Our study represents the first local analysis of a significant number of adult-onset IgE-mediated food allergic patients in Manitoba. Findings of this case-series agreed with the retrospective review by Kamdar et al at Northwestern University in 2015 regarding the most common culprit food being seafood (21). As was the case with this study, we did not attempt to analyze shellfish subtypes (mollusks vs. crustaceans). With respect to diagnoses, all patients aside from one had positive SPT to the allergens in question. The patient without positive epicutaneous testing underwent food-specific serum IgE levels, which demonstrated significant elevation. No patients included in this case-series underwent oral-food challenge, likely due to the selection bias for inclusion.
In a 2012 Canadian review, the overall prevalence of food “allergy” was 8.07%, with the most common self-reported allergens being shellfish (1.91% of those surveyed), cow’s milk (1.89%), fruits (1.61%), vegetables (1.29%), tree nuts (1.07%), wheat (0.86%), peanuts (0.78%), and hen’s egg (0.67%) (26). As with the majority of compendia on this pathophysiology, there were no qualifiers given for an adult-age of onset, and the self-reported nature of this data limits its usefulness.
A more objective 2004 cross-sectional study on German patients of all age groups found a 2.6% prevalence of adverse food reactions among a 4093 patient sample population, with confirmatory clinical testing to support this prevalence. Within this study, the most common allergens with confirmatory IgE testing were nuts, apples/pears, stone fruit, vegetables, other fruit, flour, milk, and egg (27).
Again, the issue of which allergens present with adult-onset symptoms fails to be outlined by either of the aforementioned studies. A more recent study in JAMA by Gupta et al. in the US population did include an adult-onset identifier in their patient subgroupings (8). Of the 40443 adult who completed their survey, 5.2% (4.9-5.4; 95% confidence interval) of respondents fulfilled criteria for adult-onset food allergy. The study then outlines the most common foods to which patients had allergy – shellfish (2.9%), milk (1.9%), peanut (1.8%), tree nut (1.2%), and finfish (0.9%) – however, once again the reporting of these allergens does not distinguish between adult-onset and pediatric-onset persistent food allergy (8). Further to this, no Canadian patients were included, limiting the clinical applicability.
Interestingly, of our patients found to be shellfish allergic, only 1/4 had documented epicutaneous testing to house dust mite (HDM) in their medical records, while 2 other patients were not tested for any aeroallergens as they did not have clinical AR, and 1 patient had the documented diagnosis of allergic rhinitis, but results of inhalant allergen testing were not documented in either electronic or physical medical records. HDM-shellfish cross-reactivity is a well-documented phenomenon, believed to be secondary to the high sequence homology between tropomyosin proteins of these organisms (4,5). This hypothesis stands analogous to the PFS/OAS, wherein an aeroallergen results in sensitization and subsequent food allergic reactivity, without ingestion of the food necessary to cause this sensitization. The cockroach also possesses a highly homologous tropomyosin protein to HDM, however given the low prevalence of this aeroallergen in Manitoba, it is not commonly tested on our standard aeroallergen panel.
In vitro subcutaneous dust mite therapy studies have shown worsening of mollusk allergy following initiation (28), lending credence to the thought that they possess shared allergenic proteins with the house dust mite. Further to this, Wong et al demonstrated significant homology between the allergenic epitopes on snail tropomyosin and HDM (4). Our study did not divide these patients by type of shellfish allergy due to the relatively low prevalence of documented HDM allergic rhinitis. With regard to our 2 non-AR shellfish allergic patients, we hypothesize that they became sensitized to these allergenic compounds via ingestion rather than aeroallergen cross-reactivity. This is of clinical importance as both patients presented with life-threatening anaphylaxis as their index reaction, accounting for half of the shellfish allergic patients reviewed.
