Regulatory Toxicology and Pharmacology 61 (2011) 115–118
Contents lists available at ScienceDirect
Regulatory Toxicology and Pharmacology
journal homepage: www.elsevier .com/locate /yr tph
Highest Observed Intake: Definition, regulatory uses and provisional values
John Hathcock a,⇑, Wantanee Kriengsinyos b
a Council for Responsible Nutrition, 1828 L St., NW, Suite 510, WA 20036, USAb Institute of Nutrition, Mahidol University, Phutthamonthon 4 Rd., Salaya, Phutthamonthon, Nakhon Pathom 73170, Thailand
a r t i c l e i n f o
Article history:Received 4 April 2011Available online 23 July 2011
Keywords:Highest Observed IntakeTolerable Upper Intake Level (UL)Acceptable Daily Intake (ADI)Risk assessmentToxicityRegulationsVitaminsBioactives
0273-2300/$ - see front matter � 2011 Elsevier Inc. Adoi:10.1016/j.yrtph.2011.07.001
Abbreviations: UL, Tolerable Upper Intake Level;Food and Agriculture Organization; WHO, World HealTrade Organization; SPS, Sanitary and PhytosanitaryMedicine; EFSA, European Food Safety Authority; SFood; HOI, Highest Observed Intake; NOAEL, No ObLOAEL, Lowest Observed Intake Level; EVM, ExpMinerals.⇑ Corresponding author. Fax: +1 202 204 7001.
E-mail address: [email protected] (J. Hathcock
a b s t r a c t
For nutrients and related substances, the Tolerable Upper Intake Level (UL) has become the internation-ally accepted risk assessment method. The major limitation of the UL method as applied by authoritativegroups thus far is that no UL has been set for nutrients without established adverse effects. In contrast tothe limitations inherent in the UL method, an alternative approach is available; it identifies a risk assess-ment value termed the Highest Observed Intake (HOI). In the absence of a UL, the HOI is the highestintake with adequate data to show, with acceptable confidence, the absence of adverse effects up to thatintake. With this concept defined and accepted in a report by the authoritative international organiza-tions FAO and WHO, a complete risk assessment for the nutrients and related substances would haveto identify UL values for those with known adverse effects and HOI values for those without knownadverse effects. The need for the HOI concept and value is illustrated by the unjustified policy and regu-latory actions taken in relation to vitamin B12. Regulatory utility is explained and tentative HOI valuesare identified for several vitamins and non-essential nutrients. Endorsement of the concept and develop-ment of HOI values by authoritative bodies is likely to be required for broad acceptance and use.
� 2011 Elsevier Inc. All rights reserved.
1. Introduction (or LOAEL) � UF. A common further step, often termed risk charac-
Policy and regulatory authorities have two basic approaches todetermining the safety of food products and components – (a) riskassessment based on toxicological or other appropriate data, and(b) evaluation of whether there is a ‘‘history of safe use’’. Some reg-ulatory policies specifically require risk assessment for many typesof ingredients. With food substances for which a Lowest ObservedAdverse Effect Level (LOAEL) or a No Observed Adverse Effect Level(NOAEL) can be identified, quantitative risk assessment commonlyinvolves application of an uncertainty factor to the LOAEL or NOAELto calculate an Acceptable Daily Intake (ADI) (National ResearchCouncil, 1983). For nutrients with known adverse effects, the Toler-able Upper Intake Level (UL) method has become internationallyaccepted. Application of the UL method involves a few simple steps:identification of the critical effect (the hazard occurring at the low-est intake), dose–response assessment, evaluation of uncertaintyand selection of an Uncertainty Factor (UF), and calculation of theUL (Institute of Medicine, 1998a). That is, UL = NOAEL
ll rights reserved.
UF, Uncertainty Factor; FAO,th Organization; WTO, WorldStandards; IOM, Institute ofCF, Scientific Committee onserved Adverse Intake Level;ert Group on Vitamins and
).
terization, is to compare the UL with the actual observed range ofintakes to determine whether some fraction of the population islikely be at risk of adverse effects. The major limitation of the ULmethod as applied by authoritative groups thus far is that no ULhas been set for nutrients without established adverse effects. Theabsence of an established critical effect has been interpreted as pre-cluding the derivation of a risk assessment value. Further, the ab-sence of a UL or comparable risk assessment value has been usedto argue for establishing maximums based on the RecommendedDietary Allowances (Domke et al., 2006) rather than safety as eval-uated by risk assessment.
