Clinical & Experimental Allergy, 44, 1420–1430
doi: 10.1111/cea.12393
ORIGINAL ARTICLE
© 2014 John Wiley & Sons Ltd
Allergens
Wheat allergy in children – new tools for diagnostics M. J. M€akel€a1, C. Eriksson2*, A. Kotaniemi-Syrj€anen1*, K. Palosuo1, J. Marsh3, M. Borres2, M. Kuitunen1 and A. S. Pelkonen1 1
Department of Allergology, Skin and Allergy Hospital, Helsinki University Central Hospital, Helsinki, Finland, 2Thermofisher Scientific/Phadia AB, Uppsala,
Sweden and 3Institute of Inflammation and Repair, University of Manchester, Manchester, UK
Clinical & Experimental Allergy
Correspondence: Mika J. M€akel€a, Skin and Allergy Hospital, P.O. Box 160, FIN-00029 HUS, Helsinki, Finland. E-mail: [email protected] Cite this as: M. J. M€akel€a, C. Eriksson, A. Kotaniemi-Syrj€anen, K. Palosuo, J. Marsh, M. Borres, M. Kuitunen and A. S. Pelkonen. Clinical & Experimental Allergy, 2014 (44) 1420–1430.
Summary Background The detection of wheat-specific IgE in children often leads to a suspicion of wheat allergy, but little information is available on the most reliable wheat allergens for predicting clinical reactivity. Objective To evaluate the role of allergenic components of wheat in wheat allergy diagnostics. Methods One hundred and eight children (median age 1.5 years; range 0.6–17.3 years) with suspected wheat allergy underwent open or double-blinded, placebo-controlled oral wheat challenges. Responsiveness to different allergenic components of wheat was studied by skin prick tests and by determination of serum IgE antibodies using a semi-quantitative microarray assay. Results Thirty (28%) children reacted with immediate symptoms, and 27 (25%) with delayed symptoms to ingested wheat, whereas 51 (47%) children exhibited no reactions in oral wheat challenges. Positive IgE responses to any of the 12 allergenic components of wheat was seen in 93%, 41%, and 43% of those with immediate, delayed or no reactions to ingested wheat, respectively (P < 0.001 to P < 0.05 in every comparisons between those with immediate reactions and those with no reactions). Positive IgE responses to ≥5 different allergenic components improved significantly the diagnostic accuracy (with a positive likelihood ratio (LR+) of 5.10). Alpha-amylase inhibitors (AAI), in particular dimeric AAI 0.19 (LR+ 6.12), alpha-, beta-, and gamma-gliadins (LR+ from 3.57 to 4.53), and high-molecular-weight (HMW) glutenin subunits (LR+ 4.37) were the single allergenic components of wheat differentiating most effectively those with immediate symptoms from those who did not exhibit any reactions. Conclusions and Clinical Relevance Wheat allergy diagnostics is difficult, even using sophisticated component methods. Our results confirm earlier findings about gliadins and identify the dimeric AAI 0.19, as a relevant allergen in clinically reactive patients when compared to non-reactive subjects. The accuracy of wheat allergy diagnosis may be improved by measuring IgE responses to several components of wheat. Keywords alpha-amylase inhibitor, component diagnostics, gliadins, glutenins, wheat allergy Submitted 12 February 2013; revised 7 August 2014; accepted 14 August 2014
Introduction The prevalence of food allergies varies in different regions of the world possibly due to differences in the age of introduction of solid foods into diet, eating habits, as well as sensitization to pollens. Some foods, including cow’s milk, hen’s egg, wheat, soy, peanut, and fish, however, account for the majority of the doc*Both authors contributed equally to the study.
umented food hypersensitivity in most geographical regions [1]. Wheat allergy prevalence varies depending on the age and region from 0.4% to 1% [1, 2]. In Finland, wheat allergy is an important clinical entity and is often diagnosed in children with multiple food allergies. According to a recent survey, wheat allergy was diagnosed by a physician among 0.9% of children aged 1–4 years in south-eastern Finland [3]. Skin prick tests that are commonly used in allergy diagnostics are susceptible to technical errors and may
New tools for wheat allergy diagnostics
exhibit false positive reactions in patients with dermographism, and false negative results if performed during antihistamine medication. Measuring serum specific IgE levels is an alternative diagnostic option. As regards wheat allergy diagnostics, an important obstacle has been a lack of reliable allergens in measurement of the IgE response. Although specific IgE levels can be used to predict symptomatic allergies for many foods, wheatspecific IgE is highly non-specific, detects only a part of wheat protein fractions, and the level of specific IgE required to clinical response with a certainty of 95% is high, up to 80 kU/L [4]. To improve the diagnostic value of specific IgE measurements, immune responses to different wheat proteins need to be studied in different clinical phenotypes of allergy. Wheat proteins can be broadly divided into watersoluble albumins, salt-soluble globulins, and insoluble prolamins, including the gliadins which are soluble in aqueous alcohols, and the glutenins (Fig. 1) [2, 5, 6]. Each of these fractions contains allergenic proteins which have been associated with clinical symptoms. One of the best documented form of wheat allergy is the association of sensitization to omega-5-gliadin with wheat-dependent, exercise-induced anaphylaxis (WDEIA) [7]. Other relatively well-documented allergens include alpha-amylase inhibitors (AAI), the response to which is associated with food allergy in children with atopic dermatitis [2]. However, there is generally a significant overlap of the responses to individual proteins in different disease conditions, with AAIs and gliadins both being important allergens in the respiratory allergy to wheat flour, that is baker’s asthma. In this study, 12 known wheat allergen components were used to study wheat-specific IgE responses in a well-characterized group of children with suspected wheat allergy. Materials and methods Patients The study subjects included 108 children, aged from 0.6 to 17.3 years (median 1.5 years, 73% younger than 3 years) who had been referred to the Department of Allergology, Helsinki University Central Hospital, Helsinki, Finland, because of a suspicion of wheat allergy, and who underwent oral wheat challenge (81% open, 19% double-blind placebo-controlled (DBPC)) between October 2006 and February 2009. The coeliac disease was excluded on a clinical basis, but not routinely, before the wheat challenge. The details for the open oral wheat challenge have been described before [8]. Whether wheat challenges were performed as open or DBPC was based on earlier food-related symptoms: wheat challenges were per-
1421
Wheat protein solubility fractions
Albumin/ Globulin Water/dilute saline
Alpha-amylase inhibitors (AAI)
Gluten
Monomeric gliadins Alcohol/water mixtures
Lipid transfer protein (LTP)
Avenin-like protein
Polymeric glutenins Dilute acid
Alpha-gliadin Beta-gliadin
Gamma-gliadin
High molecular weight (HMW) glutenins
Low molecular weight (LMW) glutenins
Slow-omegagliadin Omega-5gliadin
Fig. 1. Wheat proteins grouped.
formed routinely as DBPC in IgE-negative children with delayed symptoms [9]. In both the DBPC and open challenges, the challenge began with the application of a drop of wheat porridge (10% wheat, cooked in water) onto the mucosa of the inner side of the lower lip; thereafter, wheat was given orally in increasing amounts (1 g, 5 g, and 10 g; containing wheat protein 0.1 g, 0.5 g, 1 g, respectively) at 1 h intervals. The porridge used in the DBPC wheat challenges was potato-based and was prepared on site by the hospital kitchen staff not involved in the food challenge process. The supervising physician examined the child for the appearance of any symptoms. The exact time, nature, and eliciting dose of each observed adverse reaction were recorded. It was predetermined that in case of doubtful reactions, the challenge was to be continued until there are obvious signs of positive reactions. The challenge was stopped when objective signs and symptoms indicated a positive reaction [9], and when needed, the symptoms were treated with antihistamine, and/or intramuscular epinephrine, and/or oral steroids. In case the child did not exhibit immediate symptoms,
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
1422 M. J. M€akel€a et al the challenge was continued at home, with at least 20 g of wheat per day for 1 week, and the parents were advised to keep a daily symptom diary. In case of a suspected delayed skin reaction, an extra follow-up visit was organized at the time of the reaction, and skin symptoms and findings were evaluated and recorded by a study physician. Otherwise, the daily symptom diary was evaluated by a study physician during a separate follow-up visit at the end of DBPC challenges. In DBPC challenges, the wheat-containing food product and placebo were given in a random order 2 weeks apart, including one wash-out week in between the challenges. Definitions A parental history of atopy was regarded as physiciandiagnosed atopic eczema or allergic rhinoconjunctivitis in either of the parents. A parental history of asthma was regarded as physician-diagnosed asthma in either of the parents. Atopic eczema, allergic rhinitis, and asthma were defined as a specific diagnosis made by a physician or paediatrician. A skin prick test (SPT) wheal diameter of ≥3 mm was considered as a positive reaction against the tested allergen, that is wheat (in-house powdered whole grain wheat diluted with 0.9% NaCl as 1 : 10 weight/volume), and gliadin (1 mg/mL, Sigma, St. Louis, MO, USA). Histamine dihydrochloride (10 mg/mL, ALK) was used as a positive control, and physiologic saline as a negative control. In the diagnostic food challenge, symptoms appearing within the first 2 h were defined as immediate, and symptoms appearing after the first 2 h (up to 5 days) were defined as delayed [8]. ImmunoCAP analysis Serum samples, which were obtained just before the beginning of the wheat challenge, were analyzed with ImmunoCAPâ (Phadia AB, Uppsala, Sweden) for allergen-specific IgE antibodies to wheat, gluten and gliadin, and for total IgE. Samples with a specific IgE antibody concentration above the measuring range (0.1–100 kUA/L) were diluted and reanalyzed. Values lower than 0.1 kUA/L were given the value 0.05 kUA/L. Values of 0.35 kUA/L or greater were regarded as positive. Samples with a total IgE concentration above the measuring range (2–5000 kU/L) were diluted and reanalyzed. Microarray immunoassay Specific IgE to different wheat and Timothy grass pollen allergen components as well as to cross-reactive
