doseof a material and the responseit produces. In conventionaltoxicology,this
is usuallyachievedby experimentsin suitableanimalspecies. These provide
information aboutdose±response relationshipsand resultin the definition of a
no-observedadverse effectlevel(NOAEL) fromwhicha safe dosefor peopleis
calculated.Hazardcharacterisation in foodallergydiffers fromthis situation,
firstlybecauseanimalexperimentationis irrelevant,and secondly because it
can be viewedat both an individual level and a populationone.The population
dimension is probably the most relevant fromthe publichealthpointof view,
and consequently for the foodmanufacturer,whilethe individualdimension is
most relevant to the clinician advising a patient on management of the
condition.
For ethical reasons it is extremely difficultto obtain information about
individual responsiveness to differentdosesof allergen.Full characterisation,up
to the pointof the mostsevere reaction can onlyoccurinadvertently, wherethe
doseincrement usedprovesto be too large.Characterisationof the population
responseto foodallergensis, however, feasible and ethical,in the context of
helpingpatientsmanagetheircondition. This is achievedthroughstudiesusing
double blindplacebo-controlled food challenge (DBPCFC)withincreasing
dosesof allergenup to the lowestdoseproducingan objectiveresponse.Such
experimentsprovide information on the frequencyof responseto particular
doses in the populationtested. However, such studies have a number of
limitationsas toolsto identifya precise NOAEL.Firstly,because of their
logistics, theycan be performedonlyin relativelysmall numbers of allergic
individuals,whichlimitstheirstatistical power. In a typicalclinicalstudy,based
on binomial distribution, 29 or 58 individuals wouldbe challenged,and it can be
shownthat,statistically,suchnumbers give95%confidence that fewerthan
10%or 5% respectively of the populationfromwhichtheyare drawnwould
react to the dose identified as the NOAEL.^7 A further limitation is that
individualswho haveexperienceda severereaction to the allergen of interestare
excluded from the studies by some clinicians. A modelling approach to
overcome suchproblems has recentlybeenproposed.^8 Although this approach is
provingpromising, muchvalidation work is still required beforeit can be used
in risk management.
Oncethe hazardhas beencharacterised,and a NOAEL defined, exposure
assessment is required.For individualswitha food allergy,the risk arises
fromacuteingestionof oftenquitesmallamountsof allergen,ratherthan
ingestionovera periodof time.The usualmeasureto considerfor purposes
of risk assessment is the amountof allergenthat couldbe presentin a portion
of food.An importantunresolvedissueexists,however, withrespectto the
periodoverwhichthe intakeof allergenshouldbe summed, as wellas the
possibleeffecton the provokingdoseof exposure to smallamounts,unable
by themselves to stimulate an allergic reaction.Indeedsome publications
haveattributedincreased reactivity to exposure to suchdosesvia other routes
thanthe oral one.^9
Managingrisksfromallergenicresidues 365