Large US dog allergy study maps molecular sensitization patterns

A new large-scale US study is putting numbers behind something small-animal dermatologists have long suspected: allergic dogs don’t just react to “mites” or “pollens” broadly, they often show distinct molecular sensitization patterns that may matter when clinicians interpret serology and build immunotherapy plans. In a 2026 preprint that states a peer-reviewed Animals version also exists, Thierry Olivry and Ana Mas-Fontao analyzed 21,121 US canine serum samples from 2024 and 2025 using Nextmune’s Pet Allergy Xplorer, or PAX, a 238-allergen multiplex platform. The headline finding was the sheer prevalence of sensitization, with 90.7% of dogs positive to at least one allergen and 86.2% positive to at least one environmental allergen. (preprints.org)

The work builds on a recent shift in veterinary allergy testing toward molecular allergology. The PAX platform was validated in 2024 as a canine-specific multiplex macroarray derived from the human ALEX² system, designed to measure IgE against both allergen extracts and defined molecular components. That matters because older prevalence studies in dogs often used different assays and different extract compositions, making cross-study comparisons difficult. In the introduction to the new study, the authors note that despite the size of the US dog population, they found only two earlier American surveys with more than 100 dogs, both based on intradermal testing and published more than two decades ago. (pubmed.ncbi.nlm.nih.gov)

The dataset itself was large, but also filtered in a way clinicians should notice. The authors started with 32,710 PAX tests run by Nextmune’s US laboratory in Phoenix between January 1, 2024, and December 31, 2025. They then excluded 5,146 dogs with detectable IgE against cross-reactive carbohydrate determinants, or CCDs, after routine blocking, because CCD reactivity can create false-positive polysensitization, particularly for pollens and plant foods. After also removing non-US samples, the final cohort included 21,121 dogs, split fairly evenly across 2024 and 2025. For US dogs, the assay reported IgE levels for 238 allergens across 17 biological categories. (preprints.org)

The prevalence findings were striking, but not entirely surprising. Mites were by far the leading allergen category, with 79.5% of dogs sensitized to at least one mite allergen, and Dermatophagoides farinae alone positive in 76.7%. Insect venoms ranked second at 39.8%, followed by meats at 21.9%, weed pollens at 21.5%, and tree pollens at 18.6%. Among individual allergens, the top signals after D. farinae included honeybee and wasp venom components, bovine meat allergen Bos d 6, the storage mite Tyrophagus putrescentiae, and the weed pollen marker Par j 2. The clustering analysis then sorted 16,366 sensitized dogs into six broad profiles, centered on flea saliva, mammalian serum albumins and IgGs, honeybee venom, house dust mite group-2 allergens, storage mites, and PR-10-family allergens. A separate pollen and plant-food analysis identified additional clusters tied to Par j 2, ragweed, starch synthases, PR-10 allergens, and a smaller grass-pollen profile linked to Phl p 6. (preprints.org)

Outside commentary specific to this paper appears limited so far, but the broader professional discussion around molecular allergy testing has been moving in the same direction. A 2026 review on food allergy diagnostics in companion animals said PAX allows differentiation between cross-reactivity and primary sensitization, which is one of the core promises of component-resolved testing. At the same time, longstanding dermatology guidance still applies: canine atopic dermatitis remains a clinical diagnosis, and allergy tests should be used to identify candidate allergens for avoidance or immunotherapy rather than to confirm disease by themselves. AAHA’s recent client-facing overview makes a similar point, noting that intradermal or blood-based allergy testing is generally considered after flea allergy, food allergy, and other non-allergic causes have been ruled out. (mdpi.com)

Why it matters: For veterinary professionals, this study is less about proving that allergic dogs react to mites, and more about showing how deep molecular profiling may sharpen case interpretation. If a dog’s serology fits a coherent cluster, such as flea-associated, storage-mite-associated, or PR-10-associated sensitization, that could eventually help clinicians distinguish likely primary drivers from background positives and choose more targeted allergens for ASIT. The CCD exclusion step is also a practical reminder that not every positive on a broad panel is clinically meaningful. In other words, this is potentially useful intelligence for dermatology workflows, but it doesn’t replace the stepwise diagnostic process of ruling out ectoparasites, infections, adverse food reactions, and other causes of pruritus first. (preprints.org)

There are also caveats. The study population was made up of dogs already suspected of allergic disease, not the general dog population, so these prevalence figures shouldn’t be read as community-wide rates. The assay is commercial, and the samples came from a single company laboratory, which improves standardization but may limit generalizability. And while the preprint says a peer-reviewed Animals article exists, the web-accessible version surfaced most clearly in search was the July 3, 2026 preprint, so clinicians will likely want to review the final journal version, supplementary methods, and any later validation work before changing protocols. (preprints.org)

What to watch: The next key questions are whether the authors publish regional and seasonal breakdowns, whether these unsupervised clusters correlate with clinical phenotypes or response to immunotherapy, and whether molecular cluster-guided allergen selection improves outcomes compared with standard extract-based approaches. If those links emerge, component-resolved serology could become more clinically actionable in everyday veterinary dermatology, not just more data-rich. (preprints.org)