Pediatric Allergy and Immunology
ORIGINAL ARTICLE
Food allergy
Native and denatured egg white protein IgE tests discriminate hen’s egg allergic from egg-tolerant children Avigael H. Benhamou Senouf1, Magnus P. Borres2,3 & Philippe A. Eigenmann1 1
Department of Child and Adolescent, Pediatric Allergy Unit, University Hospitals, Geneva, Switzerland; 2ThermoFisher Scientific, Uppsala, Sweden; 3Department of Maternal and Child Health, Uppsala University, Uppsala, Sweden
To cite this article: Benhamou Senouf AH, Borres MP, Eigenmann PA. Native and denatured egg white protein IgE tests discriminate hen’s egg allergic from eggtolerant children. Pediatr Allergy Immunol 2015: 26: 12–17.
Keywords egg allergy; children; diagnosis; IgE testing; denatured proteins; ovomucoid; ovalbumin Correspondence Philippe A. Eigenmann, Pediatric Allergy Unit, University Hospitals of Geneva, 6, rue WillyDonze, CH-1211 Geneva 14, Switzerland Tel.: +41 22 372 4531 Fax: +41 22 372 4779 E-mail: [email protected] Accepted for publication 29 November 2014 DOI:10.1111/pai.12317
Abstract Background: Accurate diagnosis of egg allergy by IgE testing is challenged by a large number of atopic subjects sensitized, but clinically tolerant to eggs. In addition, discrimination between allergy to raw only, or raw and cooked egg allergy is important. In this study, we investigate the diagnostic performance of IgE tests to native and denatured egg proteins. Methods: According to food challenges and clinical tolerance, study subjects were randomized to the following groups: (Group A) sensitized but clinically tolerant to egg, (Group B) allergic to raw egg only, or (Group C) allergic to raw and cooked egg. Serum-specific IgE to native or reduced and oxidized egg white, ovomucoid, and ovalbumin were measured. Results: Increasing titers of specific IgE to the various proteins were found according to the degree of the egg allergy. Cut-off values for IgE testing to native egg could be determined to distinguish between raw egg allergic and egg-tolerant subjects (1.6 kU/ l), as well as raw and cooked egg allergic and egg-tolerant subjects (4.1 kU/l). ROC curves analysis showed that native ovalbumin was the best test for the diagnosis of allergy to raw and cooked egg, and native ovomucoid was best to distinguish between allergy to raw only, and allergy to raw and cooked egg. Sequential testing improved the diagnosis, when in addition to IgE to native egg white, IgE to native ovalbumin was tested for the diagnosis of raw and cooked egg allergy, and IgE to native ovomucoid for the discrimination between allergy to raw only, or to raw and cooked eggs. Conclusion: The diagnosis of egg allergy can be significantly improved using a panel of IgE tests to egg proteins in the native or denatured form. The accuracy can be improved using combined IgE testing.
Hen’s egg allergy is one of the most frequent food allergies in children (1), and egg avoidance is the primary option in these patients. Strict avoidance is difficult, as egg is commonly used worldwide in various types of dishes, as well as in industrially prepared foods. Egg might be eaten cooked, partially cooked or uncooked. However, a high proportion of egg-allergic patients can tolerate cooked eggs and need only to avoid raw or partially cooked eggs (1, 2). Children are commonly diagnosed based on the predictive positive value of specific IgE to egg white and often avoid all forms of egg (3, 4). Egg white sensitization per se is also know to be associated with an increased risk of developing further respiratory
12
allergies and asthma (5–7) and a high percentage of children suffering with atopic dermatitis are sensitized to egg white, while most of them are tolerant to eggs (8). Increasing evidences show that delayed introduction of foods in infants can led to an increased risk of atopic dermatitis (9), food allergies (10), celiac disease (11), and respiratory allergies (12, 13). Therefore, the necessity of eviction diets should be clearly assessed in any patient with a positive test. In addition, the patient will need follow-up testing for identifying natural tolerance acquisition. Currently, available IgE tests are only partially informative regarding egg allergy diagnosis and follow-up.
