Allergen cross-contacts can occur at several steps of the food production chain and the amount of allergen potentially present in the final food is variable. This is the reason why Health Canada recommends avoidance by food allergic consumers of all food and products that carry a precautionary statement warning [22]. However, with the notable exception of milk for which there is both high prevalence and concentration in chocolate-based and other food products [23–25]), several studies have shown that many food products with PAL contain no or only very low amounts of allergens [26]). Recent Canadian studies reported that 90% of products with a PAL for egg, peanut and hazelnut had absent or very low concentrations of allergens (i.e. below the limit of quantification of detection tests) [21, 23]. For hazelnut, the remaining 10% had a relatively high mean concentration (range 0.4 to 2167 ppm) but this is an allergen that presents with a risk of heterogeneous (variable amounts of allergen residues in the same product batch) and high concentration contamination [21].
As reported in previous studies [27, 28] this overuse of PAL and possible past experiences by some of consuming foods with PAL without any adverse reactions probably explain why our study reported that more than half of Canadian allergic consumers purchase products that have a PAL statement. This is a higher reported consumption of products with PAL by the food-allergic community compared to 2014 when only 15 to 37% declared consuming products with a precautionary statement [29]). This attitude might have been partly reinforced by the medical community allowing foods with PAL in some patients who have a high reactivity threshold [30].
As in several other studies [9, 29, 31], our results show that there is a lack of consensus on how consumers who manage food allergies interpret precautionary statements. Most commonly, they assume that precautionary allergen statements such as “may contain” mean that a low level of allergens may or may not be in the product. However, even with uncertainties regarding this labelling, more than half of consumers in this survey consider PAL as a useful tool; seeing PAL on food products makes them have greater trust in the brand and more confidence that the product will be safe if the allergen is on the PAL statement and not in the ingredients. In fact, many consumers dealing with food allergies have concerns about the safety of food products that do not contain any PAL believing they may not be safe. Physicians are more skeptical since half think PAL is not useful in its current form. Importantly, the combination of PAL overuse by manufacturers with potential heterogeneous presence of allergens and the increased consumption of prepackaged foods with PAL by consumers, allowed in some cases by clinicians, might lead to allergic incidents after consumption of these foods.
The ad hoc Joint FAO/WHO Expert Consultation recently proposed reference doses for common allergens in combination with an allergen risk assessment to inform the decision whether or not to apply a PAL statement on prepackaged food products (FAO/WHO work group (WG), 2021) [10, 11, 32, 33]. However, when we questioned consumers about reference doses (called thresholds in the questionnaire), most were not familiar with the concept, even if they understood that the amount of allergen that triggers an allergic reaction differs between individuals and to a lesser extent with the type of allergen and amount consumed. When it comes to managing food allergens in food manufacturing, allergic consumers are divided as to whether they believe it is possible to reduce the risk to zero, but interestingly, they are generally hesitant to buy foods with even a small amount of the allergen in the product unlikely to trigger an allergic reaction. Even if they could be assured that the allergen content of a food with PAL would only cause a mild allergic reaction in a small percentage of people with the allergy, they are unlikely to buy it (Fig. 1). On the other hand, paradoxically, some patients accept taking risks with PAL products since they assume that, generally, the allergen is unlikely to be there. The lack of willingness to consume even a low level of allergen is not surprising considering long-standing health care provider recommendations on strict allergen avoidance as well as education on the unpredictability of severe reactions that can occur even to trace amounts. This is an important observation if stakeholders want to contemplate how to integrate reference threshold doses in the management of PAL and to communicate allergen risk assessment to consumers in a regulatory framework as proposed by the FAO/WHO WG. Recognizing this paradox, educational measures would be required if reference doses were to be used as a future allergy risk management tool.
