Abstract
Food allergies are an important cause of morbidity in the pediatric population. We used a questionnaire to assess the value of in vitro testing in evaluating children for food allergy. The patient population was divided into two groups: those who tested positive (Class I or greater) for food allergy (combined food/inhalant allergy or food allergy alone) and those who tested negative for food allergy (inhalant only or negative test). Although we recognize that a significant proportion of food allergies may not be IgE-mediated, for the purposes of this study we considered a patient food allergy–positive if in vitro testing was positive and food allergy–negative if testing was negative. Thirty-eight of 77 patients tested (49%) were positive for food allergies. Those with food allergies did not have pre-treatment scores significantly different from those of patients negative for food allergies (55 vs. 65, p = 0.11). Similarly, we detected no statistically significant difference between the symptom-improvement scores of the two groups (37 vs. 40, p = 0.57). We conclude that in vitro testing for food allergies is warranted in children with allergic symptoms, for several reasons: The prevalence of IgE-mediated food allergy in children may be higher than previously thought, and children with food allergies experience the same severity of symptoms as those without. It must be remembered that in vitro testing only detects IgE levels, and therefore a negative test does not mean that the patient does not have food allergies. However, when positive, in vitro testing gives the physician the opportunity to modify the treatment plan, helping avoid time-consuming and demanding oral food challenges. Our questionnaire also showed that food allergy and non–food allergy patients had similar pretreatment and posttreatment symptom scores.
Food allergy is a significant problem in the pediatric population. It is also one of the least well-understood with regard to both cause and diagnosis. Furthermore, diagnosis requires a thorough patient history and appropriate diagnostic testing. This testing has involved the use of food challenges, elimination diets, and, more recently, in vitro testing modalities such as the modified radioallergosorbent test (mRAST) and PharmaciaCAP system.
In vitro assay offers ease and efficiency of testing, requiring just one venipuncture. It is, however, limited to determining IgE-mediated food allergy or sensitivity, which accounts for only a portion of food reactivity. 1
This study consists of pretreatment and posttreatment scores obtained by questionnaire in a pediatric population. The children had all undergone in vitro allergy testing (mRAST or CAP) after review of history revealed a possible allergic diathesis. We examined patients who had in vitro tests positive for food and compared them with children who had negative results in an attempt to evaluate symptom scores in IgE food-allergic patients vs. non–food-allergic patients and to learn whether management techniques were effective.
METHODS AND MATERIAL
Questionnaires were mailed to the caregivers of all children who underwent in vitro allergy testing at the University of Missouri Hospital and Clinics in the preceding 2 years (N = 155). Testing consisted of the modified radioallergosorbent test (mRAST; Sanofi Diagnostics Pasteur, formerly Kallestadt) or CAP (KABI/Pharmacia). The mailings comprised multiple questions regarding smoke exposure, the amount of school missed as a result of symptoms, and impact on the patient's life. Respondents for the children were also asked to rate 12 different symptoms on a visual-analog scale of 0 (absent) to 10 (most severe) (Table 1). This assessment was obtained for symptoms before and after treatment. Seventy-seven questionnaires were returned (50%), but not all were completed in every respect. This was especially evident with regard to questions on posttreatment symptoms, which were often left blank. These questions were not included in the final analysis concerning score changes with treatment. We examined treatment regimens by means of questionnaire and chart review (Table 2) and broke the main group into the various subgroups on the basis of in vitro results. The groups consisted of patients who tested positive for foods only (F), both foods and inhalants (B), and inhalants only (I), as well as negative testers (N).
Allergy symptoms
70412-2-table1.png)
Treatment modalities in allergic groups
70412-2-table2.png)
The mRAST and CAP panels used at the University of Missouri–Columbia were designed for use in the pediatric population. They include egg, milk, wheat, corn, soybean, cat, dog, cockroach, dust mites, and molds. 2
A computer database was created for organization and later analysis of data with the use of Foxpro software. Data were taken from the questionnaire and chart review. A large amount of data was obtained, permitting investigation of many possible relationships. We focused this study on the symptom scores in IgE food-allergic and non–food-allergic children.
We performed statistical analysis with the use of the Wilcoxon rank-sum and Kruskal-Wallis tests to compare independent groups. Results were obtained for combined symptom scores and individual symptom scores. Score changes from before to after treatment were also analyzed with the Wilcoxon signed-rank test. We conducted χ2 testing to check for differences in the use of the various treatment modalities for the respective groups.
The study protocol was exempt from local institutional review board approval because it only involved mailed questionnaires and retrospective chart review.
RESULTS
The mean age of the study group was 5.5 years (range 1 to 14 years). Level 1 or greater was considered a positive in vitro test and thus represented a positive test for allergy. A class I mRAST (>750 counts) was compared with class I CAP by means of the alternate scoring method (>60% reference D score). This allowed for increased sensitivity at the expense of specificity.