Pediatric Allergy and Immunology 26 (2015) 12–17 ª 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Benhamou Senouf et al.
In addition to the measurement of serum IgE to egg white, several studies suggest that measuring IgE to specific egg white proteins might be helpful for distinguishing between ‘cooked egg’ and ‘raw only’ egg-allergic patients, as well as between sensitized but tolerant, and allergic patients (2, 14, 15). Among the major hen’s egg allergens, ovalbumin is very sensitive to digestion and heating, whereas ovomucoid has proved to be more resistant to heat denaturation (16). Thus, proper discrimination of sensitization from true allergy, as well as determining the extent of the egg avoidance diet is paramount for an optimal diagnosis of egg allergy. To improve its accuracy, we investigated whether IgE measurement to native and denatured egg white, ovomucoid and ovalbumin can predict the clinical reactivity to eggs, and discriminate between raw and cooked, and only raw egg allergy.
Methods Patients We analyzed serum from patients proven by food challenge to be allergic to raw eggs only, or raw and cooked eggs; or tolerant (regularly eating raw and cooked eggs or proven tolerant by food challenge). Serum samples were correlated to food challenge results with a maximal time interval of 6 months between both procedures. Food challenges were conducted by an open standardized protocol according to the European Academy of Allergy and Clinical Immunology guidelines. A total of up to five increasing doses starting at 240 mg of egg (30 mg of egg protein) up to 24 g of egg (3000 mg of egg protein) were given (17). Egg was given either as pasteurized raw egg, or as a hardboiled egg. The tests were considered positive following published guidelines (18). According to their clinical reactivity to eggs, we divided the study subjects into three groups: (Group A) egg tolerant, these patients could eat eggs in all forms, raw and cooked; (Group B) only allergic to raw egg but tolerant to cooked egg, these patients did react to the raw egg challenge, but were able to regularly eat cooked eggs; (Group C) allergic to raw egg and cooked egg, these patients reacted to the cooked egg challenge. Cooked egg was defined as a hardboiled egg, or as egg containing products heated at 180°C for at least 30 min. Sensitized patients who had never been exposed to egg and who did not undergo an oral provocation test were excluded from the study. Ethics Committee approval was obtained prior to the study, and patients/their parents provided written consent. Materials Egg white native protein, native ovomucoid and albumin were all purchased from Sigma (Buchs, Switzerland). Reduced and oxidized (R&O) forms of proteins were prepared following the method developed by Cooke and Sampson (16). The protein pattern of the obtained preparations was controlled by Western blots performed as published
Denatured proteins to diagnose egg allergy
earlier (19). Specific IgE were measured by ImmunoCAPTM (prepared by ThermoFisher Scientific) with egg white native and R&O, ovomucoid native and R&O and ovalbumin native and R&O. Statistical analysis IgE binding to the egg white proteins used in the study was compared to egg white on ImmunoCAPTM (code f1) from ThermoFisher Scientific by linear regression analysis. IgE titers to the various protein preparations were compared between the groups by the Wilcoxon signed rank test for analysis of paired samples in non-normally distributed population. A p value 100) and to R&O egg white 5.84 kU/l (0.67->100). The individual titers to the various proteins are shown in Fig. 2, and median values and ranges are listed below the figure. Of note, R&O ovomucoid was negative in all egg-tolerant subjects but negative titers to this test were also found in patients only allergic to raw eggs, but not to cooked eggs.
Pediatric Allergy and Immunology 26 (2015) 12–17 ª 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
13
Denatured proteins to diagnose egg allergy
Benhamou Senouf et al.
patients vs. egg-tolerant subjects, a native egg test with a cutoff at 1.6 kU/l provided the best sensitivity (81%) and specificity (100%). For distinguishing raw and cooked eggallergic patients vs. egg-tolerant subjects, the native egg test with a cut-off at 4.1 kU/l provided the best sensitivity (83%) and specificity (100%). The same cut-off with native egg white was found with a good sensitivity (83%) and an average specificity (71%) for distinguishing between exclusively raw vs. raw and cooked egg allergy. In Table 2, we listed cut-off values for best sensitivity. There was not a single best test for discriminating between egg-tolerant and allergic patients, or between raw egg and raw and cooked egg allergy. Figure 1 Linear regression comparing values obtained in a sample by measuring native egg white (native egg white) on the x-axis, with the ImmunoCAPTM commercial extract of egg white (F1) on the y-axis. R = 0.97.