carbohydrate determinants (CCD) was measured with an in-house experimental semi-quantitative microarray assay based on a micro spot technique as described elsewhere [10, 11]. The resulting fluorescence intensity was measured with BioAnalyzerâ (LaVision, BioTec, Bielefeld, Germany) and evaluated with GenePixâ Pro v. 5.1 (Axon Instruments, Molecular Devices, Sunnyvale, CA, USA), and classified as negative, low positive, and high positive. Low positive and high positive fluorescence intensity were denoted as positive IgE responsiveness in final analyses. In addition, high positive fluorescence intensity was evaluated separately, denoting strongly positive IgE responsiveness. Individual cut-off levels for positive IgE responses for each allergen component were estimated based on the fluorescence intensity for pools of negative control samples from non-atopic subjects, and negative control samples spiked with 3000 kUA/L of myeloma IgE, and calculated as 2 9 (the average fluorescence intensity + 3 standard deviations). A fluorescence intensity > 5000 units was regarded as a strong positive IgE responsiveness. Allergens in microarray Most of the allergen components were essentially purified according to previously published methods. Highand low-molecular-weight (HMW and LMW) glutenins were enriched according to Verbruggen [12]. Alpha-, beta-, and gamma-gliadins were purified by preparative electrophoresis as described by Rumbo et al. [13] The electrophoresis was performed with a gel of 7% acrylamide concentration, and the running buffer and upper and lower reservoirs were acidic (pH 3.1) aluminium lactate buffer [14]. Slow omega-gliadins were purified according to Tatham and Shewry [15]. Monomeric, dimeric and tetrameric AAI were purified according to Kashlan and Richardson [16]. The purification of lipid transfer protein (LTP) was performed by extraction of wheat at pH 7.5, followed by gel filtration, hydrophobic interaction chromatography at pH 8 (containing ammonium sulphate) and ion exchange chromatography at pH 4.5. Avenin-like protein was enriched to a purity of about 75% with anion exchange chromatography at pH 8.8 followed by gel filtration. Recombinant omega-5gliadin, Phl p 4, Phl p 5b, Phl p 12, and CCD were obtained from Phadia AB. Statistical analysis To evaluate the statistical differences between the groups, v2-test or Fisher’s exact test (if the expected frequency for any cell was < 5) were used for dichotomous variables, and Mann–Whitney U-test was used to analyze continuous or ordinal variables. The kappa (j)
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
New tools for wheat allergy diagnostics
statistics was applied to compare agreement in IgE responsiveness between different wheat allergens, and the j-value of ≤0.20 was used as a criterion to test the possible combinations of wheat allergens [17]. Sensitivity, specificity, positive predictive values (PPV), negative predictive values (NPV), and likelihood ratios for positive (LR+) and negative (LR ) test results were calculated using routine equations. LR+ values of ≥5 and LR values of ≤0.20 are considered to have significant effects on pre-test probability [18]. The receiver operating characteristic (ROC) curves were applied to evaluate the optimal number of positive and strongly positive IgE responses against different wheat allergenic components to differentiate children with immediate reactions from those with no reactions in oral wheat challenge. In addition, the ROC curves were utilized to find out the allergenic components of wheat that were specifically associated with severe immediate reactions in oral wheat challenges. Two-tailed tests were used in all analyses. P-values less than 0.05 were considered statistically significant. The data were analyzed using IBM SPSS 19.0 (SPSS, Inc, Chicago, Illinois) for Windows. Ethics The study was approved by the Research Ethics Committee of Helsinki University Central Hospital (approval number: 374/E7/2004) and is registered in the Clinical Trial Registry of Hospital District of Helsinki and Uusimaa (HUS). The registration number is 214414. Before participation of the study, written informed consent was obtained from parents of all study children. Results Oral wheat challenge Baseline data for the study children are presented in Table 1. Eighty-five (79%) children had parental history of atopy or asthma, and 95 (88%) had clinical atopic manifestations or asthma. Skin symptoms were the most common symptoms related to food ingestion, in 100 (93%), followed by gastrointestinal and respiratory tract symptoms, in 57 (53%), and 38 (35%), respectively. During oral wheat challenge, 30 (28%) of the 108 study children reacted with immediate symptoms, 27 (25%) with delayed symptoms, and in 51 (47%) children, no wheat-ingestion related reaction was seen. Among the 30 children reacting with immediate symptoms, in 11 cases the maximum dose of ingested wheat was 1 g or less. Skin reactions were seen in 47 (44%), most commonly urticaria or erythema. Twenty-eight children (26%) reacted with gastrointestinal and 17 (16%) with respiratory symptoms. Altogether 22 chil-
1423
dren needed medication because of the immediate reaction during oral wheat challenge. Among these 22 children, there were 15 who received only antihistamine medication, and seven children who received antihistamine and intramuscular epinephrine, including two children who also received per oral corticosteroid medication. Those seven children who needed intramuscular epinephrine had either severe skin symptoms (i.e. generalized urticaria, generalized erythema) [9], or symptoms fulfilling diagnostic criteria for anaphylaxis [19]: severe skin symptoms in conjunction with respiratory symptoms (i.e. frequent dry cough, severe sneezing, and rhinorrhea) or in conjunction with gastrointestinal (i.e. vomiting, abdominal pain) symptoms, or respiratory symptoms in conjunction with gastrointestinal symptoms (Table 2). SPT and specific IgE measurements SPT results and IgE antibody responses are presented in Table 3. Patients with immediate symptoms differed significantly (P < 0.001 to P < 0.05) from those with delayed or no reaction to ingested wheat in every comparison, whereas there were no differences between the children with delayed or no reactions in the oral wheat challenge. Children with immediate reactions in the oral wheat challenge were nearly all positive in SPT (i.e. having a wheal diameter of ≥3 mm) with whole wheat as an allergen. However, approximately a half of the clinically non-reacting patients were also positive, rendering the specificity of wheat-SPT (with a wheal diameter of ≥3 mm) poor, 55%, and the LR+ low, 2.14. By increasing a SPT wheal diameter as a cut-off point, the specificity did improve, but in expense with reduced sensitivity. The same phenomenon was seen as regards gliadin-SPT, as well as specific IgE against wheat, gluten and gliadin, with only a little benefit of lifting specific IgE concentration up to 10- to 30-fold from the detection level (Table 3). Microarray-based specific IgE measurements A protein microarray assay was designed for analysis of specific IgE responses to allergenic components from wheat and Timothy grass. The protein microarray contained 12 wheat allergens, three pollen allergens from the Timothy grass and CCD. Allergens from Timothy grass were included in the array to identify Timothy grass pollen allergic patients that could have specific IgE antibodies to wheat allergen components due to IgE cross-reactions between homologous proteins in Timothy grass pollen and wheat grain. In this study, very few patients (< 10%) were sensitized to any of the three Timothy grass pollen proteins. The percentages of children with specific IgE to the different wheat compo-
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
1424 M. J. M€akel€a et al Table 1. Baseline characteristics of the study children. Reaction in oral wheat challenge
Baseline characteristics Age (yr) Male Parental history of atopy Parental history of asthma Atopic eczema Allergic rhinitis Asthma Skin prick test positive ever to cow’s milk, hen’s egg, and/or soya bean Parentally reported symptoms related to food ingestion ever Immediate-type, AW+S+GI / AW+S / AW+GI / S+GI Immediate-type, only S or GI Delayed-type, only S Delayed-type, predominately GI
All study children, n = 108
Immediate, n = 30
Delayed, n = 27
No reaction, n = 51
1.5 60 83 27 90 37 20 79
(0.6; 17.3) (56%) (77%) (25%) (83%) (34%) (19%) (73%)
1.8 18 23 6 27 11 7 28
(0.7; 17.3) (60%) (79%) (21%) (90%) (37%) (23%) (93%)a
1.5 19 22 8 21 9 4 16
(0.8; 11.4) (70%) (85%) (31%) (81%) (35%) (15%) (59%)
1.4 23 38 13 42 17 9 35
(0.6; 11.8) (45%) (79%) (27%) (82%) (33%) (18%) (69%)
53 18 21 16
(49%) (17%) (19%) (15%)
22 4 4 0
(73%)b (13%) (13%) (0%)
14 1 5 7
(52%) (4%) (19%) (26%)d
17 13 12 9
(33%) (26%)c (24%) (18%)e
P = 0.010 vs. No reaction. P = 0.001 vs. No reaction. c P = 0.027 vs. Delayed reaction. d P = 0.003 vs. Immediate reaction. e P = 0.023 vs. Immediate reaction. Immediate-type symptoms: AW: cough, rhinitis, shortness of breath, wheezing; S: urticaria, erythema; GI: oral pruritus, vomiting, abdominal pain. Delayed-type symptoms: S: exacerbation of eczema; GI: diarrhoea, abdominal pain. Analyses were performed using Mann–Whitney U-test or chi-squared test or Fisher’s exact test. Data are presented as median (range), or n (%). AW, airway; S, skin; GI, gastrointestinal. a
b
Table 2. Characteristics of patients with severe immediate reaction in oral wheat challenge.
Patient No.
Age
Sex
SPT+
1 2 3
0.65 3.03 6.08
M F M
Yes Yes Yes
4
1.10
F
Yes
5
3.22
F
Yes
6 7
0.87 0.84
F F
Yes Yes
Total amount of wheat (g) in OFC
Parentally reported food-related symptoms Immediate: Urticaria Immediate: Urticaria, vomiting, cough Immediate: Urticaria, vomiting, abdominal pain, cough Immediate: Urticaria, vomiting, abdominal pain Immediate: Urticaria, wheezing, shortness of breath Delayed: Exacerbation of eczema Immediate: Urticaria, wheezing
6 6 1 16 1 16 16
Reaction to wheat in OFC Generalized urticaria, erythema, vomiting Localized urticaria, cough, rhinitis Oral pruritus, rhinitis, abdominal pain, vomiting Generalized urticaria, cough Generalized urticaria, erythema Generalized erythema Generalized urticaria
M, male; F, female; SPT+, skin prick test positive ever to cow’s milk, hen’s egg or soya bean; OFC, oral food challenge.
