Incidence of anaphylaxis and accidental peanut exposure: A systematic review

Abstract Background Peanut allergy (PA), a common food allergy, is increasing in prevalence and is associated with high rates of anaphylaxis. Prevalence of food‐related anaphylaxis is higher in children and adolescents than in adults, and the pediatric incidence is increasing. We conducted a systematic literature review and meta‐analysis to determine the incidence of peanut‐induced anaphylaxis in children and/or adolescents with PA. Methods Literature searches were conducted using the PubMed database and through supplemental methods. Eligible articles for inclusion were peer‐reviewed studies published in English that reported the incidence of anaphylaxis in pediatric PA using the 2006 National Institute of Allergy and Infectious Disease/Food Allergy and Anaphylaxis Network criteria, sample size, and follow‐up duration. Incidence rates were calculated as person‐years at risk or a crude incidence rate was calculated. Meta‐analyses of pooled data were conducted using the I 2 statistic as the measure of heterogeneity. Results A total of 830 citations were screened; 8 met the study inclusion criteria and were selected for review. Pooled meta‐analysis estimates of the incidence of (1) anaphylaxis among children/adolescents with food allergies, (2) anaphylaxis among children/adolescents with PA, and (3) accidental exposure to peanuts among children/adolescents with PA were 3.72 cases per 100 person‐years (95% confidence interval [CI] = 2.35, 5.10), 2.74 cases per 100 person‐years (95% CI = 1.42, 4.05), and 12.28 cases per 100 person‐years (95% CI = 11.51, 13.05), respectively. Conclusions The risks of anaphylaxis among children with food allergies and those with PA contribute to the serious overall burden of PA and food allergy for children and their families.


| INTRODUCTION
Peanut allergy (PA) is among the most common food allergies, and is often a lifelong condition associated with risks of severe and potentially fatal allergic reactions, including anaphylaxis, due to accidental peanut exposures. [1][2][3][4] The general population prevalence of PA in Western countries is estimated to be 1%-2%, 1,[5][6][7][8][9] and some data indicate that it has increased over the past 2 decades. [9][10][11][12] The prevalence of PA is generally lower in other parts of the world, 9 although substantial knowledge gaps in this area exist for developing and emerging countries. 13 Standard management of PA, and other food allergies, consists of strict dietary avoidance of the triggering food and use of intramuscular epinephrine (adrenaline) in cases of allergic reactions due to accidental exposure. 14,15 Health-related quality of life is significantly impaired in people with PA and other food allergies and their caregivers due to the constant vigilance required to avoid accidental exposures to trigger foods and associated stress, anxiety, and dietary and lifestyle restrictions. [16][17][18][19][20] Among the food allergies, PA has been associated in multiple studies with the highest rates of severe reactions, anaphylaxis, and fatal anaphylaxis in Western nations. [21][22][23][24][25][26][27][28] The largest longitudinal study to date of the rate of accidental peanut exposures in children with PA (n = 1941), conducted in Canada, reported an annual incidence of 12.4%, and found that two-thirds of the exposures caused moderate or severe reactions. 29 Although fatal food-related anaphylaxis is rare, with an estimated incidence rate of 1.8 per million person-years overall, and 3.25 per million person-years in children, the ongoing and highly unpredictable risk of severe reactions due to accidental exposure contributes greatly to the burden of PA and other food allergies. 19,[30][31][32] In addition, studies in Western nations indicate that rates of foodrelated anaphylaxis, including fatal events and hospitalizations, are higher in children and adolescents than in older age groups, [32][33][34][35] and have been increasing markedly in pediatric age groups. [33][34][35][36][37][38][39][40][41] Determining the rate of anaphylaxis in the high-risk population of children and adolescents with PA is important to assess the burden of this risk, and to provide essential data for future analyses of PAassociated needs for healthcare and auxiliary services and support. While many studies have assessed rates of anaphylaxis in the general population, or in patients with food allergies, specific data for the PA pediatric population are limited. Obtaining such data is particularly important to better understand the effect this information has on clinical decision-making.
When food allergy is diagnosed, the first-line treatment has traditionally been avoidance and prescription of rescue medications in the event of accidental exposures. In January 2020, the first treatment for PA was approved by the US Food and Drug Administration, although other immunotherapies using food for PA have been used in phase 2 trials. [42][43][44][45] In addition to the approved product, oral immunotherapy with foods (not approved by any regulatory authorities) for various food allergies is already being offered and the benefit-risk ratio of this treatment is an important subject for analysis. 46 In addition, an epicutaneous immunotherapy for PA is in development, 47 and a range of other potential therapeutic approaches to PA, such as probiotics, biologics, and DNA vaccines, are under investigation. 48 The objective of this systematic literature review and analysis is to assess the incidence of peanut-induced anaphylaxis in children and adolescents with PA based on peer-reviewed published data.