When examining our FDEIA population, 1 of the wheat allergic patients had negative skin testing to wheat extract but positive testing to fresh flour. This phenomenon has been demonstrated previously in a small case series of wheat-dependent exercise-induced anaphylaxis (18). The pathophysiology behind this unique occurrence has yet to be elucidated, but is presumably related to allergen alteration during processing. In similar fashion, the patent who demonstrated true IgE-mediated allergy to whey had negative skin prick testing to commercial dairy extract as well as fresh cow’s milk.
It cannot be emphasized enough how significant the high proportion of anaphylaxis was in this review. The serious nature of this reaction notwithstanding, 39% of food allergic US adults report at least 1 visit to an emergency department for anaphylaxis in their lives, and 9% report at least 1 visit in the last year based on a JAMA review (8). These staggering statistics are a clear reflection of the cost – both monetary and otherwise – of severe allergic food reactions. A caveat to this point, however is that of referral bias; through which one can conclude that more significant allergic reactions were likely selectively referred to our allergy subspecialty clinic, and more minor reactions were unlikely to have been deemed severe enough to necessitate allergy consultation.
In the aforementioned study by Ruiz et al., the most common culprit foods which resulted in anaphylaxis that necessitated emergency department presentation were peanuts followed by tree nuts and shellfish (24). In our studied patients, however, culprit foods for non-exercise-dependent anaphylaxis varied, with a shellfish culprit in 2/6 patients, and soy, finfish, whey, and almonds culprits in 1 patient each. The differences between their study and ours can be explained by the low number of patients included in our review, the possible inclusion of pediatric patients and the lack of qualifiers regarding FDEIA in the Ruiz study, and by our specification that patients included have an adult-onset reaction to foods previously tolerated.
Unfortunately, data regarding whether our adult-onset PFS patients immigrated to Canada from a country with differing aeroallergens was not recorded. This is of interest, as it has been demonstrated that in-movers to a new environment demonstrate lower rates of atopic disease, with gradual increases as they become sensitized (29). This could theoretically provide an explanation as to adult development of cross-reactive allergic disease, as accounted for by the time to sensitization. The prevalence of PFS overall (1 pre-existing, 2 adult-onset) was fairly consistent with results published by Ma et al, who reported an estimated prevalence of PFS of 8% in the general population based upon a sample of 250 US allergists (30). Confounding this finding would be our small sample size and perhaps low referral rates for patients with PFS; although this has not been studied in the Canadian population. Both aforementioned adult-onset PFS patients had a history consistent with allergic rhinitis, in keeping with the cross-reactivity hypothesis of PFS (3).
The proportion of concomitant allergic rhinitis was expectedly high given the number of patients with PFS and adult-onset shrimp allergy. This was expected based upon the stated pathophysiology of shrimp allergy and PFS. One almond allergic patient also demonstrated AR with positive SPT to trees (a common cross-reactor with almonds), however their index reaction was anaphylaxis, and thus it was concluded that this was not in keeping with PFS, given the relative rarity of anaphylaxis with pure PFS, and the lack of (peri)oral tingling accompanying their reaction (7). We also felt it was important to distinguish this reaction from classically described PFS given its life-threatening nature.
Although most adult-onset allergy studies to date use the accepted legal definition of “adult” as aged 18 and older (10,16,21,23,31), we defined this with an age of 16 years or older as our cut-off. Although this resulted in only 1 additional case in our series, we believe this is a valid exclusion point, as the age 16 and older has been used in oral immunotherapy (OIT) Canadian guidelines (32). This is reflective of the loss of immune plasticity following childhood, which has been as a reduction in OIT response by 17% for each year after the age of 5 (33). It is also local practice that patients aged 16 or older are often referred for assessment by an adult allergist rather than a pediatric allergist in an academic setting.
Our conclusions are limited by the retrospective nature of this review and the fact that not all patients underwent the same diagnostic tests (i.e., not all patients were skin tested for aeroallergens). Further to this, the small number of cases included in this review may have skewed our data, and further study and extension of this data would be essential to expand our understanding of the true prevalence of this pathology.