nents are shown in Table 4. Specific IgE to all wheat allergen components differed significantly (P < 0.001 to P < 0.05) between children with immediate reactions during the oral wheat challenge and those with delayed or no reactions to wheat ingestion. By contrast, there were no differences between children with delayed or no reactions in the oral wheat challenge. The allergens differentiating most effectively children with immediate
reactions from those with no reactions were AAIs, alpha-, beta-, and gamma-gliadins, as well as HMWglutenin subunits (Table 4). In search of an optimal combination of sensitivity and specificity, AAI 0.19, in particular, emerged as an important component, with a LR+ of 6.12–6.80. We also analyzed specific IgE to wheat allergens in different combinations (based on poor agreement
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
New tools for wheat allergy diagnostics
1425
Table 3. Skin prick test results and immunoglobulin E responses with regard to the type of reaction in wheat challenge. Reaction in oral wheat challenge
Skin prick test results and IgE responses Skin prick test reaction to Wheat (mm) ≥3 mm ≥4 mm ≥5 mm ≥6 mm ≥7 mm Gliadin (mm) ≥3 mm ≥4 mm ≥5 mm Total IgE (kU/L) Specific IgE against Wheat (kUA/L) ≥0.35 kUA/L ≥3.50 kUA/L ≥10.50 kUA/L Gluten (kUA/L) ≥0.35 kUA/L ≥3.50 kUA/L ≥10.50 kUA/L Gliadin (kUA/L) ≥0.35 kUA/L ≥3.50 kUA/L ≥10.50 kUA/L
Immediate n = 30 7 29 29 27 19 15 4 24 23 12 211
LR+*
(0; 12)a,b 2.14 (97%)a,b (97%)a,b 2.24 (90%)a,b 3.06 (63%)b,c 5.38 (50%)b,c 8.50 (0; 9)a,b 2.55 (80%)a,b (77%)a,b 3.91 (40%)b,d 5.1 (12; 5027)c,e
LR *
PPV*
NPV*
0.06 0.06 0.14 0.42 0.53
56% 57% 64% 76% 83%
97% 97% 92% 80% 76%
0.29 0.29 0.65
60% 70% 75%
85% 85% 72%
18.23 (0.05; 983.40)a,b
Delayed n = 27 0 13 11 7 5 3 0 9 7 3 45
LR+†
(0; 10) (48%) 1.07 (41%) 0.94 (26%) 0.88 (19%) 1.57 (11%) 1.89 (0; 5) (33%) 1.06 (26%) 1.32 (11%) 1.42 (5; 7559)
PPV†
NPV†
0.94 1.04 1.05 0.92 0.94
36% 33% 32% 46% 50%
67% 64% 64% 67% 67%
0.97 0.92 0.96
36% 41% 43%
66% 67% 66%
LR
†
0.41 (0.05; 76.43)
29 24 18 15.26
(97%)a,b 2.35 (80%)a,b 3.40 (60%)a,b 4.37 (0.05; 814.83)a,b
0.06 0.26 0.46
58% 67% 72%
97% 87% 79%
14 6 4 0.15
(52%) 1.26 (22%) 0.94 (15%) 1.08 (0.05; 47.20)
0.82 1.02 0.99
40% 33% 36%
70% 65% 66%
28 23 17 14.25
(93%)a,b 2.16 (77%)a,b 3.01 (57%)a,b 4.82 (0.06; 868.97)a,b
0.12 0.31 0.49
56% 64% 74%
94% 84% 78%
12 6 3 0.20
(44%) 1.03 (22%) 0.87 (11%) 0.94 (0.07; 45.80)
0.98 1.04 1.01
35% 32% 33%
66% 64% 65%
0.13 0.31 0.49
53% 66% 74%
93% 85% 78%
1.09 1.02 1.01
32% 33% 33%
63% 65% 65%
28 (93%)a,b 23 (77%)a,b 17 (57%)a,b
1.90 3.26 4.82
12 (44%) 6 (22%) 3 (11%)
0.91 0.94 0.94
No reaction n = 51 0 23 22 15 6 3 0 16 10 4 79
(0; 10) (45%) (43%) (29%) (12%) (6%) (0; 5) (31%) (20%) (8%) (2; 5031)
0.25 (0.05; 1214.15) 21 (41%) 12 (24%) 7 (14%) 0.13 (0.05; 801.52) 22 (43%) 13 (26%) 6 (12%) 0.32 (0.05; 838.97) 25 (49%) 12 (24%) 6 (12%)
*Immediate vs. no reaction. † Delayed vs. no reaction. a P < 0.001 vs. delayed reaction group. b P < 0.001 vs. no reaction group. c P < 0.01 vs. delayed reaction group. d P < 0.05 vs. delayed reaction group. e P < 0.05 vs. no reaction group. Analyses were performed using Mann–Whitney U-test or chi-squared test. Data are presented as median (range), or n (%); percentages in brackets represent sensitivity for the immediate and delayed groups, and 1-specificity for the no reaction group. LR+, positive likelihood ratio; LR , negative likelihood ratio; PPV, positive predictive value; NPV, negative predictive value.
(j-value of ≤0.20) between the tested two components) in search of an optimal diagnostic tool. However, the LR+ values for different combinations were only modest (2.44–3.09), and inferior to individual allergenic components, making the combinations clinically useless. Significant differences were found in the number of wheat allergens with positive or strongly positive specific IgE responses between those with immediate reactions and those with delayed or no reactions in the oral wheat challenge (Table 4). ROC-curve revealed that positive IgE responses to ≥5 wheat allergens and
strongly positive IgE responses to ≥3 wheat allergenic components gave a moderate sensitivity (70% and 60%, respectively) with a good specificity (86% and 88%, respectively) resulting in both cases with a LR+ of 5.10 for immediate reactions in the oral wheat challenge. Severity of reactions The omega-5-gliadin-specific IgE antibody level has earlier been suggested to serve as a risk marker for reactivity in oral challenge [20]. We categorized the
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
Number of positive responses Any
(60%)a,b (33%)c,f (83%)a,b (67%)b,c
(53%)b,c (33%)c,d (60%)b,c (37%)a,b (70%)a,b (43%)b,c (63%)d,e (53%)a,b (50%)c,d (27%)
(33%)d,e (23%)f (60%)a,b (27%)d (37%)c,d (17%)d (33%)e,f (30%)d,e (47%)e,f (30%)f
7 (0; 12)a,b 28 (93%) a, b
18 10 25 20
Polymeric glutenins High-molecular-weight (HMW) glutenins Strongly positive Low-molecular-weight (LMW) glutenins Strongly positive
All
16 10 18 11 21 13 19 16 15 8
10 7 18 8 11 5 10 9 14 9
Immediate n = 30
Monomeric gliadins Alpha-gliadin Strongly positive Beta-gliadin Strongly positive Gamma-gliadin Strongly positive Slow omega-gliadin Strongly positive Omega-5-gliadin Strongly positive
Alpha-amylase inhibitors (AAI) AAI monomer 0.28 Strongly positive AAI dimer 0.19 Strongly positive AAI tetramer CM 2, 3, 16 Strongly positive Lipid transfer protein (LTP) Strongly positive Avenin-like protein Strongly positive Gluten group
Albumin/Globulin group
Positive IgE response against wheat components
2.16
4.37 2.83 2.50 3.09
4.53 8.50 4.37 6.23 3.57 5.53 2.15 3.89 2.83 2.72
5.67 3.97 6.12 6.80 4.68 NA 2.43 5.10 2.38 3.06
LR+*
0.12
0.46 0.76 0.25 0.43
0.53 0.69 0.46 0.67 0.37 0.61 0.52 0.54 0.61 0.81
0.71 0.81 0.44 0.76 0.69 0.83 0.77 0.74 0.66 0.78
LR *
Reaction in oral wheat challenge
56%
72% 63% 60% 65%
73% 83% 72% 79% 68% 77% 56% 70% 63% 62%
77% 70% 78% 80% 73% 100% 59% 75% 58% 64%
PPV*
94%
79% 69% 87% 80%
76% 71% 79% 72% 82% 73% 77% 76% 74% 68%
71% 68% 79% 69% 71% 67% 69% 70% 72% 69%
NPV*
(15%) (4%) (30%) (22%)
(11%) (0%) (19%) (0%) (19%) (4%) (30%) (7%) (11%) (7%)
(11%) (11%) (15%) (7%) (7%) (4%) (11%) (4%) (19%) (11%)
0 (0;11) 11 (41%)
4 1 8 6
3 0 5 0 5 1 8 2 3 2
3 3 4 2 2 1 3 1 5 3
Delayed n = 27
Table 4. Specific immunoglobulin E responses against wheat components with regard to the type of reaction in wheat challenge.
0.94
1.08 0.31 0.89 1.03
0.94 NA 1.35 NA 0.94 0.47 1.01 0.54 0.63 0.76
1.89 1.89 1.51 1.89 0.94 NA 0.81 0.63 0.94 1.13
LR+†
1.04
0.99 1.09 1.06 0.99
1.01 1.04 0.94 1.06 1.01 1.04 1.00 1.07 1.08 1.03
0.94 0.94 0.94 0.96 1.00 0.96 1.03 1.02 1.01 0.99
LR
†
33%
36% 14% 32% 35%
33% 0% 42% 0% 33% 20% 35% 22% 25% 29%
50% 50% 44% 50% 33% 100% 30% 25% 33% 38%
PPV†
64%
66% 63% 64% 66%
65% 65% 67% 64% 65% 64% 66% 64% 64% 65%
67% 67% 67% 66% 65% 66% 65% 65% 65% 66%
NPV†
(14%) (12%) (33%) (22%)
(12%) (4%) (14%) (6%) (20%) (8%) (29%) (14%) (18%) (10%)
(6%) (6%) (10%) (4%) (8%) (0%) (14%) (6%) (20%) (10%)
0 (0; 12) 22 (43%)
7 6 17 11
6 2 7 3 10 4 15 7 9 5
3 3 5 2 4 0 7 3 10 5
No reaction n = 51
1426 M. J. M€akel€a et al
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
(14%) (0; 9) (31%) (12%) 1.03 0.97 0.94 1.26 81% 79% 58% 75% 0.39 0.45 2.34 5.10
*Immediate vs. no reaction. † Delayed vs. no reaction. a P < 0.001 vs. delayed reaction group. b P < 0.001 vs. no reaction group. c P < 0.01 vs. delayed reaction group. d P < 0.01 vs. no reaction group. e P < 0.05 vs. delayed reaction group. f P < 0.05 vs. no reaction group. Analyses were performed using Mann–Whitney U-test or chi-squared test. Data are presented as median (range), or n (%); percentages in brackets represent sensitivity for the immediate and delayed groups, and 1-specificity for the No reaction group. LR+, positive likelihood ratio; LR , negative likelihood ratio; PPV, positive predictive value; NPV, negative predictive value; NA, not applicable.
33% 40%
65% 66%
7 0 16 6 66% 36% 0.99 1.08 0.35
≥5 Number of strongly positive responses Any ≥3
21 4 22 18
(70%)a,b (0;12)a,b (73%)b,c (60%)a,b
5.10
75%
83%
4 0 8 4
(15%) (0; 8) (30%) (15%)
LR LR * LR+* Immediate n = 30 Positive IgE response against wheat components
Table 4 (continued)
Reaction in oral wheat challenge
PPV*
NPV*
Delayed n = 27
LR+†
†
PPV†
NPV†
No reaction n = 51
New tools for wheat allergy diagnostics
1427
clinical reactions according to the severity to those who did not require epinephrine, and to those who needed epinephrine due to generalized symptoms and likely anaphylaxis (n = 7). Those who exhibited severe immediate reactions (i.e. required epinephrine) in the oral wheat challenge had the strongest IgE responsiveness against gamma-gliadin (sensitivity 86%, specificity 92%, LR+ 10.99, and LR 0.16 for strongly positive IgE responsiveness, with P < 0.001 with regard to those with no reactions to ingested wheat), and HMW-glutenin subunits (sensitivity 86%, specificity 86%, LR+ 6.26, and LR 0.17 for positive IgE responsiveness, with P < 0.001 with regard to those with no reactions to ingested wheat) (Fig. 2). Discussion The results of this study confirm the non-specific nature of whole wheat as a diagnostic marker for wheat allergy, and identify the dimeric AAI 0.19, as a relevant allergen in clinically reactive patients vs. those with specific IgE to wheat but no allergy. The accuracy of wheat allergy diagnosis could be improved by measuring IgE responses to several components of wheat. The measurement of wheat-specific IgE and its use for clinical diagnosis is problematic due to the low specificity when using whole wheat as a test allergen either in SPTs or in serum assays. Wheat-specific IgE is common among atopic children without true food allergies and this has not been taken into account in many studies. To avoid this pitfall, patient selection is of great importance. We used a real-life setting at the policlinics to which patients are referred either specifically for wheat challenge due to the detection of wheat-specific IgE and thus suspicion of wheat allergy, or because the patient is allergic to other foods and has symptoms suggestive of wheat allergy. This creates a pool of patients with negative challenge, and those with either acute or delayed symptoms, each of whom may be either IgE positive or negative for wheat proteins. We could, therefore, analyze a significant number of patients with different IgE and reactivity patterns reflecting the true conditions where the IgE tests are used. Our goal was to improve the IgE diagnostics by utilizing new, potentially more specific allergen components to identify patients likely to have acute and even severe reactions. Firstly, we characterized several glutenin components and different fractions of gliadin because these allergens have been shown to be important in WDEIA and in some cases of atopic dermatitis and anaphylaxis in adults and children [2]. Battais et al. [21] demonstrated that HMW-glutenin subunits were only minor allergens in 28 patients including 14 children with various symptoms of wheat allergy,
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
1428 M. J. M€akel€a et al 1.0
0.8
Sensitivity
0.6
0.4
Skin prick test reaction to wheat (mm). AUC 0.866 IgE to wheat (kU/L). AUC 0.874 IgE to high molecular weight glutenin (negative to strongly positive response). AUC 0.849 IgE to gamma-gliadin (negative to strongly positive response). AUC 0.881
0.2
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - Specificity
Fig, 2. Receiver operating characteristic curves for wheat allergenic proteins linked to immediate severe reactions in oral wheat challenge in comparison with ROC curves for skin prick test results and wheatspecific IgE.
