Study of induction of Tolerance to Oral Peanut: a randomized controlled trial of desensitisation using peanut oral immunotherapy in children (STOP II)

Article history

The research reported in this issue of the journal was funded by the EME programme as project number 08/99/18. The contractual start date was in January 2010. The final report began editorial review in July 2013 and was accepted for publication in April 2014. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The EME editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Andrew Clark and Pamela Ewan are inventors on a patent application covering the peanut protein dose range.

Copyright statement

© Queen’s Printer and Controller of HMSO 2014. This work was produced by Anagnostou et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Setting

We undertook this single-centre RCT at the Wellcome Trust Clinical Research Facility on the Cambridge Biomedical Campus between January 2010 and March 2013. The study was conducted according to the principles of Good Medical Practice for clinical trials.

Patient and public involvement

Patient representatives were involved from the outset. We surveyed the pilot study participants to gather feedback that informed the design of the current trial. The national patient support group, the Anaphylaxis Campaign, has reviewed the current protocol and they have ratified the design and confirmed the study end points are important and relevant to their membership. Furthermore, a representative of the Anaphylaxis Campaign joined as a member of the Trial Steering Committee.

The result will be disseminated by an article in the Anaphylaxis Campaign’s newsletter and website.

Aims

The overarching objective was to determine efficacy of OIT in desensitising children with peanut allergy. We aimed to determine whether or not the treatment was successful in the intervention group compared with the control (phase 1) and whether or not it is successful when offered to the control group (phase 2) (Figure 1).

FIGURE 1.

The Study of induction of Tolerance to Oral Peanut (STOP II) study design.

Participants

Participants were recruited both locally (allergy clinic) and nationally (through the national patient support group Anaphylaxis Campaign). Eligible participants were 7–16 years of age with an immediate-type hypersensitivity reaction after peanut ingestion, positive skin prick test (SPT) to peanut (extract from ALK-Abelló, Hørsholm, Denmark) defined by a weal of ≥ 3 mm in diameter in the presence of a negative saline and positive histamine control, and positive DBPCFC. 19 We excluded participants if they had a significant chronic illness (except for eczema, rhinitis and asthma) because this is an immunomodulatory therapy, if a care provider or current household member had suspected or diagnosed allergy to peanut, or if there was unwillingness or inability to comply with study procedures. We did not exclude participants who had a previous life-threatening reaction, tree nut allergy or a history of severe asthma.

The Cambridge Central Ethics Committee approved the study (09/H0308/154) and the guardian of each participant gave his or her informed consent. Children of an appropriate age (≥ 12 years) were encouraged to provide their own assent.

Randomisation

Subjects were randomised (1 : 1) via an audited online system (Randomizer, Medical University of Graz, Graz, Austria) to the active or control group (see Figure 1). Minimisation was used to reduce imbalance of baseline covariates, with a random element using a weighting probability of 0.8. Factors were sex, age, challenge threshold, peanut-specific serum immunoglobulin E (IgE), severity from history and presence of asthma or other food allergy. Group allocation was unblinded.

During phase 1, the active group underwent 6 months of active OIT and the control group underwent 6 months of standard care (peanut avoidance). Minimisation was used to avoid imbalance of baseline covariates, with a random element using a weighting probability of 0.8. At the end of phase 1 (6 months) all participants were assessed for peanut allergy by DBPCFC. During phase 2, participants in the control group still allergic to peanut after 6 months of standard care were given active OIT, followed by a DBPCFC after 6 further months (see Figure 1). There were no important changes to method after commencement.

Participants underwent an initial DBPCFC followed by 1 : 1 randomisation. During phase 1, participants in the active group underwent OIT for 26 weeks. Participants randomised to the control group were asked to continue normal peanut avoidance for the first 26 weeks of the study (standard care). Both groups then underwent a second DBPCFC. In phase 2, control group participants underwent active OIT for 26 weeks followed by a third DBPCFC at 52 weeks (end of phase 2).

