Regulatory Focus™ > News Articles > 2019 > 6 > Regulatory Challenges in Medical Foods: Natural Variations in Ingredients of Agricultural Origin

Feature ArticlesFeature Articles
Posted 17 June 2019 | By Laurent Ameye, PhDClaudine BlacheHeinrich Schneider, MD  | ©

Regulatory Challenges in Medical Foods: Natural Variations in Ingredients of Agricultural Origin

3243 This article examines the sources of variability in raw materials from “natural origins” and how Foods for Medical Purposes (FSMP) companies cope with and overcome the challenges posed by regulated nutrient levels. The authors examine the sources of this variability using the examples of one vitamin (B12) and one trace mineral (selenium) and consider how FSMP companies work with the challenge of narrow acceptance criteria set by the regulations. The authors propose that regulatory authorities take the natural variation of nutrients in raw material from agricultural sources into account when they review their FSMP regulations and do so in addition to considering scientific reference values which are used as the basis for determining minimum and maximum levels of nutrients in FSMP.
Foods for Special Medical Purposes (FSMP) and their US-equivalent medical foods are legally defined as specially processed or formulated foods intended for patients with diseases or medical conditions whose dietary management cannot be achieved by modification of the normal diet alone. Like foods for the general population, FSMP must be safe and nourish the individuals for whom they are intended. The quality and safety of FSMP are ensured by the application of Good Food Manufacturing Practices (GMP) and compliance with ingredient purity criteria as well as microbiological and contaminant limits. The nutritional and medical suitability of an FSMP formulation must meet accepted scientific criteria.
In addition, in several countries and regions, the composition of FSMP must comply with minimum and/or maximum levels for vitamins and minerals in the finished product.1 For example, regulations of the European Union (EU), Australia, China and India stipulate permissible (and often quite narrow) ranges of nutrient variation. Most regulations allow deviations from these levels when it is scientifically justified and necessary for the intended use and target population of the product. However, compliance with these levels poses significant challenges for companies striving to use raw materials from natural origin since those materials vary “naturally” in their content of vitamins and minerals.
Nutrient Variability in Raw Materials
The vast majority of FSMP products are composed primarily of agricultural ingredients or ingredients derived from such materials. Using agricultural ingredients versus using chemically synthesized ingredients is cost effective, environmentally friendly and meets the growing societal preference for natural ingredients. Agricultural raw materials, however, present a significant intrinsic variability in their nutrient content due to variations between plant or animal breeds within the same biological species, the influence of different soils, climate and seasonal variations and differences in agricultural practices. The variability of vitamin B12 and selenium content in cow’s milk are outlined as examples below and the regulatory implications and challenges for FSMP recipe developers are discussed.
Variability of Vitamin B12 Content in Bovine Milk
Natural sources of vitamin B12 in human diets come from animal products, especially bovine milk. A 250 ml glass of milk from a cow not receiving B-vitamin supplementation can contribute up to 56% of the required daily intake of vitamin B12.2 It has been shown that the Vitamin B12 in bovine milk concentration varies with the breed. It is lower, for example, in Jersey cows than in Holstein cows.3,4 In addition, the milk content of vitamin B12 also varies between herds of the same breed and between individuals of the same breed, even in the same herd. The milk of one particular Holstein herd was found to contain on average up to 70% more vitamin B12 than the milk of another Holstein herd.5 Within a single herd of the same breed, it was shown that the milk of some cows may contain more than three times the amount of vitamin B12 found in the milk from another cow of the herd.6 The vitamin B12 concentration in bovine milk also varies between seasons; it is higher during spring and fall than during winter and summer.7-8
The variations in vitamin B12 in cow’s milk were linked to the finding that bacteria in the bovine rumen produce this vitamin and a portion of the synthesized vitamin is secreted into cow milk. Differences in ruminal vitamin B12 synthesis are explained partly by the genotype and the influence of the animals’ diet.9-10
As a consequence of the natural variability, the content of vitamin B12 in cow’s milk proteins used to manufacture infant FSMP by food business operators varies significantly by up to a factor of 10 (internal data, Table A). The Codex Alimentarius compositional criteria for vitamin B12 content in infant FSMP accommodate such natural variations in that they allow a variation in the final product by a factor of 15 (Min-GUL* values: 0.1-1.5 µg/100 kcal).11 However, the existing compositional criteria for infant FSMP in the European Union, for example, are much narrower, allowing only a variation by a factor of 5 (Min-Max values: 0.1-0.5 µg/100 kcal).12
*Guidance Upper Levels (GUL) are for nutrients without sufficient information for a science-based risk assessment. These levels are values derived on the basis of meeting infants’ nutritional requirements and an established history of apparent safe use. The purpose of the GULs is to provide guidance to manufacturers and should not be interpreted as goal values. Nutrient contents in infant formulas should usually not exceed the GULs unless higher nutrient levels cannot be avoided due to high or variable contents in constituents of infant formulas or due to technological reasons.