2. Highest Observed Intake (HOI)
In contrast to the limitations inherent in the UL method, anexpanded alternative approach is, for nutrients for which no hazardhas been identified and no UL can be established, to identify a riskassessment value that is now officially termed the Highest ObservedIntake (HOI) (FAO/WHO, 2006), although the concept has also beencalled the Observed Safe Level (OSL) (Hathcock and Shao, 2008).Simply, in the absence of a UL, the HOI is the highest intake withadequate data to show, with acceptable confidence, of the absenceof adverse effects up to that intake, i.e., safety up to that level ofintake. The HOI is not literally the highest observed intake, but isthe highest with adequate data to support safety.
In the HOI procedure, the available data are screened to identifyany observable hazard, that is, a possible critical effect. If one is found,
116 J. Hathcock, W. Kriengsinyos / Regulatory Toxicology and Pharmacology 61 (2011) 115–118
the UL procedure is used. If none is found, the available evidence isevaluated to determine the highest intake for which the possibilityof an adverse effect is adequately excluded by the available data.The data evaluation procedures are exactly analogous to those in riskassessment for established hazards. With this concept defined andaccepted in a report by the authoritative international organizationsFAO and WHO, a complete risk assessment for the nutrients andrelated substances would identify UL values for those with knownadverse effects and HOI values for those without known adverseeffects. The demands for data should be similar for establishing ULand HOI values. If the data with toxicological utility are not sufficientto identify a UL or an HOI value, the more general evidence related touse should be evaluated to determine whether the history of usedemonstrates safety. Although multiple papers have been publishedon history of safe use (Sybesma et al., 2006; Constable et al., 2007;Health Canada, http://www.hc-sc.gc.ca/fn-an/consult/_novel_-foods/consultation_guidelines-directives09-eng.php), there is notyet any general agreement on the details of the method and typesof evidence that should be required.
The HOI has been defined by FAO/WHO (2006):
Highest Observed Intake – the highest level of intake observedor administered as reported within a stud(ies) of acceptablequality. It is derived only when no adverse health effects havebeen identified.
The HOI has been accepted by the Codex Alimentarius in itsNutritional Risk Analysis Principles and Guidelines for applicationto the Work of the Committee on Nutrition and Foods for SpecialDietary Uses (Codex Alimentarius Commission, Procedural Manual,19th Edition, 2010):
With this sanction, the HOI concept acquires global policy andregulatory importance because Codex is recognized as the preem-inent international authority on food safety by the World TradeOrganization in its Sanitary and Phytosanitary (SPS) Agreement(World Trade Organization, 2011).
To support developing international standards and regulations,the HOI concept and its acceptance by the Codex Alimentarius needsto be recognized, and generally accepted values should be devel-oped. In the reviews by the Institute of Medicine and the EuropeanCommission Scientific Committee on Food (now superseded by theEuropean Food Safety Authority, EFSA) no data were found to iden-tify any hazard related to high intakes of thiamin, riboflavin, vita-min B-12, biotin, and pantothenic acid (IOM, 1998b; SCF, 2000,b,2001a,b, 2002). By standard practice by the IOM and SCF/EFSA, aUL cannot be set if a NOAEL or LOAEL was not identified due tothe absence of established adverse effect or hazard. The UnitedKingdom’s Expert Group on Vitamins and Minerals did not definethe HOI or any similar concept, but identified an ‘‘Advisory Level’’on a similar basis for a few nutrients, including vitamin B12 (EVM,2003). They did not define or describe the procedure, but it seemsquite similar to the HOI definition.
Use of the HOI approach based exclusively on human data hasthe obvious limitation of the quantity and quality of those datasets.The choices are to use the usually more robust animal data andencounter the larger uncertainty of cross-species extrapolation,or to use the much less robust human data and avoid the uncer-tainty of this extrapolation. The net uncertainty could be reducedby combining the two approaches.
3. Application of the HOI
The need for the HOI concept and its value are illustrated by theillogical policy and regulatory actions taken in relation to vitaminB12. Apparently, the absence of a UL led to the misinterpretation
that there is insufficient data to evaluate the safety of this vitamin.This interpretation occurred despite the accumulation of largeamounts of human clinical data demonstrating a lack of adverse ef-fects of oral vitamin B12 at doses hundreds or thousands of timeshigher than the nutritional requirement or recommended intakes(USDA, 2007). Nonetheless, the absence of a UL for this vitaminhas led to the proposal or implementation of scientifically unjusti-fied and unnecessarily restrictive policies for maximum amountsthat may be incorporated into products. For example, Germanyproposed and later withdrew a regulatory limit of 9 mg for vitaminB12 (Domke et al., 2006) and France (le Ministre, 2006) imple-mented a limit of 3 lg in supplement products even though theamounts in unfortified conventional foods can range upward of100 lg per serving of some foods, e.g., beef liver (USDA, 2007).There is no public health purpose of such restrictions and the reg-ulatory actions taken by France and considered by Germany seemdisproportionate to the (lack of) hazard, and thus violate the Euro-pean Commission’s guidelines for proportionality in its documenton the precautionary principle (European Commission, 2000).