We were then interested in determining the performance of each specific test to egg components in order to discriminate between groups of patient by calculating ROC curves. When comparing Group A to Group C, all tests were highly reliable with the highest performance for native ovalbumin (area under the curve 0.985) and the lowest for R&O ovomucoid (0.925) (Fig. 3a). When comparing Group A to B, the performance of the tests was slightly lower, with R&O egg white being best (0.888) and native ovomucoid being the least powerful (0.737) (Fig. 3b). When comparing Group B to Group C, native ovomucoid was best (0.855), and R&O ovalbumin was the least powerful (0.720) (Fig. 3c). We then explored the ideal cut-off values for each test for distinguishing between various clinical situations (i.e., tolerant vs. allergic to raw and cooked egg, or raw egg only; or allergic to raw vs. raw and cooked eggs). According to the clinical situation, one might target either the highest sensitivity or the highest specificity of the tests. Table 1 lists the cut-off values for the best specificity, but by always keeping the sensitivity above 80%. In addition, when calculable, the likelihood ratio was added. For distinguishing exclusively raw egg-allergic
(a)
(b)
Discussion IgE sensitization to hen’s egg is very common in atopic individuals and may correspond to various situations going from clinical tolerance, up to allergy to all dishes containing eggs. In consequence, IgE tests allowing accurate discrimination are needed. We postulated that combining testing to various egg proteins in the native form, or with a modified allergenicity might improve the diagnosis. We showed improved diagnosis of egg allergy by selectively using a specific test according to the clinical question. In a seminal study including the diagnosis of egg allergy, Sampson and Ho have introduced the concept of threshold values to raise the accuracy for the in vitro diagnosis of egg allergy (3). Subsequent studies have confirmed that IgE titer cut-offs provide a valuable tool, nevertheless with the significant limitation of a population-based validity (20–22). Further studies have suggested an improvement of egg allergy diagnosis by combining tests to various native egg components. Ando et al. (2) have tested 108 children suspected of egg allergy by specific IgE to egg white, ovomucoid and ovalbumin. In vitro diagnosis was then correlated with the results of food challenges to raw and cooked egg. This study identified egg white and ovomucoid as the most reliable tests and provided cut-off values. Based on food challenges to raw and cooked eggs, Alessandri et al. (23) came to a similar conclusion as in
(c)
Figure 2 IgE test results to the egg white (a), ovomucoid (b), and ovalbumin (c) in the native or denatured form (R&O). The results are presented according to the clinical reactivity of the patients: Group A tolerant to all forms of egg, Group B only allergic to raw egg but tolerant to cooked egg and Group C allergic to raw egg and cooked raw. Median values and ranges for each condition are provided at the bottom of the figure.
14
Pediatric Allergy and Immunology 26 (2015) 12–17 ª 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Benhamou Senouf et al.
Denatured proteins to diagnose egg allergy
(a)
(b)
(c)
Figure 3 Receiver operating characteristic (ROC) curve illustrating the performance of IgE testing to the various egg proteins in native or denatured (R&O) forms. (a) compares Group A (tolerant to all forms of egg) to Group C (allergic to raw egg and cooked raw), (b) compares Group A to Group B (only allergic to raw egg but tolerant to cooked egg), and (c) compares Group B and Group C.
