piepttniorampt

s

lcnpgcitbtfp

crap(f

E

Regulatory Toxicology and Pharmacology 30, S103–S108 (1999)Article ID rtph.1999.1334, available online at http://www.idealibrary.com on

Identification of Risk Groups for Intake of Food ChemicalsPhilippe Verger,1 Isabelle Garnier-Sagne, and Jean-Charles Leblanc

Observatoire des Consommations Alimentaires, CNERNA, 16, rue Claude-Bernard, 75005 Paris, France

Received May 25, 1999

sdw

eso

acplcwStSsddcbSH(ot

iperc

F

fap

This paper summarizes by four recent examples therocedure of the characterization of groups at risk for

ngestion of food chemicals and its limitations. Thexamples concern two food additives (sulfites and as-artame), one nutrient (calcium), and one food con-aminant (patuline). On the one hand, results showhat the empirical description of a group at risk isowadays the best way to provide to risk managers

nformation useful for consumer protection. On thether hand, these examples show that in any case ifisk 0 does not exist, it is impossible to completelyvoid any group at risk. The responsibility of riskanagers is to decide what is the minimum size of the

opulation to protect as a function of the risk and ofhe cost of the protection.

Key Words: food intake; food chemical; monitoring;urvey.

INTRODUCTION

A group “at risk” is by definition a part of the popu-ation more exposed to hazard than the rest of theonsidered population. In the field of food safety, thisotion includes, on the one hand, factors linked to thehysiological or pathological characteristics of the sub-roup and, on the other hand, factors linked to theomposition of ingested foods. Schematically, the risks evaluated using the comparison between the inges-ion and the acceptable daily intake (ADI) or a tolera-le daily intake (TDI) which are the possible consump-ions of a chemical all along one’s lifetime without riskor health. For nutrients the intake is usually com-ared with the recommended daily allowance (RDA).The exposure assessment is based on several techni-

al approaches which depend on the level of precisionequired (FAO/WHO consultation, 1997). In any case,ccuracy depends on the size of the sample of theopulation and on the duration of the observationLowik, 1997). The most usual sample size is between 1or 20,000 and 1 for 50,000 people (1200 to 3000 per-

1 To whom correspondence should be addressed. Fax: 133/1.44.08.72.76.-mail: cnerna@inapg.inra.fr.

S103

ons for 60,000,000 inhabitants in France) and theuration of the inquiry is between 1 day and 1 yearith a most frequent period of 7 days.

MATERIALS AND METHODS

The following examples concern recent results onxposure assessment conducted in France by the Ob-ervatory of Food Consumption (OFC) and the Centref Research Foch (CRF).The French Observatory of Food Consumption man-

ges a database in which several sources of statisticsan be exploited. At first, the panel data provided by arivate institute (SECODIP) are used to represent theevel of French household consumption. These dataontain the purchase of more than 30,000 food productsithin 6000 households collected over a year. TheECODIP data are, at first, completed with data fromhe Food Survey of the French National Institute oftatistics and Economic Studies (INSEE). The INSEEurvey provides data simultaneously for home-pro-uced foods and for purchased foods. This allows theetermination of correction factors applied to pur-hased foods data from SECODIP in order to reachetter estimates of food consumption. Finally, theECODIP data are also completed with the Out ofome Food Survey managed by the CREDOC in 1994

Le Francois et al., 1996). For home production as forut of home consumption, correction factors are appliedo purchased foods data.

The Centre of Research Foch realizes specific inquir-es providing individual results of food consumption forarticular groups of consumers. The usual method tovaluate food consumption is based on 5-day dietaryecords self-administered by adults and teenagers andompleted by parents for younger children.

RESULTS

ood Additives

Three European directives imply an evaluation ofood additive intakes for all member states. The Frenchdministration and scientists have used a step-by-steprocedure (Verger et al., 1998) for 10 food additives

© 1999 International Life Sciences Institute, Washington, DCAll rights reserved.

p1r

filEw

cdpscnossmcumrUth

ePCc

da

uaspfit

tCFosataa

tiby2

a

tpftf

c

f

S104 VERGER, GARNIER-SAGNE, AND LEBLANC

reliminarily selected by the Budget Method (Hansen,979; Hallas-Moller, 1995) as to comply with theseequirements.

