Peanut and hazelnut occurrence as allergens in foodstuffs with precautionary allergen labeling in Canada

Food allergies, particularly to peanuts and tree nuts like hazelnuts, significantly impact a substantial portion of the Canadian population. In Canada, peanuts and hazelnuts are priority allergens requiring clear labeling when intentionally added. However, precautionary allergen labeling (PAL) is frequently used to indicate potential allergen presence due to cross-contamination during processing. While PAL aims to protect allergic individuals, overuse may limit food choices and lead to risky consumer behavior. Previous studies have shown a low occurrence of allergens in products with PAL, but data from Canada is limited. This study addresses this gap by analyzing a large number of Canadian food samples to determine the actual occurrence of peanuts and hazelnuts in products with PAL. This data is crucial for developing allergen management strategies and establishing Canadian-specific allergen action levels, similar to the VITAL program in Australia and New Zealand. The study also investigates lot-to-lot variability and differences between PAL statements, brand types, and product origins to provide a more comprehensive understanding of cross-contamination in Canadian food products.

Existing literature reveals a significant variation in the occurrence of allergens in products with PAL, with some studies reporting as low as 10% of products containing the declared allergen. Several studies have explored factors like lot-to-lot variability, differences between PAL statements (e.g., "may contain" vs. "may contain traces"), and brand differences. However, most previous studies either have focused on a limited range of products or lack data from Canada. The lack of Canadian-based occurrence data has hampered the development of effective allergen management tools and risk assessment strategies within the country.

The study employed a sample plan based on past occurrence studies and the Canadian Food Inspection Agency (CFIA) food recall database. Eight high-risk food categories (baked goods, baking mixes, candies, chocolate, cookies, savoury biscuits, snacks, and cereal products) were selected. A total of 871 samples were analyzed for peanuts and 863 samples for hazelnuts. Products with PAL statements, those declaring allergens as minor ingredients, and those with allergen-free claims were included. Samples were analyzed in duplicate using ELISA sandwich kits from r-Biopharm. Two spiking validation methods were performed: one using allergen solutions and another using incurred spiking in a cookie matrix. Data was analyzed using SAS Studio University Edition to compare allergen protein concentrations across food categories, PAL types, countries of origin, and brand types. Statistical tests included PROC FREQ, PROC SGPLOT, PROC MEANS, PROC UNIVARIATE, and PROC NPAR1WAY (due to non-normal data distribution).

A total of 871 samples were analyzed for peanuts, with 72% having a PAL statement, 1% listing peanuts as a minor ingredient, and 27% claiming to be peanut-free. Peanuts were detected in 4% (38/871) of samples, with concentrations ranging from 0.6 to 28.1 ppm. For hazelnut analysis, 863 samples were tested, with 94% having a PAL statement and 6% claiming to be nut-free. Hazelnut proteins were found in 9% (78/863) of samples, with concentrations ranging from 0.4 to 2167 ppm. Chocolate products showed the highest occurrence of both peanuts and hazelnuts. Lot-to-lot variability was observed for both allergens, with some products testing positive in only one or a few of the five lots analyzed. No statistically significant differences were found in allergen concentrations between different PAL statements ("may contain," "may contain traces," etc.), product origins (Canada, USA, Europe), or brand types (private vs. regular brands), although the low number of positive samples limited the statistical power. One sample with a peanut-free claim tested positive, and three samples listing peanuts as a minor ingredient did not contain detectable levels of peanut protein.

The low occurrence of peanuts (4%) and hazelnuts (9%) in samples with PAL aligns with findings from other studies. However, the high concentration range, particularly for hazelnuts (up to 2167 ppm), indicates a significant risk for allergic individuals, especially considering the potential for high dietary intake with some products. The observed lot-to-lot variability highlights the challenges in consistently preventing cross-contamination. The finding that "may contain" and "may contain traces" statements did not differ significantly in allergen concentrations suggests that these statements may not effectively communicate the varying levels of risk. The high contamination rate in chocolate products warrants specific attention, suggesting that these products should be considered high-risk and carefully managed in risk assessments and manufacturing practices. The positive result in a peanut-free product underlines the need for stringent allergen control measures to ensure the accuracy of allergen-free claims. The study’s findings can inform the development of Canadian allergen action levels and guide improvements in food manufacturing practices.

This study provides valuable Canadian-specific data on the occurrence of peanut and hazelnut allergens in products with PAL. The findings underscore the need for improved manufacturing practices to minimize cross-contamination and the development of Canadian allergen action levels to optimize PAL use. Future research could focus on more detailed analysis of intra-lot variability and exploration of the effectiveness of different allergen control strategies in various food matrices.

Intra-lot variability was not fully assessed due to the study design, which prioritized analyzing multiple lots across a wide range of products. The relatively low number of positive samples for some allergen-PAL combinations limited the statistical power of certain comparisons. The validation of ELISA kits was done using cookies as a single matrix, potentially influencing the reliability of the results for other types of products. The availability of products with specific hazelnut PAL statements instead of the broader "tree nut" statement also impacted the number of samples analyzed.