Table 1 Cut-off values for best specificity correlated with a sensitivity 80%, with likelihood ratio (LR) Raw vs. tolerant
Raw and cooked vs. tolerant
Raw vs. raw and cooked
Egg white native Egg white R&O Ovomucoid native Ovomucoid R&O Ovalbumin native
1.6 0.8 0.2 0.2 1.3
4.1 2.1 0.2 0.4 1.2
Ovalbumin R&O
0.4 kU/l (sens82%,spec58%); LR 2.0
4.1 kU/l (sens83%,spec100%) 1.2 kU/l (sens92%,spec92%); LR 11.1 0.2 kU/l (sens75%,spec35%); LR 1.1 0.4 kU/l (sens81%,spec100%) 1.2 kU/l (sens81%,spec83%); LR 4.8 (52,98) 0.5 kU/l (sens82%,spec58%); LR 2.0
kU/l kU/l kU/l kU/l kU/l
(sens81%,spec100%) (sens82%,spec58%); LR 1.9 (sens65%,spec64%); LR 1.8 (sens55%,spec100%) (sens83%,spec92%); LR 10.0
kU/l kU/l kU/l kU/l kU/l
(sens83%,spec71%); LR 2.9 (sens83%,spec54%); LR 1.8 (sens75%,spec64%); LR 2.1 (sens81%,spec69%); LR 2.6 (sens81%,spec17%); LR 1.0
0.5 kU/l (sens81%,spec22%); LR 1.0
Table 2 Cut-off values for best sensitivity correlated with a specificity 80%, with likelihood ratio (LR) Raw vs. tolerant Egg white native Egg white R&O Ovomucoid native Ovomucoid R&O Ovalbumin native Ovalbumin R&O
1.3 kU/l (sens81%,spec82%); LR 1.0 kU/l (sens73%,spec83%); LR 0.5 kU/l (sens58%,spec86%); LR 0.35 kU/l (sens35%,spec100%) 1.2 kU/l (sens83%,spec83%); LR 0.8 kU/l (sens65%,spec83%); LR
4.5 4.4 4.0 5.0 3.9
Raw and cooked vs. tolerant
Raw vs. raw and cooked
1.3 kU/l (sens95%,spec81%); LR 5.2 0.91 kU/l (sens95%,spec83%); LR 5.7 2.2 kU/l (sens46%,spec81%); LR 2.4 0.35 kU/l (sens86%,spec100%) 1.2 kU/l (sens80%,spec83%); LR 4.8 0.8 kU/l (sens71%,spec83%); LR 4.3
5.5 6.1 0.5 0.6 5.6 5.5
their study IgE to ovomucoid was positive in egg allergy. These results can be correlated to an earlier report identifying ovomucoid as the major egg allergen (24). However, a recent study by Kim et al. (25) showed variable pattern of component sensitization, according to the age of the patient. Nevertheless, all these reports used hen’s egg proteins in a native form, that is, not corresponding to the allergenicity of egg eaten in a cooked form. We postulated that by in vitro testing of modified egg proteins, corresponding to egg proteins after cooking, we might improve the diagnostic accuracy of serum-IgE tests. Indeed, heat-modified egg proteins have a reduced allergenicity as assessed not only in IgE tests, but also in-vivo and in cellular assays (26). We modified major egg white proteins according
kU/l kU/l kU/l kU/l kU/l kU/l
(sens70%,spec81%); (sens50%,spec82%); (sens69%,spec86%); (sens77%,spec81%); (sens42%,spec83%); (sens29%,spec83%);
LR LR LR LR LR LR
3.7 2.8 4.8 4.0 2.5 1.6
to a previously published methodology and used them as the allergen matrix to a widely used commercial test (ImmunoCAPTM from ThermoFisher Scientific). We found an excellent correlation between egg white measured with the commercial ImmunoCAPTM test and our egg white protein extract. Nevertheless, we emphasize that the two IgE titers obtained in a given patient are not comparable, as IgE value to our egg white extract tended to be overall higher than the commercial test. We also could demonstrate that IgE titers to the egg components clearly correlated with the degree of egg allergy in relation to tolerance or allergy to the various forms of egg. For example, subjects clinically tolerant to egg all had negative ovomucoid specific IgE, while the group of patients allergic only to raw eggs had a median titer of 0.67 kU/l (range:
Pediatric Allergy and Immunology 26 (2015) 12–17 ª 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
15
Denatured proteins to diagnose egg allergy
Benhamou Senouf et al.