4. Provisional HOI values
The quantitative values identified in this section are offered on aprovisional basis to illustrate the methodology; we recognize thatfor wider acceptance and use, HOI values would need to be estab-lished by an authoritative scientific group such as the US IOM orthe EFSA.
The HOI risk assessment method (called the Observed Safe Levelby the authors) has been used in a series of published risk assess-ments for several non-essential nutrients, i.e. bioactive food com-ponents, including carnitine, chondroitin sulfate, coenzyme Q10,glucosamine, lutein, and lycopene (Hathcock and Shao, 2006a,b,2007; Shao and Hathcock, 2006). None of these substances hasany established causally related adverse effects, although most ofthem have been tested at a range of intakes in human clinical trialsand some have extensive supporting animal data. Some of thesedatasets are quite small and also vary in quality from one to an-other of these substances. Most of the studies are relatively shortterm, and confidence would be enhanced with longer term data.Acquisition of such data is difficult for multiple reasons – humanstudies seldom have safety measures as primary endpoints, andlonger-term studies are almost always planned to detect possiblelong-term benefits.
When risk assessments based on animal data were available,such as for lutein and lycopene, the HOI values were comparedwith the animal data ADI values (Shao and Hathcock, 2006). Inthese cases, any policy or regulatory decision based on the HOI val-ues would be more conservative.
The HOI approach has also been applied in risk assessments oftaurine, L-glutamine, and L-arginine, amino acids without knownspecific toxic effects. The HOI risk assessment method indicatesthe evidence for the absence of adverse effects is strong for taurineat supplemental intakes up to 3 g/d, L-glutamine at intakes up to14 g/d, and L-arginine at intakes up to 20 g/d (Hathcock and Shao,2008).
Although the United Kingdom Expert Group on Vitamins andMinerals (EVM) did not name the procedure, it applied the termGuidance Level in 2 ways: (1) in a manner analogous to the HOIdefinition, e.g. their 2000 lg guidance level for vitamin B12 (EVM,2003), and also (2) as a cautious term where they judged the datato be weak.
In the published risk assessments of lutein and lycopene (Shaoand Hathcock, 2006), the outcomes of the HOI method applied tohuman data were compared with standard risk assessments basedon animal data. With no known adverse effects (other than
Table 1Tentative Highest Observed Intake (HOI) values.
Nutrient Tentative HOI Uncertainty Factor (UF)applied
Rationale for UF Reference
Thiamin(hydrochloride)
100 mg (described as guidancelevel)
1 No pattern of toxicity at any intake; some trial data up to5000 mg
EVM (2003)
Riboflavin 40 mg (described as guidancelevel)
10 Clinical trial at 400 mg involved small numbers ofsubjects, requiring adjustment for uncertainty
EVM (2003)
Vitamin B12(cyanocobalamin)
2000 mcg (described asguidance level)
1 No adverse effects at any intake, but large clinical trial at2000 mcg
EVM (2003)
Biotin 900 mcg (described as guidancelevel)
10 Clinical trial at 9 mg for 4 years, but small number ofsubjects
EVM (2003)
Pantothenic acid 200 mg (described as guidancelevel)
10 Clinical trial at 2000 mg involved small numbers ofsubjects
EVM (2003)
Lutein 20 mg (human data); 38 mg(animal data)
1 human data; 1000animal data
No adverse effects; animal data described as NOAELgreater than 639 mg/kg in rats—equal to 38 g in a 60 kghuman
Shao and Hathcock(2006)
Lycopene 75 mg (human data); 270 mg(animal data)
1 human data; 1000animal data
No adverse effects; animal data described as NOAELgreater than 4500 mg/kg in rats—equal to 270 g in a60 kg human
Shao and Hathcock(2006)
Chondroitin 1200 mg 1 Multiple clinical trials; no adverse effects at any level Hathcock and Shao(2007)
Glucosamine 2000 mg 1 Multiple clinical trials at 1500 mg; no adverse effects atany level
Hathcock and Shao(2007)
Coenzyme Q10 1200 mg 1 Multiple clinical trials; various unrelated diseaseconditions
Hathcock and Shao(2006a)
Carnitine 2000 mg 1 Multiple clinical trials; various unrelated diseaseconditions
Hathcock and Shao(2006b)
J. Hathcock, W. Kriengsinyos / Regulatory Toxicology and Pharmacology 61 (2011) 115–118 117
carotenodermia), the lutein and lycopene HOI were restricted so-lely by the dosages used in good quality clinical trials. The riskassessments based on animal data employed 1000-X uncertaintyfactors, and conversion to a standard human body weight of60 kg. The human data HOI for lutein indicated a value of 20 mg/d, but the animal data risk assessment suggested that 38 mgshould be safe for human adults. Similarly, the human data HOIfor lycopene led to identification of a 75 mg/d HOI, but the animaldata risk assessment indicated that 270 mg/d should be safe forhuman adults.