Exposure assessment to sulfites (E249–E250). Arst step consists in utilizing the maximum authorized

evels of additives in any food category according touropean regulation, in order to provide a screeninghich overestimates the risk.Results of exposure assessment of sulfites provide a

urve of distribution of consumption (Fig. 1) which isefined by a median and a high level (mean, 90 or 95thercentile) of exposure. The safety limit (ADI) can beuperimposed on the curve. Each level of exposureorresponds to a number of consumers and the totalumber of consumers exposed over the ADI constitutesne or more at-risk groups. Results of this first stephow that the theoretical maximum daily intake ofulfites is higher than the acceptable daily intake forore than 30,000,000 people (mean of consumers). A

orrected evaluation was performed using real levels ofse of sulfites provided by food manufacturers. Thisore refined step of evaluation shows that the cor-

ected mean is about two times lower than the TMDI.nfortunately, these corrections cannot be applied to

he high levels of consumption which are three timesigher than the ADI for 6,000,000 people in France.2

Figure 2 shows the characterization of high consum-rs for sulfites who are overrepresented in “Midi-yrenees” (southwest of France), Brittany, or “Provenceote d’Azur” and have a higher income per unit of

onsumption.

Exposure assessment to aspartame for diabetic chil-ren. Diabetics need to control their intake of sugars,nd low-calorie noncarbohydrate sweeteners may be

2 A corrected evaluation was performed using information fromhe industry. These data consist of real levels of sulfites used as aart of the considered food products. Information was not availablerom all manufacturers and corrections were applied to the mean ofhe curve of distribution regarding the importance of each correctedood on the whole market.

FIG. 1. Distribution of TMDI for sulfites.

sed in diabetic diets as partial alternatives to glucosend sucrose, allowing them to consume a wide range ofweetened foods. Since diabetics may thereby be ex-osed to higher than average amounts of sweetenersor a long time period, it is important to check thatntakes remain within the acceptable daily intake, par-icularly for children.

Intakes of sweeteners by the French general popula-ion were examined in 1994 by the Observatory of Foodonsumption and in 1997 by the Centre of Researchoch on selected populations of diabetics. Thus, thebjectives of these studies were to estimate the con-umption of three sweeteners (aspartame, saccharin,nd acesulfame K) widely used in low-calorie foods andabletop sweeteners by the French general populationnd insulin-dependent diabetic children and teenagerspriori considered as a group at risk.The average intake of aspartame calculated on the

otal population is 69.4 mg/day, the maximum level ofntake recorded being 392.6 mg. There are differencesetween girls and boys, i.e., average intakes for 15-ear-old (and over) girls exceed those for boys by up to0 mg/day.These results (Fig. 3) position this population of di-

betic children among the higher consumers in France

FIG. 2. The 10% higher consumers of sulfites (E220–E228) so-iodemographic pattern.

FIG. 3. Comparison of distribution curve for aspartame as aunction of the considered population.

f9

N

p

orpiCm

S105IDENTIFICATION OF RISK GROUPS FOR INTAKE OF FOOD CHEMICALS

or sweeteners (diabetic intake medians exceed the5th percentile indicated in French data; OCA, 1992).

utrients: Exposure Assessment to CalciumRecent results show that calcium consumption of a

art of the French population is lower than the rec-

FIG. 4. Distribution of calcium intake in a veg

mmended daily allowance. A particular group atisk for calcium deficiency is a priori the vegetarianopulation. A study of food intake and eating behav-or of vegetarians was conducted in 1997 by theentre of Research Foch. Total consumption waseasured using 5-day dietary records. Results show

rian population (2% of the French population).

eta

twftsTpt

C

gdira

S106 VERGER, GARNIER-SAGNE, AND LEBLANC

hat the distribution of calcium consumption for thehole population of vegetarians is not very different

rom that of the general population (Fig. 4). Never-heless, the study of distribution of calcium con-umption by vegans is dramatically lower (Fig. 5).his population represents 19% of the vegetarianopulation which itself represents about 2% of theotal population in France.

