Prognosis of food-induced anaphylaxis in children: A single-center real-life study

Abstract

Background:

Food allergies are known to resolve over time, but there is little information on the natural history of food-induced anaphylaxis (FIA).

Objective:

This study aimed to evaluate the natural history of FIA in children and determine the factors that affect prognosis.

Methods:

Children with FIA who were followed up for at least 3 years, between 2010 and 2020, were included. Patients' families were contacted by telephone to question their child's tolerance status and invite them for reevaluation if uncertain. The patients were grouped as tolerant or persistent according to parent reports or reevaluation results. Logistic regression analysis was performed to determine the factors that affected persistence.

Results:

The study included 185 patients (62.2% boys) with 243 anaphylactic reactions to various foods. Fifty-eight patients (31%) gained tolerance within a 3-year follow-up period. Tolerance rates were higher in patients with FIA to milk (40%) and egg (43.9%) compared with to tree nuts (18.8%), legumes (5.6%), and/or seafood (11.1%) (p < 0.001). In a multivariate analysis, risk factors for persistent FIA were multiple food anaphylaxis (odds ratio [OR] 3.755 [95% confidence interval {CI}, 1.134–12.431]; p = 0.030), total IgE > 100 kU/L (OR 5.786 [95% CI, 2.065–16.207]; p = 0.001), and skin-prick test wheal size > 10 mm (OR 4.569 [95% CI, 1.395–14.964]; p = 0 .012) at presentation.

Conclusion:

Approximately a third of the patients with FIA developed tolerance within 3 years. Clinicians should remember that children with food allergies, even anaphylaxis, may develop tolerance over time. Regular follow up and reevaluation of tolerance status are necessary to avoid unnecessary elimination.

Food-induced anaphylaxis (FIA), the most severe form of food allergy (FA), is the most common cause of anaphylaxis-related deaths in children.¹ Therefore, FIA causes anxiety both for families and clinicians in terms of preventing recurrence.^2,3 As a result, children are subjected to unnecessary or prolonged restriction diets, which, in turn, leads to social, physical, and psychological problems.² Knowledge of the natural history of FIA and predictors of persistence will improve management. In previous studies, prognostic factors for FA persistence included large wheal size with the skin-prick test (SPT),^4,5 high food specific immunoglobulin E (sIgE) levels,^4,5 allergic comorbidities,⁶ and type of culprit food.⁷ Although there have been several studies of the natural course of FAs in children,^4–7 to our knowledge, there has been no study that specifically examined the natural course of FIA. In addition, previous natural course studies generally included reactions to one specific food type.^4–6 Several studies showed that severe symptoms may be associated with FA persistence.^6,8,9 Just as the prognosis differs for asthma and severe asthma,¹⁰ it is possible that FIA may have a different natural course and different risk factors associated with persistence. Our aim in the present study was to investigate the prognosis and natural course of FIA in children and to evaluate the risk factors of persistence of FA in children with anaphylaxis.

METHODS

Study Population and Design

In this single-center study, patients with FIA who presented to our pediatric allergy and immunology clinic between 2010 and 2020 were identified by screening the patient records according to ICD-10 codes (T78.2: anaphylaxis, T78.0: food-induced anaphylaxis, T78.1: food allergy. Patients with FA who first developed anaphylaxis while being followed up in the allergy clinic were included. The parents of these patients were contacted by telephone and questioned about the patient's experiences with the culprit food, including whether the patient had tried the culprit food at home or had any accidental exposure, whether the patient had a recurrence, the timing and severity of any reactions, and autoinjector use. Patients who had not yet consumed the food at home and had no new reactions in the past year were invited to the hospital for evaluation.

Those who had no new reaction in the past year but did not come to the clinic for reevaluation and/or refused testing and patients who were evaluated as nontolerant but had < 3 years of follow up were excluded from the study (Fig. 1). Patient data were recorded in a standard form, which included demographic, atopic, and clinical characteristics. The presence of allergic comorbidities (atopic dermatitis, asthma, allergic rhinitis) diagnosed by the study physicians, as well as the baseline sIgE value and SPT wheal diameter, eosinophil count and percentage, and tryptase and total IgE levels at presentation were noted. The study protocol was approved by the Ankara City Hospital Ethics Committee (protocol E1-20–1054). Informed consent was obtained from all patients' parents during their outpatient clinic visits or during the telephone interview.

