Exacerbations Of Atopic Dermatitis

Published Date: 02 Nov 2017

Roerdink EM1, Flokstra-de Blok BM2, Blok J3, Schuttelaar MLA3, Van der Heide S4, Kukler J1, Kerkhof M5, Dubois AE1

1. University of Groningen, University Medical Center Groningen, Department of Pediatric Pulmonology and Allergy, GRIAC Research Institute, Groningen, The Netherlands. 2. University of Groningen, University Medical Center Groningen, Department of General Practice, GRIAC Research Institute, Groningen, The Netherlands. 3. University of Groningen, University Medical Centre Groningen, Department of Dermatology, Groningen, The Netherlands. 4. University of Groningen, University Medical Center Groningen, Department of Laboratory Medicine, GRIAC Research Institute, Groningen, The Netherlands. 5. University of Groningen, University Medical Center Groningen, Department of Epidemiology, GRIAC Research Institute, , Groningen, The Netherlands.

Introduction

Atopic dermatitis (AD) is a chronic inflammatory multifactorial disease with a relapsing course. It is characterized by a distinctive morphology, intense pruritus and a typically age-related distribution. Worldwide approximately 5 to 20% of children are affected by AD, causing a considerable reduction in quality of life of affected patients and their families (1). AD commonly presents during early infancy and childhood with 60% of cases appearing within the first year of life and in 85% during the first 5 years of life (2). There are several scorings systems for assessing the severity of AD. An example of a validated scoring system is the SCORAD (SCORing Atopic Dermatitis).(3).

Several factors suggest a role for IgE-mediated allergy in the pathogenesis of AD. First, AD is frequently the first manifestation of the atopic march, as 80% of patients will develop asthma or allergic rhinitis later in life (4). In addition, about two thirds of patients have a positive history for atopic symptoms. Finally, AD is associated with an increase in specific and total IgE levels (5,6).

Although the role of food allergy in eliciting and maintaining skin lesions in AD is still unclear, an increasing amount of evidence supports the fact that food allergy plays a significant role in many children with AD. Several studies showed a significant improvement of AD in patients diagnosed with food allergy who were placed on an appropriate allergen elimination diet (13;14). It has been reported that approximately 35% of children with AD have a diagnosis of food allergy (8-10), however a recent study found a prevalence of only 14.8% (11).

In general, more severe AD has been reported to be associated with higher prevalences of food allergy (12). Consequently, children with mild AD are thus often not considered to be at significant risk for food allergy and therefore it is not recommended to perform oral challenge testing.

Since reliable laboratory parameters to establish a diagnosis of food allergy are still lacking in most cases, double-blind, placebo-controlled oral food challenges (DBPCFCs) remain the golden standard in the diagnosis of food allergy (13). Three different clinical reactions have been recognized in patients with AD after food challenge (7):

1) Immediate-type reactions which commonly occur within 2 hours after ingestion of food, such as urticaria, angio-oedema and erythema etc.

2) Pruritus occurring within 2 hours after the ingestion of food with subsequent scratching leading to a deterioration of AD.

3) Late reactions, i.e. a flare of AD occurring after 6 to 48 hours. Late reactions have also been described to evolve after an immediate-type reaction.

The prevalence of food allergy in children with AD varied around 30-50%, with isolated cutaneous reactions occurring in 34-96% of positive challenges. (23,24,25,26) Immediate-type responses to certain types of food have been well defined by several studies of patients with AD. However only a few studies have examined the occurrence of isolated late AD responses and the prevalences that were found in these studies found in these studies varied. (15,19)

In order to study the pathogenic role of food allergies in AD more extensively, we performed a retrospective analysis of children with AD undergoing DBPCFCs. The objective of this study was to investigate the prevalence and nature of AD reactions observed during DBPCFC in children with AD and suspected of food allergy. We also studied the effect of the severity of AD on the outcome of DBPCFCs. Finally, we compared the frequency of positive DBPCFCs between children reporting exacerbations of eczema as their only symptom to those reporting other symptoms as well.

