Reacción a "A large study analyzes the benefits and risks of vegetarian and vegan diets in children"

The article presents acceptable methodological quality and appears to follow the standards expected in a systematic review with meta-analysis: it has prior registration, details the search strategy, explains the selection criteria, and describes the process of critical appraisal, data extraction, and statistical analysis. This does not mean that there are no areas for improvement (all reviews have them), but at least it reports the fundamental elements that allow its transparency and rigour to be assessed. My assessment is preliminary because an in-depth analysis would require more time.

This meta-analysis fits reasonably well with the existing evidence: it confirms that vegetarian and vegan children tend to consume less energy and some micronutrients than omnivores, but that does not automatically imply that their intake is insufficient. In many cases, it simply reflects that omnivores tend to consume above the recommendations, while vegetarian families tend to better adjust their diet to the recommendations.

An important nuance is that the meta-analysis's overall table mixes studies from countries with very different socioeconomic levels. In low-income contexts (especially in India, which contributes many studies), the differences observed may not be due to the vegetarian pattern itself, but to inequalities in access to food. Therefore, some of the differences in intake and biochemical markers could reflect poverty rather than an inherent risk in the diet. But again, the authors say, and I quote: ‘However, average energy and protein intake remained, in general, within the recommended ranges.’ ‘In both ovo-lacto vegetarians and vegans, most mean values, including ferritin and 25(OH)D [active form of vitamin D], remained, in general, within paediatric reference ranges.’

In high-income countries, where access to food and supplements is adequate, the results show differences in intake but few signs of overt deficiencies, except for the possible increased likelihood of anaemia in lacto-ovo-vegetarians.

The main limitation is that the meta-analysis compares average intakes between groups but does not report how many children actually fall short of the recommended values. This makes it difficult to know whether lower intake is clinically relevant or simply reflects more moderate but adequate consumption.

Another limitation is the mix of socioeconomic contexts (without access to the analysis that the authors have done but do not report in full): data from low-income countries may amplify differences and are not fully comparable with those from Spain or Europe. Biochemical markers are also not adequately broken down by income level, which is key to interpreting ferritin, vitamin A, D or E, given that these deficiencies are much more common in vulnerable populations regardless of dietary patterns. In these settings (low-income countries), being vegetarian may be associated with lower purchasing power or less access to food, which could distort the conclusions.

There are four statements made by the authors that should be given serious consideration:

• ‘Overall, our findings suggest that these diets can promote healthy growth and offer certain health advantages. However, they also present specific nutritional challenges that require careful dietary management and ongoing clinical attention.’

This statement summarises the data quite well. The meta-analysis finds no consistent evidence that vegetarian or vegan children in high-income countries grow less well or have serious clinical deficits. However, lower intakes and some reduced biomarkers are observed, which in principle do not imply disease but do justify regular monitoring and appropriate dietary planning. In other words, these diets are viable and safe, but they cannot be improvised.

• ‘While total protein intake generally meets recommendations, the quality of plant-based proteins may be lower due to suboptimal proportions of certain essential amino acids, particularly when dietary variety is limited.’

This is consistent with the evidence: vegetarian and vegan children consume sufficient protein, but not always with an optimal amino acid profile if their diet is based on very few plant foods. However, in a minimally varied diet, such as that which any family in our context can follow, this aspect does not represent a real problem. The availability of protein-rich plant foods, the regular presence of complementary food combinations, and widespread access to nutritional information make the risk of suboptimal protein intake virtually non-existent. This is not an inherent limitation of vegetarian diets, but a reminder that, like any dietary pattern, they require variety and basic planning, which is perfectly achievable in high-income societies such as ours.

• ‘Among micronutrients, vitamin B12 emerges as the most critical concern. Recent findings confirm that well-supplemented vegan children can achieve or even exceed the vitamin B12 levels of omnivores, reinforcing the need for systematic supplementation to prevent serious consequences such as megaloblastic anaemia and irreversible neurological impairment.’

This statement carries particular weight because vitamin B12 is not available in plant foods. However, rather than providing new information, this meta-analysis reinforces already well-established evidence: properly supplemented vegan children achieve adequate levels of B12 and may even have a status equal to or higher than that of omnivores. Therefore, the key point is not to discover something new, but to confirm once again that the risk does not lie in the plant-based diet itself, but in the absence of supplementation, an aspect that in our healthcare context is well understood and easy to manage.

• "Iron levels are another crucial aspect, given the essential role of this mineral in growth and cognitive development. Our analysis revealed that ovo-lacto-vegetarian and vegan children tend to have significantly higher iron intake than their omnivorous peers, but have lower ferritin concentrations and a higher probability of iron deficiency and anaemia. In our analysis, anaemia was significant in both low/middle-income and high-income countries, underscoring the need to monitor iron levels regardless of income level."

This is one of the most interesting observations of the meta-analysis, but also one that calls for caution. The authors interpret that although vegetarian and vegan children tend to consume more total iron—from non-haem iron in plant foods, which is less well absorbed than haem iron from animal sources—this lower absorption could explain the lower ferritin levels and higher likelihood of anaemia observed. The explanation makes physiological sense, but it is unclear whether the intake and ferritin data come from the same subjects. When comparing aggregate results from different studies, there is a risk of committing an ecological fallacy, i.e., assuming individual relationships based on patterns that are actually only observed at the group level (and in this case, possibly in different studies).

The meta-analysis combines studies where some report intakes and others report biomarkers, and these data are not always directly related. Added to this is another important limitation: most studies are cross-sectional, so we cannot rule out reverse causality. That is, it may be that children with low ferritin have been advised to increase their iron intake, which would produce exactly the pattern we see: higher intakes along with lower ferritin levels, but without intake being the cause of the reduced biomarker. Therefore, although the authors' interpretation is plausible, we cannot say with certainty that the higher proportion of non-haem iron explains the lower ferritin levels. Well-designed longitudinal studies or trials are needed to simultaneously analyse intake, absorption and biomarkers in the same individuals.