whereas LMW-glutenin subunits were recognized by most patients with wheat allergy. By contrast, in our study, sera from more than half of the clinically reactive children recognized HMW-glutenin subunits more specifically than they did LMW-glutenin subunits. One explanation for this difference may be incomplete purification of the allergen in our study. However, onethird of the non-reactive children also showed IgE binding to LMW-glutenin subunits, making it practically useless as a clinical tool. IgE responsiveness to HMW-glutenin subunits, however, served as an important marker of severity of the oral wheat challenge reaction. Our results were in accordance with earlier studies of omega-5-gliadin [5, 20], which is clearly a more specific marker for wheat allergy diagnostics than whole wheat proteins. In the original report by Palosuo et al. [5], the specificity was up to 100% in wheat-sensitized children. In the present study, however, only 50% of the children with acute reactions, and 18% of those with no reactions had positive IgE responses to omega5-gliadin. Our results are clearly less specific and are closer to those of Ito et al. [20], obviously due to differences in patients selection. Alpha-gliadin was actually the best allergen in the gluten fraction in the microarray analysis, in terms of both sensitivity and specificity with regard to immediate reactions. Both alpha- and gamma-gliadins were also reported as important allergens by Pastorello et al. [22]. Moreover, in our study,
IgE responsiveness to gamma-gliadin in the microarray test served as an important marker for severity. It has been earlier shown that specific IgE from those suffering from anaphylaxis or urticaria detect sequential epitopes on gliadins whereas patients with atopic dermatitis probably recognize conformational epitopes [23]. Although this topic remains poorly understood, the development of improved purification procedures for native allergens is now providing more information about native gliadins and other wheat proteins as diagnostic tools and possibly even therapy [24]. AAIs have long been recognized as major allergens in baker’s asthma [25] but their role in patients reacting to ingested wheat was described first by James et al. [26]. As then, AAIs have been characterized in more detail in different disease phenotypes. AAIs are present both in the water-soluble and water-insoluble protein fractions and are not lost during cooking [25]. Different subunits of AAIs have been shown to bind IgE from both WDEIA patients as well as from patients with atopic dermatitis or anaphylaxis [22]. Based on earlier studies, the diagnostic value of different types has been difficult to evaluate. We wanted, therefore, to study the major monomeric, dimeric and tetrameric types of AAI in the microarray system. Children with immediate reactions in oral challenge most often recognized the dimeric AAI 0.19, and even children with delayed reactions recognized this protein slightly more often than those without a reaction. As dimeric AAI 0.19 gave the highest LR+ for immediate reaction in oral wheat challenge, superior to cut-offs for wheatSPT and wheat-specific IgE with similar levels of sensitivity (60–63%), measuring specific IgE to dimeric AAI 0.19 should be considered when aiming for the optimal diagnostic tool. Although IgE responsiveness to none of the tested wheat allergen combinations proved to be clinically useful, we found that the greater the number of wheat allergens with IgE responsiveness, the more likely were the immediate symptoms in oral wheat challenges. A recent study employed phage cDNA library to enrich sequences encoding IgE-binding proteins of wheat, showing reactions to proteins which had not previously been related to wheat allergy, such as thioredoxins and uncharacterized proteins [27]. Wheat grain contains a complex mixture of proteins and over 1000 individual components have been separated by two-dimensional electrophoresis [28]. Our wheat microarray concept could therefore be a feasible tool in further studies with novel proteins. One of the weaknesses of the study was that only a part of the challenges were performed as DBPC. In addition, as there are no international consensus statements on diagnosing delayed food-related symptoms, an inhouse protocol was applied. On the other hand, our
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
New tools for wheat allergy diagnostics
in-house protocol for diagnosing delayed wheat-related symptoms included DBPC challenges in conjunction with a daily symptom diary and a follow-up visit to minimize the number of false positive cases. Lack of children without a history of food-related symptoms as a control group can also be seen as a weakness of the study. In addition, semi-quantitative nature of data obtained from microarray assay did not allow all such analyses which were possible if the data were truly quantitative and continuous. In conclusion, wheat allergy diagnostics is difficult, even using sophisticated component methods. Our results confirm earlier findings about gliadins and identify the dimeric AAI 0.19, as a relevant allergen in clinically reactive patients when compared to non-reactive subjects. The accuracy of wheat allergy diagnosis may be improved by measuring IgE responses to several components of wheat.
References 1 Zuidmeer L, Goldhahn K, Rona RJ et al. The prevalence of plant food allergies: a systematic review. J Allergy Clin Immunol 2008; 121:1210–8. 2 Inomata N. Wheat allergy. Curr Opin Allergy Clin Immunol 2009; 9:238–43. 3 Pyrh€ onen K, N€ayh€a S, Kaila M, Hiltunen L, L€a€ar€a E. Occurrence of parentreported food hypersensitivities and food allergies among children aged 1– 4 yr. Pediatr Allergy Immunol 2009; 20:328–38. 4 Sampson HA. Utility of food-specific IgE concentrations in predicting symptomatic food allergy. J Allergy Clin Immunol 2001; 107:891–6. 5 Palosuo K, Varjonen E, Kekki O-M et al. Wheat x-5 gliadin is a major allergen in children with immediate allergy to ingested wheat. J Allergy Clin Immunol 2001; 108:634–8. 6 Shewry PR, D’Ovidio R, Lafiandra D, Jenkins JA, Mills ENC, Bekes F. Wheat grain proteins. In: Khan K, Shewry PR, eds. Wheat: chemistry and technology. St Paul (MN): American Association of Cereal Chemists, 2009:223–98. 7 Palosuo K, Alenius H, Varjonen E et al. A novel wheat gliadin as a cause of exercise-induced anaphylaxis. J Allergy Clin Immunol 1999; 103:912– 7. 8 Kotaniemi-Syrj€anen A, Palosuo K, Jartti T, Kuitunen M, Pelkonen AS, M€akel€a MJ. The prognosis of wheat
9
10
11
12
13
14
15
1429
Acknowledgement We are grateful to Mr Henry Akerstr€ om for help with the immunoassay analysis and to research nurse Tuija Rito for the work with children and their families. Funding Helsinki University Central Hospital Research Funds, Sigrid Juselius Foundation, Finnish Foundation for Allergy Research. Conflict of interest MSc Camilla Eriksson and MD Magnus Borres are employed by Thermofisher Scientific/Phadia. There are no other conflicts of interest to declare by anyone of the authors.
hypersensitivity in children. Pediatr Allergy Immunol 2010; 21:e421–8. Sampson HA, van Wijk RG, BindslevJensen C et al. Standardizing doubleblind, placebo-controlled oral food challenges: American Academy of Allergy, Asthma and Immunology – European Academy of Allergy and Clinical Immunology PRACTALL concensus report. J Allergy Clin Immunol 2012; 130:1260–74. Constantin C, Quirce S, Poorafshar M et al. Micro-arrayed wheat seed and grass pollen allergens for componentresolved diagnosis. Allergy 2009; 64:1030–7. € Asarnoj A, Moverare R, Ostblom E et al. IgE to peanut allergen components: relation to peanut symptoms and pollen sensitization in 8-year-olds. Allergy 2010; 65:1189–95. Verbruggen IM, Veraverbeke WS, Vandamme A, Delcour JA. Simultaneous isolation of wheat high molecular weight and low molecular weight glutenin subunits. J Cereal Sci 1998; 28:25–32. Rumbo M, Chirdo FG, Giorgieri SA, n MC. Preparative fracFossati CA, A~ no tionation of gliadins by electrophoresis at pH 3.1 (A-PAGE). J Agric Food Chem 1999; 47:3243–7. Clements RL. A continuous acetic acid system for polyacrylamide gel electrophoresis of gliadins and other prolamines. Electrophoresis 1988; 9:90–3. Tatham AS, Shewry PR. The conformation of wheat gluten proteins. The sec-
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
16
17
18
19
20
21
22
ondary structures and thermal stabilities of a-, b-, c- and x-gliadins. J Cereal Sci 1985; 3:103–13. Kashlan N, Richardson M. The complete amino acid sequence of a major wheat protein inhibitor of a-amylase. Phytochemistry 1981; 20:1781–4. Petrie A, Sabin C. Medical statistics at a glance, 3rd edn. Oxford: Blackwell Science, 2009. Sackett DL, Richardson WS, Rosenberg W, Haynes RB. Evidence based medicine: How to practice and teach EBM. London: Churchill Livingstone, 1997. Burks AW, Jones SM, Boyce JA et al. NIAID-sponsored 2010 guidelines for managing food allergy: applications in the pediatric population. Pediatrics 2011; 128:955–65. Ito K, Futamura M, Borres MP et al. IgE antibodies to x-5 gliadin associate with immediate symptoms on oral wheat challenge in Japanese children. Allergy 2008; 63:1536–42. Battais F, Pineau F, Popineau Y et al. Food allergy to wheat: identification of immunogloglin E and immunoglobulin G binding proteins with sequential extracts and purified proteins from wheat flour. Clin Exp Allergy 2003; 33:962–70. Pastorello EA, Farioli L, Conti A et al. Wheat IgE-mediated food allergy in European patients: a-amylase inhibitors, lipid transfer proteins and low-molecular-weight glutenins. Int Arch Allergy Immunol 2007; 144:10–22.
1430 M. J. M€akel€a et al 23 Battais F, Mothes T, Moneret-Vautrin DA et al. Identification of IgE-binding epitopes on gliadins for patients with food allergy to wheat. Allergy 2005; 60:815–21. 24 Sotkovsk y P, Sklenar J, Halada P, Setinova I, Kainarova A. A new approach to the isolation and characterization of wheat flour allergens. Clin Exp Allergy 2011; 41:1031–43.
25 Tatham AS, Shewry PR. Allergens to wheat and related cereals. Clin Exp Allergy 2008; 38:1712–26. 26 James JM, Sixbey JP, Helm RM, Bannon GA, Burks AW. Wheat a-amylase inhibitor: a second route of allergic sensitization. J Allergy Clin Immunol 1997; 99:239–44. 27 Weichel M, Vergoossen NJ, Bonomi S et al. Screening the allergenic reper-
toires of wheat and maize with sera from double-blind, placebo-controlled food challenge positive patients. Allergy 2006; 61:128–35. 28 Skylas DJ, Mackintosh JA, Cordwell SJ et al. Proteome approach to the characterization of protein composition in the developing and mature wheatgrain endosperm. J Cereal Sci 2000; 32:169–88.
