Allergen immunotherapy and/or biologicals for IgE‐mediated food allergy: A systematic review and meta‐analysis

Abstract Background There is substantial interest in immunotherapy and biologicals in IgE‐mediated food allergy. Methods We searched six databases for randomized controlled trials about immunotherapy alone or with biologicals (to April 2021) or biological monotherapy (to September 2021) in food allergy confirmed by oral food challenge. We pooled the data using random‐effects meta‐analysis. Results We included 36 trials about immunotherapy with 2126 mainly child participants. Oral immunotherapy increased tolerance whilst on therapy for peanut (RR 9.9, 95% CI 4.5.–21.4, high certainty); cow's milk (RR 5.7, 1.9–16.7, moderate certainty) and hen's egg allergy (RR 8.9, 4.4–18, moderate certainty). The number needed to treat to increase tolerance to a single dose of 300 mg or 1000 mg peanut protein was 2. Oral immunotherapy did not increase adverse reactions (RR 1.1, 1.0–1.2, low certainty) or severe reactions in peanut allergy (RR 1,6, 0.7–3.5, low certainty), but may increase (mild) adverse reactions in cow's milk (RR 3.9, 2.1–7.5, low certainty) and hen's egg allergy (RR 7.0, 2.4–19.8, moderate certainty). Epicutaneous immunotherapy increased tolerance whilst on therapy for peanut (RR 2.6, 1.8–3.8, moderate certainty). Results were unclear for other allergies and administration routes. There were too few trials of biologicals alone (3) or with immunotherapy (1) to draw conclusions. Conclusions Oral immunotherapy improves tolerance whilst on therapy and is probably safe in peanut, cow's milk and hen's egg allergy. More research is needed about quality of life, cost and biologicals.


| Background
Around 6% of people live with immunoglobulin E (IgE)-mediated food allergy 1 which affects quality of life, social activities, anxiety and nutrition. 2,3 Patients are advised to avoid the allergen and use medications to relieve symptoms, but accidental exposure is common and can be life-threatening. 4 There is growing interest in allergen-specific immunotherapy (hereafter immunotherapy) and biological therapies. However, there are conflicting results, with some studies finding improved September 2021) in food allergy confirmed by oral food challenge. We pooled the data using random-effects meta-analysis.
Results were unclear for other allergies and administration routes.
There were too few trials of biologicals alone (3) or with immunotherapy (1) to draw conclusions.

Conclusions:
Oral immunotherapy improves tolerance whilst on therapy and is probably safe in peanut, cow's milk and hen's egg allergy. More research is needed about quality of life, cost and biologicals.

K E Y W O R D S
biological, food allergy, IgE-mediated, immunotherapy, peanut

G R A P H I C A L A B S T R A C T
We systematically reviewed 39 randomized controlled trials about allergen-specific immunotherapy and/or biologicals in 2244 people with IgE-mediated food allergy, mostly children. We found that oral immunotherapy is probably safe and effective in peanut, milk and egg allergy. The number needed to treat (NNT) to increase tolerance to a single dose of 300 mg or 1000 mg peanut protein was 2. Abbreviation: NNT, number needed to treat allergen tolerance and others emphasizing increased adverse reactions. 5 Past reviews sometimes combined studies of various designs and therapies, focused on only one type of food allergy or are outdated. [6][7][8][9][10] We compiled an up-to-date systematic review about the efficacy and safety of immunotherapy and/or biologicals to inform forthcoming Global Allergy and Asthma European Network (GA 2 LEN) guidelines and add to existing guidelines. 11,12

| Objectives
We prioritized the following review questions after feedback from people with food allergy, patient advocates, healthcare professionals, teachers and policymakers. No industry representatives were involved.
1. What is the efficacy, safety and cost-effectiveness of a) immunotherapy alone, b) immunotherapy with a biological or c) biologicals alone for children and adults with any IgE-mediated food allergy compared to no active treatment agent? 2. What is the efficacy, safety and cost-effectiveness of immunotherapy administered by different routes for children and adults with any IgE-mediated food allergy?

