Glutamate: Safe and Adequate Intake Levels for Infants—Should Breast Milk Be Taken Off the Market?
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
Overall Assessment
This manuscript presents a systematic critical analysis of the application of traditional toxicological risk assessment frameworks (particularly the NOAEL/BMD and uncertainty factor approach for ADI derivation) to nutrients and endogenous food components such as glutamate and human milk oligosaccharides, using breast milk and infant nutrition as the specific context. The paper is well-structured, richly supported by evidence, and logically rigorous, demonstrating strong academic value and practical relevance.
Specific Comments
The manuscript criticises EFSA's establishment of an ADI of 30 mg/kg bw/day for glutamate as overly conservative on multiple occasions (e.g., Sections 3 and 9). However, it does not sufficiently discuss the regulatory constraints under which EFSA operates (such as the requirement to base assessments on animal study data) or its precautionary approach as a regulatory authority. This criticism somewhat overlooks the tension between scientific validity and regulatory applicability. It is recommended that a paragraph be added in Section 3 or 9 to briefly explain the background and regulatory basis for EFSA's assessment, making it clear that the ADI was not derived from a disregard for nutritional science but rather from the mandatory framework for food additive evaluation. This would help balance the discussion and avoid an oversimplified critique of EFSA's work.
In the discussion on HMOs in Section 6, the authors emphasise their "unabsorbed, locally acting" characteristics, but do not adequately address the complexities of systemic exposure and long-term safety assessment. Additionally, the potential risks in specific sensitive subpopulations (e.g., infants with allergies) are not discussed. Furthermore, a more detailed methodological analysis of the alternative strategy used by EFSA in HMO assessments—comparison with breast milk exposure—would enhance the completeness of the argument.
The concept of risk–benefit assessment is mentioned in Sections 1 and 9, but is not systematically developed. Given that the central theme of this paper is a call to move beyond toxicity-driven assessment, incorporating a risk–benefit framework would help strengthen the conclusions.
Some content is repetitive, such as the description of glutamate metabolic mechanisms in Section 5 (physiological functions) and Section 8 (infant formula). Additionally, Figure 2 (page 15) is conceptually clear, but lacks detailed interpretation in the main text.
Lines 140–147: There is a slight inconsistency in the wording. The text states that EFSA identified a LOAEL of 3 g/day, but then says "did not establish a numerical UL because of limited data." Clarification is needed.
Lines 253–254: Glutamate is a non-essential amino acid, but classifying it as a "micronutrient" is inaccurate.
Line 671: The statement "This logic is further reinforced by EFSA's regulatory decisions on infant formula. Individual amino acids and amino acid mixtures are authorised as nutrient sources based on nutritional suitability and metabolic tolerance, not on derivation of additive-style acceptable daily intakes" references EFSA's regulatory decisions on infant formula but does not provide specific references. It is recommended that relevant EFSA opinions or guidance documents be cited to support this argument.
Author Response
Comment 1: The manuscript criticises EFSA's establishment of an ADI of 30 mg/kg bw/day for glutamate as overly conservative on multiple occasions (e.g., Sections 3 and 9). However, it does not sufficiently discuss the regulatory constraints under which EFSA operates (such as the requirement to base assessments on animal study data) or its precautionary approach as a regulatory authority. This criticism somewhat overlooks the tension between scientific validity and regulatory applicability. It is recommended that a paragraph be added in Section 3 or 9 to briefly explain the background and regulatory basis for EFSA's assessment, making it clear that the ADI was not derived from a disregard for nutritional science but rather from the mandatory framework for food additive evaluation. This would help balance the discussion and avoid an oversimplified critique of EFSA's work.
Response 1: Thank you for this important comment. We fully agree that EFSA operates within a clearly defined scientific and procedural framework and that it is not a regulatory authority but a risk assessment body providing scientific opinions to the European Commission and Member States, who take the final risk management decisions. We also recognise that EFSA does not automatically prioritise animal data but uses the most appropriate point of departure available, which may come from either animal toxicology or human data, depending on the substance and dataset. We have now added a paragraph in Section 3 (and briefly echoed this nuance in Section 9) to clarify that our critique is not directed at EFSA as an institution or at the scientific quality of its work, but at the limitations of applying a default food-additive style framework to nutrient-like substances such as glutamate. The added text emphasises that the group ADI of 30 mg/kg bw per day reflects methodological and mandate-driven constraints within this framework, rather than a disregard for nutritional and physiological evidence. We hope this revision makes the balance between scientific validity and regulatory applicability more explicit.
Comment 2: In the discussion on HMOs in Section 6, the authors emphasise their "unabsorbed, locally acting" characteristics, but do not adequately address the complexities of systemic exposure and long-term safety assessment. Additionally, the potential risks in specific sensitive subpopulations (e.g., infants with allergies) are not discussed. Furthermore, a more detailed methodological analysis of the alternative strategy used by EFSA in HMO assessments - comparison with breast milk exposure—would enhance the completeness of the argument.