ORIGINAL ARTICLE
Food allergy
Native and denatured egg white protein IgE tests discriminate hen’s egg allergic from egg-tolerant children Avigael H. Benhamou Senouf1, Magnus P. Borres2,3 & Philippe A. Eigenmann1 1
Department of Child and Adolescent, Pediatric Allergy Unit, University Hospitals, Geneva, Switzerland; 2ThermoFisher Scientific, Uppsala, Sweden; 3Department of Maternal and Child Health, Uppsala University, Uppsala, Sweden
To cite this article: Benhamou Senouf AH, Borres MP, Eigenmann PA. Native and denatured egg white protein IgE tests discriminate hen’s egg allergic from eggtolerant children. Pediatr Allergy Immunol 2015: 26: 12–17.
Keywords egg allergy; children; diagnosis; IgE testing; denatured proteins; ovomucoid; ovalbumin Correspondence Philippe A. Eigenmann, Pediatric Allergy Unit, University Hospitals of Geneva, 6, rue WillyDonze, CH-1211 Geneva 14, Switzerland Tel.: +41 22 372 4531 Fax: +41 22 372 4779 E-mail: [email protected] Accepted for publication 29 November 2014 DOI:10.1111/pai.12317
Abstract Background: Accurate diagnosis of egg allergy by IgE testing is challenged by a large number of atopic subjects sensitized, but clinically tolerant to eggs. In addition, discrimination between allergy to raw only, or raw and cooked egg allergy is important. In this study, we investigate the diagnostic performance of IgE tests to native and denatured egg proteins. Methods: According to food challenges and clinical tolerance, study subjects were randomized to the following groups: (Group A) sensitized but clinically tolerant to egg, (Group B) allergic to raw egg only, or (Group C) allergic to raw and cooked egg. Serum-specific IgE to native or reduced and oxidized egg white, ovomucoid, and ovalbumin were measured. Results: Increasing titers of specific IgE to the various proteins were found according to the degree of the egg allergy. Cut-off values for IgE testing to native egg could be determined to distinguish between raw egg allergic and egg-tolerant subjects (1.6 kU/ l), as well as raw and cooked egg allergic and egg-tolerant subjects (4.1 kU/l). ROC curves analysis showed that native ovalbumin was the best test for the diagnosis of allergy to raw and cooked egg, and native ovomucoid was best to distinguish between allergy to raw only, and allergy to raw and cooked egg. Sequential testing improved the diagnosis, when in addition to IgE to native egg white, IgE to native ovalbumin was tested for the diagnosis of raw and cooked egg allergy, and IgE to native ovomucoid for the discrimination between allergy to raw only, or to raw and cooked eggs. Conclusion: The diagnosis of egg allergy can be significantly improved using a panel of IgE tests to egg proteins in the native or denatured form. The accuracy can be improved using combined IgE testing.
Hen’s egg allergy is one of the most frequent food allergies in children (1), and egg avoidance is the primary option in these patients. Strict avoidance is difficult, as egg is commonly used worldwide in various types of dishes, as well as in industrially prepared foods. Egg might be eaten cooked, partially cooked or uncooked. However, a high proportion of egg-allergic patients can tolerate cooked eggs and need only to avoid raw or partially cooked eggs (1, 2). Children are commonly diagnosed based on the predictive positive value of specific IgE to egg white and often avoid all forms of egg (3, 4). Egg white sensitization per se is also know to be associated with an increased risk of developing further respiratory
12
allergies and asthma (5–7) and a high percentage of children suffering with atopic dermatitis are sensitized to egg white, while most of them are tolerant to eggs (8). Increasing evidences show that delayed introduction of foods in infants can led to an increased risk of atopic dermatitis (9), food allergies (10), celiac disease (11), and respiratory allergies (12, 13). Therefore, the necessity of eviction diets should be clearly assessed in any patient with a positive test. In addition, the patient will need follow-up testing for identifying natural tolerance acquisition. Currently, available IgE tests are only partially informative regarding egg allergy diagnosis and follow-up.