| Study eligibility criteria and search strategy
Incidence rate was selected to assess the frequency of anaphylaxis in the pediatric population, as it estimates the rate of new events occurring on a population level during a defined period of time.
Anaphylaxis is a severe, potentially life-threatening systemic hypersensitivity reaction that may or may not recur. 49 Therefore, we considered anaphylaxis incidence rates to be more informative for practical considerations, such as contribution to PA burden and healthcare needs, than other commonly used rates such as prevalence or cumulative incidence.
The reporting of our search methods, analysis, and results follow the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) criteria. 50 Literature search criteria were developed prior to the search (see Appendix Table S1) and the search was not conducted iteratively. One comprehensive screen was applied within the United States National Library of Medicine/National Institutes of Health PubMed database to identify peer-reviewed, pub-

| Definition of anaphylaxis for this meta-analysis
Sampson's criteria for anaphylaxis (see Appendix Table S2) was used to define anaphylaxis in each of the studies. In instances where Sampson's criteria were not explicitly used for the definition of anaphylaxis, studies that used one of three definitions of anaphylaxis were included: (1) an explicit description of cases as anaphylaxis; (2) reported allergic reactions by severity (mild, moderate, severe); those categorized as "severe allergic reactions" were considered to be anaphylaxis; (3) reported "anaphylactic reaction."

| Data collection/extraction
Studies were included if incidence rate data specific to the type of allergy (e.g., peanut allergy, all food allergy) were reported. If incidence rates were not directly reported, studies which provided the necessary data for the calculation of incidence rates were included.
Incidence rates were defined as the number of incident cases of anaphylaxis divided by the person-years at risk. In the studies reviewed, incidence rates were typically provided per 1000 personyears. In instances where a study did not explicitly report the incidence rate, but reported the number of incident cases of anaphylaxis, the average length of follow-up, and the sample size, a crude incidence rate was calculated as follows:

anaphylaxis cases ðaverage length of follow-upÞ � sample size
If the article did not provide average length of follow-up but provided median length of follow-up time, a crude incidence rate was calculated as follows:

anaphylaxis cases ðmedian length of follow-upÞ � sample size
A minority of the included studies provided 95% confidence intervals (CIs) with their point estimates. The following formulas 52 were used to calculate crude 95% CI for the studies that did not report 95% CI: ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi

| Meta-analysis
Pooled incidence rates were calculated among three populations of interest (children/adolescents with PA, with food allergy, and with PA who experienced accidental exposure). The pooling was performed using a random effects model. 53 Heterogeneity is presented as the I 2 statistic, which describes the percentage of variation across studies that is due to heterogeneity rather than chance. 54 As observed with the chi-squared statistic, the p-value associated with the I 2 statistic registers as statistically significant if any one study differs from another.

