We conducted a nationwide Canadian telephone survey on food allergy prevalence between February 2016 and January 2017, targeting vulnerable populations (New, Indigenous, and lower-income Canadians).
To examine the independent effect of demographic characteristics on food allergy.
Canadian households with vulnerable populations were targeted using Canadian Census data and the household respondent reported whether each household member had a perceived (self-reported) or probable (self-report of a convincing history or physician diagnosis) food allergy. The association between perceived and probable food allergy and demographic characteristics was assessed through weighted multivariable random effects logistic regressions.
Children, females, Canadian-born participants, adults with post-secondary education, and those residing in smaller households were more likely to report perceived or probable food allergy. Although immigrant parents self-reported less food allergy, Canadian-born children of Southeast/East Asian immigrant versus other immigrant or Canadian-born parents reported more food allergy.
We have demonstrated clear associations between demographic characteristics and food allergy, which may provide important clues to the environmental determinants of food allergy.
To the editor,
We conducted a nationwide Canadian telephone survey on food allergy (FA) prevalence between February 2016 and January 2017 (SPAACE [Surveying Prevalence of Food Allergy in All Canadian Environments] to SPAACE [S2S] ), targeting vulnerable populations (New, Indigenous, and lower-income Canadians) using 2006 Canadian Census data (Additional file 1). We compared prevalence between vulnerable and non-vulnerable populations  and showed (in univariable analysis) that overall prevalence was lower in immigrants and adults without post-secondary education. We now examine the independent effect of other demographic characteristics (age, sex, race/ethnicity, and household size), as well as immigrant status, education, and household income on presence of FA in multivariable analyses.
Through the use of 2006 Canadian Census data, the census tracts (CT) from within the census metropolitan areas  that contained the highest proportion of households with New, Indigenous, and lower-income Canadian were selected and converted into postal codes using the 2006 Statistics Canada postal code conversion file. A random sample of household telephone numbers with accompanying mailing addresses was then selected from these postal codes by Info-Direct (a company maintaining household telephone directory listings in Canada: Cornerstone Info-Direct, Toronto, Ontario, Canada) and ASDE Survey Sampler (Gatineau, Quebec, Canada).
Due to the methodology for targeted sampling of the vulnerable populations, the smaller Canadian provinces (Newfoundland and Labrador, Prince Edward Island, Nova Scotia, and New Brunswick) as well as the Canadian territories (Northwest, Yukon, and Nunavut) were either excluded or not proportionately represented. Although our primary research objective was to ensure adequate representation of the vulnerable populations, we also wanted to provide nationwide prevalence estimates. Hence, we sampled households from additional regions from the four smaller provinces that contained the highest proportion of the vulnerable populations in these provinces although the percentage was lower than in the larger provinces. In the territories, a sample of households was obtained from the entire region.
The eligible adult household respondent (Additional file 1) completed the Food Allergy Prevalence Questionnaire (FAPQ) [1, 4, 5] in English or French, providing the following for each household member: allergy to peanut, tree nut, fish, shellfish, sesame, milk, egg, wheat, soy, or other foods, details on the most severe reaction for the nine allergens (symptoms, interval between exposure and symptoms, mode of contact, and if the allergy was physician-diagnosed), age, sex, race/ethnicity, birthplace, years in Canada (for non-Canadian-born), education (if ≥ 18 years), and household income. Food allergy was defined as perceived or probable. An individual was defined as having a perceived FA if they were reported by the adult household respondent to have any FA; an individual was defined as having a probable FA (Additional file 1) if the household respondent reported symptoms/signs compatible with a convincing history and/or physician diagnosis of a peanut, tree nut, fish, shellfish, sesame, milk, egg, wheat, and/or soy allergy [1, 2, 5].
The FAPQ included the following race/ethnicity options: South Asian (e.g. East Indian, Pakistani, Sri Lankan), Southeast Asian (e.g. Cambodian, Filipino, Indonesian, Laotian, Vietnamese), East Asian (i.e., Chinese, Japanese, Korean), Black, Indigenous (self-identified with First Nations, Metis, or Inuit), Arab, Latin American, West Asian (e.g., Afghan, Iranian, Iraqi), white, or other. In the analysis, race/ethnicity was stratified as South Asian, Southeast/East Asian, Black, Indigenous, white, or other (Arab, Latin American, West Asian, other, multiple, and unknown race/ethnicity).