Figure 1.

Selection of patients.

Allergological Workup

Sensitization was assessed in vivo by SPT and/or in vitro by sIgE levels. The sIgE levels were measured by using Phadia Immunocap (Thermo Fisher Scientific, Uppsala, Sweden) before 2015 and by using the Immulite 2000 system (Siemens Healthcare Diagnostics, Tarrytown, NY) after 2015. The sIgE levels were regarded as positive when ≥ 0.35 kU/L. SPT was performed at least 4 to 6 weeks after the anaphylactic reaction. Antihistamines were discontinued 1 week before testing. SPT was performed by using commercial preparations (ALK, Abello, Madrid, Spain) or with the fresh form of certain foods such as vegetables, fruit, and fish by using the prick-to-prick method as per the European Academy of Allergy and Clinical Immunology (EAACI) guidelines.¹¹ SPT was performed on the back in younger children and on the volar surface of the forearm in older children. After 15 minutes, indurations with a mean diameter ≥ 3 mm than the negative control were considered positive results.

In the analysis of patients with multiple FA, (i.e., those who experience anaphylaxis to more than one food), the largest SPT diameter and highest sIgE value were used if they reacted to multiple foods. Patients allergic to more than one tree nut counted as multiple FIA. Oral food challenge (OFC) was performed during reevaluation if the patients had not tolerated the culprit food yet and/or there was no new reaction in the past year. Children whose caregivers provided informed consent for open OFC with the culprit food were challenged in our allergy unit under the supervision of an allergist as recommended in the EAACI guidelines.¹² A stepwise OFC was preferred for milk (baked, then yogurt, then milk) and eggs (baked, then boiled). In the open OFC, freshly prepared culprit food was given orally with stepwise increasing doses every 15 minutes. OFCs were considered positive if objective symptoms occurred within 2 hours of the last challenge dose. Patients with negative results were observed for at least 2 hours after OFC, then were told to continue receiving the suspected food at home and present to the hospital in case of any reaction.

Definition of FIA

FIA was defined according to the criteria of the EAACI task force position papers¹² on the management of anaphylaxis in childhood.¹² Patients who met one of the following two criteria were accepted as anaphylaxis: (1) Involvement of two or more systems (dermatologic, respiratory, cardiovascular, gastrointestinal, and/or neurologic) with rapid onset (within minutes to several hours) after exposure to a suspected food allergen, or (2) hypotension that occurred minutes to several hours after exposure to a known food allergen. Anaphylaxis severity was determined according to the EAACI position paper¹² based on the patient's medical records and parent reports. Moderate and severe reactions were combined and compared with mild reactions.

Definition of Tolerance

Tolerance was defined as the absence of a reaction after consuming the culprit food. All the patients were instructed not to consume the culprit food. However, if the patient did ingest the food at home, then the first reported instance of reaction-free intake was accepted as the date of resolution. If the patient had a previous negative OFC result, then the date of OFC was accepted as the date of resolution. Patients were classified as nontolerant if their parent(s) reported another reaction to the culprit food within the past year or if their OFC failed during reevaluation. Those with SPT and/or sIgE test results above the thresholds were also regarded as nontolerant without OFC. The patients with multiple FAs were considered nontolerant if they were still unable to tolerate one or more foods and had persistent allergic sensitivity. The patients who had ingested only the baked form of the food (as in the ingredients of any baked food) or trace amounts of the food (defined as the amount transferred through shared packaging, use of the same production and/or preparation equipment, or by air contact) were reassessed for tolerance of other forms. Those who did not tolerate other forms successfully were considered nontolerant. The patients with milk allergy who had symptom-free intake of other uncooked forms of milk such as cheese, yogurt, and ice cream were regarded as tolerant.⁴

Statistical Analysis

We used SPSS version 22.0 (IBM Corp, Armonk, NY). Categorical values were not normally distributed; therefore, the data were presented as median and interquartile range (IQR). For comparisons between the tolerant and nontolerant groups, we used the χ² and Fisher exact tests for qualitative variables. The Mann-Whitney U and Wilcoxon rank sum tests were used for quantitative variables. Prognostic factors were analyzed with univariate and multivariate analysis. Multivariate analysis included variables with p < 0.2 in the univariate analysis and factors thought to be associated with persistence. The Cox proportional hazard regression models were used to predict persistence, and the Kaplan-Meier method was used to demonstrate the natural course of FIA with cumulative resolution curves. The p values < 0.05 were considered significant.