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
doi: 10.1111/cea.12393
ORIGINAL ARTICLE
© 2014 John Wiley & Sons Ltd
Allergens
Wheat allergy in children – new tools for diagnostics M. J. M€akel€a1, C. Eriksson2*, A. Kotaniemi-Syrj€anen1*, K. Palosuo1, J. Marsh3, M. Borres2, M. Kuitunen1 and A. S. Pelkonen1 1
Department of Allergology, Skin and Allergy Hospital, Helsinki University Central Hospital, Helsinki, Finland, 2Thermofisher Scientific/Phadia AB, Uppsala,
Sweden and 3Institute of Inflammation and Repair, University of Manchester, Manchester, UK
Clinical & Experimental Allergy
Correspondence: Mika J. M€akel€a, Skin and Allergy Hospital, P.O. Box 160, FIN-00029 HUS, Helsinki, Finland. E-mail: [email protected] Cite this as: M. J. M€akel€a, C. Eriksson, A. Kotaniemi-Syrj€anen, K. Palosuo, J. Marsh, M. Borres, M. Kuitunen and A. S. Pelkonen. Clinical & Experimental Allergy, 2014 (44) 1420–1430.
Summary Background The detection of wheat-specific IgE in children often leads to a suspicion of wheat allergy, but little information is available on the most reliable wheat allergens for predicting clinical reactivity. Objective To evaluate the role of allergenic components of wheat in wheat allergy diagnostics. Methods One hundred and eight children (median age 1.5 years; range 0.6–17.3 years) with suspected wheat allergy underwent open or double-blinded, placebo-controlled oral wheat challenges. Responsiveness to different allergenic components of wheat was studied by skin prick tests and by determination of serum IgE antibodies using a semi-quantitative microarray assay. Results Thirty (28%) children reacted with immediate symptoms, and 27 (25%) with delayed symptoms to ingested wheat, whereas 51 (47%) children exhibited no reactions in oral wheat challenges. Positive IgE responses to any of the 12 allergenic components of wheat was seen in 93%, 41%, and 43% of those with immediate, delayed or no reactions to ingested wheat, respectively (P < 0.001 to P < 0.05 in every comparisons between those with immediate reactions and those with no reactions). Positive IgE responses to ≥5 different allergenic components improved significantly the diagnostic accuracy (with a positive likelihood ratio (LR+) of 5.10). Alpha-amylase inhibitors (AAI), in particular dimeric AAI 0.19 (LR+ 6.12), alpha-, beta-, and gamma-gliadins (LR+ from 3.57 to 4.53), and high-molecular-weight (HMW) glutenin subunits (LR+ 4.37) were the single allergenic components of wheat differentiating most effectively those with immediate symptoms from those who did not exhibit any reactions. Conclusions and Clinical Relevance Wheat allergy diagnostics is difficult, even using sophisticated component methods. Our results confirm earlier findings about gliadins and identify the dimeric AAI 0.19, as a relevant allergen in clinically reactive patients when compared to non-reactive subjects. The accuracy of wheat allergy diagnosis may be improved by measuring IgE responses to several components of wheat. Keywords alpha-amylase inhibitor, component diagnostics, gliadins, glutenins, wheat allergy Submitted 12 February 2013; revised 7 August 2014; accepted 14 August 2014
Introduction The prevalence of food allergies varies in different regions of the world possibly due to differences in the age of introduction of solid foods into diet, eating habits, as well as sensitization to pollens. Some foods, including cow’s milk, hen’s egg, wheat, soy, peanut, and fish, however, account for the majority of the doc*Both authors contributed equally to the study.
umented food hypersensitivity in most geographical regions [1]. Wheat allergy prevalence varies depending on the age and region from 0.4% to 1% [1, 2]. In Finland, wheat allergy is an important clinical entity and is often diagnosed in children with multiple food allergies. According to a recent survey, wheat allergy was diagnosed by a physician among 0.9% of children aged 1–4 years in south-eastern Finland [3]. Skin prick tests that are commonly used in allergy diagnostics are susceptible to technical errors and may
New tools for wheat allergy diagnostics
exhibit false positive reactions in patients with dermographism, and false negative results if performed during antihistamine medication. Measuring serum specific IgE levels is an alternative diagnostic option. As regards wheat allergy diagnostics, an important obstacle has been a lack of reliable allergens in measurement of the IgE response. Although specific IgE levels can be used to predict symptomatic allergies for many foods, wheatspecific IgE is highly non-specific, detects only a part of wheat protein fractions, and the level of specific IgE required to clinical response with a certainty of 95% is high, up to 80 kU/L [4]. To improve the diagnostic value of specific IgE measurements, immune responses to different wheat proteins need to be studied in different clinical phenotypes of allergy. Wheat proteins can be broadly divided into watersoluble albumins, salt-soluble globulins, and insoluble prolamins, including the gliadins which are soluble in aqueous alcohols, and the glutenins (Fig. 1) [2, 5, 6]. Each of these fractions contains allergenic proteins which have been associated with clinical symptoms. One of the best documented form of wheat allergy is the association of sensitization to omega-5-gliadin with wheat-dependent, exercise-induced anaphylaxis (WDEIA) [7]. Other relatively well-documented allergens include alpha-amylase inhibitors (AAI), the response to which is associated with food allergy in children with atopic dermatitis [2]. However, there is generally a significant overlap of the responses to individual proteins in different disease conditions, with AAIs and gliadins both being important allergens in the respiratory allergy to wheat flour, that is baker’s asthma. In this study, 12 known wheat allergen components were used to study wheat-specific IgE responses in a well-characterized group of children with suspected wheat allergy. Materials and methods Patients The study subjects included 108 children, aged from 0.6 to 17.3 years (median 1.5 years, 73% younger than 3 years) who had been referred to the Department of Allergology, Helsinki University Central Hospital, Helsinki, Finland, because of a suspicion of wheat allergy, and who underwent oral wheat challenge (81% open, 19% double-blind placebo-controlled (DBPC)) between October 2006 and February 2009. The coeliac disease was excluded on a clinical basis, but not routinely, before the wheat challenge. The details for the open oral wheat challenge have been described before [8]. Whether wheat challenges were performed as open or DBPC was based on earlier food-related symptoms: wheat challenges were per-
1421
Wheat protein solubility fractions
Albumin/ Globulin Water/dilute saline
Alpha-amylase inhibitors (AAI)
Gluten
Monomeric gliadins Alcohol/water mixtures
Lipid transfer protein (LTP)
Avenin-like protein
Polymeric glutenins Dilute acid
Alpha-gliadin Beta-gliadin
Gamma-gliadin
High molecular weight (HMW) glutenins
Low molecular weight (LMW) glutenins
Slow-omegagliadin Omega-5gliadin
Fig. 1. Wheat proteins grouped.
formed routinely as DBPC in IgE-negative children with delayed symptoms [9]. In both the DBPC and open challenges, the challenge began with the application of a drop of wheat porridge (10% wheat, cooked in water) onto the mucosa of the inner side of the lower lip; thereafter, wheat was given orally in increasing amounts (1 g, 5 g, and 10 g; containing wheat protein 0.1 g, 0.5 g, 1 g, respectively) at 1 h intervals. The porridge used in the DBPC wheat challenges was potato-based and was prepared on site by the hospital kitchen staff not involved in the food challenge process. The supervising physician examined the child for the appearance of any symptoms. The exact time, nature, and eliciting dose of each observed adverse reaction were recorded. It was predetermined that in case of doubtful reactions, the challenge was to be continued until there are obvious signs of positive reactions. The challenge was stopped when objective signs and symptoms indicated a positive reaction [9], and when needed, the symptoms were treated with antihistamine, and/or intramuscular epinephrine, and/or oral steroids. In case the child did not exhibit immediate symptoms,
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
1422 M. J. M€akel€a et al the challenge was continued at home, with at least 20 g of wheat per day for 1 week, and the parents were advised to keep a daily symptom diary. In case of a suspected delayed skin reaction, an extra follow-up visit was organized at the time of the reaction, and skin symptoms and findings were evaluated and recorded by a study physician. Otherwise, the daily symptom diary was evaluated by a study physician during a separate follow-up visit at the end of DBPC challenges. In DBPC challenges, the wheat-containing food product and placebo were given in a random order 2 weeks apart, including one wash-out week in between the challenges. Definitions A parental history of atopy was regarded as physiciandiagnosed atopic eczema or allergic rhinoconjunctivitis in either of the parents. A parental history of asthma was regarded as physician-diagnosed asthma in either of the parents. Atopic eczema, allergic rhinitis, and asthma were defined as a specific diagnosis made by a physician or paediatrician. A skin prick test (SPT) wheal diameter of ≥3 mm was considered as a positive reaction against the tested allergen, that is wheat (in-house powdered whole grain wheat diluted with 0.9% NaCl as 1 : 10 weight/volume), and gliadin (1 mg/mL, Sigma, St. Louis, MO, USA). Histamine dihydrochloride (10 mg/mL, ALK) was used as a positive control, and physiologic saline as a negative control. In the diagnostic food challenge, symptoms appearing within the first 2 h were defined as immediate, and symptoms appearing after the first 2 h (up to 5 days) were defined as delayed [8]. ImmunoCAP analysis Serum samples, which were obtained just before the beginning of the wheat challenge, were analyzed with ImmunoCAPâ (Phadia AB, Uppsala, Sweden) for allergen-specific IgE antibodies to wheat, gluten and gliadin, and for total IgE. Samples with a specific IgE antibody concentration above the measuring range (0.1–100 kUA/L) were diluted and reanalyzed. Values lower than 0.1 kUA/L were given the value 0.05 kUA/L. Values of 0.35 kUA/L or greater were regarded as positive. Samples with a total IgE concentration above the measuring range (2–5000 kU/L) were diluted and reanalyzed. Microarray immunoassay Specific IgE to different wheat and Timothy grass pollen allergen components as well as to cross-reactive
