Retrospective analysis on the agreement between skin prick test and serum food specific IgE antibody results in adults with suspected food allergy
Our study shows that in adult patients being evaluated for a possible food allergy, agreement between the SPT and sIgE of common food allergens is at least moderate or better. Of the analyses that reached statistical significance, SPT and sIgE results showed near perfect agreement for cashew and salmon; substantial agreement for pistachio, codfish, shimp and egg white; moderate agreement for peanut, walnut, pecan, lobster. The value of objective testing in the diagnosis of IgE mediated food allergy is belied by the fact that the history of the reaction was frequently unclear and therefore not helpful. In this population, there appears to be substantial barriers to performing oral food challenges as very few were done. Our data did not supply any particular reason for the low number of oral challenges performed at our site. In general, both patients and physicians contribute to the reluctance to perform oral challenge. While clinicians often have financial or logistical barriers, patients generally contend with fear of experiencing a reaction or the significant time a challenge requires.
In our analysis, a significant p value (p 0.05) rejects the null hypothesis that the kappa analysis result was due to chance. Peanut, walnut, pistachio, cashew, pecan, salmon, shrimp, lobster and egg white analyses reached a significant p value. Hazelnut, almond, crab, clam, milk, soy, wheat, sesame did not reach a significant p value probably because of small sample sizes and therefore their kappa analyses cannot be interpreted.
There is a paucity of literature examining the concordance of SPT and sIgE in adult populations. Asha’ari et al. showed that when tested for a predetermined panel of food allergens including peanut, egg, flour and chicken without clinical correlation, the agreement of SPT and sIgE was between fair to good [5]. This is comparable to our study in that there were no instances of poor agreement. Our data demonstrated a relatively stronger agreement between the SPT and sIgE to common food allergens. This may be due to the fact that the study population was tested for foods to which there was a history of symptoms after ingestion. Given the targeted approach to selecting food allergens to test for, our study population was less likely to produce discordant SPT and sIgE results due to asymptomatic sensitization.
The major strength of this study is that it is reflective of the routine clinical practice. Our study population was specific to patients in whom food allergy was already suspected because a reaction had occurred: highly representative of daily practice. The food allergens they were tested for at our allergy clinic reflected their clinical history or concern, making the results clinically relevant.
There are several limitations to our study. Inevitably, there was some heterogeneity to the data because this was a retrospective analysis; we could only abstract the data available on the medical records. We could also not obtain oral food challenges to validate the diagnosis of food allergy. We did not aim to correlate the agreement or the individual test results with an OFC validated diagnosis and it is not possible to do so from our data. Additionally, not every patient underwent both SPT and sIgE testing to the same panel of allergens because the referring physician had frequently ordered a broad panel of sIgE titres and at our allergy clinic a smaller selection of foods were skin prick tested. This reduced the number of patients who could be included in the kappa analysis and subsequent statistical analysis of several food allergens was not possible due to the limited sample size.
The literature in pediatric food allergy has established the utility of SPT and sIgE in predicting the result of oral food challenges and therefore their use in diagnosing food allergy. SPT wheal size and sIgE titre cutoffs that predict oral food challenge response have been characterized in children. This is so for peanut, fish, egg and milk while wheat and soy remain a challenge [9–12]. This advancement has likely led to greater roles of SPT and sIgE in diagnosing food allergy in children. In adults however, the utility and validity of sIgE is less well studied despite its frequent use. So far, there are no validated SPT or sIgE values that can predict a reaction on oral challenge test in adults as there are in children. It is further unclear whether the predictive values from the pediatric population can carry over to the adult population. The objective of our study was not to identify such cut offs, but the identification of these threshold values should be the direction of future research. The strength of the agreement between SPT and sIgE for common food allergens demonstrated by our study represents the first step of characterizing the utility of these tests in adults with suspected food allergy.
The data from this study may serve to reassure clinicians when both testing modalities are not available concurrently that the results will likely agree and that these two tests are possibly interchangeable. It may be useful for non-allergists who evaluate patients with complaints suggestive of IgE mediated food allergy to obtain sIgE to the foods of concern. Therefore when a patients presents with a history strongly suggestive of IgE mediated food allergy and a positive sIgE to the food of concern, referring physicians will be in a more confident position to triage the patient to specialist care or to counsel the patient appropriately while waiting for a specialist’s evaluation.
Our study also found a substantial minority of patients in whom a diagnosis of food allergy is still unclear after a thorough history and SPT or sIgE testing while the number of food challenges performed remained low. This finding should encourage allergists to use the oral challenge where appropriate and the provincial health services to remove any logistical or incentive barriers that discourage their use.