RESULTS

There were 185 patients with a total of 243 reactions who met the inclusion criteria and were included in the analysis (Fig. 1). The median (IQR) age of the patients was 7 years (5–13 years), and 62.2% were boys (n = 115). The median (IQR) age at onset (first anaphylactic reaction) was 12 months (6–33 months). Culprit foods were milk (n = 74), egg (n = 41), tree nuts (n = 54), legumes (n = 30), seafood (n = 9), and others (n = 35) (sesame, sunflower, poppy, fruits and/or vegetables, coconut, and honey). Among the patients with tree nut allergy, eight reacted to two different nuts and one reacted to three different tree nuts. All the patients with tree nut allergy reacted to hazelnut. The baseline characteristics of the patients are summarized in Table 1. FIA completely resolved in a total of 58 patients (31%) within 3 years.

Table 1.

Demographic characteristics of the patients (N = 185)

IQR = Interquartile range; IgE = immunoglobulin E.

Of 140 patients with a single FIA, 55 (39.2%) gained tolerance to the culprit food. Of the 45 patients with multiple FIA, 3 (6%) gained tolerance to all culprit foods, 15 patients gained tolerance to at least one food, and 27 patients did not develop tolerance to any food. The comparison of patients with single and multiple FIA is presented in Supplemental Table 1. When analyzed by food type, the rate of tolerance was 40% for milk (30/74), 43.9% for egg (18/41), 16.6% for nuts (9/54), 6% for legumes (2/30) (16.6% for peanuts [1/6] in the legume groups), 11.1% for seafood (1/9), and 11.4% for other foods (4/35) (Fig. 2). Resolution rates were higher in milk and egg allergies than in tree nut, legume, and seafood allergies (p < 0.001). The median (IQR) age at resolution was 24 months (14–30 months) (Fig. 3). Children in the nontolerant group were older (median age, 108 versus 60 months; p < 0.001). The age at onset was lower in the tolerant group (median [IQR], 10 months [6–24 months]) (median [IQR], 12 months [6.5–36 months]) but the difference was not significant (p = 0.22). Infant anaphylaxis was more common in the tolerant group (p = 0.030). Symptoms and reaction severity did not differ significantly between the groups (p > 0.05), but moderate and severe reactions were less frequent among the children with reactions to milk and egg (88/115 [76%]) (112/128 [87.5%]) (p = 0.025).

Figure 2.

Tolerance rates at 3 years according to food type.

Figure 3.

The time to resolution of food-induced anaphylaxis (FIA) with follow-up time (months).

Multiple food induced anaphylaxis was reported in 45 children (24%) and was more common in the nontolerant group (42/127 [33%]) than in the tolerant group (3/58 [5%]) (p < 0.01). The children who were nontolerant had larger SPT induration, higher sIgE levels, and higher total IgE levels at presentation (p < 0.01). Recurrent anaphylaxis (before or after diagnosis) was reported in 50% of the patients overall and was more prevalent in the nontolerant group (p < 0.01). The comparison of the tolerant and nontolerant groups is shown in Table 2. Clinical characteristics of the patients in the tolerant and nontolerant groups according to culprit food are presented in Supplemental Table 2. In multivariate logistic regression analysis, anaphylaxis to multiple foods, total IgE > 100 kU/L, and SPT wheal diameter > 10 mm were prognostic factors of persistent FIA. The presence of asthma was also included in the analysis, but it was not found to be a significant prognostic factor (Table 3).

Table 2.

Comparison of patients who were tolerant and nontolerant

IQR = Interquartile range; SPT = skin-prick test; sIgE = specific immunoglobulin E.

Data in bold represent statistically significant difference. n is number of patients.

Table 3.

Predictive factors for persistence

OR = Odds ratio; CI = confidence interval; IgE = immunoglobulin E; SPT = skin-prick test.