Methods

Study population

Between 2001 and 2011, children underwent DBPCFCs for one or more suspected food allergies as part of regular care. Children were excluded from DBPCFC in case of unstable comorbidity that could compromise safety, or if the children or parents were unwilling to undergo the test, which was the case in less than 2% of patients. Descriptive characteristics of the children and details of their DBPCFC were collected in an electronic database. Only the first challenge performed per child for any given food, was used in the analyses. This study was exempt from medical ethical approval, as the DBPCFCs were performed as a routine diagnostic test.

Double-blind placebo-controlled food challenges

A history of the patients undergoing DBPCFCs was taken, including the nature and severity of previous reactions. DBPCFCs were performed as previously described (16). Briefly, for at least six weeks prior to the DBPCFC, the suspected food was eliminated from the diet. The challenge was performed on two separate days, with an interval of at least two weeks. The order of the placebo or the active challenge days (verum days) was randomized. The challenge was perfomed dubble-blind: both patient and the physicians performing the challenge, did not know the order of the verum and placebo days. The suspected food was masked in a food matrix which the patient could tolerate. Blinding of the recipes used was verified by sensory testing in a laboratory setting. Increasing doses of the test food were given to patients in six steps with an interval of 30 minutes. This was done according to previously described protocols (16). During the DBPCFC, the occurrence of symptoms was registered. The DBPCFC was discontinued when one or more objective symptoms [1] occurred or when subjective symptoms [2] reoccurred twice after ingestion of the same dose.

Immediate symptoms were defined as symptoms occurring during the challenge, or within 2 hours after the last dose. During these 2 hours, the patient was still under medical supervision. Two days after each day of the DBPCFC, late symptoms were registered by a telephone follow-up. Symptoms were considered ‘intermediate’ when they occurred between 2-6 hours and ‘late’ when they occurred between 6-48 hours after the last challenge dose.

Two days after the second challenge day, the result of the DBPCFC was determined (positive or negative DBPCFC). Patients with a positive DBPCFC were advised to continue the avoidance of the tested food. Conversely, patients with a negative DBPCFC underwent an open challenge, and/ or were advised to reintroduce the tested food. Tolerance was confirmed by telephone interview one month later (16).

AD measurement during the DBPCFC

Children and/ or parents were questioned about previous and/or current AD. Prior to the DBPCFCs the children were examined and the presence of AD was quantified using the SCORAD by the allergy nurse performing the test. The SCORAD was obtained prior to the challenge and prior to each dose of the challenge. The children were carefully examined for the presence and extent of AD. The severity of AD was categorized as mild (SCORAD <15), moderate (SCORAD15-40), or severe (SCORAD >40).

The different clinical reactions in patients with AD after food challenge were categorised as follows:

Clinical reaction in AD patients after food challenge

Immediate reactions (< 2 hrs after last dose)

a. Exacerbation of AD only

b. Only symptoms other than AD

c. Both exacerbation of AD and other symptoms

Intermediate reactions* (2 – 6 hrs after last dose)

a. Exacerbation of AD after immediate reaction

b. Exacerbation of AD without previous immediate reaction

Late reactions (>6 – 48 hrs after last dose)

a. Exacerbation of AD after immediate reaction

b. Exacerbation of AD without previous immediate reaction

These reactions were studied, using the SCORAD as a quantative measurement for AD. An increase of 2 points or more in the SCORAD was considered an increase of AD (exacerbation of AD). The outcome of the DBPCFC was studied in children with only an exacerbation of AD in their history and compared to those reporting other symptoms as well.

*Symptoms occurring between 2-6 hours after ingestion of food, are not included in the classification of clinical symptoms described in the introduction. (7) However, AD reactions may occur in this time period and are therefore included in this study.

Statistical analysis

The prevalence of the clinical reactions was determined by subtracting the percentage of reactions on placebo days from the percentage on verum days. P-values were considered significant under 0.05. Multiple logistic regression analyses were used to determine: the association between current AD and the outcome of the DBPCFC, a previous history of AD and the outcome of the DBPCFC and between the severity of AD and the outcome of the DBPCFC and was corrected for gender, age, suspected food, sIgE, asthma and rhinoconjuntivitis. To correct for type of food, the different foods were categorized in the following categories: peanut, cow’s milk, hen’s egg, nuts and a group with remaining foods. The severity of AD was analysed as a continuous independent variable in logistic regression analysis, because a better model fit was obtained than using a categorical variable. To correct for IgE we used the natural logarithm of IgE. The placebo-controlled prevalence was calculated by subtracting the prevalence of placebo reactions from the prevalence of verum reactions.

