Benefits
Triglyceride reduction
An American Heart Association advisory found that 4 g a day of prescription EPA plus DHA lowers triglycerides by 30% or more in people with very high levels, a well-established effect at these doses, and that EPA-only and EPA plus DHA products lower triglycerides by similar amounts in people with moderately high levels. The AHA had earlier advised 2 to 4 g a day for people with high triglycerides, under medical care.
Mood support with EPA-rich formulas
A pooled analysis of 26 placebo-controlled trials (2,160 people) found omega-3s eased depressive symptoms overall, with benefit from EPA-rich formulas (60% EPA or more) at up to 1 g of EPA a day but not from DHA-rich ones. One 8-week trial of 1 g a day in 196 adults found no difference from placebo.
Liver fat and liver enzyme support
A pooled review of 22 trials found omega-3s reduced liver fat versus placebo in people with fatty liver. A newer review of 20 trials found improvements in GGT and in liver fat seen on ultrasound, but no change in ALT, AST or liver fat measured by MRI, and more side effects. The overall picture is uncertain, and diet and weight loss remain the basics.
Joint pain, morning stiffness and inflammation markers
A pooled analysis of 17 randomized trials in inflammatory joint conditions found 3 to 4 months of omega-3s reduced patient-reported joint pain, morning stiffness, tender joints and NSAID use. An umbrella review of 32 pooled analyses found omega-3s lowered blood CRP, TNF-alpha and IL-6. The effects were small to moderate and used alongside usual care.
DHA needs during pregnancy
DHA builds up in the baby's brain and retina during pregnancy, and EFSA advises an extra 100 to 200 mg of DHA a day on top of the usual 250 mg of EPA plus DHA. Trials on the timing of birth are mixed: a review of 70 trials found fewer births before 34 weeks (2.7% vs 4.6%), but a later trial of 5,486 pregnancies found no difference. Ask your prenatal care provider.
Cognitive aging in older adults
In a 5-year trial of 3,501 older adults, 1 g a day of EPA plus DHA did not slow cognitive decline compared with placebo. In 174 people with mild to moderate Alzheimer disease who completed a 6-month trial, DHA-rich omega-3 did not slow decline overall, with a possible effect in a subgroup of 32 with very mild symptoms. These trials found no benefit for memory from the supplements.
Dry eye and macular health trial results
In a 12-month trial of 535 people with moderate to severe dry eye, 3 g a day of EPA plus DHA was no better than an olive oil placebo for symptoms or eye-surface signs. In a 5-year trial of 4,203 people aged 50 to 85 at risk of advanced macular degeneration, 1 g a day of EPA plus DHA did not slow progression. Current trial evidence shows no benefit for either.
Fewer skin blemishes in a small trial
In a 10-week trial of 45 people with mild to moderate acne, 2,000 mg a day of EPA plus DHA was followed by fewer inflamed and non-inflamed blemishes, and skin samples showed less inflammation. A control group was included. This was a single small trial, so the evidence is preliminary.
Focused attention in youths with low EPA levels
In a 12-week double-blind trial in 92 youths aged 6 to 18 with ADHD, 1.2 g a day of EPA improved one measure of focused attention more than placebo (effect size 0.38), with larger attention and vigilance gains in those with the lowest starting EPA levels. Impulsivity improved less than on placebo, especially in youths with high EPA levels. This is a single trial of EPA alone.
Mechanism of action
Triglycerides, heart rate and blood pressure
A cardiology review concluded that EPA and DHA lower blood triglycerides, resting heart rate and blood pressure, and may improve blood vessel function. They act on cell membranes, ion channels and gene expression, and are converted into signaling molecules.
Inflammation signaling
EPA and DHA change the mix of eicosanoids the body makes from fats, which shifts signaling away from inflammation. In people, pooled trials show modest drops in CRP, TNF-alpha and IL-6.
DHA in brain and retina
DHA is a building block of cell membranes in the brain and the retina. Higher intake has not translated into better memory or eye outcomes in large supplement trials of adults who were not deficient.
Clinical trials
Randomized, double-masked trial of 1 g/day EPA plus DHA and/or lutein/zeaxanthin versus placebo, with cognitive tests every 2 years for 5 years. (Chew et al. 2015, JAMA)
3,501 older adults (mean age 72.7) at risk of advanced macular degeneration who completed cognitive testing.
The yearly change in composite cognitive score did not differ between the omega-3 and no-omega-3 groups (P = .63). Null primary outcome.
Randomized, double-blind, placebo-controlled trial of 4 g/day icosapent ethyl (a prescription purified EPA) versus placebo in statin-treated adults with raised triglycerides, median 4.9 years. (Bhatt et al. 2019, N Engl J Med)
8,179 adults with heart disease or diabetes plus other risk factors.
The primary end point occurred in 17.2% on EPA versus 22.0% on placebo (hazard ratio 0.75). Hospitalization for atrial fibrillation or flutter was higher with EPA (3.1% vs 2.1%), and serious bleeding was 2.7% vs 2.1% (P = 0.06).
