Critically appraised topic on adverse food reactions of companion animals (4): can we diagnose adverse food reactions in dogs and cats with in vivo or in vitro tests?

Background The gold standard to diagnose adverse food reactions (AFRs) in the dog and cat is currently an elimination diet with subsequent provocation trials. However, those trials are inconvenient and client compliance can be low. Our objective was to systematically review the literature to evaluate in vivo and in vitro tests used to diagnose AFR in small animals. Results We searched three databases (CAB Abstracts, MEDLINE and Web of Science) for pertinent references on September 16, 2016. Among 71, 544 and 41 articles found in the CAB Abstract, MEDLINE and Web of Science databases, respectively, we selected 22 articles and abstracts from conference proceedings that reported data usable for evaluation of tests for AFR. Serum tests for food-specific IgE and IgG, intradermal testing with food antigens, lymphocyte proliferation tests, fecal food-specific IgE, patch, gastroscopic, and colonoscopic testing were evaluated. Conclusions Testing for serum food-specific IgE and IgG showed low repeatability and, in dogs, a highly variable accuracy. In cats, the accuracy of testing for food-specific IgE was low. Lymphocyte proliferation tests were more frequently positive and more accurate in animals with AFR, but, as they are more difficult to perform, they remain currently a research tool. All other reported tests were only evaluated by individual studies with small numbers of animals. Negative patch test reactions have a very high negative predictability in dogs and could enable a choice of ingredients for the elimination diet in selected patients. Gastroscopic and colonoscopic testing as well as food-specific fecal IgE or food-specific serum IgG measurements appear less useful. Currently, the best diagnostic procedure to identify AFRs in small animals remains an elimination diet with subsequent provocation trials.


Background
Elimination diets with subsequent provocation trials are uniformly recommended to diagnose adverse food reactions (AFRs) in dogs and cats [1][2][3][4][5]. However, performing home-cooked elimination diets and monitoring of clinical changes during the diet and subsequent provocation tests are work-intensive and time consuming and pet and client compliance is variable [1,6]. For owners, laboratory tests of blood, saliva, and hair from patients offer an easier way to achieve a diagnosis of AFRs.

Clinical scenario
Consider the example of two patients: a six-month-old female intact Labrador retriever and a five-year-old female spayed Domestic Shorthair cat. Both animals exhibit pruritus that manifests by year-round scratching. The dog also suffers from flatulence and occasional episodes of vomiting. The cat has several patches of selfinduced hair loss on the abdomen and flanks and an indolent ulcer on the left upper lip. You inform the owners of both patients that you suspect that all clinical signs might be caused by a reaction to a component of their pet's diet and advise that an elimination diet is indicated for eight weeks to evaluate potential food involvement [7]. The owners ask you if there is an easier way to identify the role of food antigens such as, for example, a blood test.

Structured question
Can we diagnose AFRs in dogs and cats with in vivo or in vitro tests?

Search strategy
We searched the Web of Science (Core Collection), MEDLINE and CAB Abstract databases on September 16, 2016 using the following string: (dog* or canine or cat* or feline) and (food* or diet*) and test* and (allerg* or hypersens* or adverse) not (human* or child* or adult*). We limited the search to journal articles published from 1980 to present; there were no language restrictions. Bibliographies from selected articles and proceedings of recent specialized veterinary dermatology and internal medicine conferences were also searched.

