measure costs
Counteracting effects of the same measure.
| Marker | Effect | Latency | Mode of action | Outcome | Evidence | Condition | Sources |
|---|---|---|---|---|---|---|---|
| LDL cholesterol | costs. Raises Direction known, size not established the low bound is mixed EPA+DHA supplements, the high bound is DHA monotherapy | not specified | chronic | unclear | Evidence grade A | applies regardless of the starting value | No source on file |
Moves without implying progress
Informational or no established effect: the marker moves or was checked, and that is not progress.
| Marker | Effect | Latency | Mode of action | Outcome | Evidence | Condition | Sources |
|---|---|---|---|---|---|---|---|
| HDL cholesterol | moves, and does not count. Raises Direction known, size not established the low bound is mixed EPA+DHA supplements, the high bound is DHA monotherapy Small and driven by DHA; EPA barely changes HDL. Not an outcome-relevant effect, and too small to “fix” low HDL. | not specified | chronic | no | Evidence grade A | applies regardless of the starting value | |
| HbA1c | checked, nothing happens. No established effect Across 20 randomised trials in type 2 diabetes, omega-3 supplementation produced no marked change in HbA1c, fasting glucose or postprandial glucose, while lowering triglycerides by 0.24 mmol/L. A deliberate no-effect row. | not specified | chronic | unclear | Evidence grade A | applies regardless of the starting value | |
| high-sensitivity CRP | moves, and does not count. Lowers Direction known, size not established A dose-response meta-analysis of 40 randomised trials found a significant non-linear reduction in CRP up to 1200 mg/day of EPA and DHA in cardiometabolic disease. No pooled mg/L figure is given in the abstract, so no magnitude is entered. | not specified | chronic | unclear | Evidence grade B | applies regardless of the starting value | |
| HOMA-IR | checked, nothing happens. No established effect confidence interval of the estimate, not the range one person can expect The largest systematic review (83 RCTs, high AMSTAR-2 quality) found no effect of long-chain omega-3 on HOMA-IR (+0.06, 95 % CI -0.21 to 0.33), fasting insulin or HbA1c and a trivial fasting glucose rise (+0.04 mmol/L); an earlier meta-analysis of 11 RCTs was also null overall. Included as a 'no effect' row: omega-3 is not an insulin-sensitiser. | 24–24 weeks | chronic | no | Evidence grade A | applies regardless of the starting value |
Food sources
What you would have to eat to reach the stated order of magnitude, and what comes along with it.
| Food | Content per 100 g | Amount per week | Added burden |
|---|---|---|---|
| Algal oil The source without a fish burden: no purines, no mercury, no retinol. No BLS row; EPA and DHA content is product-dependent. 25 g per 100 g remains a typical label figure, not a table value. | 25.0 g | 56 gequals 11 × | no relevant added burden on file |
| Cod liver oil BLS Q730000, cod-liver oil. 17667 µg retinol per 100 g: 5 g carries 883 µg of preformed vitamin A. Unsuitable in pregnancy. No upper-intake-level source was opened in this batch. | 23.0 g | 61 gequals 12 × |
|
| Farmed salmon BLS T410132, poached salmon. In Germany this is typically Atlantic farmed salmon; BLS does not split farmed from wild on this code. Purines 249.9 mg/100 g from the 2024 USDA analysis, Atlantic farmed salmon raw. The FDA mercury row does not separate farmed from wild. | 1.9 g | 725 gequals 4.8 × |
|
| Wild salmon BLS T417152, stewed wild salmon — Pacific salmon, not Atlantic wild. Purines 259.8 mg/100 g from the 2024 USDA analysis, sockeye raw. The FDA mercury row does not separate wild from farmed. | 1.7 g | 828 gequals 5.5 × |
|
| Anchovies BLS T104702, anchovy in oil, canned, drained. 4951 mg sodium per 100 g; a 10 g portion still carries 495 mg. Purines 321 mg/100 g from Clifford 1976 (canned). | 0.9 g | 1.6 kgequals 54 × |
|
| Herring BLS T102132, poached herring. Purines 139.2 mg/100 g from Kaneko 2014 (raw) in the USDA purine database. Mercury from the FDA herring row. | 2.0 g | 700 gequals 5.4 × |
|
| Mackerel BLS T107182, mackerel fried without fat. Purines 204.8 mg/100 g from Kaneko 2014 (raw). Other mackerel species in the same FDA table run higher in mercury (chub 0.088 ppm, Spanish 0.182 ppm). | 1.4 g | 986 gequals 6.6 × |
|
| Sardines in oil BLS T105702, sardine in oil, canned, drained. Purines 399 mg/100 g from Clifford 1976 (canned) in the USDA purine database. | 2.4 g | 574 gequals 6.4 × |
|
| Rainbow trout BLS T420132, poached trout. Purines 181 mg/100 g from Kaneko 2014, rainbow trout raw. | 0.9 g | 1.5 kgequals 10 × |
|
| Canned light tuna BLS T121902, tuna canned in its own juice, drained. Sodium in this BLS row is calculated near zero (no added salt). Little EPA and DHA, so this is lean protein rather than an omega-3 source. Purines 211.8 mg/100 g from the 2024 USDA analysis, drained canned solids. Canned albacore is about three times higher in mercury in the same FDA table. | 0.1 g | 10.8 kgequals 72 × |
|
Sources
Full details of every study the rows above rely on.