carbohydrate determinants (CCD) was measured with an in-house experimental semi-quantitative microarray assay based on a micro spot technique as described elsewhere [10, 11]. The resulting fluorescence intensity was measured with BioAnalyzerâ (LaVision, BioTec, Bielefeld, Germany) and evaluated with GenePixâ Pro v. 5.1 (Axon Instruments, Molecular Devices, Sunnyvale, CA, USA), and classified as negative, low positive, and high positive. Low positive and high positive fluorescence intensity were denoted as positive IgE responsiveness in final analyses. In addition, high positive fluorescence intensity was evaluated separately, denoting strongly positive IgE responsiveness. Individual cut-off levels for positive IgE responses for each allergen component were estimated based on the fluorescence intensity for pools of negative control samples from non-atopic subjects, and negative control samples spiked with 3000 kUA/L of myeloma IgE, and calculated as 2 9 (the average fluorescence intensity + 3 standard deviations). A fluorescence intensity > 5000 units was regarded as a strong positive IgE responsiveness. Allergens in microarray Most of the allergen components were essentially purified according to previously published methods. Highand low-molecular-weight (HMW and LMW) glutenins were enriched according to Verbruggen [12]. Alpha-, beta-, and gamma-gliadins were purified by preparative electrophoresis as described by Rumbo et al. [13] The electrophoresis was performed with a gel of 7% acrylamide concentration, and the running buffer and upper and lower reservoirs were acidic (pH 3.1) aluminium lactate buffer [14]. Slow omega-gliadins were purified according to Tatham and Shewry [15]. Monomeric, dimeric and tetrameric AAI were purified according to Kashlan and Richardson [16]. The purification of lipid transfer protein (LTP) was performed by extraction of wheat at pH 7.5, followed by gel filtration, hydrophobic interaction chromatography at pH 8 (containing ammonium sulphate) and ion exchange chromatography at pH 4.5. Avenin-like protein was enriched to a purity of about 75% with anion exchange chromatography at pH 8.8 followed by gel filtration. Recombinant omega-5gliadin, Phl p 4, Phl p 5b, Phl p 12, and CCD were obtained from Phadia AB. Statistical analysis To evaluate the statistical differences between the groups, v2-test or Fisher’s exact test (if the expected frequency for any cell was < 5) were used for dichotomous variables, and Mann–Whitney U-test was used to analyze continuous or ordinal variables. The kappa (j)
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
New tools for wheat allergy diagnostics
statistics was applied to compare agreement in IgE responsiveness between different wheat allergens, and the j-value of ≤0.20 was used as a criterion to test the possible combinations of wheat allergens [17]. Sensitivity, specificity, positive predictive values (PPV), negative predictive values (NPV), and likelihood ratios for positive (LR+) and negative (LR ) test results were calculated using routine equations. LR+ values of ≥5 and LR values of ≤0.20 are considered to have significant effects on pre-test probability [18]. The receiver operating characteristic (ROC) curves were applied to evaluate the optimal number of positive and strongly positive IgE responses against different wheat allergenic components to differentiate children with immediate reactions from those with no reactions in oral wheat challenge. In addition, the ROC curves were utilized to find out the allergenic components of wheat that were specifically associated with severe immediate reactions in oral wheat challenges. Two-tailed tests were used in all analyses. P-values less than 0.05 were considered statistically significant. The data were analyzed using IBM SPSS 19.0 (SPSS, Inc, Chicago, Illinois) for Windows. Ethics The study was approved by the Research Ethics Committee of Helsinki University Central Hospital (approval number: 374/E7/2004) and is registered in the Clinical Trial Registry of Hospital District of Helsinki and Uusimaa (HUS). The registration number is 214414. Before participation of the study, written informed consent was obtained from parents of all study children. Results Oral wheat challenge Baseline data for the study children are presented in Table 1. Eighty-five (79%) children had parental history of atopy or asthma, and 95 (88%) had clinical atopic manifestations or asthma. Skin symptoms were the most common symptoms related to food ingestion, in 100 (93%), followed by gastrointestinal and respiratory tract symptoms, in 57 (53%), and 38 (35%), respectively. During oral wheat challenge, 30 (28%) of the 108 study children reacted with immediate symptoms, 27 (25%) with delayed symptoms, and in 51 (47%) children, no wheat-ingestion related reaction was seen. Among the 30 children reacting with immediate symptoms, in 11 cases the maximum dose of ingested wheat was 1 g or less. Skin reactions were seen in 47 (44%), most commonly urticaria or erythema. Twenty-eight children (26%) reacted with gastrointestinal and 17 (16%) with respiratory symptoms. Altogether 22 chil-
1423
dren needed medication because of the immediate reaction during oral wheat challenge. Among these 22 children, there were 15 who received only antihistamine medication, and seven children who received antihistamine and intramuscular epinephrine, including two children who also received per oral corticosteroid medication. Those seven children who needed intramuscular epinephrine had either severe skin symptoms (i.e. generalized urticaria, generalized erythema) [9], or symptoms fulfilling diagnostic criteria for anaphylaxis [19]: severe skin symptoms in conjunction with respiratory symptoms (i.e. frequent dry cough, severe sneezing, and rhinorrhea) or in conjunction with gastrointestinal (i.e. vomiting, abdominal pain) symptoms, or respiratory symptoms in conjunction with gastrointestinal symptoms (Table 2). SPT and specific IgE measurements SPT results and IgE antibody responses are presented in Table 3. Patients with immediate symptoms differed significantly (P < 0.001 to P < 0.05) from those with delayed or no reaction to ingested wheat in every comparison, whereas there were no differences between the children with delayed or no reactions in the oral wheat challenge. Children with immediate reactions in the oral wheat challenge were nearly all positive in SPT (i.e. having a wheal diameter of ≥3 mm) with whole wheat as an allergen. However, approximately a half of the clinically non-reacting patients were also positive, rendering the specificity of wheat-SPT (with a wheal diameter of ≥3 mm) poor, 55%, and the LR+ low, 2.14. By increasing a SPT wheal diameter as a cut-off point, the specificity did improve, but in expense with reduced sensitivity. The same phenomenon was seen as regards gliadin-SPT, as well as specific IgE against wheat, gluten and gliadin, with only a little benefit of lifting specific IgE concentration up to 10- to 30-fold from the detection level (Table 3). Microarray-based specific IgE measurements A protein microarray assay was designed for analysis of specific IgE responses to allergenic components from wheat and Timothy grass. The protein microarray contained 12 wheat allergens, three pollen allergens from the Timothy grass and CCD. Allergens from Timothy grass were included in the array to identify Timothy grass pollen allergic patients that could have specific IgE antibodies to wheat allergen components due to IgE cross-reactions between homologous proteins in Timothy grass pollen and wheat grain. In this study, very few patients (< 10%) were sensitized to any of the three Timothy grass pollen proteins. The percentages of children with specific IgE to the different wheat compo-
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
1424 M. J. M€akel€a et al Table 1. Baseline characteristics of the study children. Reaction in oral wheat challenge
Baseline characteristics Age (yr) Male Parental history of atopy Parental history of asthma Atopic eczema Allergic rhinitis Asthma Skin prick test positive ever to cow’s milk, hen’s egg, and/or soya bean Parentally reported symptoms related to food ingestion ever Immediate-type, AW+S+GI / AW+S / AW+GI / S+GI Immediate-type, only S or GI Delayed-type, only S Delayed-type, predominately GI
All study children, n = 108
Immediate, n = 30
Delayed, n = 27
No reaction, n = 51
1.5 60 83 27 90 37 20 79
(0.6; 17.3) (56%) (77%) (25%) (83%) (34%) (19%) (73%)
1.8 18 23 6 27 11 7 28
(0.7; 17.3) (60%) (79%) (21%) (90%) (37%) (23%) (93%)a
1.5 19 22 8 21 9 4 16
(0.8; 11.4) (70%) (85%) (31%) (81%) (35%) (15%) (59%)
1.4 23 38 13 42 17 9 35
(0.6; 11.8) (45%) (79%) (27%) (82%) (33%) (18%) (69%)
53 18 21 16
(49%) (17%) (19%) (15%)
22 4 4 0
(73%)b (13%) (13%) (0%)
14 1 5 7
(52%) (4%) (19%) (26%)d
17 13 12 9
(33%) (26%)c (24%) (18%)e
P = 0.010 vs. No reaction. P = 0.001 vs. No reaction. c P = 0.027 vs. Delayed reaction. d P = 0.003 vs. Immediate reaction. e P = 0.023 vs. Immediate reaction. Immediate-type symptoms: AW: cough, rhinitis, shortness of breath, wheezing; S: urticaria, erythema; GI: oral pruritus, vomiting, abdominal pain. Delayed-type symptoms: S: exacerbation of eczema; GI: diarrhoea, abdominal pain. Analyses were performed using Mann–Whitney U-test or chi-squared test or Fisher’s exact test. Data are presented as median (range), or n (%). AW, airway; S, skin; GI, gastrointestinal. a
b
Table 2. Characteristics of patients with severe immediate reaction in oral wheat challenge.
Patient No.
Age
Sex
SPT+
1 2 3
0.65 3.03 6.08
M F M
Yes Yes Yes
4
1.10
F
Yes
5
3.22
F
Yes
6 7
0.87 0.84
F F
Yes Yes
Total amount of wheat (g) in OFC
Parentally reported food-related symptoms Immediate: Urticaria Immediate: Urticaria, vomiting, cough Immediate: Urticaria, vomiting, abdominal pain, cough Immediate: Urticaria, vomiting, abdominal pain Immediate: Urticaria, wheezing, shortness of breath Delayed: Exacerbation of eczema Immediate: Urticaria, wheezing
6 6 1 16 1 16 16
Reaction to wheat in OFC Generalized urticaria, erythema, vomiting Localized urticaria, cough, rhinitis Oral pruritus, rhinitis, abdominal pain, vomiting Generalized urticaria, cough Generalized urticaria, erythema Generalized erythema Generalized urticaria
M, male; F, female; SPT+, skin prick test positive ever to cow’s milk, hen’s egg or soya bean; OFC, oral food challenge.