| ME THODS
A task force of allergy specialists, patient representatives, primary care doctors, psychologists, other clinicians, teachers and methodologists from 19 countries undertook the review.
The full methods are available in the protocol (International Prospective Register of Systematic Reviews PROSPERO registration: CRD42021250940) so methods are only briefly summarized here.

| Eligibility criteria
Studies meeting the following criteria were eligible:

| Study selection
Clinicians, patients and information specialists developed a search strategy (Supplement S1). Methodologists (CS, DdS) searched six databases (CINAHL, Cochrane Library, EMBASE, ISI Web of Science, MEDLINE, Scopus), reviewed bibliographies of reviews, guidelines and identified studies and contacted experts in the field for additional research.
Two methodologists independently screened the titles, abstracts and full text of potentially relevant studies (CS, DdS). Shortlisted studies were rescreened by the task force for consensus. Some studies included in previous reviews did not meet our stringent criteria (S2).

| Data extraction
Pairs of methodologists (MG, DdS, CS) and task force members (all authors) extracted study characteristics and outcomes independently. Clinicians independently extracted additional data about adverse reactions and thresholds (PRdR, EK, GR, PB, AWN, ADG, SA, ME, TZ), including the different definitions used (S3). A senior clinician acted as an arbitrator if needed, but there was consensus (GR).
We had most data about peanut allergy so we compiled the proportion able to tolerate 300 mg (level at which there is a substantial reduction in risk from products labelled with 'may contain') 13 and 1000 mg peanut protein as both a single last tolerated and cumulative dose. Our approach was conservative, so when someone was reported to tolerate 300 mg to 600 mg peanut protein we classified them as tolerating a 300 mg dose. We used a range of 1000 mg to 2000 mg for the 1000 mg threshold. We used these ranges because an individual's threshold may vary by at least a factor of two when they are challenged twice within a short time. 14

| Risk of bias in individual studies
Four methodologists and clinicians independently assessed the risk of bias in individual studies using the Cochrane Risk of Bias tool 2 (ROB2) 15 (pairs of all authors). Arbitration was available if needed from a senior clinician (GR) but there was agreement in the risk of bias assessments.

| Synthesis of results
We used the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach to synthesize data about each outcome. 16 We pooled intention-to-treat data using random-effects Mantel-Haenszel meta-analysis (Revman 5.4) because the studies included different populations, regimes and time periods and to avoid overweighting large but imprecise studies.
We divided studies based on the food allergy and therapy route. We undertook subgroup analysis based on risk of bias, age, allergy severity, comparator and threshold tolerated. In sensitivity analysis, we used a continuity correction (adding 0.05 to numerators and denominators) where there were no events in each study arm for severe or life-threatening events, anaphylaxis and adrenaline use.
We used funnel plots to help assess publication bias. We quantified the heterogeneity of studies using the I2 statistic, with values <25% indicating low heterogeneity. 17 All task force members developed conclusions by consensus, recognizing any potential conflicts of interest, which were declared in advance. We used standardized GRADE statements to summarize the conclusions 18 (S4 lists our wording conventions).
We used the word 'may' to represent low certainty evidence, 'probably' for moderate certainty and 'resulted in' for high certainty evidence.

| Study characteristics
We included 36 trials of immunotherapy with 2126 participants, mainly children, and 3 trials about biological monotherapy with 118 participants, mainly teenagers and adults. ( Table 1 and The GRADE certainty of evidence was generally low to moderate (S6). The certainty of evidence was mainly downgraded due to risk of bias (S5), indirectness and imprecision. About one quarter of studies were at low risk of bias (28%), half at moderate risk (47%) and quarter at high risk (25%).
Tables 2-4 list the key findings, with details in S6 and S7. Trends were broadly similar for people with mild/moderate or severe allergy and studies at lower and higher risk of bias (S7). There was insufficient information to be able to distinguish immunotherapy with raw vs. cooked foods or to draw conclusions about the most effective regimen or duration of treatment.