With regard to all symptoms combined, the food-allergic group (F or B, n = 37) had a mean pretreatment score of 55, compared with 65 for the non–food-allergic group (I or N, n = 37; p = 0.11) (Table 3). The food-allergic group exhibited a mean change of −37(n = 29) in combined symptom score, compared with −40 (n = 33) for the non–food-allergic group (p = 0.57) (Table 3). A negative change in symptom score represented an improvement in symptoms per questionnaire report.
The groups were further broken down into food-allergic–only (F, n = 17), both food- and inhalant-allergic (B, n = 20), inhalant-allergic only (I, n = 22), and negative for allergy (N, n = 15). Mean pretreatment combined-symptom scores were 47 (F), 61 (B), 68 (I), and 61 (N) (p = 0.08) (Table 4). With regard to mean changes in combined-symptom scores, results were −28 (F), −46 (B), −42 (I), and −36 (N) (p = 0.08) (Table 4).
We used the χ2 test of independence to determine whether treatment type was related to the results of allergy testing. Avoidance was used less often in the negative group (B, 85%; F, 94%; I, 83%; N, 38%; p < 0.001). Medication was used less often in the food-only group (B, 85%; F, 50%; I, 87%; N, 88%; p = 0.01). Immunotherapy was used more often in the inhalant-only group (B, 0%; F, 6%; I, 26%; N, 6%; p = 0.03).
On examining the two larger groups (food allergy or non–food allergy), we found avoidance was used more often in the food allergy group (B or F, 89%; I or N, 64%; p = 0.009). Medication was used more often in the non–food allergy group (B or F, 68%; I or N, 87%; p = 0.05). Immunotherapy was used more in the non–food allergy group also (B or F, 3%; I or N, 18%; p = 0.03). On examining the four groups individually and grouped into food and non-food groups, we noted that the use of surgery was not significantly different among them (B, 45%; F, 44%; I, 35%; N, 56%; p = 0.62) (B or F, 45%; I or N, 44%; p = 0.92).
Finally, we examined skin-symptom scores. We found no significant differences on four-group (B, 3.1; F, 1.2; I, 3.1, N, 2.2; p = .24) or two-group analysis (B or F, 2.2; I or N, 2.8; p = 0.39).
DISCUSSION
Thirty-eight of 77 patients returning questionnaires (49%) were positive for food allergy, thus showing a significant amount of IgE-mediated food allergy in our patient population. It must be noted that a difference exists between true food allergy and food intolerance or sensitivity. Food allergy is an abnormal immunologic response to the ingestion of food, either immediate or delayed, involving one or all of the Gell and Coombs types. 3 – 5 The best-understood type of these is IgE-related, or Gell and Coombs type I.
“Sensitivity,” or “intolerance,” implies a nonimmunologic reaction, possibly to the additives or chemicals in the offending food. Obviously this would not involve an IgE-type reaction. The symptom complex, however, may be very similar to that of an allergic syndrome.
Overall, when we examined combined-symptom scores before treatment and changes with treatment, we found no significant difference between the two groups. This held true even when the groups were further broken down (F vs. B vs. I vs. N). In other words, patients with IgE food allergies were comparable to food-negative patients in the severity of symptoms, as well as in symptom improvement with treatment.
However, when the population was broken down into four groups, some differences did appear. These differences were not statistically significant but were suggested by the data. Food allergy–only patients had the lowest pretreatment scores and demonstrated the least improvement. This may have been a result of the fact that the questionnaire may not have addressed the symptoms found disproportionately in food allergy (e.g., gastrointestinal symptoms). Such symptoms may have been more common in the food groups and therefore their inclusion would have increased symptom scores. The questionnaire did include questions on skin changes, a symptom sometimes considered worse in food allergy sufferers. Regarding pretreatment skin-symptom scores, we found no significant difference among any of the groups (B, F, I, N and food vs. nonfood). Also, in vitro testing checks for IgE food allergy. Patients may have other immunologic causes of symptoms (i.e., not all food allergy is IgE-mediated). Skin and gastrointestinal symptoms may be more significant in the food-intolerant individual than in the truly allergic one and therefore incorrectly associated with the latter as well.
The various groups of children underwent different treatment regimens and were then analyzed. Interestingly enough, when the population was broken down into four groups (F, B, I, N), some differences did surface. As would be expected, avoidance was used more in those who tested positive for any allergy than it was in the group of negative patients. Immunotherapy was used more in the inhalant group than in any of the others. Medications were ordered less in the food-only group, and those who tested negative for allergy frequently received medication recommendations. This was partly due to the fact that food allergies were often treated with avoidance and may not have needed additional treatment for control. Also, as a result of clinical suspicion, patients often were started on antihistamines before getting the results of in vitro testing.
