select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to observational study on types of diet and COVID-19 infection

A study published in BMJ Nutrition, Prevention & Health looks at vegetarian and plant based diets and the incidence of COVID-19.


Prof Margaret Rayman, Professor of Nutritional Medicine, University of Surrey, said:

“The authors say that “we recommend the practice of following plant-based diets or vegetarian dietary patterns”. I think that this study can’t provide substantial evidence for this recommendation with regards to COVID-19 infection. This is an important limitation, particularly in terms of people who may have higher nutritional needs than most, such as pregnant women.  Furthermore, I disagree that predominantly plant-based diets provide more nutrients that boost the immune system and help to fight viral infections”. In fact, the opposite is the case as shown below.

“A recent study assessed 141 observational/intervention studies published since 2000 that looked at nutrient intake and status in adult populations consuming vegetarian, vegan or meat-eating diets (1).  Intake of vitamin B12, iodine, iron, zinc, calcium and vitamin D, and bone turnover markers were generally lower in plant-based vs. meat-eater dietary patterns as these micronutrients are mostly found in animal foods or, in the case of iron, zinc and calcium, have a lower bioavailability in plant foods (1). The important n-3 long-chain essential fatty acids (n-3 LC-PUFA), EPA and DHA were also found to be much lower in vegans and vegetarians (1).

“Similar data come from the German Nutritional Evaluation (NuEva) study of similar numbers of omnivores, flexitarians, vegetarians, and vegans (total n=258) which revealed an insufficient dietary intake of selenium, zinc, potassium, iron (women), calcium, vitamin B12, n-3 LC-PUFA, and vitamin D particularly in vegetarians and vegans. Analysis of biomarkers showed deficiency of vitamin B12, selenium, zinc and iodine and low ferritin (i.e. low iron stores) (2,3). 

“The issue with metal ions is that food factors such as phytate (myo-inositol hexaphosphate), soyabean protein, polyphenols, calcium, oxalic acid and dietary fibre inhibit their absorption (4).  Phytate is the primary storage form of phosphorus in plants and is found in cereals, legumes, oil seeds and nuts. It can chelate iron, zinc, calcium and magnesium, reducing their absorption (4). Polyphenols (e.g. tannins) and oxalates in food matrices also bind metal ions (4).

”These are important studies because they indicate deficiency of a number of nutrients in the plant-based diet that have been shown to reduce the effect of COVID-19 infection and the likelihood of death.  There is evidence showing that a higher status of selenium, zinc and copper are associated with a better cure rate or recovery from COVID-19.  Both selenium and zinc are known to be important to immunity and viral recovery (5,6).  The majority of non-survivors from COVID-19 were characterized by a combined selenium and zinc deficit, and both trace elements recovered in blood during hospital stay and towards discharge were indicators of high survival chances (see below). The following are some examples.

“My colleagues and I found a significant association between COVID-19 cure rate and background selenium status (published hair-selenium concentration) in Chinese cities that was very highly cited (7); this was supported by a later paper (8). Furthermore, studies in Germany compared serum selenium and selenoprotein P concentrations in COVID-19 patients with reference data from the large European EPIC cross-sectional study (control) and found that selenium status was significantly higher in serum samples from surviving COVID patients than from non-survivors (8,9,10). Apart from the studies conducted in Germany (9,10,11), a number of other studies have assessed selenium in serum/plasma in COVID-19 finding the same effect (12,13,14).

“While these studies are heavily based on selenium (my main research area), a search of the internet will show a further list of studies based on zinc.

The authors of the study say that “those following a predominantly plant-based or “vegetarian/vegan diet were 39% less likely to become infected than the omnivores” which is due to their healthier weight and lifestyle.  My explanation for their result is that insufficient adjustments were made between the groups to take account of the differences, a case of unaccounted confounding.”  