The McNemar test was performed tot test whether verum reactions occurred more frequently than placebo reactions within the same subjects for different types of clinical reactions.. A chi-square test was used to test for the difference in the placebo-controlled prevalences in children with positive DBPCFCs and negative DBPCFCs. A chi-square test for trend was used to test for the difference in the placebo-controlled prevalences in children with mild, moderate and severe AD. SPSS statistics 20 was used to analyse the data.

Results

In total, 1186 DBPCFCs were performed in 682 children. Of these 682 children, 381 children (55.9%) had a current history for AD, 282 children (41.3%) did not have a current history for AD and for 19 children (2.8%) data were missing. A previous history of AD (at some point in their lives), was present in 578 children (84.8%). Of all children with current AD, 182 children (47.8%) also had a current history of asthma and 148 (38.8%) had also a current history of rhinoconjunctivitis. A previous history of asthma was present in 193 children (50.7%) and a previous history of rhinoconjunctivitis in 156 children (40.9%).

The percentage male/female, the mean age, the mean IgE results and the percentage positive DBPCFCs were comparable between children with and without current AD.

See table. 1 for descriptive data on these children with and without current AD.

Association between outcome DBPCFC and AD

Out of the 381 children with a current history of AD, 196 (51.4%) children had a positive DBPCFC. Of all children without a current history of AD, 48.7% had positive DBPCFCs. No association between the result of the DBPCFCs and the presence of AD was found (odds ratio 0,90 [0.600; 1.351] p=0.612). In the group of 381 children with AD, 196 verum reactions occurred, and 83 placebo reactions (placebo-controlled prevalence of 113). In the group of 189 children without AD, 92 verum reactions occurred and 24 placebo reactions (placebo-controlled prevalence is 68). No significant difference between the placebo-controlled prevalence of children with and without AD could be found (p=0.180).

Of all children with a previous history of AD, 50.5% had a positive DBPCFC. No assocation was found between a previous history of AD and the outcome of the DBPCFC (odds ratio= 0.825 [0.468; 1.455] p= 0.506).

Severity of AD and rate of positive DBPCFCs

Out of the 381 children with AD at the start of the challenge, 277 children (72.7%) were categorized as having mild AD, 70 (18.4%) as having moderate AD and 4 (1.0%) as having severe AD. Positive DBPCFCs were seen in 141 (50.9%) of the children with mild AD, in 36 children (51.4 %) with moderate AD, and 3 (75.0%) with severe AD.

A higher SCORAD, was not significantly associated with the occurrence of a positive DBPCFCs (p= 0.865).

In the group of patients with mild AD ,150 were verum reactions after DBPCFC occurred and 49 placebo reactions.

In the group with moderate AD, 38 verum reactions and 19 placebo reactions occurred. In the group with severe AD 3 verum and 0 placebo reactions occurred.

This results in a placebo-controlled prevalence of 101 (mild), 19 (moderate) and 3 (severe). These prevalences did not differ significantly, using a chi-square test for Trend (χ² 0.241 p=0.624).

Clinical reactions in patients with AD

Clinical reactions with different time courses following DBPCFC are shown in table 2. Immediate reactions, consisting of an exacerbation of AD only, occurring within 2 hrs after ingestion of the last dose, were rarely seen. An exacerbation of AD only, occurred in 2 patients out of 326 (0.6%) on a verum day and in 1 patient on a placebo day.

Immediate reactions with symptoms other than AD, were seen more frequently. On verum days, 129 children (39.6%) showed symptoms other than AD, such as gastrointestinal symptoms or respiratory symptoms. On placebo days, this prevalence was only 26 (6.8%). Using Mc Nemar’s test, the difference prevalence of these immediate symptoms other than AD between verum and placebo days, was significant. (p=<0.01). Also, the prevalence of both an exacerbation of AD together with the occurrence of other symptoms, was significantly higher on verum days (p=0.000). (see table 2) Fourteen children (4.3%) had an exacerbation of AD and other symptoms on the verum day in contrast to 0 children on placebo days.