Data in bold represent statistically significant difference. n is number of patients.

DISCUSSION

In this study, we reported the natural course of FIA over a 3-year follow-up in children. Tolerance to the culprit food was observed in 31%. The resolution time varied according to the culprit food and was higher in milk and egg allergies than in tree nut, legume, and seafood allergies. Higher total IgE levels, anaphylaxis to multiple foods, and larger SPT wheal sizes were identified as risk factors for persistent FIA. In a study on tolerance in self-reported FA, Gupta et al.¹³ reported that 26.6% of patients with FA (n = 1245) had the FA resolved at a mean age of 5.4 years, and FA resolution was less frequent in children who had severe reactions (25.9% versus 44.1%). Bock¹⁴ reported that only three of the nine children acquired tolerance within 3 to 9 years. In our study, including only severe reactions, the resolution rate was 31% at 3 years. The age at onset is reported to affect the prognosis of FA being related to severe reactions in anaphylaxis.^6,15,16 Elizur et al.⁶ demonstrated that resolution of cow's milk allergy was significantly delayed in infants whose first reaction occurred before 1 month of age. In contrast, Topal et al.¹⁶ showed that presentation after 6 months of age was associated with persistence of cow's milk allergy. Gupta et al.¹³ reported that the probability of tolerance was significantly higher in children whose first reactions occurred early.

Similarly, we showed that resolution was less common among children whose first reactions were after 12 months of age; as a result, infant anaphylaxis was more frequent in the tolerant group of our study. Results of studies showed that milk and egg allergies resolve more frequently and earlier than allergies to peanuts and nuts.^7,9No information was provided about the prognosis of patients who had food anaphylaxis. We observed tolerance in ∼40% of patients with milk, 43.9% with egg, 16.9% with nut, 6% with legume, and 11% with seafood anaphylaxis. FIA resolution rates vary according to food types; although the reason has not been clarified.¹⁷ Results of some studies suggested that resolution rates are lower in patients with more severe initial reactions.^6,8,9 In our study, in accordance with previous study,¹⁸ anaphylaxis due to milk and egg allergies was milder than with other foods, and this may be one of the reasons for earlier tolerance. However, detailed studies that illuminate the epitopes and the immune response to allergens at the molecular level are required to clarify this subject. Allergen sIgE and SPT wheal diameters have also been shown to predict the prognosis of FA.^4,5,19 Accordingly, the results of our multivariate analyses indicated that wheal size > 10 mm in SPT at presentation may predict persistence. In addition, contrary to some studies,^16,20 we found that total IgE levels > 100 kU/L at presentation may predict persistence.

Allergic comorbidities have also been reported to affect the prognosis of FA.^6,16,21 As reported by Wood et al.,⁴ asthma was more common in our persistent group (64.5% versus 48%). However, allergic diseases other than asthma were not associated with tolerance in our study. Frequency of mite sensitivity has been reported to be higher in patients with persistent FA.^22,23 We found that sensitivity to pollen, another aeroallergen, was higher in the nontolerant group. In accordance with previous studies on FA,^13,15,16,22 multiple FIA were more frequent in the persistent group in our study (33% versus 5%). In our study, the total IgE level (p = 0.042) and highest SPT wheal size (p = 0.006) were greater in those with multiple FAs, which suggested a higher level of allergen sensitivity in those with multiple FAs, which contributes to the longer persistence. One of the limitations of this study was the small sample size and coming from a single center. A second limitation was that the patients could not be checked regularly at 6-month intervals, so it was not possible to determine exactly when they developed tolerance. The age at which the patient tolerated the culprit food at home or passed OFC was accepted as the date of tolerance. An additional limitation was the possibility of missing patients who did not meet inclusion criteria when first evaluated but subsequently experienced anaphylaxis.

CONCLUSION

The allergies of approximately one-third of the patients with FIA in this study resolved at the end of a 3-year follow-up. An SPT wheal size > 10 mm, total IgE level > 100 kU/L, and multiple food anaphylaxis were identified as risk factors that predict persistence. When considering that FAs may resolve over time, even in patients with anaphylaxis, evaluating patients for tolerance at regular intervals (at least 1 year) is important to prevent unnecessary food restrictions.