Patients were not blinded to allocation groups; however, the primary outcome was assessed using an objective blinded measure, i.e. ‘no reaction’ during a DBPCFC. Challenge assessors were blinded to the challenge placebo/active arm, but not to the treatment allocation. Study personnel were the chief investigator, a clinical fellow, a study nurse and a scientist. All study personnel were masked to the randomised active/placebo assignment in the challenge except the scientist who prepared the challenge material but had no interaction with the participant or study team. Allocation was saved in a locked database accessible only to the unblinded scientist.

Procedures

The primary end point was the proportion of desensitised subjects at the end of phase 1. Desensitisation was defined as ‘no reaction’ during peanut DBPCFCs, with a cumulative dose of 1400 mg of peanut protein.

Statistical analysis

The full analysis population was all subjects who were randomised and participated in at least one post-baseline assessment. The criteria for per-protocol analysis of the outcome of peanut challenge at the end of immunotherapy consisted of desensitisation and continuation of immunotherapy up to the maintenance dose of 800 mg of protein.

Fisher’s exact test was used to compare the proportion of those with desensitisation to peanut after 6 months in the active group with the control group at the end of phase 1. Multiple logistic regression was used to adjust the odds of desensitisation for baseline characteristics. Secondary analyses tested for treatment differences with Fisher’s exact test (proportion response to treatment in active group and control group at end of phase 2), paired and independent sample t-tests (absolute and fold change in threshold), and Mann–Whitney U-tests (change in QoL scores, basophil AUC of percentage of CD63 and MFI, peanut-specific serum IgE and SPT weal diameter). All statistical tests described in this section use a two-sided 5% significance level. We assessed analysis on the intention-to-treat population that included all those who were randomised and participated in at least one post-baseline assessment.

Sample size was based on Fisher’s exact test with 90% power and 5% significance (two-sided). A sample size of 49 in each group is sufficient to detect proportions of participants, with desensitisation to peanut of 0.64 and 0.30 in the active and control groups, respectively, at the end of phase 1. Allowing for a 5% dropout increased the required sample size to 52 participants in each group and 104 subjects overall. Based on the above, we would expect 35 waiting list group patients to proceed to the active intervention in phase 2. Non-parametric tests may be used instead of parametric tests if the assumptions are not appropriate.

A statistical analysis plan, which embodies all the calculations performed, was agreed and signed off before any analysis was undertaken on the database (see Appendix 1).

Results

We enrolled 104 children, aged 7–16 years (median 12.4 years), to the study. Five children did not react during their baseline peanut challenge, thus not meeting the inclusion criteria, and were excluded from further participation in the study. Therefore, 99 children were randomised: 49 out of 99 to the active group and 50 out of 99 to the control group (Figure 2).

FIGURE 2.

The CONSORT diagram of patient flow through the study.

One child was discontinued and five withdrew from the active group during phase 1. In the control group, four children withdrew and one was discontinued during phase 1. Two further children withdrew from the control group when they underwent the intervention in phase 2.

There were no significant differences in baseline characteristics between allocation groups (Table 1).

Clinical end points

Primary objective

There was a significant difference between the numbers of patients who tolerated 1400 mg of peanut protein during DBPCFC at the end of phase 1 in the active (24/39) and the control groups (who underwent peanut avoidance for 24 weeks) (0/41) (p < 0.001) (Table 2). The proportion desensitised during phase 1 after 24 weeks was 0.62 (range 0.45–0.78) and 0 (range 0–0.091) in the active and control groups, respectively.

TABLE 2 - Clinical end points for phase 1 and phase 2

CI, confidence interval; NOAEL, no observed adverse effect level.

First entry is for the control group fold change and the second entry is for the active group fold change.

p-value obtained from Fisher’s exact test.

p-values are from Wilcoxon signed-rank tests.

p-value from Mann–Whitney U-test.

QoL end points, FAQLQ-PF – parent form for 5–12 years. Scale 0 (best) to 6 (worst); median change in FAQLQ-PF scores from baseline to post treatment.

Desensitised defined as tolerates 1400 mg of protein DBPCFC.

Proportion able to tolerate daily ingestion defined as able to ingest 800 mg of protein regularly up to a total of 26 weeks’ OIT.

