Epicutaneous immunotherapy for the treatment of peanut allergy in children and young adults

Study design and participant selection

This multicenter, randomized, double-blind, placebo-controlled, phase II study compared 2 doses of Viaskin Peanut versus placebo in children and young adults with peanut allergy. The primary end point was the proportion of participants with a successful outcome after 52 weeks of blinded treatment, with treatment success defined as either passing a double-blind, placebo-controlled oral food challenge (OFC) with 5044 mg of peanut protein at week 52 or by a 10-fold or greater increase in the successfully consumed dose (SCD) of peanut protein compared with the baseline OFC. Secondary end points included comparison of the 100- and 250-μg Viaskin Peanut doses, safety, and modulation of immune parameters.

Inclusion criteria included the following: (1) 4 to 25 years of age, (2) physician-diagnosed peanut allergy or a convincing clinical history of peanut allergy, (3) positive skin prick test (SPT) response to peanut (wheal size ≥3 mm greater than that elicited by the saline control) or peanut-specific IgE level of greater than 0.35 kilounits of antibody (kU_A)/L, and (4) positive baseline OFC result to a cumulative dose of 1044 mg or less peanut protein. Subjects with a history of severe anaphylaxis (previous hypotension, neurologic compromise, or mechanical ventilation) to peanut were excluded. See Table E1 in this article’s Online Repository at www.jacionline.org for detailed inclusion/exclusion criteria.

Enrollment and randomization

Participants were randomly assigned to double-blind peanut EPIT by using Viaskin Peanut 100 μg (VP100), Viaskin Peanut 250 μg (VP250), or placebo (1:1:1) at each of 5 clinical Consortium of Food Allergy Research (CoFAR) sites (75 total participants). The study was blinded through 52 weeks (Fig 1). Enrollment and randomization of younger participants (ages 4-<6 years) was paused after the first 10 participants were enrolled for a predetermined interim Data Safety Monitoring Board (DSMB) safety review after 35 days of dosing.

Study product

The Viaskin Peanut patch used for this study is comprised of an epicutaneous delivery system containing a dry deposit of a formulation of peanut protein extract manufactured by DBV Technologies SA. The peanut extract is an unmodified lyophilized product derived from the extraction and freeze-drying of defatted peanut flour made from raw peanuts. A liquid formulation of the extract is then deposited on the backing of an occlusive chamber by using electrospraying. The Viaskin patch has a diameter of 26 mm, with an inner diameter of 18 mm containing the peanut protein. The matching Viaskin placebo is the same device devoid of any peanut protein but containing excipients included in the active patch.

EPIT dosing protocol

The Viaskin patch, plus optional Tegaderm covering, was placed on the upper arm (age >11 years) or the interscapular space (4–11 years) in a clockwise rotation by using 1 of 6 application sites at 24-hour intervals. Graduated dosing was performed with the same strength patch by increasing the time worn as follows: week 1, 3 h/d; week 2, 6 h/d; and week 3, 12 h/d. This was followed by patch application for 24 h/d beginning on day 22.

Participants were monitored in the research unit on days 1 and 2 for adverse reactions. If significant local reactions (ie, grade 3 or grade 4 skin reactions; see Table E2 in this article’s Online Repository at www.jacionline.org for grading criteria) occurred, participants were instructed to remove the patch immediately and contact the study team for further instructions regarding subsequent patch application. For persistent patch-site reactions, the patch was removed, and the participant was instructed to apply the patch for the length of time that it was tolerated for the following 3 days, followed by an increase in duration of patch application every 3 to 4 days until tolerated for a 24-hour period.

Usual medications, including topical corticosteroids and calcineurin inhibitors, were continued but not within 1 inch of the patch site. Oral and topical antihistamines and topical 1% hydrocortisone were approved for the treatment of patch-site reactions, with more potent topical steroids reserved for limited use with more bothersome reactions.

Adherence and safety assessments

Participants were contacted by telephone monthly and returned to the research unit at the start of weeks 2 to 4 and at completion of weeks 12, 24, 36, and 52 to review tolerability of the study drug, adherence, and any adverse events.

Adherence to daily dosing was assessed by using 2 methods. Participants maintained daily diary logs, recording the date and time of patch application and removal during the first 6 months of therapy. Thereafter, dosing logs were only used to record missed doses, doses removed prematurely, or doses associated with adverse symptoms. Dosing logs were reviewed by study personnel at each visit. Participants were also instructed to return all used and unused patches at each visit.