In the definition and policy context described, authoritativegovernment groups need to review the nutrients to identify HOIas well as UL values. Since no authoritative group has yet publishedsuch work, we provide the provisional HOI values from previouspublications. See Table 1 for several guidance levels identified bythe EVM, and OSL (equivalent to HOI) values from the peer-re-viewed literature. We emphasize that these HOI values are the re-sult of risk assessment; the data were assessed for evidence ofadverse effects and none was found, but standard dose–responseevaluation indicated that the data should be judged reliable onlyup to the levels suggested as HOI values.
5. Regulatory uses of the HOI
In development of regulatory values, it may be argued that nomaximums should be set on ingredients that have no establish ad-verse effects even though there is a lot of relevant human data, e.g.,vitamin B12. Nonetheless, some policy makers and regulatoryauthorities prefer to establish maximums for all added ingredients.For substances with substantial data and no known adverse effects,any maximums should be based on the HOI, and not on the RDA orany other measure of nutritional need in order to be consistentwith the Codex guideline for vitamin and mineral food supple-ments (Codex Alimentarius Commission, 2005).
If a nutrient has no UL, the setting of a maximum on the basis ofthe RDA instead of establishing and applying a HOI value couldlead to the illogical situation of greater restrictions for such a vita-min than for another with a clearly established adverse effect. Forexample, vitamin B6 (pyridoxine) has official UL values of 100down to 10 mg/day (IOM, 1998b; EVM, 2003; SCF, 2000c). The
lowest of these UL values is more than 5� the RDA. In contrast,use of a low, simple multiple of the RDA for vitamin B12 is morerestrictive (approximately 1� for 3 l and 3� for 9 lg) (Domkeet al., 2006; Le Ministre de l’Economie, des Finances et de l’Indust-rie, 2006). Other regulatory authorities should avoid such unscien-tific outcomes. Note: The French Council of State recently annulledthe vitamin B12 limit and a few other provisions of the ‘‘arrête’’ of2006, but did not replace these maximums with any other values(Conseil d’Etat, 2011).
The definitions and guidelines exist for recognition and applica-tion of the HOI concept when there is a regulatory decision toestablish maximums for nutrients with no known adverse effects.Authoritative national, regional, or international scientific commit-tees that conduct reviews and analyses to establish UL valuesshould also identify HOI values, where applicable. In the absenceof such actions, we suggest consideration of the values in Table1. The level of confidence in an HOI value will depend on therobustness of the available dataset. Thus, the HOI values identifiedwill need a proportional uncertainty value. The uncertainty factorssuggested in Table 1 range from 1.0 to 10.
It must be recognized that both the UL and HOI values (evenafter adjustment for uncertainty related to the strength of the data-set) are risk assessment values. We acknowledge that risk manage-ment decisions must consider the multiple sources of many foods,including natural occurrence, food fortification, and the contents ofsupplements. Regulatory maximums developed from either HOI orUL values may be less than the corresponding risk assessmentvalues.
Conflict of interest statement
Hathcock – employed by Council for Responsible Nutrition;Kriengsinyos – employed by Mahidol University – no paymentfor work on manuscript.
References
Codex Alimentarius Commission, 2005. Guidelines on vitamin and mineral foodsupplements, CAC GL 55 – 2005. <www.codexalimentarius.net/download/standards/10206/cxg_055e.pdf> (accessed 28.06.11).
Codex Alimentarius Commission, 2010. Procedural Manual, 19the Edition, PDF page123.
118 J. Hathcock, W. Kriengsinyos / Regulatory Toxicology and Pharmacology 61 (2011) 115–118
Constable, A., Jonas, D., Cockburn, A., Davi, A., Edwards, G., Hepburn, P., Herouet-Guicheney, C., Knowles, M., Moseley, B., Oberdörfer, R., Samuels, F., 2007 Dec.History of safe use as applied to the safety assessment of novel foods and foodsderived from genetically modified organisms. Food Chem. Toxicol. 45 (12),2513–2525.