Intermediate reactions, occurring within 2-6 hrs after ingestion of the last dose, were rarely seen and were not seen more often on verum days than on placebo days. This was the case for an exacerbation of AD after an immediate reaction (2.5% on verum days vs. 1.8% on placebo days), as well as for an exacerbation of AD without a previous immediate reaction (1.5% on verum days vs. 1.2% on placebo days).

Late eczematous reactions, occurring after 6-48 hrs after ingestion of the last dose, were also rarely seen. The total prevalence of late eczematous reactions was 1.5% on verum days and 1.2% on placebo days.

History of exacerbation of AD with and without other symptoms

Of all 317 children with a history relating to the suspected food, 19 children had a history of only an exacerbation of AD. The other children had a history of an exacerbation of AD and other symptoms or a history of only non-eczematous symptoms. In these children 6 verum reactions and also 6 placebo reactions occurred.

Discussion

This study shows that a current or previous history of AD in children suspected of food allergy, does not increase the risk of having a true food allergy confirmed by a positive DBPCFC. Also, the severity of AD is not associated with higher prevalences of food allergy, i.e. children with severe AD are not more at risk for testing positive for food allergy, than children with mild AD. Isolated eczematous reactions, both early and late occurred rarely during DBPCFC in children with AD.

In our study we did not find a significant association between a previous and/or current history of AD and food allergy, confirmed by a positive DBPCFC. The relationship between AD and food allergy has been theorized to be due to increased skin permeability, resulting in increased exposure of the immune system to food allergens resulting in sensitization. This is sometimes referred to as the outside-in theory and could explain why children with AD have higher prevalences of food allergies. (20) This is also suggested by a study of Lack et al, where it was shown that children who used skin preparations containing peanut oil, had an increased chance of developing a food allergy for peanut at the age of 5 years. (21) In another study Hill et al (22) found in a large cohort study, conducted in 12 countries, that IgE sensitization to food is more frequent in children with AD. Furthermore, IgE sensitization is commonly associated with food allergies. While these data support the notion that AD may promote sensitization to food and thus contribute to a relatively high prevalence of food allergy in patients with AD, our results show that AD has no additional effect on the development of clinical food allergy in sensitized AD patients. Also an association was found between FLG-mutations and peanut allergy, after controlling for the presence of AD.by Brown et al. (..)

Children with more severe forms of AD are said to be more at risk for food allergy than children with milder AD. For example, Hill et al. found that children with severe forms of AD had a relative risk of 5.9 for having a IgE-mediated food allergy compared to children without AD. (12) Guillet found that milder forms of AD are not influenced by allergic factors, as opposed to severe AD (17).

We did not find a higher prevalence of food allergy in children with severe AD as opposed to children with milder forms of AD. In our study population all children had a history suspected of food allergy. This suggests that in children with AD and a history suggestive of food allergy, the severity of AD does not affect the risk of having true food allergy, as confirmed by a positive DBPCFC.

The clinical reactions after a DBPCFC can be classified as early and late reactions. In this study we made a more detailed classification, so we could accurately identify the true course of reactions and compare these to those reported by other authors.

We also compared the reactions after verum and placebo, to correct for exacerbations unrelated to the challenge. We found that exacerbations of AD without other symptoms occurring at any time were rare. Reactions on verum days did not occur significantly more often than on placebo days. If an eczematous reaction occurred, other symptoms almost always occurred as well.

In 2006, Rowlands (18) found that eczematous reactions after food challenge were rare in agreement with our study. Children had no specific history of food allergy. They performed open challenges and only when results were equivocal, children underwent a double blind challenge. Only one positive eczematous reaction occurred and this was within 2 hrs. They concluded that AD is rarely induced even in the highest-risk patients. Including patients without a history suggestive of food allergy may have contributed to the low prevalence of food related AD found in this study. Although our patients all had a history suggestive of food allergy, we also found a low prevalence of isolated eczematous reactions. However, reactions including exacerbations of AD, together with other symptoms, were not uncommon.