Response 2: Thank you for this constructive comment. We agree that, although many HMOs act predominantly in the gut lumen, a fraction is systemically absorbed and detectable in plasma and urine of breastfed infants and formula‑fed infants receiving HMO‑fortified products, which has implications for long‑term safety evaluation. We have therefore revised Section 6 to acknowledge the evidence for systemic exposure, to briefly mention how EFSA’s assessments consider both local and systemic effects, and to highlight that the available clinical data to date have not identified safety concerns under the authorised conditions of use. We also agree that sensitive subpopulations, including infants with allergies or atopic predisposition, warrant explicit mention. The revised text now notes that EFSA’s evaluations consider allergenicity and immunological effects in line with its broader guidance and that no specific safety signals have emerged so far, while emphasising the importance of continued post‑marketing monitoring. Finally, we have expanded the methodological description of EFSA’s alternative strategy in HMO assessments, clarifying that safety is evaluated by comparing anticipated intakes from novel foods and formulas with the distribution of naturally occurring HMO intakes in exclusively breastfed infants, using updated human milk concentration databases and conservative exposure scenarios. We hope these additions strengthen the nuance and completeness of our discussion.
Comment 3: The concept of risk–benefit assessment is mentioned in Sections 1 and 9, but is not systematically developed. Given that the central theme of this paper is a call to move beyond toxicity-driven assessment, incorporating a risk–benefit framework would help strengthen the conclusions.
Response 3: Thank you for this helpful suggestion. We agree that risk–benefit assessment is a relevant framework for substances that have both essential physiological roles and potential adverse effects at high intakes, and that it aligns with our call to move beyond toxicity‑driven assessment alone. We have therefore clarified in Section 1 that risk–benefit assessment provides a complementary paradigm to classical toxicological risk assessment, and we have expanded the final part of Section 9 to indicate how the cases of vitamin C, iodine, glutamate, and HMOs naturally lend themselves to such an integrated approach. These additions explicitly acknowledge existing guidance on human health risk–benefit assessment of foods and emphasise that, for nutrients and nutrient‑like substances in infant nutrition, both risks and benefits should be weighed within a single coherent framework.
Comment 4: Some content is repetitive, such as the description of glutamate metabolic mechanisms in Section 5 (physiological functions) and Section 8 (infant formula).
Response 4: Thanks you for noting the overlap between section 5 and section 8. In response, we have reduced redundancy by concentrating the detailed physiological explanation of glutamate handling and ADI interpretation in section 5, and by streamlining the corresponding text in section 8. We have nevertheless retained a brief, partially overlapping example in section 8, because infant formula (particularly hydrolysed and amino acid–based products) provides a central, real‑world illustration of our argument and is important for readers who focus primarily on the regulatory and formula sections. We hope the reviewer agrees that this limited repetition improves readability while preserving the clarity of the key message for different readership interests.
Comment 5: Additionally, Figure 2 (page 15) is conceptually clear, but lacks detailed interpretation in the main text.
Response 5: Thank you for this comment. As you indicate that the implications of Figure 2 are not fully elaborated in the text, We have now added a short paragraph in the Discussion that explains how Figure 2 positions vitamin C, iodine, glutamate and HMOs within the spectrum between toxicity‑driven and context‑driven assessment, and how this schematic supports the main argument of the paper.
Comment 6: Lines 140–147: There is a slight inconsistency in the wording. The text states that EFSA identified a LOAEL of 3 g/day, but then says "did not establish a numerical UL because of limited data." Clarification is needed.
Response: 6: Thank you for pointing this out. We understand that the current wording may appear confusing. EFSA indeed identified a LOAEL of about 3 g/day for vitamin C based on gastrointestinal effects, but concluded that the available data did not allow a sufficiently robust derivation of a numerical UL for the general population. We have now clarified this sequence more explicitly in the text to avoid any impression of inconsistency.
Comment 7: Lines 253–254: Glutamate is a non-essential amino acid, but classifying it as a "micronutrient" is inaccurate.
Response 7: Thank you for this observation. We agree that referring to glutamate as a “micronutrient” could be misleading. We have now revised the wording to describe glutamate simply as a non‑essential amino acid and major metabolic substrate, without using the micronutrient terminology.
Comment 8: Line 671: The statement "This logic is further reinforced by EFSA's regulatory decisions on infant formula. Individual amino acids and amino acid mixtures are authorised as nutrient sources based on nutritional suitability and metabolic tolerance, not on derivation of additive-style acceptable daily intakes" references EFSA's regulatory decisions on infant formula but does not provide specific references. It is recommended that relevant EFSA opinions or guidance documents be cited to support this argument. We also exchanged EFSA by EC: the EC takes the risk management decisions.
