The VITAL study, supported by the National Institutes of Health and headed by JoAnn Manson, the Michael and Lee Bell Professor of Women’s Health at Harvard Medical School and a professor of epidemiology at the Harvard Chan School, sought to clear up confusion around two popular dietary supplements through an analysis whose design has long been considered the gold standard for clinical trials.
Randomized and placebo-controlled, VITAL followed more than 25,000 people age 50 and older who took daily supplements containing vitamin D, omega-3 fatty acids, or placebo. The Gazette spoke with Manson about the results, which were published this month in The New England Journal of Medicine.
GAZETTE: What gaps in knowledge was the study intended to fill?
MANSON: It helped to fill several knowledge gaps. For omega-3 fatty acids, the previous randomized trials had largely been in high-risk populations with a history of cardiovascular disease (CVD) or with selected risk factors for CVD.
VITAL is the first large-scale randomized trial of marine omega-3s in a general population at “usual risk” of CVD. It’s also one of the first randomized trials of these supplements in a racially and ethnically diverse study population. Assessing the role of these supplements in a general population free of cardiovascular disease and cancer at baseline fills an important gap in knowledge.
There have been few large-scale randomized trials of moderate- to high-dose vitamin D in the prevention of cancer and CVD. Most of the vitamin D trials have been for bone health, preventing fractures. These tend to be smaller trials, and some of them have tested much lower doses, only 400 to 800 international units a day. We tested 2,000 IUs a day.
GAZETTE: Why don’t we discuss omega-3 and vitamin D separately, since it seems there were several findings for each. Were the omega-3 findings clearer?
MANSON: For the omega-3s, there were several findings pointing to a coronary benefit, though for the prespecified, primary endpoint of “major cardiovascular events” — heart disease plus stroke plus total CVD mortality — there was only a small, statistically nonsignificant 8 percent reduction. It’s likely that the lack of benefit for stroke diluted the results for the primary endpoint.
But those who had low fish intake — fish is the primary source of marine omega-3s in the diet — had a statistically significant reduction in the primary endpoint, compared to placebo. The lower half of the distribution had a 19 percent reduction with treatment and, for heart attack alone, a 40 percent reduction.
For the overall study population, we had several prespecified secondary endpoints, including looking at heart attack, stroke, and CVD mortality separately. We did see a 28 percent reduction in heart attack, but no significant reductions in stroke or CVD mortality. We also saw that African-Americans appeared to benefit the most in terms of heart attack reduction — a 77 percent reduction compared to the placebo group. Again, you have to interpret this cautiously.
“Previous randomized trials of omega-3 have rarely done diet assessments. And they’ve rarely looked at the modifying effect of fish consumption on the omega-3 benefit.”
GAZETTE: And what about vitamin D?
MANSON: There weren’t significant reductions in the primary endpoints of major cardiovascular events — combined heart attacks and stroke — or the total incidence of cancer. But we saw a signal for reduction in cancer death over time.
The trial was 5.3 years, which is usually sufficient for CVD outcomes but is relatively short for cancer, due to its long latency period. It may take longer to see the effect on cancer incidence. Those cancers diagnosed in the early years of the trial were undoubtedly already present, subclinical, at the time the participants were enrolled. Previous studies suggest that vitamin D may work at later stages of cancer. What we saw, after accounting for latency period by excluding early follow-up, was a 25 percent reduction in cancer deaths.
Laboratory and clinical studies suggest that vitamin D may affect tumor biology, making tumors less invasive, less aggressive, and less likely to metastasize. And if that’s the effect, once there’s already a tumor — diagnosed or not, clinically detected or not — you might see a reduction in cancer death over the course of a five-year trial.
There were some findings by body mass index. If you were average weight or below, there did seem to be more of a benefit of the vitamin D on cancer incidence. If you have a higher body mass index, there may be a need for higher doses of vitamin D or maybe there’s a vitamin D resistance similar to insulin resistance. We really don’t know. But this warrants further study.
I think these findings do need to be interpreted cautiously — they’re subgroup findings — but these are intriguing signals.
GAZETTE: Is there a lesson here about the importance of considering different populations when you’re analyzing scientific research?
MANSON: Although it’s biologically plausible that some groups will benefit more than others, the problem is that you always have to interpret subgroup analyses cautiously.
With dietary supplements, it’s important to understand the nutrient status of the study participants and to see if those with lower dietary intake or lower nutrient status are more likely to benefit from the supplements. With omega-3s, of course you want to understand dietary intake, particularly fish consumption. The finding that those with lower fish consumption in VITAL were more likely to benefit from marine omega-3s adds to the biological plausibility.
But here’s what’s really surprising. Previous randomized trials of omega-3 have rarely done diet assessments. And they’ve rarely looked at the modifying effect of fish consumption on the omega-3 benefit, in terms of who benefits and who doesn’t. So this is a strong advantage of the VITAL trial, that we had the diet assessment and we could stratify participants by diet and low-fish consumption.
GAZETTE: Analysis of this data is continuing. Will you be looking at effects on other conditions?
MANSON: Yes. We’re looking at many other health outcomes. We have 23 ancillary studies. We’re looking at diabetes, cognitive function, mood, depression, autoimmune disorders, lung diseases, infections, heart failure, and many other clinical conditions.
GAZETTE: Any idea what the timing is on those?
MANSON: We’re saying six to 12 months to have these ancillary studies published. And that will help to inform the benefit-risk profile of these supplements.
Some people, if they’re not in a group that may particularly benefit — such as those with low fish consumption — they may want to stay tuned for the results of the ancillary studies, because these studies will help with their decision-making.
GAZETTE: You mentioned follow-up studies as well.
MANSON: We’re continuing to follow the study participants for at least two more years, and we’re going to apply for continued funding to follow for several more years after that. The cancer results in particular, due to the latency period, may take a while to become fully apparent.
GAZETTE: What would be your recommendations to the public?
MANSON: If you’re already taking these supplements, we don’t find clear reasons for stopping, but we caution against mega-dosing. The safety that we demonstrated for these doses — 2,000 IUs vitamin D, one gram a day of omega-3s — may not apply to much higher doses.
And, if you fall into a group that appears to gain greater benefit, such as — with the omega-3s — those with low-fish consumption and African-Americans, you may want to talk to your health care provider about whether to take a supplement.
But for most people, stay tuned, because we’re going to have results for ancillary studies — diabetes, cognition, depression, all these other outcomes.
Also, it’s possible that over the next couple of years, medical and public health authorities will look at these findings and at the results of other randomized trials that have been done recently and see if clinical guidelines regarding these supplements should be updated.
Interview was edited for clarity and length.