The association between: (1) any perceived and (2) any probable FA and multiple demographic characteristics was assessed through weighted univariable and multivariable random effects logistic regressions, which allowed for correlated observations within households. For both outcomes, the most informative multivariable model was specified by selecting those demographic predictors for which an association remained statistically significant at the 95% confidence level, after eliminating all others, starting with the least likely to be associated with the outcome (backward stepwise selection). The Bonferroni method was also applied to assess the robustness of the associations. The association between each individual perceived and probable FA and demographic characteristics was examined using the same multivariable model selection process as for any perceived and probable FA. In addition, the influence of parental birthplace on any perceived and probable FA in Canadian-born children was assessed through the same multivariable model selection process. In the first regression, which examined the association between any perceived and probable FA and parental birthplace, the sample was restricted to include only parents with at least one Canadian-born child. In the second regression, which examined the association between any perceived and probable FA and parental birthplace, the sample included only Canadian-born children. In addition to the predictors considered in the model for perceived and probable FA which involved the full sample, the following predictors were also considered in the models with the restricted samples: South Asian-born, Southeast/East Asian-born, and non-Asian-born immigrant parent versus Canadian-born parent.
The Research Ethics Boards of the Universities of Calgary and Waterloo approved the study.
Of the 19,286 households targeted, 5874 of 11,592 eligible households completed the FAPQ, representing a 50.7% household response rate. Of 15,322 individuals participating, 14,818 provided sufficient information to be included in this analysis. Our sample consisted of 20.0% children, 48.4% of non-white race/ethnicity, 33.4% immigrants, 51.9% of adults without post-secondary education, and 22.7% resided in lower-income households (Table 1). Based on 2016 Canadian Census data , 20.0% of the general Canadian population were children, 27.2% were of non-white race/ethnicity, 21.9% were immigrants, 44.8% of adults did not have post-secondary education, and 12.8% resided in lower-income households. Given our targeted sampling strategy, the percentage of those of non-white race/ethnicity, immigrants, adults without post-secondary education, and residing in lower-income households exceeded the Canadian population.
In multivariable analyses, adults ≥ 45 years versus participants 0–44 years (OR 0.69, 95% confidence interval (CI) 0.56, 0.86), New Canadians versus Canadian-born (OR 0.51, 95% CI 0.38, 0.69), those immigrating to Canada ≥ 10 years prior versus Canadian-born (OR 0.75, 95% CI 0.62, 0.92) and those residing in larger households (OR 0.82, 95% CI 0.75, 0.90) were less likely to report any perceived FA (Table 2). Females (OR 1.49, 95% CI 1.27, 1.74) and adults with versus without post-secondary education (OR 1.20, 95% CI 1.02, 1.43) were more likely to report perceived FA, although this latter association did not remain significant when the Bonferroni correction was applied.
New Canadians versus Canadian-born (OR 0.46, 95% CI 0.30, 0.68), those immigrating ≥ 10 years prior versus Canadian-born (OR 0.64, 95% CI 0.49, 0.82), and those residing in larger households (OR 0.85, 95% CI 0.77, 0.94) were less likely to report probable FA. Children versus adults aged ≥ 18 years (OR 1.95, 95% CI 1.38, 2.75), females (OR 1.49, 95% CI 1.22, 1.82) and adults with versus without post-secondary education (OR 1.55, 95% CI 1.23, 1.96) were more likely to report probable FA (Table 2).
Race/ethnicity and household income were not associated with either perceived or probable FA in multivariable models.
The association between demographic characteristics and the individual perceived and probable FA is reported in Additional file 1: Table S1A, B. Although many of the demographic characteristics associated with any perceived and probable FA were also associated with some of the individual FA, we also observed that race/ethnicity was associated with some individual FA. Those of Southeast/East Asian versus all other race/ethnicities were more likely to report perceived (OR 2.14, 95% CI 1.31, 3.50) and probable peanut allergy (OR 2.30, 95% CI 1.33, 3.97) and perceived shellfish allergy (OR 2.19, 95% CI 1.45, 3.30). Those of Southeast/East Asian versus those who were neither of Southeast/East Asian or Indigenous race/ethnicity were more likely to report perceived (OR 2.42, 95% CI 1.30, 4.53) and probable fish allergy (OR 4.06, 95% CI 1.87, 8.84) and perceived (OR 2.37, 95% CI 1.27, 4.41) and probable egg allergy (OR 2.41, 95% CI 1.26, 4.63). Those of South Asian versus non-Asian race/ethnicity (OR 0.25, 95% CI 0.08, 0.83) or of Southeast/East Asian versus non-Asian race/ethnicity (OR 0.32, 95% CI 0.11, 0.94) were less likely to report perceived wheat allergy. Those of Indigenous race/ethnicity versus those who were neither of Southeast/East Asian or Indigenous race/ethnicity were less likely to report perceived (OR 0.18, 95% CI 0.04, 0.81) and probable fish allergy (OR 0.19, 95% CI 0.04, 0.86) and perceived (OR 0.12, 95% CI 0.03, 0.55) and probable egg allergy (OR 0.12, 95% CI 0.03, 0.59). The associations of Asian race/ethnicities with perceived wheat allergy and of Indigenous race/ethnicity with perceived and probable fish allergy did not remain significant when the Bonferroni correction was applied.
