Mediterranean diet’s heart health benefits hold up despite intense scrutiny

Mediterranean diet’s heart health benefits hold up despite intense scrutiny

Mediterranean diet

A landmark paper has been retracted from the prestigious New England Journal of Medicine.

The high-profile study, a large multi-site trial of the Mediterranean diet in middle-aged adults with heart disease risk, has weathered intense scrutiny since its publication in 2013.

The paper reported that a traditional Mediterranean diet diminished risk of heart attack, stroke or related deaths by 28-30 percent compared to a low-fat diet. But the researchers discovered that one of their eleven sites had strayed from the randomisation protocol, potentially jeopardising these findings.

They reran their analyses to adjust for this dent in their methodology and the results held up. The updated findings were republished this week.

“We are happy with the opportunity to show the robustness of the PREDIMED results after so many ancillary and sensitivity analyses,” says senior researcher Professor Miguel Ángel Martínez-González from the University of Pamplona in Spain.

“We took the initiative to contact the editors of NEJM to report two small departures from the protocol. They affected only a small subset of the trial. The results remained intact, after correction or removing this small subset and the conclusions are the same.

“After such an intense scrutiny all of us are more convinced than ever of the soundness and robustness of the PREDIMED results.”

The original study enrolled 7,447 participants and assigned them to a Mediterranean diet supplemented with extra virgin olive oil or nuts, or a low-fat diet. The primary results were published after around five years of follow-up.

In June 2017, Dr John Carlisle of Torbay Hospital in England analysed more than 5000 published trials and listed the study as one with distributions deviating from those that would be expected from true randomisation. When Martínez-González spotted this, he notified the NEJM editors.

Randomly allocating study participants to different treatments ensures that by the law of probability, all things other than the intervention like age, sex and education are equal. If the procedure is not random, this can introduce bias because other participant characteristics could influence the outcomes.

The researchers, with the help of NEJM editors, spent 12 months meticulously reviewing each site’s randomisation procedures. In one site, they identified instances where family members and whole clinics had been allocated to the same treatment, accounting for 14 percent of the entire study population.

This digression was not done to bias the study, rather to help participants follow the diet. Analyses that excluded this subsample strengthened rather than weakened the study’s findings.

The investigators also compared participants who strictly adhered to the Mediterranean diet with those who strictly followed the reduced-fat diet. “In this causal comparison, the results were even better for the Mediterranean diet, with more than 50% relatively lower risk of cardiovascular events,” Martínez-González reports.

Twitter has been inundated with mixed opinions about the retraction. Some applaud the investigators for rectifying the analyses. Others have jumped on the retraction with comments like, “Now the investigators concede [the study] was flawed,” that the “follow-up clinical trial, PREDIMED Plus was also fishy,” and that it proved the benefits were “exaggerated.”

Although the results did not change, the retraction has shaken public opinion. “Sad day for credibility of major medical journals and nutritional science,” pronounces one Twitter post.

Others remain openly sceptical of the study. Dr Barnett Kramer from the US National Cancer Institute, for one, told the New York Times, “Nothing they have done in this re-analyzed paper makes me more confident.”

Karen Murphy, senior research fellow at the University of South Australia, emphasises “how difficult it is to conduct trials of this caliber, with a whole diet intervention across 11 different sites.

“Despite having a published protocol, sometimes deviations occur. But I think this provides the scientific and non-scientific community confidence that we can absolutely trust this data given its scrutiny from the PREDIMED investigators and NEJM.”

Rarely has a trial has been so subjected to such close inspection, according to Martínez-González. Why is this so, for an intervention that carries no risk of harm?

“While nutrition research will certainly benefit from an increased focus on scientific rigour,” says Professor Felice Jacka from the Food & Mood Centre at Deakin University, “critics should recognise the extreme challenges in conducting nutrition research and provide constructive input and feedback, rather than purely seeking to discredit. This tendency is something I see a lot in the field.”

Indeed, food and nutrition research is plagued by politics and heated debate. Editor-in-chief of the BMJ, Fiona Godlee, said at an international conference, “There are few areas of health that are more important than nutrition,” and “few that are more fraught with controversy, conflict of interest and confusion.”

Yet at the heart of it all lies an indisputable kernel of wisdom. As Jacka says, “there is no doubt that diet is related to health and that a very large body of evidence supports the contention that a diet focused on whole foods at the expense of junk and processed foods results in better health outcomes.”


Disclosure: I have conducted research on the Mediterranean diet; I know some of the interviewees in this story, and have visited the PREDIMED team. I have not met a research group more genuinely dedicated to scientific rigour and public health.

Show Buttons
Hide Buttons