| Desensitization
Most studies about peanut allergy enrolled children ( Table 2). Oral immunotherapy resulted in a large increase in the proportion able to tolerate peanut during therapy (absolute difference 62%, relative risk (RR) 9.9, 95% confidence interval (CI) 4.5-21.4, high certainty).
Those receiving oral immunotherapy were six times more likely than controls to tolerate a single dose of 300 mg peanut (roughly one peanut, number needed to treat (NNT) 2) and 17 times more likely to tolerate 1000 mg (NNT 2). Before immunotherapy, the mean maximum tolerated dose was only about 34 mg (S5).
Epicutaneous immunotherapy probably increased the proportion able to tolerate peanut during therapy (absolute difference 29%, RR 2.6, 95% CI 1.8-3.8, moderate certainty). Those receiving epicutaneous immunotherapy were twice as likely as controls to be able to tolerate a single dose of 300 mg peanut (NNT 6).
Sublingual immunotherapy may result in a large increase in the proportion able to tolerate peanut during therapy (absolute difference 55%, RR 4.7, 95% CI 1.6-13.8, low certainty). It is unclear whether subcutaneous immunotherapy had any impact because the certainty of evidence was very low.

| Sustained unresponsiveness
Peanut oral immunotherapy may increase the proportion of children able to tolerate peanut after stopping therapy (absolute difference 31%, RR 8.8, 05% CI 1.2-61.6, low certainty). No data were available about sustained unresponsiveness using other administration routes.

| Adverse reactions
There was no statistically significant increase in the proportion of people who had adverse reactions due to peanut immunotherapy, No deaths were reported due to immunotherapy, and there was no increase in anaphylaxis or life-threatening reactions (S7).

| Cow's milk allergy (Table 3)
All but one study about cow's milk allergy enrolled children ( Table 3).
An increased proportion may have experienced adverse reactions (absolute difference 67%, RR 3.9, 95% CI 2.1-7.5, low certainty), but few studies focused solely on reactions caused by the therapy (S7) and there was no significant difference in the 3 studies comparing with placebo.
The most common adverse reactions were mild gastrointestinal symptoms. A small number experienced respiratory symptoms. It was unclear whether oral immunotherapy increased severe reactions because the certainty of evidence was very low. Reports of severe reactions, anaphylaxis, use of adrenaline and discontinuation were rare. No hospital admissions or deaths were reported due to treatment.
There was insufficient evidence to draw conclusions about other administration routes in cow's milk allergy.

| Hen's egg allergy (Table 4)
All studies about hen's egg allergy were in children and all but one compared with an elimination diet ( Anaphylaxis or symptoms that were severe enough to discontinue immunotherapy were rare.

| Other food allergies and outcomes
We identified few studies about other food allergies or directly comparing immunotherapy routes to present meaningful results (see S6 and Note: Rounded to nearest whole number. Absolute difference is % active minus control. Adverse reactions and severe adverse reactions are all reactions, not necessarily linked to the therapy. Multiple definitions of severe reactions were used (see S3). In general these involved marked impact on participants (eg symptomatic bronchospasm, hypotension) and need for medical intervention. Online supplement S7 contains additional data such as the proportion of doses experiencing reactions and adverse events linked to therapy. 'Comparisons' means the number of comparisons between active and control groups. In most cases this is the total number of trials contributing the meta-analysis but some studies compared different doses with placebo. Each comparison was included separately in the metaanalysis. Data about threshold reactivity at 1000 mg peanut protein were only available for oral immunotherapy.
Abbreviations: CI, confidence interval; NA, not applicable as no statistically significant difference; NNT, number needed to treat. We did not find a clear relationship between the duration of treatment and desensitization, sustained responsiveness or safety (S7).