Participants were also monitored for patch-site reactions during scheduled visits and as needed if symptoms were reported. Skin changes at the patch site were scored as grade 0 to 4 by using a standardized scoring system (see Table E2). Symptoms extending outside of the patch site or involving systemic reactions were recorded, and the severity of allergic reactions was reported by using a customized grading system (see Table E3 in this article’s Online Repository at www.jacionline.org).

Predetermined rules for potential discontinuation of dosing included occurrence of systemic reactions during any stage of dosing, occurrence of any grade 4 patch-site reaction, more than 3 episodes of grade 3 patch-site reactions, or 2 or more consecutive grade 3 patch-site reactions. Adverse events, serious adverse events, and accidental exposures to peanut were reported throughout the study.

OFCs

At study entry, an OFC was conducted to a cumulative amount of 1044 mg of peanut protein administered in doses every 15 minutes by using a modified PRACTALL Protocol.²⁹ The OFC was repeated at week 52 to a cumulative dose of 5044 mg of peanut protein (see the Methods section in this article’s Online Repository at www.jacionline.org).

SPTs

SPTs using the GREERPick device with peanut extract (Greer Laboratories, Lenoir, NC) and saline and histamine controls were performed at enrollment and 24 and 52 weeks after study entry, as previously described.¹⁰

In vitro assays

Mechanistic studies were conducted to assess the immunomodulatory effect of peanut EPIT by using serial testing of a variety of immune parameters. Serum peanut-specific IgE and IgG₄ levels were measured by using the ImmunoCAP 250 (Thermo Fisher Scientific, Waltham, Mass). Basophil activation was measured based on CD63 upregulation by using flow cytometry in response to peanut extract stimulation of whole blood.¹⁰ Peanut-specific T-cell activation and phenotype were assessed by using flow cytometry with CD154 as an activation marker and intracellular staining for IL-4, IL-13, IFN-γ, and IL-10 (see the Methods section in this article’s Online Repository).

Ethics

Institutional review boards at each clinical site approved the protocol and consent forms. The study was conducted under a US Food and Drug Administration investigational new drug application and monitored by the National Institute of Allergy and Infectious Diseases DSMB. Written informed consent was obtained from parents/guardians, with assent of those more than 7 years of age.

Statistical analysis

The target sample size of 75 participants (randomized 1:1:1 and stratified by site) was selected to provide 95% power, assuming a 5% success rate for the primary end point in the placebo arm compared with 50% in each of the active arms. Power was determined by using a 1-sided exact unconditional binomial test (Barnard) with an α value of .0125 for each comparison of active to placebo treatment. Alternate success definitions were also compared between the active and placebo arms by using the Barnard test. Continuous variables were contrasted between treatment groups by using the Kruskal-Wallis test, followed by Wilcoxon rank sum tests for pairwise group comparisons. Safety data were contrasted between treatment groups by using the percentage of doses per participant and performed by using the Kruskal-Wallis test, followed by Wilcoxon rank sum tests for pairwise group comparisons.

Immunologic, activated basophil, and T-cell studies were contrasted between treatment groups over time by using repeated-measures models, accounting for within-participant correlation by using a Toeplitz covariance structure. Log₁₀ transformations were applied as needed.

Prespecified exploratory analyses were performed to assess the effect of age on treatment effect by using logistic regression models for binary outcomes and Spearman correlations and linear regression models for continuous outcomes. The primary end point (VP250 vs placebo and VP100 vs placebo) was assessed at the .0125 significance level, mechanistic analyses were assessed at the .01 significance level to control for the multiplicity of analyses, and all other exploratory analyses were assessed at the .05 level. Primary end point P values were computed with StatXact (version 10; Cytel, Cambridge, Mass). All other analyses were performed with SAS (version 9.3 or higher; SAS Institute, Cary, NC).

Study population

The CONSORT diagram is represented in Fig 1: 169 participants were screened, 84 had a baseline OFC, 75 were randomized, and 74 received study treatment, with 1 participant withdrawing after randomization but before treatment initiation. The analysis population consists of 74 treated participants (25 in the placebo group, 24 in the VP100 group, and 25 in the VP250 group). As shown in Table I, the majority of participants were male (62.2%), and the median age was 8.2 years (range, 4–20 years). There were no significant differences in baseline demographic characteristics, comorbid atopic diseases, or immunologic measurements across treatment groups. The median baseline peanut SPT response was 12.8 mm, the median peanut IgE level was 78.2 kU_A/L, and the median SCD was 44 mg of peanut protein.