  1. Neufingerl and Eilander, Nutrients 2022, 14: 29.
  2. Dawczynski C et al., Nutrient Intake and Nutrition Status in Vegetarians and Vegans in Comparison to Omnivores – the Nutritional Evaluation (NuEva) Study. Front Nutr.2022, 9: 819106;
  3. Klein L et al., Selenium, Zinc, and Copper Status of Vegetarians and Vegans in Comparison to Omnivores in the Nutritional Evaluation (NuEva) Study. Nutrients 2023, 15: 3538.
  1. Gibson RS, Perlas L, Hotz C. Improving the bioavailability of nutrients in plant foods at the household level. Proc Nutr Soc 2006; 65(2): 160-8.
  2. Rayman MP et al. The relevance of selenium to viral disease with special reference to SARS-CoV-2 and COVID-19. Proc Nutr Soc (2023), 82, 1–12.
  3. Wessels I, Rolles B, Rink L. The Potential Impact of Zinc Supplementation on COVID-19 Pathogenesis. Front Immunol. 2020 Jul 10;11:1712. doi: 10.3389/fimmu.2020.01712. eCollection 2020.PMID: 32754164.
  1. Zhang J, Taylor EW, Bennett K, Saad R,  Rayman  Association between regional selenium status and reported outcome of COVID-19 cases in China. AJCN 2020 Jun 1;111(6):1297-1299.
  2. Zhang HY et al. BMC Infectious Diseases 2021, 21:452.
  3. Moghaddam A, Heller RA, Sun Q et al. (2020) Selenium deficiency is associated with mortality risk from COVID-19. Nutrients 12, 2098.
  1. Heller RA, Sun Q, Hackler J et al. (2021) Prediction of survival odds in COVID-19 by zinc, age and selenoprotein P as composite biomarker. Redox Biol 38, 101764.
  2. Hackler J, Heller RA, Sun Q et al. (2021) Relation of serum copper status to survival in COVID-19. Nutrients 13, 1898.
  3. Du Laing G, Petrovic M, Lachat C et al. (2021) Course and survival of COVID-19 patients with comorbidities in relation to the trace element status at hospital admission. Nutrients 13, 3304.
  4. Fakhrolmobasheri M, Mazaheri-Tehrani S, Kieliszek M et al. (2022) COVID-19 and selenium deficiency: a systematic review. Biol Trace Elem Res, 200, 3945–3956.
  5. Younesian O, Khodabakhshi B, Abdolahi N et al. (2022) Decreased serum selenium levels of COVID-19 patients in comparison with healthy individuals. Biol Trace Elem Res 200, 1562–1567.


Dr Gavin Stewart, Senior Lecturer in Evidence Synthesis, Newcastle University, said:

“This work presents interesting data but the authors conclusions do not adequately reflect the uncertainty inherent in small observational studies that are not designed to assess causal relationships.  The conclusion that plant-based diets have a preventative role in COVID-19 infection is premature and not warranted.”


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

“This is a small observational study which reports the incidence of self-reported COVID-19 infection and dietary pattern. It suggests that those who followed a plant based diet (both vegetarian and flexitarian) had a lower incidence of self-reported COVID-19. It provides only a potential association between diet and risk of infection with COVID-19 and no causal evidence.

“This study is observational and does not include confirmation of diagnosis, and although the researchers corrected for ethnicity, education and other factors associated with risk of COVID-19 infection, the small sample size might could have meant that any association could be a statistical error. It was also noticeable that there are data reporting errors, with the paper suggesting the incorrect percentage of participants reporting as white.

“Although diet quality has an important role to play in reducing risk of disease including potential of infections like COVID-19, and this is widely considered to be a dietary pattern based on foods from plant sources e.g. vegetables, fruit, nuts, seeds, beans, peas, lentils along with wholegrain and modest amounts of dairy and meats (if that is a person’s dietary preferences). It is important to replace nutrients found in animal products e.g. iron, iodine, B12, calcium and omega 3 fatty acids with sources from plants to reduce the risk of nutritional deficiencies. 

“Also the suggestion that antioxidants might be a mechanism has largely been dismissed in the scientific literature for over a decade, it is thought that the variety of nutrients and other components in plants may reduce our risk of disease by a number of other mechanisms, including up regulating protective pathways in our liver and enhancing the health benefits of our gut microbiome. The antioxidant theory although easily demonstrated in the test tube is considered an epiphenomenon in humans.”


Prof Kevin McConway, Emeritus Professor of Applied Statistics, Open University, said:

“As usual with observational studies investigating possible associations between diet and health outcomes, this one needs considerable care in interpretation.

“The main problem is that, despite the finding that people who ate plant-based diets had lower incidence of Covid infection than people who ate omnivorous diets, we just can’t be sure whether this correlation is caused by the type of diet they say they eat, or by something else. The press release and the research paper clearly acknowledge this. The doubt arises because there are many differences between the diet groups, other than the actual diets that people reported. According to data given in the research paper, the groups differed in terms of ethnicity, educational level, whether they had pre-existing medical conditions, their BMI, and how much physical  activity they did – and very probably they differed in other ways too. So it’s possible that the observed difference in their chance of contracting Covid during the period of study (March to July 2022) was caused by one or more of these other factors, and not by the difference in their type of diet.

“Obviously the researchers are aware of this issue, and they made statistical adjustments to allow for the other factors I have mentioned, plus others. After these adjustments, the correlation between diet type and the risk of infection remained. But one cannot be sure that everything relevant has been adjusted for, and no adjustment could be made for factors on which the researchers had no data. This doesn’t rule out the possibility that diet type does, to some extent at least, cause differences in infection risk, but it does very clearly mean that we just can’t be sure about cause and effect from this study.