| Biological therapies
There were too few studies to draw meaningful conclusions about biologicals alone or added to immunotherapy (S6). In those aged 13+ years, there was a trend towards improved tolerance of peanut during biological monotherapy, with few side effects. However the cer- Epicutaneous immunotherapy may also be safe and effective amongst children with peanut allergy, but less is known about this for other food allergies, or about sublingual or subcutaneous immunotherapy for any food allergy. We found insufficient evidence to draw conclusions about benefits or harms from biologicals.

| Comparison with previous research
Many studies, reviews and opinion pieces have explored immunotherapy for people with food allergy, sometimes with inconsistent results. 22 Our review differs from others because it includes the highest quality and most up-to-date evidence, divided by type of food allergy and administration route. We excluded trials that did not use food challenges to confirm food allergy before treatment, as this may introduce bias into absolute event rates. We examined tolerated thresholds in detail, which we believe is the first analysis of its kind. Our review supports others suggesting that clinicians could consider immunotherapy for some groups of patients. 6,7

| Implications for research
There is now a good pool of studies about oral immunotherapy for peanut allergy, but we can only draw limited conclusions about other food allergies, administration routes and biologicals due to sparse good quality evidence.
An important area not explored in our review is patient preferences and motivation to use immunotherapy and/or biologicals. Our findings suggest that (oral) immunotherapy may improve tolerance whilst people continue the therapy, but gains may not be sustained after stopping therapy. Therefore, people need to consider whether they are willing to continue therapy for an extended period. Future research using standardized threshold levels for each allergen may help patients make more informed decisions.
Future studies could create core outcome sets 23  Standardized immunotherapy regimens would also assist comparisons.
More high quality trials are needed to directly compare different administration routes and to explore the impacts of adding biologicals.

| Strengths and limitations
A strength of our review is that it was conducted by a large group of clinicians, allied health professionals, patient representatives, teachers and researchers from different parts of the world.
By analyzing detailed data about desensitization thresholds, we showed that only 2 children needed to be treated with peanut oral immunotherapy to increase 1 child's reactivity threshold up to 300 mg and 1000 mg. This would reduce the chance of reactions from accidental consumption of packaged food containing peanut. 11 Given variability in reactivity thresholds, 12 achieving tolerance to a single dose of 1000 mg is likely protective to real-life exposures.
However, our conclusions are limited by methodological variations in the available studies and a lack of eligible studies about many topics. We found that adverse reactions tended to increase, but we do not know whether these reactions were due to the therapy, accidental exposure or unrelated causes. This limits conclusions about the safety profile. Some studies excluded patients with previous anaphylaxis, which may also affect generalizability.
A number of trials have pragmatic designs and do not confirm food allergy at the outset using oral food challenges. We excluded these studies to ensure that participants definitely had food allergy, but there is an argument for including such research in decision-making if the selection criteria reflect real-life clinical practice.

| CON CLUS IONS
There may be merit in allergen-specific immunotherapy and/or biologicals for a carefully selected group of patients, but there are also downsides including (mainly mild) reactions to manage, relatively long treatment periods and benefits potentially only lasting whilst people maintain the regimen.
Immunotherapy is also specific to the culprit food(s) and requires exposure to allergens, which may be off-putting for some. We know little about how patient motivation to sustain the therapy may differ in various parts of the world or based on patient characteristics.
When making patient-centred shared decisions about whether to offer immunotherapy and/or biologicals to an individual, clinicians and patients need to weigh up factors such as patient preferences, motivation, accessibility and cost alongside the evidence about safety and effectiveness.

ACK N OWLED G EM ENTS
The Global Allergy and Asthma European Network (GA2LEN) contributed funding towards the systematic review process. Authors and contributors volunteered their time.

CO N FLI C T O F I NTE R E S T
Debra de Silva, none declared. Organisation received GA 2 LEN funding to support review process. Pablo Rodríguez del Río,