End points	Control	Active	Control vs. active p-valuea
Clinical end points
	n = 46	n = 39
Phase 1
Number and proportion desensitised and able to tolerate daily ingestion after phase 1
Desensitised	0	24	< 0.001b
Not desensitised	46	15
Proportion desensitised (95% CI)	0 (0 to 0.091)	0.62 (0.45 to 0.78)	–
Proportion able to tolerate daily ingestion (95% CI)	0	0.84 (0.70 to 0.93)	–
Within-patient changes in NOAEL (mg), absolute and fold change
Median (range) absolute change in NOAEL (mg)	0 (–95 to 45)	1345 (45–1400)	0.002, < 0.001c
Median (range) fold change in NOAEL (mg)	0.81 (0.05–1.82)	25.5 (1.82–280)	0.003, < 0.001c
NOAEL (mg) after phase 1
NOAEL (mg) median (range)	5 (5–400)	1400 (100–1400)	< 0.001d
Median difference in NOAEL (mg) between groups
Median (95% CI)	1395 (395 to 1395)		< 0.001d
Phase 2
Proportion desensitised and able to tolerate daily ingestion after phase 2 (treatment)
Proportion desensitised (95% CI)	0.54 (0.35 to 0.72)	–	–
Proportion able to tolerate daily ingestion (95% CI)	0.91 (0.79 to 0.98)	–	–
QoL end points
	n = 20	n = 19
Median (range) change in score from baseline to post-treatment	–1.41 (–4.83 to 1.38)	–1.61 (–4.87 to 0.24)	< 0.001, < 0.001

Secondary objectives

Response to treatment, defined by the ability to tolerate daily doses of 800 mg of peanut protein after 24 weeks of immunotherapy, occurred in 0.84 [95% confidence interval (CI) 0.70 to 0.93] of the active group at the end of phase 1 and 0.91 (95% CI 0.79 to 0.98) of the control group at the end of phase 2 (post immunotherapy). There was a significant increase in peanut no observed adverse effect level (NOAEL) in the active group after phase 1, with a median change in threshold of 25.5-fold (p < 0.001) (Figure 3) compared with a small positive change in peanut threshold (NOAEL) in the control group during phase 1 (0.81, range 0.05–1.82).

FIGURE 3.

Mean and 95% CI peanut protein NOAEL (mg) by treatment group. NOAEL, highest dose of peanut protein tolerated in mg of protein during challenge or OIT (active treatment); difference at 6 months, p < 0.001. p-value from the Mann–Whitney U-test.

The proportions of patients who achieved desensitisation (0.54, 95% CI 0.35 to 0.72) or response to treatment (0.91, 95% CI 0.79 to 0.98) in the control group after 24 weeks of OIT (end of phase 2) were similar to that in the active group.

Quality-of-life end points

Quality-of-life scores assessed by the FAQLQ-PF measure were similar between active and control groups at baseline. FAQLQ-PF (aged 5–12 years) was available from 19 children in the active group and 20 in the control group, before and after treatment. After treatment in both groups there was a similar and clinically meaningful improvement (decrease) in QoL scores (see Table 2).

Immunological end points

Immunological assessments revealed a significant small reduction in median SPT weal diameter (–1 mm, range –8 to 4 mm; p = 0.0015) and an increase in peanut-specific serum IgE [12.7 kilounit (kU)/l, range –18.6 to 1359 kU/l; p < 0.001] after 24 weeks’ OIT in the active group. Basophil stimulation data were expressed as the AUC of plots of MFI and percentage of CD63-positive cells against concentration of peanut protein. No significant within-patient differences were found after treatment for AUC of MFI or percentage of CD63, although there was a reduction in MFI and percentage of CD63 at the lower peanut doses (Table 3 and Figure 4).

TABLE 3 - Immunological end points

SPT performed with single-lancet technique using peanut extract, normal saline and histamine controls (ALK-Abelló; Hørsholm, Denmark).

First entry is for the control group fold change and the second entry is for the active group fold change.

Peanut-specific IgE measured by ImmunoCAP (Thermo Scientific).