| Publication quality and bias assessment
We used the Q-Coh scale to assess the quality of individual studies and their risk of bias. 55 This scale encompasses 26 individual items organized in eight major domains: study design, representativeness, comparability of groups, maintenance of comparability, exposure measure, outcome measure, attrition, and statistical analyses. Studies were deemed to be of "good" quality if at least 6 domain items were scored positively and of "acceptable" quality if 4-5 domain items were scored positively. Studies that were considered of less than "good" or "acceptable" quality were excluded from this review.

| Literature search
As shown in Figure 1, which provides the results of the search strategy described above, the initial database search of PubMed yielded 830 total records. Since the search strategy was intended to be broad and included multiple common search terms, the majority of the records identified were not eligible for inclusion. Screening of titles and abstracts led to exclusion of 822 for failure to meet eligibility criteria ( Figure 1). As a result, eight articles were included in the systematic review and meta-analysis of pooled data. All of these were of "good quality" according to the Q-Coh scale.
As described in Section 2, three search outputs were collected.
Tables 1-3 summarize the anaphylaxis incidence data captured from one or more studies identified with each output.
3.1.1 | Search Output 1. Incidence of anaphylaxis among children/adolescents with food allergies The incidence of anaphylaxis among children/adolescents with food allergies was either reported in or derived from seven studies conducted in Canada, Spain, the United Kingdom, or the United States.  Table S2) were used to define anaphylaxis in each of the studies. In instances where Sampson's criteria were not explicitly used for the definition of anaphylaxis, studies that used one of three definitions of anaphylaxis were included: (1) an explicit description of cases as anaphylaxis; (2) reported allergic reactions by severity (mild, moderate, severe); those categorized as "severe allergic reactions" were considered to be anaphylaxis; (3)

| Search Output 2. Incidence of anaphylaxis among children/adolescents with peanut allergy
Of the eight studies included in Search Output 1, four studies were conducted exclusively among children with PA.  (Figure 3).

| Search Output 3. Incidence of accidental exposure to peanuts among children/adolescents with peanut allergy
The incidence rate of accidental exposure to peanuts among children/ adolescents with PA was either reported in or derived for the three studies described in Table 3.
Incidence rates ranged from 11.94 accidental exposures per 100 person-years 60 to 14.34 per 100 person-years. 59 The pooled meta-analysis of the incidence rate of accidental exposures to peanuts among children/adolescents with PA is 12.28 cases per 100 person-years (95% CI = 11.51, 13.05), with 0.0% heterogeneity ( Figure 4).

| DISCUSSION
Anaphylaxis is a distressing and potentially fatal event; people may develop stress disorders and psychiatric comorbidity symptoms after experiencing anaphylaxis, which may then impact the way they and their families cope with food and PA. [63][64][65] The risk of anaphylaxis is associated with physical, cognitive, and behavioral aspects of anxiety that must be addressed in order to ensure optimal psychological as well as medical outcomes. [63][64][65] More information regarding anaphylaxis can help the health professional to better support and manage the patient and their caregivers suffering from these events in an evidence-based and cost-effective manner. [63][64][65][66][67] This comprehensive systematic literature review and metaanalysis summarizes the reported incidence of anaphylaxis in children and adolescents with food allergies, PA, and accidental exposure to PA in the literature following the formulation of guideline-based T A B L E 2 Characteristics of studies that report the incidence of anaphylaxis among children/adolescents with peanut allergies  The studies included in this meta-analysis showed a higher range of peanut-induced anaphylaxis cumulative incidence in preschool and older children (14.7%-15.5%) than in adolescents (6.9%) with food allergy; the overall range (children and adolescents combined) was approximately 1%-9% (Table 1). Two other studies of note that . 58 This study also did not investigate anaphylaxis stated as such but did report events that were "potentially life-threatening-throat tightness and angioedema, angioedema in the mouth, cough, wheeze, chest tightness, shortness of breath, noisy breathing, tachypnea, voice change." The study found that 22 children (

| Strengths and limitations
Our study helps to confirm and clarify these previously reported data with the use of more consistent and well-validated criteria for peanutinduced anaphylaxis and meta-analysis using person-year rates of incidence