Three placebo-treated participants withdrew/discontinued dosing (2 because of anxiety before the week 52 OFC and 1 because of noncompliance), as did 3 participants from the VP100 group (1 because of grade 3/4 patch reactions, 1 because of non–study-related syncopal episodes, and 1 because of non–study-related illness). All of these participants were considered failures for the primary end point.

Efficacy of peanut EPIT

Table II presents results for the primary end point. For the placebo group, 3 (12.0%) participants met the primary end point compared with 11 (45.8%) for the VP100 group and 12 (48.0%) for the VP250 group. Only 1 participant (placebo) passed the week 52 OFC. Comparison of the treatment groups revealed significant differences between the placebo-treated participants and both active treatment arms (P =.005 and P =.003, respectively), with no difference between the VP100 and VP250 groups (P =.48).

Post hoc analyses were undertaken to assess 2 additional efficacy end points (Table II). First, we compared the proportion of participants in each group who had an SCD of at least 1044 mg of protein at the week 52 OFC, which was achieved in 3 (12.0%) placebo-treated participants, 3 (12.5%) VP100-treated participants, and 7 (28.0%) VP250-treated participants (P = not significant for all comparisons). Second, we compared the number of participants who had an SCD of at least 1044 mg of protein plus at least a 10-fold increase in SCD at the week 52 OFC, revealing that only 2 (8.0%) placebo-treated participants, 2 (8.3%) VP100-treated participants, and 4 (16.0%) VP250-treated participants met this stricter definition of success (P = not significant for all comparisons).

Table III shows the SCD for the week 52 OFC, as well as the change in SCD from baseline (Fig 2, A). The placebo group had a median change in SCD of 0 mg of protein (interquartile range [IQR], −40.0 to 1.0) compared with median changes of 43 mg of protein (IQR, 0.0 to 140) in the VP100 group and 130 mg of protein (IQR, 30 to 600) in the VP250 group. Median change in SCD was significantly different among the 3 treatment groups (P = 0.003, Kruskal-Wallis test), as well as between the placebo and VP100 and VP250 groups (placebo vs VP100, P =.014; placebo vs VP250, P =.003; VP100 vs VP250, P =.41).

As a preplanned exploratory analysis, we assessed the potential effects of age on outcomes (Fig 2, B, and Table IV and see Table E4 in this article’s Online Repository at www.jacionline.org). We fit a model with the primary end point as the outcome with age as a continuous variable and with age as a dichotomous variable when comparing participants 11 years or younger with those older than 11 years. Both approaches revealed a statistically significant age-by-treatment interaction, with a successful outcome being more common in younger participants (P = .03, dichotomous analysis; P = .006, continuous model). In the subgroup of participants 11 years or younger, treatment success was achieved in 1 (6%) placebo-treated child, 10 (59%) VP100-treated children, and 11 (61%) VP250-treated children (P = .0006 and P = .0003, respectively, compared with placebo; VP100 vs VP250, P = .98).

Logistic regression analysis was performed to determine whether any additional baseline factors other than age predicted treatment success (see Table E5 in this article’s Online Repository at www.jacionline.org). Only an SCD of less than 44 mg at baseline was statistically associated with a successful outcome (P = .0001). This association might only reflect that a lower baseline SCD results in easier attainment of the primary end point; baseline SCD was not significantly correlated with change in SCD from baseline to week 52. Notably, the presence or severity of atopic dermatitis at baseline was not predictive of treatment response.

Safety and adherence

Table V presents dosing symptoms by dose, participant, and percentage of doses per participant for each treatment. Overall, 14.4% of placebo doses resulted in a reaction compared with 79.8% of VP100 and VP250 doses. The majority of reactions were mild and limited to the patch site. Grade 2 or greater patch-site reactions occurred with 1.6% of placebo doses (no grade 3 or 4 reactions) compared with 18.7% of VP100 doses and 23.4% of VP250 doses. One grade 4 patch-site reaction occurred with the VP100 dose in a 12-year-old participant 34 days after enrollment. Reactions extending past the patch site occurred with 1.5% of placebo doses, 8.9% of VP100 doses, and 16.2% of VP250 doses.