Immunological end points	Control (n = 46)	Active (n = 39)	p-valuea
Within-participant changes in SPT weal diameter (mm)
Median (range) change from baseline to post treatment	–1.5 (–8 to 9)	–1 (–8 to 4)	0.0693, 0.0015
Within-participant changes in total IgE (kU/l) by group
Median (range) change from baseline to post treatment	24 (–303 to 1582)	99 (–164 to 1411)	0.0109, < 0.001
Within-participant changes in peanut-specific serum IgE (kU/l)
Median (range) change from baseline to post treatment	74.5 (–55 to 637)	12.7 (–18.6 to 1359)	< 0.001, < 0.001

FIGURE 4.

(a) In vitro basophil activation by peanut before and after desensitisation as percentage of CD63-positive cells. Heparinised whole blood was stimulated with a range of peanut protein concentrations (0.001–100 µg/ml) and flow cytometry was used to assess CD63 expression (%) within the basophil population. Differences in AUCs were not significant. (b) In vitro basophil activation by peanut before and after desensitisation as CD63 MFI (MFI of CD63 marker). Heparinised whole blood was stimulated with a range of peanut protein concentrations (0.001–100 µg/ml) and flow cytometry was used to assess CD63 expression (MFI) within the basophil population. Differences in AUCs were not significant.

Logistic regression

Logistic regression revealed several baseline covariates had an influence on the final NOAEL (the highest amount of peanut tolerated after OIT) (Table 4). Treatment, log-baseline NOAEL, age, family history and log-transformed peanut-specific IgE have a statistically significant effect on log-transformed NOAEL at 6 months. On average, patients in the OIT group have a log-NOAEL at 6 months 4.12-fold higher than those in the waiting list group. For every unit increase in log-transformed baseline NOAEL, the log-transformed NOAEL at 6 months, on average, increases 0.33-fold. For every year increase in age at baseline the log-transformed NOAEL at 6 months decreases, on average, by 0.17. Patients with a family history of peanut allergy have, on average, 0.64 lower log-transformed NOAEL at 6 months than patients who do not have a family history of peanut allergy.

TABLE 4 - Tobit regression model22 exponentiated estimates – dependent variable natural log-6-month NOAEL

WAO, World Allergy Organization score.

NOAEL the highest amount of peanut protein (mg) tolerated after OIT. The continuous variables can be interpreted as the percentage change in NOAEL expected from a unit increase of that variable when all other covariates are fixed. The categorical variable estimates can be interpreted as the percentage change compared with the reference group, when all other covariates are fixed. For logged covariates, the expected percentage change in 6-month NOAEL with a 10% increase in the logged covariate can be calculated as 1.10 log(exponentiated estimate). For example, for a 10% increase in baseline NOAEL (mg) we would expect a [1.1 log(1.40) = 1.032] 3.2% increase in 6-month NOAEL (mg). Similarly, for a 10% increase in baseline peanut-specific IgE (kU/l) we would expect a 4.8% decrease in 6-month NOAEL (mg).

Covariates	Estimate	95% CI	p-value
OIT	105.5	67.73 to 164.40	< 0.001
Log (baseline NOAEL + 1)	1.40	1.15 to 1.69	< 0.001
Age	0.76	0.65 to 0.89	< 0.001
Female	1.42	0.87 to 2.31	0.16
Weight	1.04	1.01 to 1.06	0.004
QoL	1.12	0.89 to 1.42	0.32
Asthma	1.09	0.65 to 1.74	0.79
Eczema	0.82	0.53 to 1.26	0.36
Rhinitis	0.69	0.46 to 1.05	0.09
Other food allergy	0.84	0.52 to 1.36	0.48
Family history of peanut allergy	0.41	0.24 to 0.71	0.001
WAO grade 2	2.88	1.65 to 5.01	< 0.001
WAO grade 3	0.86	0.44 to 1.68	0.66
WAO grade 4	0.62	0.21 to 1.82	0.40
Peanut SPT weal diameter	1.02	0.94 to 1.11	0.60
Other nut SPT weal diameter > 3 mm	1.37	0.81 to 2.31	0.23
Tryptase	1.06	0.95 to 1.19	0.32
Log (peanut-specific IgE + 1)	0.60	0.51 to 0.71	< 0.001
Log (total IgE + 1)	1.74	1.35 to 2.23	< 0.001
Log (basophil activation CD63 MFI AUC)	0.98	0.73 to 1.32	0.91

For every unit increase in log-transformed baseline peanut-specific IgE, the log-transformed NOAEL at 6 months decreases, on average, by 0.31.