Identified evidence
Our literature search identified 71, 544 and 41 articles in the CAB Abstract, MEDLINE and Web of Science (Core Collection) databases, respectively. Abstracts of relevant titles were screened and any potentially useful manuscript was downloaded and scrutinised in detail. The bibliography of these articles was examined further for additional pertinent citations. In addition, proceedings of recent veterinary dermatology or internal medicine conferences were evaluated.
Altogether, we selected 23 papers [1, 3-6, 8-25] and one abstract from conference proceedings [26] that reported results of various laboratory tests in dogs or cats where an AFR was definitely diagnosed or ruled-out. We excluded studies where the diagnosis of AFR was not confirmed or the results of the individual laboratory tests could not be attributed to a specific patient. The chosen publications were mostly case-control studies, and there were two case series [11,12] and one each a single case report [10] and a prospective cohort study [13]. In all, there were twelve studies testing foodspecific IgE in the serum of dogs [1, 5, 6, 9, 10, 12-14, 16, 18, 23, 24] and three in cats [15,17,21]. Four studies also evaluated canine food-specific IgG [1,5,16,27]. Lymphocyte proliferation tests were assessed in four studies in dogs [11,14,18,20] and one in cats [15]. In dogs, intradermal testing and gastroscopic food testing were reported in six [3,4,9,10,13,14] and three studies [8,19,28] respectively. There were two studies for patch testing in dogs [5,25], and one study each for gastroscopic food testing in cats [17], colonoscopic testing in dogs [22], determination of canine fecal IgE [8] and hair and saliva testing in dogs [26]. Some studies evaluated several different tests in dogs [1,4,5,10,11,13,14,16,18,23] and in cats [15]. Studies were reported from 1991 [4] to 2017 [24,25]}. All papers were in English except one, which was in German [1]. The number of animals and type of test performed in each paper are listed in Table 1.

Evaluation of evidence
Calculations of the accuracy, positive and negative predictabilities of the various tests for a positive food challenge in dogs and cats with naturally occurring AFRs are reported in Tables 2 and 3, respectively.
These parameters could not be evaluated in nine of the selected studies, mainly because of the lack of performance of provocation tests with individual food items [1,11,12,16,18,19,21,23,28]. One report only provided details of the individual positive, but not negative provocation trials-and this prevented the calculation of the tests' accuracy [17]. Some studies evaluated the tests only in laboratory dogs [8,9,13,20,22], and it is not clear if the pathophysiology of AFRs in sensitized laboratory animals mirrors that of the naturally-occurring disease. In most studies using laboratory dogs, the tests were more accurate, presumably because the more controlled environment and food intake might have minimized the impact of other environmental factors that could be influencing the development of clinical signs. Some studies had only six or fewer of dogs [8,10,20] or cats [15] with AFRs included. While most reports were of animals with cutaneous AFRs, dogs [19,22,28] and cats [15,17] with gastrointestinal disease were also included in some articles.
While testing for allergen-specific IgE is well established for environmental allergens in humans, dogs and cats [29], it is also offered for food allergens in many countries; this explains while most of our included studies evaluated serum food-specific IgE testing. Two studies showed a low repeatability of serum food-specific IgE and IgG testing when different aliquots of the same sample were evaluated in a blinded fashion [1,16], the authors then concluded that these tests were unsuitable for clinical use. One study found a high concentration of food-specific serum IgE in a large number of dogs that had environmental atopic dermatitis and that had signs that did not improve after being fed an elimination diet [12]. Similar results were obtained in other studies in which dogs with AFRs were compared to apparently healthy dogs [5,6,14,16,18,23]. When the serum test results for food-specific IgE were correlated with food provocation outcomes in dogs with AFRs [4-6, 9, 10, 14, 24], the tests' accuracy and positive and negative predictabilities varied highly.
Intradermal testing with food antigens in laboratory dogs sensitized to specific foods usually yielded concordant positive reactions [9,13]. When allergic patients in clinical practice were tested, however, dogs with environmental, but non-food-induced atopic dermatitis also exhibited numerous positive reactions to food antigens [3], while dogs with AFRs often had no positive results [3,4,14].
With lymphocyte proliferation tests [11,14,15,18,20], the accuracy was generally higher, but this test is   technically more difficult to conduct and blood specimens need to be processed very quickly after sampling, two reasons why this test is generally not offered by commercial laboratories.
In the two studies assessing the usefulness of patch testing with food antigens, the accuracy and negative predictability of patch testing were satisfactory and excellent respectively (particularly for protein sources), but the positive predictability was low [5,25]. As a result, this test cannot be used for the diagnosis of AFR but it could be useful as a tool to identify suitable ingredients for the elimination diet in selected dogs.

Conclusion and implication for practitioners
Patch testing with food ingredients might be useful in some selected dogs to choose the ingredients for an elimination diet. Currently, all other tests cannot be recommended for the clinical diagnosis of AFRs in dogs and cats. Although serum IgE testing for food-specific IgE is offered by many laboratories in many countries as a tool for the diagnosis of AFRs, it is not reliable in dogs and cats. At this time, the best diagnostic procedure to identify AFRs in small animals remains an elimination diet with subsequent provocation trials.