Response 8: Thank you for this comment. We agree that the statement on EC’s decisions regarding infant formula and amino acid sources could be supported by explicit references. We have now added citations to EFSA’s scientific opinion on the essential composition of infant and follow‑on formulae, which emphasises nutritional suitability and clinical evaluation of formulas, including those based on protein hydrolysates, as well as to EFSA’s guidance on applications for infant formulae manufactured from protein hydrolysates. These documents illustrate that safety and suitability of protein sources and amino‑acid–based formulations are determined based on growth, tolerance and compositional criteria, rather than on derivation of additive‑style acceptable daily intakes.
Reviewer 2 Report
Comments and Suggestions for Authors
Foods-4245064 Glutamate: Safe and Adequate Intake Levels for Infants—Should Breast Milk Be Taken Off the Market
This manuscript addresses very important issues related to the setting of limits for food ingredients by EFSA. The authors use three examples to establish limits for vitamin C, iodine, and glutamate. The authors managed to almost mock the classic toxicological approach to setting limits in the case of glutamate, which is contained in breast milk completely naturally. It is worth recalling the 16th century statement by Paracelsus: "Only the dose makes the poison."
General questions
- Lines 319-325 generate question “What is valid glutamate dose when I loose position of breast-feed infant?
- Whole manuscript: EFSA generates limits for individual food additives. Who predicts limits for mixture these additives. We have on markets ultra-processed foods! See e.g. https://www.bbc.co.uk/food/articles/cpv8zpyzdnjo
- I suffered during reading by so many abbreviations in the text. I discovered list at the end of the text. I hope that some remark about list place can be done at the manuscript front page.
- EFSA behavior is very slow compared FDA. US nation like nontoxic food and FDA enables innovations more quickly than EFSA.
I did not find any other serious errors in the text, only some formal ones, and these are mentioned in the following list.
- Line 799 no doi code.
- Lines 831, 834, 837, 839, 844, 944 no doi codes
Author Response
Comment 1: Lines 319-325 generate question “What is valid glutamate dose when I loose position of breast-feed infant?
Response 1: Thanks you for this question. Our aim in lines 319–325 was not to propose a new numerical ADI, but to highlight that the current group‑ADI of 30 mg/kg bw per day is inconsistent with physiologically normal intakes, which are substantially higher in healthy breast‑fed and formula‑fed infants without evidence of harm. In our view, the most relevant benchmark for “valid” glutamate exposure in early life is therefore the intake range observed in these infant populations, combined with nutritional and clinical outcomes, rather than a single, lower ADI‑type cut‑off derived from animal neurotoxicity data. Hence we propose to keep the text as it stands.
Comment 2: Whole manuscript: EFSA generates limits for individual food additives. Who predicts limits for mixture these additives. We have on markets ultra-processed foods! See e.g. https://www.bbc.co.uk/food/articles/cpv8zpyzdnjo
Response 2: Thanks you for raising the broader issue of mixtures and ultra‑processed foods. Our manuscript focuses on glutamate, vitamin C, iodine and HMOs in early‑life nutrition, comparing EFSA reference/upper levels with the intakes observed in the physiological mixture of nutrients and bioactives in human milk and infant formula, rather than with complex additive mixtures in ultra‑processed foods. We note that EFSA has begun to develop approaches for combined exposure and mixture risk assessment, especially for substances with similar modes of action, but a detailed evaluation of such additive mixtures lies beyond the scope of this infant‑focused, nutrient‑centred review. Hence we propose to keep the text as it stands.
Comment 3: I suffered during reading by so many abbreviations in the text. I discovered list at the end of the text. I hope that some remark about list place can be done at the manuscript front page.
Response 3: Thanks you for this helpful remark. We agree that the number of abbreviations can make the text demanding to read, particularly for readers who are less familiar with this specific field, for whom we hope the review will be informative. To facilitate this, we explicitly point readers to the list of abbreviations early in the manuscript at the end of the introduction.
Comment 4: EFSA behavior is very slow compared FDA. US nation like nontoxic food and FDA enables innovations more quickly than EFSA.
Response 4: Thank you for this comment. Differences in timelines and procedures between regulatory authorities such as EFSA and FDA are well recognised and reflects, among other factors, distinct legal mandates, decision‑making structures and approaches to precaution and innovation. A comparative analysis of EFSA and FDA regulatory behaviour, however, falls outside the scope of the present manuscript, which focuses on how health‑based guidance values for specific substances relate to physiologically normal intakes in early life. We therefore prefer to keep our discussion centred on the scientific and regulatory principles underpinning the EFSA assessments that are directly relevant to the topics covered in this review.
Comment 5: Lines 799, 831, 834, 837, 839, 844, 944 no doi codes
Response 5: We thank the reviewer for carefully checking the references. For the first five items mentioned, no DOI codes are available because these are older reports/monographs that were not assigned DOIs; for these references we have now added the official access links and indicated them as “available online: …”. For the remaining two items, DOI codes are available and have been added to the reference list in the revised manuscript.