For every unit increase in log-transformed baseline total IgE, the log-transformed NOAEL at 6 months increases, on average, by 0.36.

A Tobit regression model was used to fit the data, showing OIT treatment and log-baseline NOAEL was associated with a statistically significant increase in log-transformed NOAEL after 24 weeks of OIT (see Table 4). On average, patients in the OIT group had a log-NOAEL at 6 months 4.66-fold higher than those in the control group after phase 1.

A Tobit model allows for censoring in positive dependent variables, i.e. in this study the NOAEL (peanut threshold) was censored at 1400 mg, and the maximum cumulative dose of peanut protein administered to patients was 1400 mg and patients who did not react at 1400 mg may not react at even higher cumulative doses of peanut protein. A total of 24 out of the 85 patients included in this analysis achieved a NOAEL of 1400 mg at 6 months.

Adverse events

The number of adverse events was similar in both groups after treatment (Table 5). The majority of events were gastrointestinal, with oral itching being the most common (occurring after 6.3% of all doses). Cutaneous events were uncommon, appearing after only 0.16% of doses. Wheezing occurred after 0.41% of doses and was treated with inhaled beta-2 agonists or oral antihistamines in all cases, except for one participant who also received intramuscular epinephrine on two occasions, with rapid resolution of his symptoms. There were no serious adverse reactions and no cardiovascular events.

TABLE 5 - Adverse events during treatment presented per participant (n = 94) and per dose (total doses = 17,793)

IM, intramuscular; ITU, intensive care unit; SAR, serious adverse reaction; SUSAR, serious unexpected suspected adverse reaction.

Events	n (%) of participants who experienced an adverse event	n (%) of adverse events per dose of OIT
Symptoms
Mouth itch	76 (81)	1121 (6.30)
Abdominal pain	54 (57)	460 (2.59)
Nausea	31 (33)	393 (2.21)
Vomiting	31 (33)	134 (0.75)
Diarrhoea	1 (1)	5 (0.03)
Urticaria	12 (13)	29 (0.16)
Angiooedema	18 (19)	71 (0.40)
Erythema	20 (21)	41 (0.23)
Rhinitis	23 (24)	65 (0.37)
Wheezing	21 (22)	73 (0.41)
Laryngeal oedema	1 (1)	1 (0.01)
Cardiovascular collapse or fainting	0 (0)	0 (0.00)
Outcome
Admission to ITU/SAR/SUSAR	0 (0)	0 (0.00)
Use of inhaled beta-2 agonist	18 (19)	63 (0.35)
Use of IM epinephrine	1 (1)	2 (0.01)

Discussion

Daily doses of peanut OIT up to a maximum dose of 800 mg of protein had a clinically meaningful effect on the disease, demonstrated by a high incidence of desensitisation, large absolute and fold increases in NOAEL threshold and a significant improvement in QoL score.

These results provide the first well-controlled and accurate estimate of the efficacy, benefits and risks of desensitisation using peanut OIT. Previous studies were either uncontrolled or too underpowered to estimate the effect size of the intervention. 20,21

Peanut allergy is a common long-lived disease with onset in childhood. Fear of severe reactions and the effect this has on social behaviour reduces QoL, which can often be lower than it is for children with other chronic diseases such as type 1 diabetes or rheumatological conditions. 6,7 Parental perception of disruption in daily activities is mainly due to the perceived risk of their children’s death,6 leading to anxiety related to making food choices inside and outside the home. 7 Secondary effects of anxiety lead to socially disadvantageous behaviour in some, for example avoiding eating outside the home. 7

This study confirms the improvement in QoL shown by an earlier uncontrolled study of peanut OIT. 24 Following OIT, our patients reported that they no longer avoid eating foods with precautionary labelling and they now eat out at restaurants without checking ingredients.