Combined symptom scores for food-allergic and non–food-allergic patients
70412-2-table3.png)
Combined symptom scores for individual groups
70412-2-table4.png)
For further analysis, the two groups (B or F and I or N) were then studied. Again, we noted some differences in treatment. Avoidance was more common in the food allergy group — again, consistent with the common treatment for food allergy. Medication and immuno-therapy were more commonly used in the nonfood group, also consistent with the usual modalities in inhalant allergy therapy. The incidence of surgical intervention was no different in any of the groups when they were compared.
Food allergy is often difficult to diagnose. It involves multiple organ systems 6 , 7 and may be difficult to differentiate from sensitivity on the basis of history alone. The food diary is an attempt to obtain an objective record of intake with any resulting symptoms: type, duration, and time of onset after ingestion. Such monitoring is quite time-consuming, especially with the wide variety of food in the American diet. The patient must be highly motivated and must not withhold information on eating habits.
Once food allergy is suspected, several ways of testing may be used to confirm its presence. The gold standard is still believed to be the double-blind, placebo-controlled food challenge. 8 It can also be used to verify any foods suspected from other testing. This is somewhat labor-intensive, but it can be performed in the office. Elimination diets are an alternative way for the patient to evaluate possible offending foods at home.
Progressive-dilution skin testing for food can be a useful adjunct in the evaluation of a patient for food allergy. However, it is also time-consuming and requires trained personnel to administer and interpret the test.
Several in vitro testing modalities have been devised to make testing easier. In one type, the release of mediators from sensitized cells is examined: examples include the leukocyte histamine release test, the human basophil degranulation test, and the leukocyte leukotriene release test. 6 The usefulness of these tests has not been fully elucidated. Total IgE has also been used as a screening tool for food allergy. 9
The more popular in vitro testing modalities are the mRAST and PharmaciaCap, which were used in this study. Although they only identify the IgE food-allergic patient, they are quick and easy tests to perform and are popular choices for use in the pediatric population. Several possible limitations do exist, though.
One limitation concerns the cutoff for a “positive” test. No specific level has been found to be diagnostic of allergy purely on the basis of testing results. Furthermore, as the “positive” class score is decreased (e.g., from 2 to 1), sensitivity increases but specificity decreases. Actually, we have shown, in a previous unpublished study, that with regard to inhalant allergens, the RAST/PharmaciaCAP class in no way correlated with subjective symptom scores. Because we saw no difference in symptom scores (i.e., a class 1 had symptoms as severe as class 4 or 5 when we considered inhalants only), we arbitrarily chose 1 as our “positive” level in this study, extrapolating this result to foods as well as inhalants.
Further concerns about in vitro testing include the need to obtain extract exactly like the suspect food. This may be difficult to do, especially in rare foods. Also, it is important to keep in mind the possibility of cross-reactivities of the various food types.
This study in no way is an attempt to evaluate these modalities. Rather, we examined IgE food allergy in a pediatric population, as determined from specific IgE in vitro testing (mRAST and PharmaciaCAP).
Any study involving a questionnaire has areas of concern. These concerns, addressed in a previous study, 10 include the need for a thorough yet understandable form, as well as a form that is not too long. Obviously there is also concern over possible recall bias. This concern could be alleviated if patients were given questionnaires to complete before being tested and then questioned again after treatment.
In this study we examined IgE-mediated food allergy. It is important to realize that allergy is truly a clinical diagnosis and that some who test positive for foods may not have problems ingesting the food. Furthermore, symptoms that arise when a patient eats certain foods are not necessarily immunologic in nature and therefore may not have positive results on immunologic testing modalities. A food challenge may be important to confirm clinical symptoms to any food testing showing positive on in vitro assay.
In our practice setting, we do not routinely perform food challenges after positive in vitro assays in children. This is at least in part the result of the overall excellent results we have obtained clinically by treating them. The children with food allergy showed no difference in symptoms compared with the non-food group. This would lead one to believe that, at least with regard to IgE results, it may not be easy to differentiate food vs. inhalant allergens purely on the basis of the nature and severity of symptoms. The accuracy of the history could be improved with a food diary to document any temporal relationship between ingestion and symptoms; however, not all symptoms represent true allergy.
CONCLUSION
We believe that in vitro testing modalities—specifically mRAST and PharmaciaCAP—are useful in diagnosing pediatric allergy. Not all food-allergic patients have IgE-mediated disease, but those who do benefit from treatment as much as those with inhalant allergies. In those children with multiple food allergens, challenge testing to evaluate for clinical symptoms would be beneficial because it is often difficult to eliminate multiple items from the pediatric diet.
The questionnaire, with emphasis on individual symptom scores, is an effective way of monitoring patients with allergies. Because allergy truly is a clinical disease, improvement in the symptom complex of the allergy sufferer is the true mark of successful intervention.
We thank Ms. Jane Johnson of the Department of Biostatistics at the University of Missouri School of Medicine for her assistance in the preparation of this manuscript.