Several authors reported on the prevalence of food allergy in children with AD and the prevalence of skin symptoms after DBPCFC. Sampson and McCaskill found that 56% of children with severe AD, had positive food challenges. In 84% of 101 positive food challenges, cutaneous symptoms occurred. In 34 % the skin symptoms were the only symptoms. Skin symptoms were defined as a diffuse erythematous macular of morbiliform rash and pruritus. 23 In a prospective study, Eigenmann et al. found that about one third of the children with moderate to severe AD, have a clinical reactivity to foods. Children were considered allergic, based on positive challenge results, a convincing history and positive specific IgE-results (using the CAP system), a convincing history and positive specific IgE-results and/or skin prick tests, or a CAP-system value of greater than the 95% predictive cut-off. Children reacted with cutaneous symptoms in 95% of positive challenges, and in 83% isolated cutaneous reactions occurred.Children enrolled in this study were not selected on having a history of reacting to food. 24 Another study on the prevalence of food hypersensitivity in children with AD, was performed by Burks et al. They evaluated 46 patients with mild to severe AD for food allergy. In total, 65 DBPCFCs were performed, of which 27 were considered as positive. All symptoms occurred within 2 hours. Skin symptoms (diffuse erythematous macular of mrbiliform rash and pruritus) occurred in 96% of all positive challenges. Isolated cutaneous symptoms were seen in 30 % of positive reactions. Children were not specifically suspected of food allergy. 25 In a similar, second study, Burks et al. found positive test results in 38.7% of 266 DBPCFCs. All symptoms occurred within 2 hours after ingestion. Skin symptoms were seen in 78% of all positive challenges. Isolated cutaneous symptoms were seen in about 50% of positive challenges. 26

These prevalences of isolated skin symptoms in these studies, were considerably higher than the prevalences found in our study.

Two more detailed studies have been performed on the clinical reactions in children with AD, suspected of food allergy (15,19).These retrospective studies included children with mild to severe (15) and moderate to severe (19) AD. In the study by Breuer et al. (15) in total 49 reactions were positive on verum days. Only 2 placebo reactions occurred. Immediate-type reactions were defined as symptoms occurring within 6 hrs after ingestion of the last dose. Immediate-type AD reactions (isolated or combined), occurred in 43 of 49 (88%) positive reactions of which 21 were isolated immediate reactions. There were 22 late eczematous reactions after immediate reactions and 6 isolated late eczematous reactions occurred. Niggeman et al. (19) found that on verum days, immediate isolated eczematous reactions occurred in 47 of 81 immediate reactions (40.5%). Only 1 positive placebo reaction occurred. AD in combination with urticaria occurred in 19 of 81 early reactions (16.4%). Late eczematous reactions (after 2 hrs) occurred in 22 out of 29 late reactions (18.9%), eczematous reactions combined with urticaria occurred in 3 out of 29 late reactions (2.6%). In 6 cases immediate and late reaction occurred, of which 5 were eczematous reactions and AD and urticaria occurred in 1 out of 6.

The studies of Breuer et al. and Niggemann et al. show higher percentages of immediate and late isolated eczematous reactions after DBPCFC than our study. False-positive challenge results seem an unlikely explanation for this discrepancy, as both studies showed very few placebo reactions. Unforeseen forms of selection bias may have occurred and given rise to these differences, but there is no indication that the study population in these reports are different from those described here. However, time-related factors may be involved. The main authors of these reports were interviewed and both thought the prevalences of isolated eczematous reactions to be currently lower than previously reported. Both authors described such reactions as being "rare". (personal communication Bodo Niggemann and Thomas Werfel) This is further supported by the finding that patients presenting with a history of isolated eczematous reactions had a placebo-controlled prevalence of food allergy of nihil: the number of children that reacted on verum days and placebo days was equal.

In conclusion, we found that a (previous) history of AD does not significantly increase the chance of testing positive with DBPCFC .In children with AD and a history suggestive of food allergy, the severity of AD does not affect the risk of having true food allergy, as confirmed by a positive DBPCFC. Such children should therefore be considered for food allergy testing, regardless of the severity of their AD. However, in our population of children with AD and a history suggestive of food allergy, isolated eczematous reactions were seen rarely. When AD exacerbations occurred, they almost always occurred in combination with other symptoms. Furthermore, the placebo-controlled prevalence of challenge proven food allergy in children with histories of isolated eczematous reactions to foods in this study was nihil. Our results thus suggest that children with an exacerbation of AD in the absence of other symptoms in their history are unlikely to be food allergic. These results also suggest that patients with chronic AD without a history suggestive of food allergy are unlikely to be food allergic.