Randomized, double-blind trial of 4 g/day omega-3 carboxylic acids (EPA plus DHA) versus corn oil in statin-treated adults at high heart risk. (Nicholls et al. 2020, JAMA)
13,078 adults with high cardiovascular risk, raised triglycerides and low HDL.
Stopped early for likely futility: the primary end point occurred in 12.0% versus 12.2% (hazard ratio 0.99). Stomach side effects were more common with omega-3 (24.7% vs 14.7%). Null primary outcome.
Multicenter, double-blind, randomized trial of fish-oil capsules with 900 mg/day omega-3 versus vegetable-oil capsules, from before 20 weeks to 34 weeks of pregnancy, in six Australian centers. (Makrides et al. 2019, N Engl J Med)
5,544 pregnancies in 5,517 women; 5,486 analyzed.
Early preterm delivery occurred in 2.2% with omega-3 and 2.0% with control (relative risk 1.13, not significant). More infants in the omega-3 group were very large for gestational age. Null primary outcome.
Cochrane systematic review of randomized trials comparing omega-3 supplements or foods during pregnancy with placebo or no omega-3. (Middleton et al. 2018, Cochrane Database Syst Rev)
70 trials with 19,927 women.
Birth before 37 weeks (RR 0.89) and before 34 weeks (4.6% vs 2.7%, RR 0.58) were both lower with omega-3, rated high-quality evidence. Pregnancies lasting past 42 weeks were probably more common.
Randomized, double-blind, placebo-controlled 8-week trial of EPA-enriched or DHA-enriched omega-3 (1,000 mg/day each) as stand-alone treatment. (Mischoulon et al. 2015, J Clin Psychiatry)
196 adults with major depressive disorder; 154 completed.
All three groups improved, but neither omega-3 preparation separated from placebo on the main depression scale; response and remission rates were similar. Null primary outcome.
Meta-analysis of double-blind, placebo-controlled trials of omega-3 for depressive symptoms. (Liao et al. 2019, Transl Psychiatry)
26 trials with 2,160 participants.
Omega-3 had an overall benefit (SMD -0.28). EPA-pure and EPA-major formulas (60% EPA or more) at 1 g/day of EPA or less showed benefit, while DHA-pure and DHA-major formulas did not.
Randomized, double-blind, placebo-controlled 12-week trial comparing EPA 1.2 g/day with placebo. (Chang et al. 2019, Transl Psychiatry)
92 youths aged 6 to 18 with ADHD.
EPA improved focused attention more than placebo (effect size 0.38), with larger gains in those with low starting EPA levels; impulsivity improved less than on placebo, especially in youths with high EPA levels. EPA did not change hs-CRP or BDNF.
Randomized, double-blind, placebo-controlled trial of DHA-rich omega-3 for 6 months, then open treatment to 12 months. (Freund-Levi et al. 2006, Arch Neurol)
174 patients with mild to moderate Alzheimer disease.
Cognitive decline did not differ between groups at 6 months. In a subgroup of 32 with very mild impairment (MMSE above 27), decline was slower with omega-3. Treatment was safe and well tolerated. Null primary outcome.
Randomized, double-blind trial of high-dose fish oil (5.5 g/day EPA plus DHA) versus low-dose control (0.4 g/day) added to triple drug therapy. (Proudman et al. 2015, Ann Rheum Dis)
Adults with rheumatoid arthritis of less than 12 months' duration who had not taken disease-modifying drugs.
Failure of triple drug therapy was lower with high-dose fish oil (hazard ratio 0.28) and first ACR remission was more frequent (hazard ratio 2.17). DAS28 activity scores, disability scores and adverse events did not differ.
Meta-analysis of randomized controlled trials of omega-3 for joint pain in rheumatoid arthritis, inflammatory bowel disease and painful periods. (Goldberg et al. 2007, Pain)
17 randomized controlled trials.
After 3 to 4 months, omega-3 reduced patient-reported joint pain, minutes of morning stiffness, the number of painful or tender joints and NSAID use. Physician-assessed pain and the Ritchie articular index did not change.
Randomized, double-blind, placebo-controlled trial of perioperative fish oil. (Mozaffarian et al. 2012, JAMA)
1,516 adults having cardiac surgery.
Postoperative atrial fibrillation occurred in 30.0% with fish oil and 30.7% with placebo (odds ratio 0.96). No increase in bleeding or serious adverse events. Null primary outcome.
Multicenter, randomized, double-blind trial of 3,000 mg/day fish-derived EPA plus DHA versus olive oil placebo for 12 months. (Asbell et al. 2018, N Engl J Med)
535 adults with moderate to severe dry eye disease.
Symptom scores improved similarly in both groups (difference -1.9 points, P = 0.21), and eye-surface signs and tear tests did not differ. Null primary outcome.
Meta-analysis of randomized controlled trials of omega-3 supplements in adults with nonalcoholic fatty liver disease. (Kim et al. 2025, Clin Nutr)
20 trials with 1,615 participants.
Omega-3 improved GGT and liver fat seen on ultrasound, with possible publication bias, but did not change AST, ALT, MRI-measured liver fat, liver stiffness or biopsy findings. Side effects were more common with omega-3.