Factors affecting the determination of threshold doses for - AInotes

26 Taylor et al J ALLERGY CLIN IMMUNOLJANUARY 2002the offending food, including milk in 9 incidents at levelsranging from 0.04% to 1.1% (wt/wt), egg in 3 incidentsat levels ranging from 0.003% to 0.16% (wt/wt),wheat gluten in 2 incidents at levels ranging from 0.3%to 1.3% (wt/wt), soy protein in 2 incidents at levels rangingfrom 0.5% to 7.0% (wt/wt), and hazelnut in 1 incidentat a level of 0.2% (wt/wt). Quantitative exposureinformation was available in several of these cases. Fatalanaphylaxis occurred in a consumer with milk allergyafter eating approximately 100 g of a sausage productcontaining 0.06% undeclared casein that equated to 60mg of casein. 8 An asthmatic reaction occurred in an individualwith hazelnut allergy after ingestion of about 3 to6 g of a chocolate confection containing 0.2% of undeclaredhazelnut that equated to 6 to 12 mg of hazelnut. 8Although such episodes reveal the hazards posed byundeclared residues of allergenic foods contaminatingother foods, they are not particularly suitable for establishingthe lowest provoking dose. For example, the circumstancesof exposure, including the amount and typeof foods eaten during the episode, can only be determinedanecdotally. Also, the reliability of the analyticresults can be questioned because validated, collaborativelystudied, standard methods are not yet available.CLINICAL APPROACHES TO THRESHOLDDOSE ESTIMATIONThe best estimates of the threshold dose for variousallergenic foods can be obtained from controlled clinicalchallenge trials. In only a few cases were such trials intendedspecifically to determine the threshold dose. 5,9 More frequently,challenges have been conducted for diagnosticpurposes rather than for determining the lowest provokingdose. However, typical protocols involve starting at dosesthat are one half or less of the amount of the offending foodestimated by the patient to provoke symptoms. 10Because low doses of allergenic foods are sometimesused in the diagnosis of food allergies, especially inpatients with histories of serious allergic reactions, thepossibility exists that data on the threshold doses forcommon allergenic foods might exist in certain allergyclinics. Consequently, in September 1999, the FoodAllergy Research and Resource Program at the Universityof Nebraska, with sponsorship from the food industry,convened a roundtable conference entitled “ThresholdDoses for Allergenic Foods: How Much Is Too Much?”RESULTS OF THE ROUNDTABLESeveral clinical groups did possess potentially usefulinformation relating to the threshold doses for commonlyallergenic foods derived from the results of past doubleblind,placebo-controlled, food challenges (DBPCFCs)used for diagnostic purposes. The most useful data wereavailable on peanuts, eggs, and cows’ milk and will bereported here. Some data were also available on fish andmustard seed. Data were also available for soybeans,wheat, sesame seed, tree nuts, and crustaceans, but thenumber of patients tested was too few, and in some cases,the nature of the challenge materials was not preciselyknown (eg, the variety of tree nuts, the form of soybean,and the method of processing or species of crustacean).Table I contains the results of the accumulated data onthe lowest provoking doses in DBPCFCs for the variousinvestigators involved in the roundtable conference forpeanuts, eggs, and cows’ milk. For peanut, results wereshared on DBPCFCs (and some single-blind, placebocontrolledfood challenges [SBPCFCs] and open challenges)of 306 patients. The lowest provoking dose was 1mg of peanut, which was seen with 4 patients; all of thesepatients were from the clinic of Dr Fabienne Rance inFrance. The range of lowest provoking doses for thesepatients was 1 mg to 5 g. For eggs, results were shared onDBPCFCs (and some SBPCFCs and open challenges) of281 patients. The lowest provoking dose was 1 mg of liquidwhole egg, which was seen with 2 patients, both fromthe clinic of Dr Rance in France. The range of lowestprovoking doses for these patients was 1 mg to 5 g. Forcows’ milk, the results are much more difficult to interpretbecause several different forms of cows’ milk wereused in the challenges: liquid cows’ milk, nonfat drymilk, and infant formula. For cows’ milk, results wereshared on DBPCFCs (and a few SBPCFCs and openchallenges) of 299 patients. The lowest provoking dosewas 0.02 mL of milk, which was seen in 21 patients; allof these patients were from the clinic of Dr David Hillfrom Australia. The range of lowest provoking doses forthese patients was 0.02 mL to greater than 100 mL. Forfish, the lowest provoking dose was 5 mg of either cod orherring, as determined from the results of DBPCFCs(and some SBPCFCs) on 32 patients. However, thespecies of fish used in these challenges were variable,and this appeared to affect the results, as noted fromcomparative information obtained on 14 patients for 4different species by Dr Carsten Bindslev-Jensen in Denmark.Fifteen individuals with mustard allergy have beenevaluated by means of DBPCFCs in France, and the mostsensitive individual reacted to 1 mg of ground mustardseed. To apply this information in the context of a typicalserving of food containing undeclared residues of anallergenic food, 1 mg of an allergenic food in a typical50-g serving would be equivalent to 20 mg of allergenicfood per kilogram of the ingested food product (20 ppm).DISCUSSIONClearly, sufficient results are available to conclude thatthe threshold doses for commonly allergenic foods arefinite, measurable, and above zero. However, attemptingto reach consensus on the threshold doses for peanut,egg, cows’ milk, fish, and mustard on the basis of theexisting data would probably be premature for a numberof reasons. The number of patients who have been subjectedto DBPCFCs with low doses of the offending foodis probably sufficient to estimate the threshold dose atleast for peanut, egg, and cows’ milk. However, thenature of the data suggests that it may be premature to