Pediatric Allergy and Immunology 26 (2015) 12–17 ª 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Benhamou Senouf et al.
In addition to the measurement of serum IgE to egg white, several studies suggest that measuring IgE to specific egg white proteins might be helpful for distinguishing between ‘cooked egg’ and ‘raw only’ egg-allergic patients, as well as between sensitized but tolerant, and allergic patients (2, 14, 15). Among the major hen’s egg allergens, ovalbumin is very sensitive to digestion and heating, whereas ovomucoid has proved to be more resistant to heat denaturation (16). Thus, proper discrimination of sensitization from true allergy, as well as determining the extent of the egg avoidance diet is paramount for an optimal diagnosis of egg allergy. To improve its accuracy, we investigated whether IgE measurement to native and denatured egg white, ovomucoid and ovalbumin can predict the clinical reactivity to eggs, and discriminate between raw and cooked, and only raw egg allergy.
Methods Patients We analyzed serum from patients proven by food challenge to be allergic to raw eggs only, or raw and cooked eggs; or tolerant (regularly eating raw and cooked eggs or proven tolerant by food challenge). Serum samples were correlated to food challenge results with a maximal time interval of 6 months between both procedures. Food challenges were conducted by an open standardized protocol according to the European Academy of Allergy and Clinical Immunology guidelines. A total of up to five increasing doses starting at 240 mg of egg (30 mg of egg protein) up to 24 g of egg (3000 mg of egg protein) were given (17). Egg was given either as pasteurized raw egg, or as a hardboiled egg. The tests were considered positive following published guidelines (18). According to their clinical reactivity to eggs, we divided the study subjects into three groups: (Group A) egg tolerant, these patients could eat eggs in all forms, raw and cooked; (Group B) only allergic to raw egg but tolerant to cooked egg, these patients did react to the raw egg challenge, but were able to regularly eat cooked eggs; (Group C) allergic to raw egg and cooked egg, these patients reacted to the cooked egg challenge. Cooked egg was defined as a hardboiled egg, or as egg containing products heated at 180°C for at least 30 min. Sensitized patients who had never been exposed to egg and who did not undergo an oral provocation test were excluded from the study. Ethics Committee approval was obtained prior to the study, and patients/their parents provided written consent. Materials Egg white native protein, native ovomucoid and albumin were all purchased from Sigma (Buchs, Switzerland). Reduced and oxidized (R&O) forms of proteins were prepared following the method developed by Cooke and Sampson (16). The protein pattern of the obtained preparations was controlled by Western blots performed as published
Denatured proteins to diagnose egg allergy
earlier (19). Specific IgE were measured by ImmunoCAPTM (prepared by ThermoFisher Scientific) with egg white native and R&O, ovomucoid native and R&O and ovalbumin native and R&O. Statistical analysis IgE binding to the egg white proteins used in the study was compared to egg white on ImmunoCAPTM (code f1) from ThermoFisher Scientific by linear regression analysis. IgE titers to the various protein preparations were compared between the groups by the Wilcoxon signed rank test for analysis of paired samples in non-normally distributed population. A p value 100) and to R&O egg white 5.84 kU/l (0.67->100). The individual titers to the various proteins are shown in Fig. 2, and median values and ranges are listed below the figure. Of note, R&O ovomucoid was negative in all egg-tolerant subjects but negative titers to this test were also found in patients only allergic to raw eggs, but not to cooked eggs.
Pediatric Allergy and Immunology 26 (2015) 12–17 ª 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
13
Denatured proteins to diagnose egg allergy
Benhamou Senouf et al.
patients vs. egg-tolerant subjects, a native egg test with a cutoff at 1.6 kU/l provided the best sensitivity (81%) and specificity (100%). For distinguishing raw and cooked eggallergic patients vs. egg-tolerant subjects, the native egg test with a cut-off at 4.1 kU/l provided the best sensitivity (83%) and specificity (100%). The same cut-off with native egg white was found with a good sensitivity (83%) and an average specificity (71%) for distinguishing between exclusively raw vs. raw and cooked egg allergy. In Table 2, we listed cut-off values for best sensitivity. There was not a single best test for discriminating between egg-tolerant and allergic patients, or between raw egg and raw and cooked egg allergy. Figure 1 Linear regression comparing values obtained in a sample by measuring native egg white (native egg white) on the x-axis, with the ImmunoCAPTM commercial extract of egg white (F1) on the y-axis. R = 0.97.