nents are shown in Table 4. Specific IgE to all wheat allergen components differed significantly (P < 0.001 to P < 0.05) between children with immediate reactions during the oral wheat challenge and those with delayed or no reactions to wheat ingestion. By contrast, there were no differences between children with delayed or no reactions in the oral wheat challenge. The allergens differentiating most effectively children with immediate
reactions from those with no reactions were AAIs, alpha-, beta-, and gamma-gliadins, as well as HMWglutenin subunits (Table 4). In search of an optimal combination of sensitivity and specificity, AAI 0.19, in particular, emerged as an important component, with a LR+ of 6.12–6.80. We also analyzed specific IgE to wheat allergens in different combinations (based on poor agreement
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
New tools for wheat allergy diagnostics
1425
Table 3. Skin prick test results and immunoglobulin E responses with regard to the type of reaction in wheat challenge. Reaction in oral wheat challenge
Skin prick test results and IgE responses Skin prick test reaction to Wheat (mm) ≥3 mm ≥4 mm ≥5 mm ≥6 mm ≥7 mm Gliadin (mm) ≥3 mm ≥4 mm ≥5 mm Total IgE (kU/L) Specific IgE against Wheat (kUA/L) ≥0.35 kUA/L ≥3.50 kUA/L ≥10.50 kUA/L Gluten (kUA/L) ≥0.35 kUA/L ≥3.50 kUA/L ≥10.50 kUA/L Gliadin (kUA/L) ≥0.35 kUA/L ≥3.50 kUA/L ≥10.50 kUA/L
Immediate n = 30 7 29 29 27 19 15 4 24 23 12 211
LR+*
(0; 12)a,b 2.14 (97%)a,b (97%)a,b 2.24 (90%)a,b 3.06 (63%)b,c 5.38 (50%)b,c 8.50 (0; 9)a,b 2.55 (80%)a,b (77%)a,b 3.91 (40%)b,d 5.1 (12; 5027)c,e
LR *
PPV*
NPV*
0.06 0.06 0.14 0.42 0.53
56% 57% 64% 76% 83%
97% 97% 92% 80% 76%
0.29 0.29 0.65
60% 70% 75%
85% 85% 72%
18.23 (0.05; 983.40)a,b
Delayed n = 27 0 13 11 7 5 3 0 9 7 3 45
LR+†
(0; 10) (48%) 1.07 (41%) 0.94 (26%) 0.88 (19%) 1.57 (11%) 1.89 (0; 5) (33%) 1.06 (26%) 1.32 (11%) 1.42 (5; 7559)
PPV†
NPV†
0.94 1.04 1.05 0.92 0.94
36% 33% 32% 46% 50%
67% 64% 64% 67% 67%
0.97 0.92 0.96
36% 41% 43%
66% 67% 66%
LR
†
0.41 (0.05; 76.43)
29 24 18 15.26
(97%)a,b 2.35 (80%)a,b 3.40 (60%)a,b 4.37 (0.05; 814.83)a,b
0.06 0.26 0.46
58% 67% 72%
97% 87% 79%
14 6 4 0.15
(52%) 1.26 (22%) 0.94 (15%) 1.08 (0.05; 47.20)
0.82 1.02 0.99
40% 33% 36%
70% 65% 66%
28 23 17 14.25
(93%)a,b 2.16 (77%)a,b 3.01 (57%)a,b 4.82 (0.06; 868.97)a,b
0.12 0.31 0.49
56% 64% 74%
94% 84% 78%
12 6 3 0.20
(44%) 1.03 (22%) 0.87 (11%) 0.94 (0.07; 45.80)
0.98 1.04 1.01
35% 32% 33%
66% 64% 65%
0.13 0.31 0.49
53% 66% 74%
93% 85% 78%
1.09 1.02 1.01
32% 33% 33%
63% 65% 65%
28 (93%)a,b 23 (77%)a,b 17 (57%)a,b
1.90 3.26 4.82
12 (44%) 6 (22%) 3 (11%)
0.91 0.94 0.94
No reaction n = 51 0 23 22 15 6 3 0 16 10 4 79
(0; 10) (45%) (43%) (29%) (12%) (6%) (0; 5) (31%) (20%) (8%) (2; 5031)
0.25 (0.05; 1214.15) 21 (41%) 12 (24%) 7 (14%) 0.13 (0.05; 801.52) 22 (43%) 13 (26%) 6 (12%) 0.32 (0.05; 838.97) 25 (49%) 12 (24%) 6 (12%)
*Immediate vs. no reaction. † Delayed vs. no reaction. a P < 0.001 vs. delayed reaction group. b P < 0.001 vs. no reaction group. c P < 0.01 vs. delayed reaction group. d P < 0.05 vs. delayed reaction group. e P < 0.05 vs. no reaction group. Analyses were performed using Mann–Whitney U-test or chi-squared test. Data are presented as median (range), or n (%); percentages in brackets represent sensitivity for the immediate and delayed groups, and 1-specificity for the no reaction group. LR+, positive likelihood ratio; LR , negative likelihood ratio; PPV, positive predictive value; NPV, negative predictive value.
(j-value of ≤0.20) between the tested two components) in search of an optimal diagnostic tool. However, the LR+ values for different combinations were only modest (2.44–3.09), and inferior to individual allergenic components, making the combinations clinically useless. Significant differences were found in the number of wheat allergens with positive or strongly positive specific IgE responses between those with immediate reactions and those with delayed or no reactions in the oral wheat challenge (Table 4). ROC-curve revealed that positive IgE responses to ≥5 wheat allergens and
strongly positive IgE responses to ≥3 wheat allergenic components gave a moderate sensitivity (70% and 60%, respectively) with a good specificity (86% and 88%, respectively) resulting in both cases with a LR+ of 5.10 for immediate reactions in the oral wheat challenge. Severity of reactions The omega-5-gliadin-specific IgE antibody level has earlier been suggested to serve as a risk marker for reactivity in oral challenge [20]. We categorized the
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
Number of positive responses Any
(60%)a,b (33%)c,f (83%)a,b (67%)b,c
(53%)b,c (33%)c,d (60%)b,c (37%)a,b (70%)a,b (43%)b,c (63%)d,e (53%)a,b (50%)c,d (27%)
(33%)d,e (23%)f (60%)a,b (27%)d (37%)c,d (17%)d (33%)e,f (30%)d,e (47%)e,f (30%)f
7 (0; 12)a,b 28 (93%) a, b
18 10 25 20
Polymeric glutenins High-molecular-weight (HMW) glutenins Strongly positive Low-molecular-weight (LMW) glutenins Strongly positive
All
16 10 18 11 21 13 19 16 15 8
10 7 18 8 11 5 10 9 14 9
Immediate n = 30
Monomeric gliadins Alpha-gliadin Strongly positive Beta-gliadin Strongly positive Gamma-gliadin Strongly positive Slow omega-gliadin Strongly positive Omega-5-gliadin Strongly positive
Alpha-amylase inhibitors (AAI) AAI monomer 0.28 Strongly positive AAI dimer 0.19 Strongly positive AAI tetramer CM 2, 3, 16 Strongly positive Lipid transfer protein (LTP) Strongly positive Avenin-like protein Strongly positive Gluten group
Albumin/Globulin group
Positive IgE response against wheat components
2.16
4.37 2.83 2.50 3.09
4.53 8.50 4.37 6.23 3.57 5.53 2.15 3.89 2.83 2.72
5.67 3.97 6.12 6.80 4.68 NA 2.43 5.10 2.38 3.06
LR+*
0.12
0.46 0.76 0.25 0.43
0.53 0.69 0.46 0.67 0.37 0.61 0.52 0.54 0.61 0.81
0.71 0.81 0.44 0.76 0.69 0.83 0.77 0.74 0.66 0.78
LR *
Reaction in oral wheat challenge
56%
72% 63% 60% 65%
73% 83% 72% 79% 68% 77% 56% 70% 63% 62%
77% 70% 78% 80% 73% 100% 59% 75% 58% 64%
PPV*
94%
79% 69% 87% 80%
76% 71% 79% 72% 82% 73% 77% 76% 74% 68%
71% 68% 79% 69% 71% 67% 69% 70% 72% 69%
NPV*
(15%) (4%) (30%) (22%)
(11%) (0%) (19%) (0%) (19%) (4%) (30%) (7%) (11%) (7%)
(11%) (11%) (15%) (7%) (7%) (4%) (11%) (4%) (19%) (11%)
0 (0;11) 11 (41%)
4 1 8 6
3 0 5 0 5 1 8 2 3 2
3 3 4 2 2 1 3 1 5 3
Delayed n = 27
Table 4. Specific immunoglobulin E responses against wheat components with regard to the type of reaction in wheat challenge.
0.94
1.08 0.31 0.89 1.03
0.94 NA 1.35 NA 0.94 0.47 1.01 0.54 0.63 0.76
1.89 1.89 1.51 1.89 0.94 NA 0.81 0.63 0.94 1.13
LR+†
1.04
0.99 1.09 1.06 0.99
1.01 1.04 0.94 1.06 1.01 1.04 1.00 1.07 1.08 1.03
0.94 0.94 0.94 0.96 1.00 0.96 1.03 1.02 1.01 0.99
LR
†
33%
36% 14% 32% 35%
33% 0% 42% 0% 33% 20% 35% 22% 25% 29%
50% 50% 44% 50% 33% 100% 30% 25% 33% 38%
PPV†
64%
66% 63% 64% 66%
65% 65% 67% 64% 65% 64% 66% 64% 64% 65%
67% 67% 67% 66% 65% 66% 65% 65% 65% 66%
NPV†
(14%) (12%) (33%) (22%)
(12%) (4%) (14%) (6%) (20%) (8%) (29%) (14%) (18%) (10%)
(6%) (6%) (10%) (4%) (8%) (0%) (14%) (6%) (20%) (10%)
0 (0; 12) 22 (43%)
7 6 17 11
6 2 7 3 10 4 15 7 9 5
3 3 5 2 4 0 7 3 10 5
No reaction n = 51
1426 M. J. M€akel€a et al
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
(14%) (0; 9) (31%) (12%) 1.03 0.97 0.94 1.26 81% 79% 58% 75% 0.39 0.45 2.34 5.10
*Immediate vs. no reaction. † Delayed vs. no reaction. a P < 0.001 vs. delayed reaction group. b P < 0.001 vs. no reaction group. c P < 0.01 vs. delayed reaction group. d P < 0.01 vs. no reaction group. e P < 0.05 vs. delayed reaction group. f P < 0.05 vs. no reaction group. Analyses were performed using Mann–Whitney U-test or chi-squared test. Data are presented as median (range), or n (%); percentages in brackets represent sensitivity for the immediate and delayed groups, and 1-specificity for the No reaction group. LR+, positive likelihood ratio; LR , negative likelihood ratio; PPV, positive predictive value; NPV, negative predictive value; NA, not applicable.
33% 40%
65% 66%
7 0 16 6 66% 36% 0.99 1.08 0.35
≥5 Number of strongly positive responses Any ≥3
21 4 22 18
(70%)a,b (0;12)a,b (73%)b,c (60%)a,b
5.10
75%
83%
4 0 8 4
(15%) (0; 8) (30%) (15%)
LR LR * LR+* Immediate n = 30 Positive IgE response against wheat components
Table 4 (continued)
Reaction in oral wheat challenge
PPV*
NPV*
Delayed n = 27
LR+†
†
PPV†
NPV†
No reaction n = 51
New tools for wheat allergy diagnostics
1427
clinical reactions according to the severity to those who did not require epinephrine, and to those who needed epinephrine due to generalized symptoms and likely anaphylaxis (n = 7). Those who exhibited severe immediate reactions (i.e. required epinephrine) in the oral wheat challenge had the strongest IgE responsiveness against gamma-gliadin (sensitivity 86%, specificity 92%, LR+ 10.99, and LR 0.16 for strongly positive IgE responsiveness, with P < 0.001 with regard to those with no reactions to ingested wheat), and HMW-glutenin subunits (sensitivity 86%, specificity 86%, LR+ 6.26, and LR 0.17 for positive IgE responsiveness, with P < 0.001 with regard to those with no reactions to ingested wheat) (Fig. 2). Discussion The results of this study confirm the non-specific nature of whole wheat as a diagnostic marker for wheat allergy, and identify the dimeric AAI 0.19, as a relevant allergen in clinically reactive patients vs. those with specific IgE to wheat but no allergy. The accuracy of wheat allergy diagnosis could be improved by measuring IgE responses to several components of wheat. The measurement of wheat-specific IgE and its use for clinical diagnosis is problematic due to the low specificity when using whole wheat as a test allergen either in SPTs or in serum assays. Wheat-specific IgE is common among atopic children without true food allergies and this has not been taken into account in many studies. To avoid this pitfall, patient selection is of great importance. We used a real-life setting at the policlinics to which patients are referred either specifically for wheat challenge due to the detection of wheat-specific IgE and thus suspicion of wheat allergy, or because the patient is allergic to other foods and has symptoms suggestive of wheat allergy. This creates a pool of patients with negative challenge, and those with either acute or delayed symptoms, each of whom may be either IgE positive or negative for wheat proteins. We could, therefore, analyze a significant number of patients with different IgE and reactivity patterns reflecting the true conditions where the IgE tests are used. Our goal was to improve the IgE diagnostics by utilizing new, potentially more specific allergen components to identify patients likely to have acute and even severe reactions. Firstly, we characterized several glutenin components and different fractions of gliadin because these allergens have been shown to be important in WDEIA and in some cases of atopic dermatitis and anaphylaxis in adults and children [2]. Battais et al. [21] demonstrated that HMW-glutenin subunits were only minor allergens in 28 patients including 14 children with various symptoms of wheat allergy,
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
1428 M. J. M€akel€a et al 1.0
0.8
Sensitivity
0.6
0.4
Skin prick test reaction to wheat (mm). AUC 0.866 IgE to wheat (kU/L). AUC 0.874 IgE to high molecular weight glutenin (negative to strongly positive response). AUC 0.849 IgE to gamma-gliadin (negative to strongly positive response). AUC 0.881
0.2
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - Specificity
Fig, 2. Receiver operating characteristic curves for wheat allergenic proteins linked to immediate severe reactions in oral wheat challenge in comparison with ROC curves for skin prick test results and wheatspecific IgE.