Non–patch-site reactions were uncommon, reported in 0.2% of placebo and VP100 doses and 0.1% of VP250 doses. One participant in the VP100 dose group experienced systemic hives that lasted 2 to 4 hours and responded to oral antihistamines. The most commonly reported treatment was topical corticosteroids, followed by oral antihistamines. No epinephrine was used for the treatment of dosing symptoms.

The median percentage of doses per participant with a patch-site reaction was 1.6% for placebo-treated participants compared with 92.8% for VP100-treated participants and 96.1% for VP250-treated participants, whereas for non–patch-site reactions, the median was 0% doses per participant for all groups. The median percentage of doses per participant with a treated reaction was 0% for the placebo group compared with 8.9% for the VP100 group and 16.2% for the VP250 group.

Significant differences were observed for any dosing reaction, patch-site reactions, duration of reaction, doses requiring treatment, and severity of the patch-site reaction. Pairwise group comparisons identified all of the above as being lower in the placebo group compared with the VP100 and VP250 treatment groups. No statistically significant differences were observed between the VP100 and VP250 groups (see Table E6 in this article’s Online Repository at www.jacionline.org). Three unrelated severe adverse events were observed during the study: syncopal episodes, abdominal pain, and migraine headache.

Reported compliance with treatment was overall excellent. A total of 26,372 doses were expected, with 25,611 (97.1%) administered: 97.0% in the 4- to 11-year-olds and 97.4% in those older than 11 years.

Immunologic outcomes

Fig 3 shows immunoglobulin results by treatment at baseline and weeks 12, 24, and 52. When assessing global treatment effects over time, significant differences were observed between treatment groups for log₁₀ peanut IgG₄ levels (P < .0001) and log₁₀ peanut IgG₄/peanut IgE ratios (P < .0001). In particular, participants receiving active treatment had increases in both peanut IgG₄ levels (Fig 3, B) and IgG₄/IgE ratios(Fig 3, C) when compared with those receiving placebo. No differences over time between treatments were seen for log₁₀ peanut IgE levels (P = .37), total IgE levels (P = .54), or percentage of peanut IgE (P = .23).

Fig 4 shows peanut SPT results by treatment at baseline and weeks 24 and 52. A significant difference was not observed between treatment groups (P = .17). However, when the change in SPT response was examined from baseline to week 52, an apparent dose effect was noted, with a reduction in SPT size in the VP250 group (median, −2.5 [IQR, −7.5 to 0.5]; P = .02) but not within the VP100 (median, −3.25 [IQR, −7.0 to 3.0]; P = .07) or placebo (median, −2.0 [IQR, −5.0 to 1.5]; P =.27) groups.

When assessing global treatment effects on peanut-induced basophil activation, significant differences were observed at a stimulant dose of 0.01 μg (P < .0001) but not at higher doses. These data are consistent with a shift in threshold of reactivity to peanut rather than a loss of reactivity to peanut. This effect at a dose of 0.01 μg was evident beginning at 12 weeks for both the VP100 and VP250 treatment groups (see Fig E1 in this article’s Online Repository at www.jacionline.org).

T-cell studies are summarized in Table E7 in this article’s Online Repository at www.jacionline.org. At baseline, 50% and 42% of peanut-responsive CD154⁺CD4⁺ T cells were positive for IL-4 and IL-13, respectively, compared with 3% positive for IFN-γ and 4% positive for IL-10. Statistical analysis for these studies applied a more stringent P value of .01 because of the number of tests performed. No T-cell results reached this level of significance, but a global treatment effect over time on IL-4– and IL-13–producing cells trended toward significance (P = .059 and P = .040 for IL-4 and IL-13, respectively). Median frequencies of IL-4– and IL-13–producing T cells were lower compared with those in placebo-treated subjects at the VP250 dose but not at the VP100 dose.

Finally, data were analyzed to assess for relationships between baseline age and mechanistic outcomes at week 52. Independent of treatment group, lower age at baseline was correlated with an increasing peanut IgG₄/IgE ratio (rho = −0.31, P = .010), as well as with larger decreases from baseline in percentages of CD63⁺ cells for stimulant levels of 0.1 μg and 0.01 μg (rho = 0.33 and 0.31, respectively; P ≤ .01). Within groups, for VP100 participants, lower age at baseline correlated with higher week 52 peanut IgG₄ levels (rho = −0.57, P = .005) and greater change from baseline to week 52 in peanut IgG₄/IgE ratios (rho = −0.56, P = .007). Correlations between baseline age and other mechanistic factors at week 52 were not significant for the other treatment groups.