99% of the batches have a vitamin B12 content within the range of the mean ± 2.58 SD. As shown in Table A, a variability of ± 70% around the mean B12 content accounts for 99% of the batches.
Table A. Variability of Vitamin B12 Concentrations in Commercial Whey Powder Isolates (Internal Data)
Whey Powder Isolate:
Number of Batches
Vitamin B12
Mean Content (ppb)
Vitamin B12
Standard Deviation
Vitamin B12
Mean ± 2.58 SD
Vitamin B12
99% Confidence Interval
2.58 SD/Mean
Vitamin B12
Minimum Measured Value
Vitamin B12
Maximum Measured Value
2.31- 13.09
Variability of Selenium Content in Bovine Milk
The major sources of selenium in the human diet are bread, grains, meat, poultry, fish and eggs.13 The content of selenium in bovine milk varies between region and countries. For example, the selenium concentration in the milk of New Zealand cows is typically lower (5 mg/L)12 than that in the milk of cows in the United Kingdom (10 mg/L)14 or cows in the US (20 to 30 mg/L).15 The selenium content in cows’ milk is dependent on the content of selenium in the cows’ diet.16 In areas with plants deficient in selenium, the selenium content in milk from non-supplemented cows is low and dietary fortification increases the selenium milk content linearly.17 Most of the selenium in milk is associated with milk proteins (55–75% of the selenium in milk is associated with caseins, 17–38% with whey protein) and only 7% with milk fat.18 Seasonal variations in milk selenium content are also wide, reported to vary between 27 and 106 nmol/L in Frisian cows.19
As a result of natural variability, the content of selenium in whey proteins isolated from cows’ milk and used to manufacture FSMP by food business operators with a broad supplier base, also varies significantly, by up to a factor of more than 100 between measurement extremes (internal data, Table B). This natural variability is much wider than the accepted range for infant FSMP in the European Union that allows for a factor of less than 3 (Min-Max values: 3-8.6 µg/100 kcal)20 or by the Codex Alimentarius that allows for a factor of 9 (Min-GUL values: 1-9 µg/100 kcal).21
99% of the batches have a selenium content within the range of the mean ± 2.58 SD. As shown in Table B, a variability of ± 54% around the mean selenium content accounts for 99% of the batches.
Table B. Variability of Selenium Concentrations in Commercial Whey Powder Isolates (Internal Data)
Whey Powder Isolate:
Number of Samples Tested
Mean Content (ppb)
Standard Deviation
99% Confidence Interval
Mean ± 2.58 SD (ppb)
2.58 SD/ Mean
Minimum Measured Value
Maximum Measured Value
How to Reconcile Natural Variations With Narrow Acceptance Criteria
The examples above illustrate that natural variability in the composition of agricultural ingredients may result—if no interventions are taken—in a nutrient variability in FSMP larger than the variability permitted by regulatory levels. The following describes the measures FSMP companies can implement to manage this challenge.
One way to decrease the impact of high nutrient variabilities in agricultural raw materials on the final FSMP composition is to fortify the product with a concentrated source of the nutrients concerned. Because the amount of nutrient contribution from fortification is tightly controlled, fortification allows a decrease in variability in the final product for those nutrients displaying highly variable content in agricultural raw materials. The higher the contribution of the fortification, compared to the contribution of the raw materials, the lower the nutrient variability will be in the final product. However, this is only a feasible solution if the amount of the nutrient coming from the agricultural raw materials in the product is consistently low, thereby avoiding the risk that fortification pushes the content of the nutrient concerned in the final formulation above the permitted maximum limit.