Most peanut-allergic patients are able to avoid accidentally eating large amounts of peanut; however, if untreated, they are at constant risk of reacting to peanut hidden in foods. Without OIT, the attitude of both patients and families to selecting potentially contaminated foods varies widely and is determined mostly by subjective judgements and previous experience. 25 Consequently, accidental reactions are not uncommon, occurring in 14–55% per year. 9 A recent study combined published data on peanut thresholds, peanut contamination of chocolate and patterns of population food consumption using probabilistic modelling to estimate the absolute risk of reacting to precautionary-labelled foods. 26 The risk of a peanut-allergic child reacting after eating a chocolate with precautionary labelling was 0.61% (95% CI 0.47% to 0.81%). The authors showed that 36% (95% CI 31% to 42%) of chocolate bars for which peanut was not listed in the ingredients, sourced from France and the USA, contained detectable peanut protein, with a median of 8.25 mg/kg protein (95% CI 6.54 mg/kg to 10.54 mg/kg). 26 A study of 62 catering establishments in Northern Ireland found that 21% served a peanut-contaminated meal, despite a request for a peanut-free meal. 27 The highest level of contamination was found in a chicken curry containing 4.795 mg of peanut protein. 27 Our data show that peanut OIT can raise the reactive threshold by 25.5-fold so that 83–91% of patients can eat 800 mg of peanut protein regularly without reacting (approximately equal to four to six whole peanuts). In addition, 54–64% of children could tolerate a challenge with 1400 mg of protein without reacting, providing protection to approximately 10 peanuts. Raising the reactive threshold for patients is a key part of this treatment, protecting patients from greater amounts than they are likely to encounter on a day-to-day basis in contaminated snacks or meals.

Our primary end point was the proportion of patients desensitised after 6 months of OIT. Peanut allergy may resolve in up to 20% of children over several years,4 but there are no data over a short period such as 6 months. There is currently no disease-modifying treatment for peanut allergy; therefore, the control group underwent the current best treatment for peanut allergy – peanut avoidance – to identify the proportion whose allergy resolved spontaneously over the 6-month period of phase 1, to reduce type 1 error.

We did not use a placebo during the control period because the risk of including it outweighed any potential benefit. During the pilot study, participants receiving peanut OIT reported that they significantly relaxed their peanut avoidance behaviour during the first 6 months of treatment, despite advice to the contrary. Therefore, we could not exclude the likelihood that a significant proportion of placebo-treated patients in this phase 2 study would relax their avoidance practice, guessing they were taking active treatment and, therefore, risking a severe reaction. Placebos in general are included to reduce the risk of type 1 error (i.e. false-positive results) in studies for which the primary end point can be influenced by knowing one is receiving a treatment, attention from health-care professionals, and the expectations of a treatment’s effectiveness by those running the research study. However, the risk of a type 1 error in the context of this trial was acceptably low because the primary end point was measured using a blinded objective measure (i.e. ‘no reaction’ to a large amount of peanut during DBPCFC) that could not be influenced by these factors. A limitation of this study is that secondary outcomes such as QoL scores may have been influenced by knowledge of treatment allocation.

There was a small risk of bias as participants who knew they were receiving active treatment may have under-reported minor symptoms at the higher challenge doses, although these symptoms would have been infrequent and subjective. Additionally, during phase 1, 10 subjects did not undergo a post OIT challenge because they had withdrawn or not reached the target maintenance OIT dose at 6 months. It is probable that in phase 1 the true response rate is lower than estimated; however, in phase 2, for which there were few dropouts, we still observed a large effect. A very conservative sensitivity analysis was performed in which all the unobserved subjects in the active group were imputed as not desensitised and all the unobserved subjects in the control group were imputed as desensitised. The analysis gave a risk difference of 0.41 (exact 95% CI 0.21 to 0.58), which supports the conclusions of the main analysis.

It is probable that long-term peanut protein ingestion will be required to provide continued protection from accidental exposure and that long-term desensitisation rather than clinical tolerance is a realistic end point for treatment. During OIT, there is evidence of basophil and mast cell desensitisation (i.e. reduced SPT weal size), accompanied by a gradual change in peanut-specific T-cell surface markers from Th2 to Th1 phenotype (Dr Katherine Anagnostou, University of Cambridge, 2010, personal communication). Recent data from our group show that elevated peanut-specific serum IgE levels persist for several years after starting OIT, despite apparent clinical desensitisation (Dr Katherine Anagnostou, personal communication).