A single observational study can, if put together with other studies of various kinds, contribute to “knowledge of what’s cause and effect, and the researchers do mention another three observational studies in different places and with different groups of participants that also found an association between plant-based diets and lower risk of Covid infection. However, one would really need other data – perhaps some empirical information on how diets could, within the body, reduce the infection chance – before coming to anything like a conclusion on the matter. The research paper does mention some possible ways in which eating a plant-based diet could act in the body to reduce infection chance, but the study itself provides no data on this. So a very considerable doubt remains.

“The researchers provide only a relative measure of the size of the association they found. The press release says that “those following a predominantly plant-based or vegetarian/vegan diet were 39% less likely to become infected than the omnivores”, and that indeed corresponds to what is in the paper (after the adjustments for other factors). Translating this into absolute risks, they researchers found that, of a group of 100 people, like those in the study and on an omnivorous diet, 52 would have been infected during the research period. In another group of 100 people, who matched the first 100 in terms of the factors adjusted for, but were on a predominantly plant-based or vegetarian/vegan diet, an estimated 39 would have been infected, so considerably fewer. There is a margin of statistical error around that estimate, though – the number infected in the plant-based group could plausibly be anywhere between 33 and 47, so it’s possible that the difference in risk is quite a lot smaller or bigger than the simple estimate indicates. Also, I must emphasise that we still can’t be sure that it’s the difference in diet group that causes the difference in infection risk – it could still be due to some factor or factors that were not adjusted for.

“It’s also worth mentioning that the nature of diets in the population where the study took place (Brazil) might be different from in other countries, such as the UK. If the typical vegetarian, or omnivorous, diets of the people who participated in this study were very different than typical vegetarian, or omnivorous diets here, perhaps the findings would be different here. And indeed the diets of people who volunteered to be in this research study might not even be typical of Brazilian diets generally. The researchers list several other limitations of their study in their research paper.

“There are, of course, many reasons why people might want to increase the amount of plant-based food in their diets, apart from any effect that might possibly have on Covid infection risk. For example, there are concerns about the welfare of animals produced for human food, and about effects of animal production on greenhouse gases and climate change. But, because of the lack of certainty about cause and effect in this study, I’m not sure whether it adds much to those reasons.

“For clarity, I’ll mention that what is meant by the different diet types in this study might not match the ways the same words are often used in the UK. In the category of “plant-based diets” these researchers included what they call flexitarian (or semi-vegetarian) diets, which they use to describe people who do eat meat, but at most three times a week. Also the research couldn’t distinguish between people who don’t eat meat or fish but do eat eggs and/or dairy products, and vegans who eat only plant-based substances, because there weren’t enough vegans amongst their participants, so they are all put together into the vegetarian category.

“I would disagree with one specific issue in the paper, where the researchers write, “It is interesting that in our study when we analysed three dietary groups, we found that the lower incidence of COVID-19 in the plant-based group compared with omnivorous group was due mainly to the vegetarian group”, and then try to draw some conclusions from that. I don’t agree that their data supports this at all. In their fully adjusted model, the reduction in infection odds was estimated to be 39% in the vegetarian group compared to the omnivores, and very similar, a 38% reduction in odds, in the flexitarian group compared to the omnivores. Given the inevitable and considerable statistical uncertainty in these estimates, that’s not evidence of a difference in reduction between the vegetarian and flexitarian groups. It’s true that the difference, compared to the omnivores, is statistically significant for the vegetarians, but not quite statistically significant for the flexitarians. But comparing two subgroups using statistical significance in this way does not make good sense statistically. It looks very much as if both the vegetarians and the flexitarians have much the same reduction in Covid infection risk, compared to the omnivores, who eat meat more than three times a week, even though (as the paper points out) the food habits of the flexitarians are more similar to those of the omnivores than are the food habits of the vegetarians. But how this might be interpreted does depend in any case on what is causing what, and we can’t be sure of that anyway.”


‘Vegetarian and plant-based diets associated with lower incidence  of COVID-19’ by Júlio César Acosta-Navarro et al. was published in BMJ Nutrition, Prevention & Health at 23:30 UK time on Tuesday 9th January.


DOI: 10.1136/bmjnph-2023-000629



Declared interests

Prof Margaret Rayman: I have no interests which might be regarded by a reasonable and objective third party as giving rise to a conflict with your role as an expert in this story.

Dr Gavin Stewart: No conflicts of interest to declare

Dr Duane Mellor: I have no commercial conflicts of interest, but I choose for personal reasons (not health) to follow a vegetarian diet.

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  My quote above is in my capacity as an independent professional statistician.”

in this section

filter RoundUps by year

search by tag