From the allergists’ perspective, reaction threshold on an OFC was perceived as an important deciding factor when guiding patients towards avoidance or introduction of foods with PAL. Interestingly, a 100 mg peanut protein reaction threshold was empirically used by some clinicians to allow introduction of foods with PAL, although most did not specify a specific cut-off. A reference dose of 2 mg of peanut protein was proposed by the FAO/WHO WG to decide whether a PAL is required on a prepackaged product (FAO/WHO WG, 2021) [10, 11, 32, 33], knowing that this amount of peanut corresponds to a quantity of peanut protein slightly lower than the ED05 [13, 34]. These results raise the question of whether “high-threshold” reactors might safely consume these products with PAL in the context where recommendations proposed by the FAO/WHO WG apply and could assure a low level of allergens in prepackaged foods displaying a PAL. If PAL would mean a low level of allergen, a single dose OFC at a defined threshold (e.g. corresponding to an ED20, 25 mg peanut protein) would be an interesting avenue to allow introduction of foods with PAL in patients who do not react at this dose, thereby increasing available food options. On the other hand, if the food challenge is positive, these patients would be advised to pursue avoidance of foods with PAL. Most allergists from the survey were open to performing single-dose food challenges to stratify patients’ risk regarding PAL introduction, as previously proposed by some clinicians in clinical practice [35, 36]. The most significant obstacles foreseen with this approach included limited access to OFCs in Canada and variability of thresholds with external cofactors (Table E2)[35].
One of the limitations of the consumer survey is that a majority (72%) of the surveyed people were Food Allergy Canada members who were likely more aware of allergic disease and better informed about food allergen labelling than consumer non-members. In addition, 12% of the surveyed patients did not have a diagnosis confirmed by a physician, although this is a relatively low rate considering the large population sampled. Furthermore, the survey shows that a high proportion of allergy diagnoses (76%) were confirmed by allergists. Regarding the limitations of the allergist survey, there is a possible selection bias with mainly CSACI members participating (ie. no survey of non-members who are largely community based) although provincial associations were also contacted. Though there is an overrepresentation of Ontario and Quebec, this is where the majority of allergists in Canada are concentrated. Finally, the large representation of full or part-time academic practitioners in the sample could explain the high rate of clinicians performing OFCs in the survey.
As proposed by the FAO/WHO WG in 2022, the use by manufacturers of reference doses combined with allergen risk assessments for the management of PALs could be an option which would allow a reduction in the use of PALs on prepackaged food products and would contribute to reducing the reported variability of allergen concentrations in products with PAL. However, while this approach would standardize and significantly improve the approach manufacturers take to the application of PAL, our study shows it is challenging for the allergic consumer. With limited understanding of the concept of thresholds, including familiarity with their own threshold of reactivity, they are not willing to buy foods that contain even a small amount of allergen in the product, which might only cause mild reactions in a small minority of allergic individuals. They are more comfortable with the uncertainty of a PAL. With almost half of the respondents believing zero risk is possible there is a lack of understanding of the manufacturing process and the approach some companies take to mitigate risks. This makes it even more difficult to then have consumers contemplate how set reference doses would create a better system to determining PAL. Furthermore, if these international recommendations were to be regulated in Canada, transparency and consumer education on this approach is required to explain the rare possibility of food allergic reactions in the context of an unlabelled presence of allergen in a food product. In addition, for the 5% with a reactivity threshold below the ED05, the FAO/WHO recommendations would not improve their QoL because the absence of PAL could mean either an absence of allergen or an amount of allergen sufficient to trigger an allergic reaction in these very sensitive individuals. For these patients, options will have to be proposed.
Given the complexity of introducing and adopting a risk-based approach to the application of PAL with set reference doses, a multi-stakeholder approach, including regulators, industry, allergists and patients is necessary to determine the best way to move towards implementation. Allergists and patient organizations can begin by helping patients understand threshold concepts and a risk management approach to PAL. However, in an environment where the true risk is unknown and variable, waiting for these concepts to be unilaterally understood has risk considering the continued proliferation in the use of PAL. Of all the possible short-term options, we believe that a risk-based approach for PAL mainly supported by GMPs, combined with guidance for the industry would ultimately help allergic consumers, in a shared decision with their allergists, to better deal with PAL. Support of this approach by regulators as well as further information and education around allergen risk assessment for consumers but also for clinicians could give more confidence in PAL for the majority of consumers with food allergies ; this would be a significant improvement compared to what they currently face.