The Most Trusted Pill in the Cabinet
Fish oil has held a place of honor in American medicine cabinets for decades: the amber bottle promising cleaner arteries and a quieter heart. The story was plausible, even romantic. 1970s research linked fish-heavy Inuit diets to low heart disease; early trials seemed to confirm it; in 2002 the American Heart Association recommended fish oil for coronary heart disease. By 2012 nearly 19 million American adults were taking it. The randomized trials, however, kept coming back empty.
25,871 People, Five Years, One Gram a Day
The Vitamin D and Omega-3 Trial, known as VITAL, was built to settle the question at a scale no one could dispute. JoAnn Manson's team randomized 25,871 healthy adults to a daily gram of marine omega-3 (460 mg EPA plus 380 mg DHA) or placebo for a median of 5.3 years. The trial was double-blind, NIH-funded, and adherence topped 83 percent, and then the numbers came in: a major cardiovascular event, a heart attack, a stroke, or cardiovascular death, struck 386 people in the omega-3 group and 419 in the placebo group, for a hazard ratio of 0.92 with a 95 percent confidence interval of 0.80 to 1.06 and a p-value of 0.24, which is statistically no effect. Manson reported the null plainly: supplementation with n-3 fatty acids did not result in a lower incidence of major cardiovascular events than placebo.
The 112,000-Person Verdict
VITAL was the anchor, but the full weight of evidence runs deeper. In 2018 a Cochrane review pooled 79 randomized trials covering more than 112,000 people and found high-quality evidence that increasing EPA and DHA intake has little or no effect on all-cause mortality and cardiovascular events, concluding supplemental omega-3 fats are probably not useful for preventing or treating cardiovascular disease. A 2020 update granted a possible small coronary benefit while leaving the core finding intact. The UK ASCEND trial gave 15,480 diabetic patients the same one-gram dose for 7.4 years. Serious vascular events: rate ratio 0.97, confidence interval 0.87 to 1.08. Lead investigator Louise Bowman said no justification remained for recommending it. Three mega-trials, three misses at the standard dose. The AHA's 2017 advisory had already retreated from its 2002 position. The capsule kept selling anyway.
The Two-Billion-Dollar Arithmetic
Nobody ran the population math, so here it is. A capsule costs roughly twenty cents a day, or $73 a year, and with 18.8 million American users the national outlay nears $1.4 billion annually for a heart intervention with no measured heart benefit. Grant the most generous reading: VITAL's secondary endpoint showed a hazard ratio of 0.72 for total heart attack and a number needed to treat of 250, meaning 250 people must take the capsule for 5.3 years to prevent one. Multiply 250 by 5.3 by $73 and the price is about $96,700 per heart attack prevented, for a secondary endpoint the trial never powered. For the primary endpoint the number needed to treat is undefined, because the effect was not significant. Zero in the denominator. That is what billions a year buy: a hope, a habit, and a rounding error in the event rate.
The Strongest Case Against
The deepest objection is REDUCE-IT, and it deserves its full strength. In 2018 Deepak Bhatt's team randomized 8,179 statin-treated patients with high triglycerides to icosapent ethyl, a highly purified prescription EPA at 4 grams a day, and found a 25 percent relative risk reduction in major ischemic events: hazard ratio 0.75, number needed to treat 28. Its defenders call the earlier failures a dosing error: standard trials used 840 milligrams of EPA plus DHA while REDUCE-IT used nearly five times that in purified EPA. The counter to the counter is the placebo. REDUCE-IT used mineral oil, and critics led by Steven Nissen noted that LDL cholesterol and inflammatory markers rose in the placebo group, suggesting the comparison was against something mildly harmful, though regulators judged the placebo effect small. Then came STRENGTH, the cleanest possible test of that dosing-error defense, giving 13,078 high-risk patients 4 grams a day of EPA plus DHA against corn oil, a genuinely neutral comparator, and stopping early for futility at hazard ratio 0.99, which means that if the mineral oil artifact explained everything, the neutral comparator should have revealed the true EPA benefit, and it did not. The prescription-strength EPA story is genuinely unresolved, which is why it cannot rescue the over-the-counter capsule: at the one-gram dose sold in every drugstore, VITAL, ASCEND, and Cochrane agree. One more objection deserves air: a 2019 meta-analysis of 13 trials and 127,477 people found an 8 percent reduction in heart attack and coronary death above 840 milligrams a day, and Manson noted VITAL's heart-attack signal, hazard ratio 0.72, was strongest in low fish eaters and Black participants. Something may exist at higher doses, in sicker patients, with purified EPA. None of it describes the capsule in your cabinet.
What We Didn't Prove
This evidence does not touch prescription icosapent ethyl for its approved use. If you take Vascepa for high triglycerides on a statin, that is a different drug at four times the dose with a positive trial behind it. It says nothing about high-dose purified EPA in primary prevention, which no trial has tested. Eating fish is not taking capsules: the Cochrane reviewers found little trial evidence on oily fish itself, and the AHA still recommends one to two seafood meals a week for the protein, vitamin D, selenium, and iodine fish delivers. One risk signal deserves mention, because a 2021 meta-analysis of seven trials and 81,210 patients found omega-3 supplementation raised atrial fibrillation risk by 25 percent, concentrated above 1.5 grams a day in high-risk patients. The typical one-gram capsule is unlikely to be the danger zone, but the curve argues against taking more when one shows no effect. Finally, VITAL tested a single dose in adults averaging 67. The honest summary is narrower than the headline. At the standard supplement dose, in the general adult population, fish oil does not protect the heart.
The Bottom Line
The fish oil story shows how a plausible mechanism, a beautiful observational finding, and a few early positive trials can sustain a multi-billion-dollar habit long after the randomized evidence says stop. Three mega-trials and 79 pooled studies could not find the heart benefit at the dose everyone takes. The capsule is not poison. It is something more ordinary and more expensive: a well-marketed null. REDUCE-IT keeps the door open for prescription-strength purified EPA in the right patients. That door does not lead back to the supplement aisle.
What You Can Do
First, if you take fish oil for your heart, re-examine the habit. The evidence at your dose is essentially zero. Do not stop prescription icosapent ethyl without talking to your doctor. Second, keep eating fish: one to two seafood meals a week remains the recommendation. Third, do not mega-dose, since atrial fibrillation risk rises with dose. Fourth, redirect the money: blood pressure control, a statin if your risk warrants one, and regular exercise carry randomized-trial evidence. The capsule had its trial, and it failed.