When the sample was restricted to parents with at least one Canadian-born child, immigrant parents versus Canadian-born parents were less likely to report any perceived FA (South Asian immigrant: OR 0.24, 95% CI 0.11, 0.54; Southeast/East Asian immigrant: OR 0.52, 95% CI 0.28, 0.97); non-Asian immigrant: OR 0.46, 95% CI 0.28, 0.75). Similarly, immigrant parents versus Canadian-born parents were less likely to report any probable FA (South Asian immigrant: OR 0.23, 95% CI 0.08, 0.69; Southeast/East Asian immigrant: OR 0.32, 95% CI 0.12, 0.90; non-Asian immigrant: 0.42, 95% CI 0.22, 0.80) (Additional file 1: Table S2). When the Bonferroni correction was applied, the OR for Southeast/East Asian immigrant parents was no longer significant. However, Canadian-born children of Southeast/East Asian immigrant parents versus Canadian-born children of other immigrant parents or Canadian-born parents were more likely to report any perceived (OR 2.53, 95% CI 1.58, 4.04) and probable FA (OR 2.81, 95% CI 1.69, 4.69).
We have shown that while children, females, and adults with post-secondary education were more likely to report at least one perceived or probable FA and adults ≥ 45 years, immigrants, and those in larger households were less likely to report FA, Southeast/East Asian and Indigenous race/ethnicity were associated with perceived or probable allergy to specific foods. It is likely that the association between FA and higher education and Canadian birthplace is attributable to a variety of factors, including increased FA awareness and better healthcare access, which facilitates both self-reported and physician diagnosis of FA. Further, differing racial/ethnic, cultural, and socioeconomic influences on dietary practices and environmental exposures may also contribute [7,8,9]. Our observed association between larger household size and decreased FA supports the hygiene hypothesis, whereby increased exposure to siblings and the resultant increase in early childhood infections promotes a protective immune response .
Our findings that females and adults with higher education were more likely to report FA, whereas older adults were less likely to report FA are consistent with Gupta’s research . Also in keeping with her work , we have shown in univariable analysis in a previous publication on this data that children in lower-income households were less likely to report any perceived FA (prevalence in lower income versus higher income households: 6.4% versus 9.2%, difference − 2.8%, 95% CI − 5.5%, − 0.1%) , whereas there was little association between household income and FA in adults. An Australian study also reported a higher prevalence of nut allergy in children of mothers with higher education and socio-economic index . Our finding that certain FA (i.e., peanut, fish, shellfish, and egg) occurred more frequently in those of Asian race/ethnicity has also been reported by Gupta [11, 12] and others [14, 15]. A small Canadian study involving 151 respondents reported a higher prevalence of shellfish allergy in Historically Underrepresented (i.e., Indigenous, Asian, Black, and East Indian) versus white participants . American studies have also reported a higher frequency of overall FA in Asian children [12, 15] and adults  although they have not examined South and Southeast/East Asians separately as we did. Food-induced anaphylaxis has been reported to occur more frequently in non-white children in both the United States  and the United Kingdom . In contrast to the research by Gupta [11, 12] and others [15, 18,19,20,21,22], we did not observe an association of FA with Black race/ethnicity, potentially because of our small sample of this race/ethnicity (6.7%). Further, we observed that immigrants were less likely to report FA, whereas Gupta did not observe any association, potentially because her sample had a lower percentage of immigrants (8.4% versus 33.4% in ours) .
We observed lower odds of any perceived and probable FA in immigrant versus Canadian-born parents of Canadian-born children. Interestingly, for the the Southeast/East Asian immigrant parents, we observed higher odds of perceived and probable FA in their Canadian-born children versus Canadian-born children of other immigrant or Canadian-born parents. Consistent with our findings, an Australian study reported a higher frequency of peanut allergy in Australian-born infants of East Asian-born parents versus infants of Australian/British/European-born parents , while the prevalence of allergic disease was lower in Asian-born parents. A second Australian study reported a higher frequency of nut allergy in Australian-born children of both South and East Asian-born mothers versus children of non-Asian mothers. However, children born in Asia who migrated to Australia versus Australian-born children had a decreased risk of nut allergy . Further, American-born children of immigrant versus American-born parents have been reported to have a higher odds of FA sensitization . These observations suggest that early life environmental exposures, such as climate, dietary, and microbial, exert a differential effect on FA possibly depending on genetic background.
Although our nationwide sampling frame precluded the use of food challenges to diagnosis FA and only included households with landlines and speaking either English or French, and nonresponse bias may have influenced our results, we have demonstrated clear associations between demographic characteristics and FA, potentially important clues to environmental determinants.
Availability of data and materials
The data are not publicly available due to privacy or ethical restrictions.