Association Between Omega-3 Fatty Acid Intake and Dyslipidemia: A Continuous Dose-Response Meta-Analysis of Randomized Controlled Trials
Wang et al. · Journal of the American Heart Association · 2023 · Meta-analysis
Omega-3 Fatty Acids for the Management of Hypertriglyceridemia: A Science Advisory From the American Heart Association
Skulas-Ray et al. · Circulation · 2019 · Guideline
Eicosapentaenoic acid and docosahexaenoic acid containing supplements modulate risk factors for cardiovascular disease: a meta-analysis of randomised placebo-control human clinical trials
AbuMweis et al. · Journal of Human Nutrition and Dietetics · 2018 · Meta-analysis
Effect of High-Dose Omega-3 Fatty Acids vs Corn Oil on Major Adverse Cardiovascular Events in Patients at High Cardiovascular Risk: The STRENGTH Randomized Clinical Trial
Nicholls et al. · JAMA · 2020 · Randomised trial
Benefits of fish oil supplementation in hyperlipidemia: a systematic review and meta-analysis
Eslick et al. · International Journal of Cardiology · 2009 · Meta-analysis
Omega-3 polyunsaturated fatty acid and insulin sensitivity: a meta-analysis of randomized controlled trials
Akinkuolie et al. · Clinical Nutrition · 2011 · Meta-analysis
Effects of eicosapentaenoic acid versus docosahexaenoic acid on serum lipids: a systematic review and meta-analysis
Wei & Jacobson · Current Atherosclerosis Reports · 2011 · Meta-analysis
Omega-3, omega-6, and total dietary polyunsaturated fat for prevention and treatment of type 2 diabetes mellitus: systematic review and meta-analysis of randomised controlled trials
Brown et al. (PUFAH Group) · BMJ · 2019 · Meta-analysis
Cardiovascular Risk Reduction with Icosapent Ethyl for Hypertriglyceridemia
Bhatt et al. (REDUCE-IT Investigators) · New England Journal of Medicine · 2019 · Randomised trial
Mercury Levels in Commercial Fish and Shellfish (1990-2012)
US Food and Drug Administration · FDA food-safety data table · 2012 · Reference data
Consumption of fish and vascular risk factors: a systematic review and meta-analysis of intervention studies
Alhassan et al. · Atherosclerosis · 2017 · Meta-analysis
Dose-dependent effects of omega-3 polyunsaturated fatty acids on C-reactive protein concentrations in cardiometabolic disorders: a dose-response meta-analysis of randomized clinical trials
Amlashi et al. · Inflammopharmacology · 2025 · Meta-analysis
Effects of Omega-3 Fatty Acid Supplementation on Glucose Control and Lipid Levels in Type 2 Diabetes: A Meta-Analysis
Chen et al. · PLoS One · 2015 · Meta-analysis
Bundeslebensmittelschlüssel (BLS), Version 4.0 – Deutsche Nährstoffdatenbank
Max Rubner-Institut · 2025 · Reference data
Official German nutrient database, open data CC BY 4.0. Values per 100 g edible portion. Attribute the Max Rubner-Institut as publisher.
Long-chain omega-3 fatty acids eicosapentaenoic acid and docosahexaenoic acid and blood pressure: a meta-analysis of randomized controlled trials
Miller et al. · American Journal of Hypertension · 2014 · Meta-analysis
USDA and ODS-NIH Database for the Purine Content of Foods, Release 2.0
Heydorn KC, Wu X, Garrett TJ, et al. · 2025 · Reference data
Free purine milligrams per 100 g. Includes 2024 USDA analyses and selected literature, including Kaneko 2014 (source b). Not a German official source; Souci-Fachmann-Kraut is licensed and was not used.