We were then interested in determining the performance of each specific test to egg components in order to discriminate between groups of patient by calculating ROC curves. When comparing Group A to Group C, all tests were highly reliable with the highest performance for native ovalbumin (area under the curve 0.985) and the lowest for R&O ovomucoid (0.925) (Fig. 3a). When comparing Group A to B, the performance of the tests was slightly lower, with R&O egg white being best (0.888) and native ovomucoid being the least powerful (0.737) (Fig. 3b). When comparing Group B to Group C, native ovomucoid was best (0.855), and R&O ovalbumin was the least powerful (0.720) (Fig. 3c). We then explored the ideal cut-off values for each test for distinguishing between various clinical situations (i.e., tolerant vs. allergic to raw and cooked egg, or raw egg only; or allergic to raw vs. raw and cooked eggs). According to the clinical situation, one might target either the highest sensitivity or the highest specificity of the tests. Table 1 lists the cut-off values for the best specificity, but by always keeping the sensitivity above 80%. In addition, when calculable, the likelihood ratio was added. For distinguishing exclusively raw egg-allergic
(a)
(b)
Discussion IgE sensitization to hen’s egg is very common in atopic individuals and may correspond to various situations going from clinical tolerance, up to allergy to all dishes containing eggs. In consequence, IgE tests allowing accurate discrimination are needed. We postulated that combining testing to various egg proteins in the native form, or with a modified allergenicity might improve the diagnosis. We showed improved diagnosis of egg allergy by selectively using a specific test according to the clinical question. In a seminal study including the diagnosis of egg allergy, Sampson and Ho have introduced the concept of threshold values to raise the accuracy for the in vitro diagnosis of egg allergy (3). Subsequent studies have confirmed that IgE titer cut-offs provide a valuable tool, nevertheless with the significant limitation of a population-based validity (20–22). Further studies have suggested an improvement of egg allergy diagnosis by combining tests to various native egg components. Ando et al. (2) have tested 108 children suspected of egg allergy by specific IgE to egg white, ovomucoid and ovalbumin. In vitro diagnosis was then correlated with the results of food challenges to raw and cooked egg. This study identified egg white and ovomucoid as the most reliable tests and provided cut-off values. Based on food challenges to raw and cooked eggs, Alessandri et al. (23) came to a similar conclusion as in
(c)
Figure 2 IgE test results to the egg white (a), ovomucoid (b), and ovalbumin (c) in the native or denatured form (R&O). The results are presented according to the clinical reactivity of the patients: Group A tolerant to all forms of egg, Group B only allergic to raw egg but tolerant to cooked egg and Group C allergic to raw egg and cooked raw. Median values and ranges for each condition are provided at the bottom of the figure.
14
Pediatric Allergy and Immunology 26 (2015) 12–17 ª 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Benhamou Senouf et al.
Denatured proteins to diagnose egg allergy
(a)
(b)
(c)
Figure 3 Receiver operating characteristic (ROC) curve illustrating the performance of IgE testing to the various egg proteins in native or denatured (R&O) forms. (a) compares Group A (tolerant to all forms of egg) to Group C (allergic to raw egg and cooked raw), (b) compares Group A to Group B (only allergic to raw egg but tolerant to cooked egg), and (c) compares Group B and Group C.