FIG. 5. Distribution of calcium intake in a v

ontaminants: Exposure Assessment to Patuline

Patulin is a toxic substance with suspected carcino-enic properties. Even though there is no formal evi-ence of risk, JECFA set a provisional tolerable dailyntake (pTDI) of 0.4 mg/kg body wt in 1995. This cor-espond to a daily maximum of 24 mg of patuline for andult weighting 60 kg, 8 mg for a child weighting 20 kg,

n population (0.4% of the French population).

ega

aoc5i1

afyi

m(op

dcTptpatpptebmopap

traTtbcuro

S107IDENTIFICATION OF RISK GROUPS FOR INTAKE OF FOOD CHEMICALS

nd 4 mg for a child of 10 kg. Human contamination cannly be envisaged from processed fruit and more pre-isely from apple juice. Considering a contamination at0 mg/liter of apple juice, an overstepping of the pTDIs possible by ingestion of 80 ml by a 10-kg child and of60 ml by a 20-kg child.Figures 6 and 7 show the consumption of apple juice

s a function of the considered population and as aunction of the duration of the study. The possibility foroung consumers to consume apple juice at a high levels clearly demonstrated.

In the population of children younger than 30onths (1.5 million children), 3% consume fruit juices

45,000 children); less than 50% are single consumersf apple juice (22,000 children) and 2.5% consume ap-le juice at a high level (550 children).

FIG. 6. Apple juice consumption as

FIG. 7. Apple juice consumption as a function of the method-logy.

CONCLUSIONS

This paper shows by four practical exercises that theetermination of group(s) at risk for consumption ofhemicals is technically difficult but certainly possible.he first example concerning sulfites shows that therocedure used for this evaluation has consistentlyried to characterize the group at risk for an overstep-ing of the ADI, a posteriori, without presumptionbout a particular target population. We concludedhat this procedure, based on the sociodemographicattern of consumers of sulfites, does not provide veryrecise information to risk managers in order to pro-ect the population. The second example of aspartamexposure assessment shows that a definition a priori,y an empirical assumption of the group at risk, isore powerful and more efficient than the previous

ne. The example of calcium shows the limits of thisrocedure a priori because in that case the real groupt risk is only a part of the initially defined targetopulation.The aim of the last example concerning Patuline is

o show the limits of risk assessment. The rule of theisk assessor is to provide useful tools to risk man-gers using scientific and transparent procedures.he rule of the risk manager is to take into accounthe scientific opinion, in order to reduce the proba-ility of the risk at a politically and financially ac-eptable level or to decrease the size of groups at risknder the same conditions. It seems fundamental toemember that just as risk “0” does not exist, it is

nction of the considered population.

a fu

itar

E

E

F

HH

Le Francois, P., Calamassi-Tran, G., Hebel, P., Renault, C., Lebre-

LP

M

O

V

W

S108 VERGER, GARNIER-SAGNE, AND LEBLANC

mpossible to avoid any group at risk. Moreover, thewo decisions concerning the acceptable level of risknd the acceptable size of the population at risk mustemain political choices.

REFERENCES

uropean Communities (1994a). European Parliament and CouncilDirective No. 94/35/EC of 30 June 1994 on sweeteners for use infoodstuffs. Official Journal No. L 237; 10.9.94, pp. 3–12.uropean Communities (1995). European Parliament and CouncilDirective No. 95/2/EC of 20 February 1995 on food additives otherthan colours and sweeteners. Official Journal No. L 61; 18.3.95, pp.1–40.

ood consumption and exposure assessment of chemicals. Report ofa FAO/WHO Consultation, 10–14 February 1997.allas-Møller, T. (1995). Personal communication.ansen, C. S. (1979). Conditions for use of food additives based on abudget for an acceptable daily intake. J. Food Protect. 42(5), 429–434.

ton, S., and Volatier, J. L. (1996). Food and nutrient intake outsidethe home of 629 French people of 15 years and over. Eur. J. Clin.Nutr. 50, 826–831.

owik, M. R. H. (1997). Personal communication.enttila, P. L., Salminen, S., and Niemi, E. (1988). Estimates on theintake of food additives in Finland. Z. Lebensm. Unter. Forsch.186, 11–15.AFF (1993). Dietary intake of food additives in the U.K.: Initialsurveillance, Food Surveillance Paper No. 37. HMSO, London.CA-CREDOC (1992). La consommation d’edulcorants de syntheseen France Evolution de 1989 a 1992, Collection des rapports del’Observatoire des Consommations Alimentaires, Rapport n° CP001.erger, Ph., Babayou, P., Chambolle, M., Le Breton, S., and Volatier,J. M. (1998). Estimation of the distribution of the maximum the-oretical consumption for 10 food additives, in France. Food Addit.Contam., in press.agstaffe, P. J. (1996). The assessment of food additives usage andconsumption: The Commission perspective. Food Addit. Contam.13(4), 397–403.