March 14, 2023

expert reaction to study looking at Mediterranean diet and women’s cardiovascular disease and death risk

A study published in Heart looks at a Mediterranean diet and cardiovascular disease in women.

Prof George Davey Smith FRS FMedSci, Professor of Clinical Epidemiology, University of Bristol, said:

“The press release for this paper is headed “’Mediterranean diet cuts women’s cardiovascular disease and death risk by nearly 25%”, which is an unmerited jump from observational studies to what would happen if women changed their diets.

“Observational studies of nutrition have been notoriously misleading.  For example observational studies of vitamin E supplementation showed a near 40% lower risk of coronary heart disease (CHD) in those using supplements in two studies, reported in the lead story in the New York Times as “Vitamin E greatly reduces risk of heart disease, studies suggest”.  Long-term large scale randomised controlled trials established there is no meaningful benefit of vitamin E supplementation with respect to CHD.  Nutritional measures are strongly patterned by social factors and behavioural dispositions, leading to very substantial confounding (where the exposure and outcome are both influenced by the same factor, leading to a non-causal association being seen).  The same story – of observational studies suggesting large benefit, and randomised trials showing no benefit or even harm – has been seen for many other nutritional factors, including vitamin C, beta-carotene, vitamin D, selenium and several others.

“The study the press release is about simply calculates a weighted average of estimates from published observational studies, there are no new data here.  All the studies included likely suffer from the same confounding as seen in the examples above of misleading observational epidemiological studies of nutritional factors.  The only large randomized trial of Mediterranean diet is sadly not trustworthy, since randomization was not consistently applied, as is well documented.  Such a methodological flaw casts doubt on all aspects of the trial.

“With the greater scientific literacy generated by the pandemic one hoped that probably meaningless findings from observational studies would not be presented in this way.”

Victoria Taylor, Senior Dietitian at the British Heart Foundation, said:

“It’s long been known that eating a Mediterranean-style is good for your heart, but it’s encouraging to see this research suggest that when we look at women separately from men, the benefits remain.  Heart disease is often seen as a male problem, but coronary heart disease kills more than twice as many women as breast cancer in the UK each year.  Sex-specific research like this is vital for reducing the heart disease gender gap and improving women’s care.

“Whatever your gender, a healthy lifestyle which includes a balanced diet like the Mediterranean-style diet can help you to lower your risk of developing heart and circulatory diseases and the risk factors for them, such as type 2 diabetes, obesity, high blood pressure and high cholesterol.  It’s easy to do – make sure you are eating plenty of fruit and vegetables, beans, lentils, wholegrains, fish, nuts and seeds, along with some low-fat dairy and fat from unsaturated sources like olive oil.  It’s also important to eat less processed meat, salt and sweet treats.

“There are limitations to this study, including the fact the studies included were observational, so can’t show cause and effect, and that they generally relied on food intake questionnaires which were self-reported, a regular problem with dietary studies that can affect reliability of results.  There are also limitations to the number of previous studies the researchers could draw on.  These are acknowledged by the authors, who also recommend that we should interpret their findings with caution, but it also supports the need for more sex-specific research.”

Prof Nita Forouhi, MRC Epidemiology Unit, University of Cambridge, said:

“In itself this is a worthwhile message from this research study that women, as well as men, would potentially benefit from adherence to the Mediterranean diet.  However, it is disappointing that this study focused on women but failed to look at some women specific factors.  The authors did subgroup analyses by sex and ethnicity, but not by whether women were pre- or post-menopausal, were using hormonal replacement therapy or not, using the oral contraceptive pill or not, had a history of gestational diabetes or not, and so on.

“Moreover, it is a major limitation that they included only the 16 studies where data were available disaggregated for women and they lost data from 66 studies where data existed but not in disaggregated form.  Some of those 66 studies may have included analyses of an interaction between sex and Mediterranean diet in relation to the health endpoints, but the authors did not report on that, which is a loss of potentially useful information.”

Dr Duane Mellor, Registered Dietitian and Senior Lecturer, Aston Medical School, Aston University, said:

“This is an interesting study that provides a systematic review and meta-analysis of previously published data looking at the potential effects of a Mediterranean Diet on risk of cardiovascular disease and an early death.  We need to be careful with headlines like ‘cuts women’s risk of cardiovascular disease and death by nearly 25%’ for two reasons: one, the data only shows and association and not causality; and secondly this only refers to the length of time individuals were included in the respective studies, after all eventually death comes to us all!

“The score used for assessing a Mediterranean Diet was also quite simplistic.  It only looked at 10 factors in total, 6 of which were deemed beneficial in terms of vegetables, legumes (beans, peas, and lentils), fruits, nuts, cereal and fish, where higher than median (average intake for the population) were each given one point; then three food groups which were deemed potentially detriment which were meat, poultry and dairy produce where also given one point where someone ate less than median intake for the population.  Finally alcohol was given a point if intake was deemed moderate, with no points for non-drinkers or those drinking excessively.

“This might seem simple and could miss a number of foods which might be key parts of a Mediterranean Diet e.g. Olive oil, and also it does not link to how much of each type of food we should be encouraging people to eat.  For example it does not suggest we should be trying to eat at least five different types of fruit and vegetables a day, only more than the average for the population.

“So, although this study does suggest an association of reduced risk of cardiovascular disease and early mortality with eating a diet based on many of the foods that are recommended as at the building blocks of a healthy diet, it does not help in suggesting how much in terms of vegetables, legumes, fruit, nuts, cereals and fish we should eat, and how little meat, poultry and dairy produce we should consume.  Also other studies not included in this one, which sought to assess dietary quality using different ways of measuring Mediterranean Diet pattern, would suggest some types of dairy produce for example unsweetened yoghurts may be beneficial for heart health.

“This brings us to the challenge of this type of research.  Due to the limitations of how diet was assessed using food frequency questionnaires and then how Mediterranean Diet was measured using this simplistic 10 point score, even though this is an analysis of a number of studies this means that it has a limited ability to suggest recommendations on how we should eat.  What we already encourage in dietary recommendations is basing our diet on foods from plants – vegetables, pulses, nuts, seeds and fruit with wholegrain cereals and fish, along with modest amounts of meats, poultry and dairy.  Also when extending these types of dietary recommendation beyond the Mediterranean it is important to adapt them to different cuisines and food cultures.”

‘Primary prevention of cardiovascular disease in women with a Mediterranean diet: systematic review and meta-analysis’ by Anushriya Pant was published in Heart at 22:30 UK time on Tuesday 14 March 2023.

DOI: 10.1136/heartjnl-2022-321930

Declared interests

Prof George Davey Smith: “I have no COIs.”

Victoria Taylor: “I do not have any interests to declare.”

Prof Nita Forouhi: “None.”

Dr Duane Mellor: “No conflicts of interest to declare.”