Table 1 Cut-off values for best specificity correlated with a sensitivity 80%, with likelihood ratio (LR) Raw vs. tolerant
Raw and cooked vs. tolerant
Raw vs. raw and cooked
Egg white native Egg white R&O Ovomucoid native Ovomucoid R&O Ovalbumin native
1.6 0.8 0.2 0.2 1.3
4.1 2.1 0.2 0.4 1.2
Ovalbumin R&O
0.4 kU/l (sens82%,spec58%); LR 2.0
4.1 kU/l (sens83%,spec100%) 1.2 kU/l (sens92%,spec92%); LR 11.1 0.2 kU/l (sens75%,spec35%); LR 1.1 0.4 kU/l (sens81%,spec100%) 1.2 kU/l (sens81%,spec83%); LR 4.8 (52,98) 0.5 kU/l (sens82%,spec58%); LR 2.0
kU/l kU/l kU/l kU/l kU/l
(sens81%,spec100%) (sens82%,spec58%); LR 1.9 (sens65%,spec64%); LR 1.8 (sens55%,spec100%) (sens83%,spec92%); LR 10.0
kU/l kU/l kU/l kU/l kU/l
(sens83%,spec71%); LR 2.9 (sens83%,spec54%); LR 1.8 (sens75%,spec64%); LR 2.1 (sens81%,spec69%); LR 2.6 (sens81%,spec17%); LR 1.0
0.5 kU/l (sens81%,spec22%); LR 1.0
Table 2 Cut-off values for best sensitivity correlated with a specificity 80%, with likelihood ratio (LR) Raw vs. tolerant Egg white native Egg white R&O Ovomucoid native Ovomucoid R&O Ovalbumin native Ovalbumin R&O
1.3 kU/l (sens81%,spec82%); LR 1.0 kU/l (sens73%,spec83%); LR 0.5 kU/l (sens58%,spec86%); LR 0.35 kU/l (sens35%,spec100%) 1.2 kU/l (sens83%,spec83%); LR 0.8 kU/l (sens65%,spec83%); LR
4.5 4.4 4.0 5.0 3.9
Raw and cooked vs. tolerant
Raw vs. raw and cooked
1.3 kU/l (sens95%,spec81%); LR 5.2 0.91 kU/l (sens95%,spec83%); LR 5.7 2.2 kU/l (sens46%,spec81%); LR 2.4 0.35 kU/l (sens86%,spec100%) 1.2 kU/l (sens80%,spec83%); LR 4.8 0.8 kU/l (sens71%,spec83%); LR 4.3
5.5 6.1 0.5 0.6 5.6 5.5
their study IgE to ovomucoid was positive in egg allergy. These results can be correlated to an earlier report identifying ovomucoid as the major egg allergen (24). However, a recent study by Kim et al. (25) showed variable pattern of component sensitization, according to the age of the patient. Nevertheless, all these reports used hen’s egg proteins in a native form, that is, not corresponding to the allergenicity of egg eaten in a cooked form. We postulated that by in vitro testing of modified egg proteins, corresponding to egg proteins after cooking, we might improve the diagnostic accuracy of serum-IgE tests. Indeed, heat-modified egg proteins have a reduced allergenicity as assessed not only in IgE tests, but also in-vivo and in cellular assays (26). We modified major egg white proteins according
kU/l kU/l kU/l kU/l kU/l kU/l
(sens70%,spec81%); (sens50%,spec82%); (sens69%,spec86%); (sens77%,spec81%); (sens42%,spec83%); (sens29%,spec83%);
LR LR LR LR LR LR
3.7 2.8 4.8 4.0 2.5 1.6
to a previously published methodology and used them as the allergen matrix to a widely used commercial test (ImmunoCAPTM from ThermoFisher Scientific). We found an excellent correlation between egg white measured with the commercial ImmunoCAPTM test and our egg white protein extract. Nevertheless, we emphasize that the two IgE titers obtained in a given patient are not comparable, as IgE value to our egg white extract tended to be overall higher than the commercial test. We also could demonstrate that IgE titers to the egg components clearly correlated with the degree of egg allergy in relation to tolerance or allergy to the various forms of egg. For example, subjects clinically tolerant to egg all had negative ovomucoid specific IgE, while the group of patients allergic only to raw eggs had a median titer of 0.67 kU/l (range:
Pediatric Allergy and Immunology 26 (2015) 12–17 ª 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
15
Denatured proteins to diagnose egg allergy
Benhamou Senouf et al.