Clarke AE, Elliott SJ, St Pierre Y, Soller L, La Vieille S, Ben-Shoshan M. Temporal trends in prevalence of food allergy in Canada. J Allergy Clin Immunol Pract. 2020;8(4):1428–30.
Clarke AE, Elliott SJ, St. Pierre Y, Soller L, La Vieille S, Ben-Shoshan M. Comparing food allergy prevalence in vulnerable and non-vulnerable Canadians. J Allergy Clin Immunol Pract. 2020;8(7):2425–30.
Soller L, Ben-Shoshan M, Harrington DW, Knoll M, Fragapane J, Joseph L, St Pierre Y, La Vieille S, Wilson K, Elliott SJ, Clarke AE. Prevalence and predictors of food allergy in Canada: a focus on vulnerable populations. J Allergy Clin Immunol Pract. 2015;3(1):42–9.
Davis CM, Apter AJ, Casillas A, Foggs MB, Louisias M, Morris EC, et al. Health disparities in allergic and immunologic conditions in racial and ethnic underserved populations: a work group report of the AAAAI Committee on the underserved. J Allergy Clin Immunol. 2021;147(5):1579–93.
Panjari M, Koplin JJ, Dharmage SC, Peters RL, Gurrin LC, Sawyer SM, et al. Nut allergy prevalence and differences between Asian-born children and Australian-born children of Asian descent: a state-wide survey of children at primary school entry in Victoria, Australia. Clin Exp Allergy. 2016;46(4):602–9.
Abrams EM, Gerdts JD, Gruber J, Lemoine-Courcelles C, Simons E, Protudjer JLP. Race/ethnicity, but not income, are associated with increased odds of shellfish allergy. J Allergy Clin Immunol Pract. 2021;9(1):550–2.
Wang Y, Allen KJ, Suaini NHA, McWilliam V, Peters RL, Koplin JJ. The global incidence and prevalence of anaphylaxis in children in the general population: a systematic review. Allergy. 2019;74(6):1063–80.
Keet CA, Savage JH, Seopaul S, Peng RD, Wood RA, Matsui EC. Temporal trends and racial/ethnic disparity in self-reported pediatric food allergy in the United States. Ann Allergy Asthma Immunol. 2014;112(3):222–9.
Joseph CLM, Zoratti EM, Ownby DR, Havstad S, Nicholas C, Nageotte C, et al. Exploring racial differences in IgE-mediated food allergy in the WHEALS birth cohort. Ann Allergy Asthma Immunol. 2016;116(3):219–24.
Mahdavinia M, Fox SR, Smith BM, James C, Palmisano EL, Mohammed A, et al. Racial differences in food allergy phenotype and health care utilization among US children. J Allergy Clin Immunol Pract. 2017;5(2):352–7.
Mahdavinia M, Tobin MC, Fierstein JL, Andy-Nweye AB, Bilaver LA, Fox S, et al. African American children are more likely to be allergic to shellfish and finfish: findings from FORWARD, a multisite cohort study. J Allergy Clin Immunol Pract. 2021. https://doi.org/10.1016/j.jaip.2020.12.026.
Koplin JJ, Peters RL, Ponsonby A-L, Gurrin LC, Hill D, Tang MLK, et al. Increased risk of peanut allergy in infants of Asian-born parents compared to those of Australian-born parents. Allergy. 2014;69(12):1639–47.
AEC, SJE, YSP, LS, SLV, and MBS designed the study. AEC, SJE, YSP, and LS conducted the research. YSP performed the statistical analysis. AEC drafted the manuscript and all authors contributed to the data interpretation and revised each draft for important intellectual content. AEC had primary responsibility for the final content. The corresponding author (AEC) attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. All authors read and approved the final manuscript.
This study was reviewed and received ethics clearance from the University of Waterloo Research Ethics Committee (ORE #20868), and the Conjoint Health Research Ethics Board at the University of Calgary (REB15-0688), and all participants provided informed consent prior.
Consent for publication
L. Soller participates in research sponsored by DBV Technologies and S. La Vieille is an employee of Health Canada. The rest of the authors declare that they have no relevant competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Additional methodological details on selection of study population, participant recruitment, telephone survey, and definition of probable food allergy. Table S1A. Multivariable logistic regression models: demographic characteristics associated with individual perceived food allergies. Table S1B. Multivariable logistic regression models: demographic characteristics associated with individual probable food allergies. Table S2. Multivariate logistic regression models: association of parental immigrant status and perceived and probable food allergy.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Clarke, A.E., Elliott, S.J., St. Pierre, Y. et al. Demographic characteristics associated with food allergy in a Nationwide Canadian Study.
Allergy Asthma Clin Immunol17, 72 (2021). https://doi.org/10.1186/s13223-021-00572-z