Conseil d’Etat. Décision n�295235, date du 27 avril 2011.Domke, A., Grossklaus, R., Niemann, B., Przyrembel, H., Richter, K., Schmide, E.,
Weissenborn, A., Wörner, B., Ziegenhagen, R., 2006. Use of vitamins in foods.Federal Institute for Risk Assessment, Berlin.
European Commission. Communication from the Commission on the precautionaryprinciple. Brussels, 2000.
Expert Group on Vitamins and Minerals. Safe Upper Levels for Vitamins andMinerals. Food Standards Agency, London, UK, 2003.
FAO/WHO. A model for establishing upper levels of intake for nutrients and relatedsubstances. Geneva: FAO/WHO Technical Workshop on Nutrient RiskAssessment, 2006.
Hathcock, J., Shao, A., 2008. Expanded approach to Tolerable Upper Intakeguidelines for nutrients and bioactive substances. J. Nutr. 138, 1992S–1995S.
Hathcock, J.N., Shao, A., 2007. Risk assessment for glucosamine and chondroitinsulfate. Regul. Toxicol. Pharmacol. 47 (1), 78–83.
Hathcock, J.N., Shao, A., 2006a. Risk assessment for coenzyme Q10 (Ubiquinone).Regul. Toxicol. Pharmacol. 45 (3), 282–288.
Hathcock, J.N., Shao, A., 2006b. Risk assessment for carnitine. Regul. Toxicol.Pharmacol. 46 (1), 23–28.
Institute of Medicine, 1998a. Dietary reference intakes: a risk assessment model forestablishing upper intake levels for nutrients. National Academy Press,Washington, DC.
Institute of Medicine, 1998b. Dietary reference intakes: thiamin, riboflavin, niacin,vitamin B-6, folate, vitamin B-12, pantothenic acid, biotin and choline. NationalAcademy Press, Washington, DC.
Le Ministre de l’Economie, des Finances et de l’Industrie. Decrets, aretes,circulaires–textes generaux. Arrete de 9 mai 2006 relatif aux nutrimentspouvant etre employes dan la fabrication de complements alimentarires.Journal Officiel de la Republic Francxaise; 2006.Texte 7 sur 62.
National Research Council, 1983. Risk assessment in the federal governmentmanaging the process. National Academy Press, Washington, DC.
Scientific Committee on Food. Opinion of the Scientific Committee on Food on theTolerable Upper Intake Level of Vitamin B1. European Commission SCF/CS/NUT/UPPLEV/46 Final, 2001a.
Scientific Committee on Food. Opinion of the Scientific Committee on Food on theTolerable Upper Intake Level of Vitamin B2. European Commission SCF/CS/NUT/UPPLEV/33 Final, 2000a.
Scientific Committee on Food. Opinion of the Scientific Committee on Food on theTolerable Upper Intake Level of Vitamin B12. European Commission SCF/CS/NUT/UPPLEV/42 Final, 2000b.
Scientific Committee on Food. Opinion of the Scientific Committee on Food on theTolerable Upper Intake Level of Vitamin B6. European Commission SCF/CS/NUT/UPPLEV/16 Final, 2000c.
Scientific Committee on Food. Opinion of the Scientific Committee on Food on theTolerable Upper Intake Level of Biotin. European Commission SCF/CS/NUT/UPPLEV/55 Final, 2001b.
Scientific Committee on Food. Opinion of the Scientific Committee on Food on theTolerable Upper Intake Level of Pantothenic Acid. European Commission SCF/CS/NUT/UPPLEV/61 Final, 2002.
Shao, A., Hathcock, J.N., 2006. Risk assessment for the carotenoids lutein andlycopene. Regul. Toxicol. Pharmacol. 45 (3), 289–298.
Sybesma, W., Hugenholtz, J., de Vos, W.M., Smid, E.J., 2006. Safe Use of geneticallymodified lactic acid bacteria in food. Bridging the gap between consumers,green groups, and industry. Elec. J. Biotechnol. 9, 424–448.
USDA, National nutrient database for standard reference, release 20. 2007 VitaminB-12 (mcg) content of selected foods per common measure. Available from:<http://www.nal.usda.gov/fnic/foodcomp/Data/SR20/nutrlist/sr20w418.pdf>(accessed 28.06.11).
World Trade Organization. Available from: <http://www.wto.org/english/tratop_e/sps_e/spsagr_e.htm> (accessed 28.06.11).