In the theoretical example shown in Table C, it is assumed that milk proteins contribute 80% and fortification contributes 20% to the overall target content of 0.3 µg/100 kcal of vitamin B12 in an infant FSMP recipe for the European Union. In this example, the vitamin premix used for fortification displays a 15% variability in vitamin B12. The variability of vitamin B12 in the milk proteins is assumed to be 70% on the basis of our data shown in Table A. As shown in Table C, the fortification with vitamin B12 narrows the resulting variability in vitamin B12 content from 70% to 59% in the final FSMP product, thereby achieving compliance with EU infant FSMP levels for vitamin B12. However, as shown in the table, with such a variability, the minimum and maximum contents are very close to the regulatory levels (even if one excludes other sources of variability such as process and shelf-life losses and variations in analytical methods). Therefore, the only way to comply with these levels is to target a protein content in the middle of the allowed range. However, it is impossible to design a product with a lower or higher protein content, while complying with the selenium levels. When regulatory levels are defined, one should therefore, not only consider whether these levels accommodate the natural variability of raw materials, but also that sufficient room remains, once the natural variability is taken into account, to develop FSMP recipes with specific nutrient content to meet the diverse nutritional requirements of the patients.
Table C. Example of Contribution of Milk Proteins and Vitamin Fortification to the Total Natural Variability in Vitamin B12 of a Theoretical Infant FSMP Formulation
  Vitamin B12 Target
(µg/100 kcal)
Vitamin B12
Variability (%)
Vitamin B12
(µg/100 kcal)
Vitamin B12
(µg/100 kcal)
Vitamin B12
EU FSMP (iFSMP) Regulation 2016
(µg/100 kcal)
Milk Proteins 0.24 70% 0.072 0.408 NA
Fortification 0.06 15% 0.051 0.069 NA
Total 0.3 59% 0.123 0.477 0.1-0.5
Narrowing the raw Material Supplier Base
Another way to decrease the impact of high nutrient variabilities in agricultural raw materials on the final FSMP composition is to pre-select the raw materials. This can be done in a variety of ways. The product manufacturer can select raw material suppliers who can provide agricultural ingredients with a lower variability of the concerned nutrient. If the country of origin is an important source of the variability, for example, the selenium content in cow milk, a supplier of milk proteins sourced from a single country or region can provide milk proteins with a narrower range of selenium content than a supplier sourcing this raw material globally. The product manufacturer also may agree on tighter nutrient specifications with its suppliers. In such cases, it will be up to the supplier to ensure that narrower specifications can be compliant on the basis of its experience and knowledge of the ingredient, its source and production process.
Batch-to-Batch Selection (“Cherry Picking” Batches)
In the event that no supplier can be identified as being able to guarantee narrow specification limits for an agriculturally sourced raw material, using only those batches with narrower test results and rejecting the remaining batches is a last resort option for FSMP manufacturers. The obvious downsides for this option are significant cost implications and potential wastage of raw material from a valuable resource.
There is often a conflict between the FSMP manufacturers’ use of raw materials from agricultural origin and regulatory requirements, which set compositional criteria. In the case of the regulatory minimum and maximum levels for vitamins and minerals, fortification of the FSMP is usually the only viable way to solve this challenge. However, even this often fails if the amount of a vitamin or mineral coming from the agricultural raw material is high in a particular batch and the fortification leads to levels above the permitted maximum limit. The other potential approaches discussed in this article, i.e., narrowing the raw material supplier base and batch-to-batch selection, are not truly sustainable solutions due to their economic and environmental downsides.
We propose that regulatory authorities take the natural variation of nutrients in raw material from agricultural sources into account when they review their FSMP regulations, in addition to scientific reference values which are used as the basis for determining minimum and maximum levels of nutrients in FSMP. FSMP manufacturers and their suppliers should seek an exchange of information with regulators and other stakeholders to share their data and recommendations to inform this process. For example, where the current maximum level of a nutrient is the cause of the issue, this level should be increased where the negative consequences of non-compliance, because of natural variations in the raw material, outweigh the theoretical risk of harmful effects in a small number of patients. In particular, when No-Observed-Adverse-Effect Level (NOAEL) has been determined for a nutrient, regulatory maximum levels may not be justified. In the future, the outcome of this exchange of information between regulators and manufacturers should:
  • avoid the rejection and subsequent destruction of product batches made from valuable agricultural resources
  • allow FSMP manufacturers to have a broader supplier base, including local sourcing from both smaller and larger agricultural companies
  1. Ruthsatz M and Morck T. “Medical Food/Food for Special Medical Purposes: Global Regulatory Challenges and Opportunities.” Regulatory Focus. August 2016. RAPS.
  2. Graulet B. “Ruminant Milk: A Source of Vitamins in Human Nutrition.” Anim. Front. 4;2014:24–30.
  3. Anthony WB, Couch JR and Rupel IW. “Vitamin B12 in Blood of Newborn and Colostrum-fed Calves and in Colostrum and Normal Milk of Holstein and Jersey Cows.” J. Dairy Sci. 34;1951:749–753.