whereas LMW-glutenin subunits were recognized by most patients with wheat allergy. By contrast, in our study, sera from more than half of the clinically reactive children recognized HMW-glutenin subunits more specifically than they did LMW-glutenin subunits. One explanation for this difference may be incomplete purification of the allergen in our study. However, onethird of the non-reactive children also showed IgE binding to LMW-glutenin subunits, making it practically useless as a clinical tool. IgE responsiveness to HMW-glutenin subunits, however, served as an important marker of severity of the oral wheat challenge reaction. Our results were in accordance with earlier studies of omega-5-gliadin [5, 20], which is clearly a more specific marker for wheat allergy diagnostics than whole wheat proteins. In the original report by Palosuo et al. [5], the specificity was up to 100% in wheat-sensitized children. In the present study, however, only 50% of the children with acute reactions, and 18% of those with no reactions had positive IgE responses to omega5-gliadin. Our results are clearly less specific and are closer to those of Ito et al. [20], obviously due to differences in patients selection. Alpha-gliadin was actually the best allergen in the gluten fraction in the microarray analysis, in terms of both sensitivity and specificity with regard to immediate reactions. Both alpha- and gamma-gliadins were also reported as important allergens by Pastorello et al. [22]. Moreover, in our study,
IgE responsiveness to gamma-gliadin in the microarray test served as an important marker for severity. It has been earlier shown that specific IgE from those suffering from anaphylaxis or urticaria detect sequential epitopes on gliadins whereas patients with atopic dermatitis probably recognize conformational epitopes [23]. Although this topic remains poorly understood, the development of improved purification procedures for native allergens is now providing more information about native gliadins and other wheat proteins as diagnostic tools and possibly even therapy [24]. AAIs have long been recognized as major allergens in baker’s asthma [25] but their role in patients reacting to ingested wheat was described first by James et al. [26]. As then, AAIs have been characterized in more detail in different disease phenotypes. AAIs are present both in the water-soluble and water-insoluble protein fractions and are not lost during cooking [25]. Different subunits of AAIs have been shown to bind IgE from both WDEIA patients as well as from patients with atopic dermatitis or anaphylaxis [22]. Based on earlier studies, the diagnostic value of different types has been difficult to evaluate. We wanted, therefore, to study the major monomeric, dimeric and tetrameric types of AAI in the microarray system. Children with immediate reactions in oral challenge most often recognized the dimeric AAI 0.19, and even children with delayed reactions recognized this protein slightly more often than those without a reaction. As dimeric AAI 0.19 gave the highest LR+ for immediate reaction in oral wheat challenge, superior to cut-offs for wheatSPT and wheat-specific IgE with similar levels of sensitivity (60–63%), measuring specific IgE to dimeric AAI 0.19 should be considered when aiming for the optimal diagnostic tool. Although IgE responsiveness to none of the tested wheat allergen combinations proved to be clinically useful, we found that the greater the number of wheat allergens with IgE responsiveness, the more likely were the immediate symptoms in oral wheat challenges. A recent study employed phage cDNA library to enrich sequences encoding IgE-binding proteins of wheat, showing reactions to proteins which had not previously been related to wheat allergy, such as thioredoxins and uncharacterized proteins [27]. Wheat grain contains a complex mixture of proteins and over 1000 individual components have been separated by two-dimensional electrophoresis [28]. Our wheat microarray concept could therefore be a feasible tool in further studies with novel proteins. One of the weaknesses of the study was that only a part of the challenges were performed as DBPC. In addition, as there are no international consensus statements on diagnosing delayed food-related symptoms, an inhouse protocol was applied. On the other hand, our
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
New tools for wheat allergy diagnostics
in-house protocol for diagnosing delayed wheat-related symptoms included DBPC challenges in conjunction with a daily symptom diary and a follow-up visit to minimize the number of false positive cases. Lack of children without a history of food-related symptoms as a control group can also be seen as a weakness of the study. In addition, semi-quantitative nature of data obtained from microarray assay did not allow all such analyses which were possible if the data were truly quantitative and continuous. In conclusion, wheat allergy diagnostics is difficult, even using sophisticated component methods. Our results confirm earlier findings about gliadins and identify the dimeric AAI 0.19, as a relevant allergen in clinically reactive patients when compared to non-reactive subjects. The accuracy of wheat allergy diagnosis may be improved by measuring IgE responses to several components of wheat.
References 1 Zuidmeer L, Goldhahn K, Rona RJ et al. The prevalence of plant food allergies: a systematic review. J Allergy Clin Immunol 2008; 121:1210–8. 2 Inomata N. Wheat allergy. Curr Opin Allergy Clin Immunol 2009; 9:238–43. 3 Pyrh€ onen K, N€ayh€a S, Kaila M, Hiltunen L, L€a€ar€a E. Occurrence of parentreported food hypersensitivities and food allergies among children aged 1– 4 yr. Pediatr Allergy Immunol 2009; 20:328–38. 4 Sampson HA. Utility of food-specific IgE concentrations in predicting symptomatic food allergy. J Allergy Clin Immunol 2001; 107:891–6. 5 Palosuo K, Varjonen E, Kekki O-M et al. Wheat x-5 gliadin is a major allergen in children with immediate allergy to ingested wheat. J Allergy Clin Immunol 2001; 108:634–8. 6 Shewry PR, D’Ovidio R, Lafiandra D, Jenkins JA, Mills ENC, Bekes F. Wheat grain proteins. In: Khan K, Shewry PR, eds. Wheat: chemistry and technology. St Paul (MN): American Association of Cereal Chemists, 2009:223–98. 7 Palosuo K, Alenius H, Varjonen E et al. A novel wheat gliadin as a cause of exercise-induced anaphylaxis. J Allergy Clin Immunol 1999; 103:912– 7. 8 Kotaniemi-Syrj€anen A, Palosuo K, Jartti T, Kuitunen M, Pelkonen AS, M€akel€a MJ. The prognosis of wheat
9
10
11
12
13
14
15
1429
Acknowledgement We are grateful to Mr Henry Akerstr€ om for help with the immunoassay analysis and to research nurse Tuija Rito for the work with children and their families. Funding Helsinki University Central Hospital Research Funds, Sigrid Juselius Foundation, Finnish Foundation for Allergy Research. Conflict of interest MSc Camilla Eriksson and MD Magnus Borres are employed by Thermofisher Scientific/Phadia. There are no other conflicts of interest to declare by anyone of the authors.
hypersensitivity in children. Pediatr Allergy Immunol 2010; 21:e421–8. Sampson HA, van Wijk RG, BindslevJensen C et al. Standardizing doubleblind, placebo-controlled oral food challenges: American Academy of Allergy, Asthma and Immunology – European Academy of Allergy and Clinical Immunology PRACTALL concensus report. J Allergy Clin Immunol 2012; 130:1260–74. Constantin C, Quirce S, Poorafshar M et al. Micro-arrayed wheat seed and grass pollen allergens for componentresolved diagnosis. Allergy 2009; 64:1030–7. € Asarnoj A, Moverare R, Ostblom E et al. IgE to peanut allergen components: relation to peanut symptoms and pollen sensitization in 8-year-olds. Allergy 2010; 65:1189–95. Verbruggen IM, Veraverbeke WS, Vandamme A, Delcour JA. Simultaneous isolation of wheat high molecular weight and low molecular weight glutenin subunits. J Cereal Sci 1998; 28:25–32. Rumbo M, Chirdo FG, Giorgieri SA, n MC. Preparative fracFossati CA, A~ no tionation of gliadins by electrophoresis at pH 3.1 (A-PAGE). J Agric Food Chem 1999; 47:3243–7. Clements RL. A continuous acetic acid system for polyacrylamide gel electrophoresis of gliadins and other prolamines. Electrophoresis 1988; 9:90–3. Tatham AS, Shewry PR. The conformation of wheat gluten proteins. The sec-
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
16
17
18
19
20
21
22
ondary structures and thermal stabilities of a-, b-, c- and x-gliadins. J Cereal Sci 1985; 3:103–13. Kashlan N, Richardson M. The complete amino acid sequence of a major wheat protein inhibitor of a-amylase. Phytochemistry 1981; 20:1781–4. Petrie A, Sabin C. Medical statistics at a glance, 3rd edn. Oxford: Blackwell Science, 2009. Sackett DL, Richardson WS, Rosenberg W, Haynes RB. Evidence based medicine: How to practice and teach EBM. London: Churchill Livingstone, 1997. Burks AW, Jones SM, Boyce JA et al. NIAID-sponsored 2010 guidelines for managing food allergy: applications in the pediatric population. Pediatrics 2011; 128:955–65. Ito K, Futamura M, Borres MP et al. IgE antibodies to x-5 gliadin associate with immediate symptoms on oral wheat challenge in Japanese children. Allergy 2008; 63:1536–42. Battais F, Pineau F, Popineau Y et al. Food allergy to wheat: identification of immunogloglin E and immunoglobulin G binding proteins with sequential extracts and purified proteins from wheat flour. Clin Exp Allergy 2003; 33:962–70. Pastorello EA, Farioli L, Conti A et al. Wheat IgE-mediated food allergy in European patients: a-amylase inhibitors, lipid transfer proteins and low-molecular-weight glutenins. Int Arch Allergy Immunol 2007; 144:10–22.
1430 M. J. M€akel€a et al 23 Battais F, Mothes T, Moneret-Vautrin DA et al. Identification of IgE-binding epitopes on gliadins for patients with food allergy to wheat. Allergy 2005; 60:815–21. 24 Sotkovsk y P, Sklenar J, Halada P, Setinova I, Kainarova A. A new approach to the isolation and characterization of wheat flour allergens. Clin Exp Allergy 2011; 41:1031–43.
25 Tatham AS, Shewry PR. Allergens to wheat and related cereals. Clin Exp Allergy 2008; 38:1712–26. 26 James JM, Sixbey JP, Helm RM, Bannon GA, Burks AW. Wheat a-amylase inhibitor: a second route of allergic sensitization. J Allergy Clin Immunol 1997; 99:239–44. 27 Weichel M, Vergoossen NJ, Bonomi S et al. Screening the allergenic reper-
toires of wheat and maize with sera from double-blind, placebo-controlled food challenge positive patients. Allergy 2006; 61:128–35. 28 Skylas DJ, Mackintosh JA, Cordwell SJ et al. Proteome approach to the characterization of protein composition in the developing and mature wheatgrain endosperm. J Cereal Sci 2000; 32:169–88.
© 2014 John Wiley & Sons Ltd, Clinical & Experimental Allergy, 44 : 1420–1430