However, preliminary work suggests that extending the dose interval to 800 mg once weekly after 2–3 years of daily treatment is well tolerated, resulting in ongoing protective desensitisation (Dr Katherine Anagnostou, personal communication).

We chose a single maintenance treatment dose of 800 mg, pragmatically, as the largest amount of peanut that would be feasible/acceptable to take on a daily basis. From the limited data available in published studies, it is apparent that the rapidity of the up-dosing schedule has a greater effect on safety and efficacy than the magnitude of the maintenance dose, with semi-rush regimes showing poor efficacy and more frequent reactions. The starting dose of 2 mg of protein was developed from dose-ranging work in our pilot studies. 21,22

Prognostic factors were explored using logistic regression revealing baseline covariates related to a change in log-transformed NOAEL threshold post OIT. Treatment with OIT was not surprisingly the most influential factor. Age, family history and peanut-specific serum IgE were associated with a significant decrease in log-transformed NOAEL after 24 weeks of OIT, implying that these might predict a less favourable outcome.

Not surprisingly, there were many more allergic events during active treatment than during periods of peanut avoidance. The safety data in this trial show that most adverse events were mild and due to gastrointestinal symptoms (e.g. oral itching), as expected from the route of administration. Skin reactions were uncommon (urticaria after 0.16% of doses). Reactions involving wheezing occurred after 0.41% of doses or approximately one-fifth of participants. Although wheezing could be taken as a sign of a more severe reaction, in all but one participant it was mild and responded to standard doses of inhaled bronchodilator drugs. A single subject self-administered epinephrine at home with good results on two occasions for wheezing after his peanut OIT doses; thus, he was withdrawn from the study. No one experienced hypotension.

There is no disease-modifying treatment for peanut allergy, meaning that families have to rely on avoidance practice and reduced QoL, leading to accidental reactions and carrying emergency medication. Peanut OIT is a promising novel treatment that shows good efficacy and an acceptable side effect profile. As this is the first study of its type, the findings are relevant to the population studied, but safety and efficacy will require confirmation using other patient subgroups and phase 3 trials. Because of the significant risks involved, OIT should be restricted to specialist centres.

Conclusions

We performed the first study with a design and size appropriate to derive an accurate estimate of the effect size of the treatment. We have demonstrated good efficacy and safety for peanut OIT, with a large effect size. Importantly, we studied a representative population of peanut allergic children, without excluding children with a history of severe reactions. The implication is that this treatment will be suitable for children with any severity of peanut allergy.

Peanut allergy is a highly homogeneous disease, with well-validated diagnostic tests and minimal variation in phenotypic characteristics. Future phase 2 and 3 confirmatory trials are desirable, including other doses but, given the highly homogeneous nature of the illness and the strong effect size observed in this study, such trials are unlikely to require large numbers of patients. This cohort of patients, followed to 6 months, requires longer-term follow-up to determine the long-term adverse event profile and frequency of administration required to continue a state of desensitisation. Indicators of tolerance to peanut should be studied.

Contributions of authors

Katherine Anagnostou performed immunotherapy challenges and interpretation, and drafted, appraised and approved the submitted version.

Sabita Islam performed blinding of challenges and all laboratory work; prepared immunotherapy doses, and drafted, appraised and approved the submitted version.

Yvonne King performed recruitment, immunotherapy challenges and interpretation, and drafted, appraised and approved the submitted version.

Loraine Foley prepared immunotherapy doses, and drafted, appraised and approved the submitted version.

Laura Pasea designed and performed statistical analysis, and drafted, appraised and approved the submitted version.

Chris Palmer determined trial design, and drafted, appraised and approved the submitted version.

Simon Bond was responsible for oversight and design of statistical analysis; and drafted, appraised and approved the submitted version.

Pamela Ewan contributed to study design, and drafted, appraised and approved the submitted version.

Andrew Clark conceived the study, defined the primary objective, performed immunotherapy challenges, and drafted, appraised and approved the submitted version.

Disclaimers

This report presents independent research. The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, MRC, NETSCC, the EME programme or the Department of Health. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the EME programme or the Department of Health.