  4. Miller J, Wentworth J and McCullough ME. “Effects of Various Factors on Vitamin B12 Content of Cows' Milk.” J. Agric. Food Chem. 14;1996:218–221.
  5. Duplessis M, Pellerin D, Cue RI and Girard CL. Factors Affecting Vitamin B12 Concentration in Milk of Commercial Dairy Herds: An Exploratory Study.” J. Dairy Sci. 99;2016:4886–4892.
  6. Ibid.
  7. Ibid.
  8. Scott KJ, Bishop DR, Zechalko A and Edwards-Webb JD. “Nutrient Content of Liquid Milk I. Vitamins A, D3, C and of the B Complex in Pasteurized Bulk Liquid Milk.” J. Dairy Res. 51;1984:37–50.
  9. Rutten MJM, Bouwman AC, Sprong RC, van Arendonk JAM and Visker MHPW. “Genetic Variation in Vitamin B12 Content of Bovine Milk and its Association With SNP Along the Bovine Genome.” PLoS ONE. 8;2013:e62382.
  10. Castagnino DS, Seck M, Beaudet V, Kammes KL, Voelker Linton JA, Allen MS, Gervais R, Chouinard PY and Girard CL. “Effects of Forage Family on Apparent Ruminal Synthesis of B Vitamins in Lactating Dairy Cows.” J. Dairy Sci. 99;2016:1884–1894.
  11. CODEX Standard for Infant Formula and Formulas for Special Medical Purposes Intended for Infants. CODEX STAN 72-1981. Accessed 14 June 2019.
  12. Chun OK, Floegel A, Chung SJ, Chung CE, Song WO and Koo SI. “Estimation of Antioxidant Intakes From Diet and Supplements in US Adults.” J Nutr. 140;2010:317-24.
  13. Grace ND, Lee J, Mills, et al. “Death. Influence of Se Status on Milk Se Concentrations in Dairy Cows.” N.Z. J. Agric. Res. 40;1997:75–78.
  14. Givens DI, Cottrill RAB and Blake JS. “Enhancing the Selenium Content of Bovine Milk Through Alteration of the Form and Concentration of Selenium in the Diet of the Dairy Cow.” J Sci Food Agric. 84;2004:811–817.
  15. Maus RW, Martz FA, Belyea RL and Weiss MF. “Relationship of Dietary Selenium to Selenium in Plasma and Milk from Dairy Cows.” J Dairy Sci. 63;1978:532-537.
  16. Aspila P. “Metabolism of Selenite, Selenomethionine and Feed Incorporated Selenium in Lactating Goats and Dairy Cows.” J Agric Sci Finl. 63;1991:1–74.
  17. Op cit 15.
  18. Van Dael P, Vlaemynck G, Van Renterghem R and Deelstra H. “Selenium Content of Cows Milk and its Distribution in Protein Fractions.” Z Lebensm Untersuch Forsch. 192;1991:422–426.
  19. Knowles SO, Grace ND, Wurms K and Lee J. “Significance of Amount and Form of Dietary Selenium on Blood, Milk, and Casein Selenium Concentrations in Grazing Cows.” J Dairy Sci. 82;1999:429–437.
  20. Commission Delegated Regulation (EU) 2016/128 of 25 September 2015 Supplementing Regulation (EU) 609/2013 of the European Parliament and of the Council as in Regard to Specific Compositional and Information Requirements for Food for Special Medical Purposes. Accessed 14 June 2018.
  21. Op cit 11.
About the Authors
Laurent Ameye, PhD, is the global regulatory compliance manager at Nestlé Health Science, Switzerland with 20 years of experience in the infant and medical food industry in research and development, clinical research and regulatory affairs. He can be contacted at
Claudine Blache is head of the ingredient and recipe management group at Nestlé Health Sciences Switzerland, with 30 years of experience in food industry in product development and manufacturing. She can be contacted at
Heinrich Schneider, MD, is head of global regulatory affairs at Nestle Health Science in Epalinges, Switzerland with 30 years of experience in regulatory and medical affairs, clinical research and product safety. He can be contacted at Heinrich.Schneider@
Cite as: Ameye L , Blache C and Schneider H. “Regulatory Challenges in Medical Foods: Natural Variations in Ingredients of Agricultural Origin.” Regulatory Focus. June 2019. Regulatory Affairs Professionals Society.


© 2022 Regulatory Affairs Professionals Society.

Discover more of what matters to you

No taxonomy
No data found