Feasibility of sustained response through long-term dosing in food allergy immunotherapy

This LTFU study received ethical clearance from Stanford IRB, and was open for enrollment to all the participants, who had successfully completed a phase 1 study under the Investigational New Drug (IND) (Trial registration: Clinicaltrial.gov NCT01490177 [11]) in our clinic. The details on the original phase 1 study protocol and results have been published on a subset of peanut-allergic individuals [11]. Furthermore, Quality of Life studies have been published for this study [20] and a similar OIT trial with 16 weeks of adjunct omalizumab [21, 22]. After achieving successful maintenance dosing in the initial phase 1 study (i.e. on reaching the maintenance dose of 2 g for the first food in their OIT in the original phase 1 trial), the participant was enrolled in the LTFU study. At that point, the clinical team performed skin tests, blood tests and reviewed the participant’s past research data. As long as there were no safety issues (i.e. no moderate to severe allergic reactions in the last 6 months), and the skin tests and blood specific IgE tests showed decreases, the clinical team, together with the patient and family, made a team decision to allow the participant to either continue to ingest the “high” maintenance dose (median 2 g protein of each food allergen in their OIT) or decrease to the “low” (median 300 mg of each food allergen in their OIT) maintenance dose after completion of the phase 1 trial. In summary, the long-term maintenance dose was chosen as a team approach between the participant and the clinical team rather than by personal preference. The major potential source of bias in this observational, not a prospective, randomized, controlled study was -low vs high LTFU dose. Reasons for decreasing doses to a low dose of each food allergen included ease of dosing, improved compliance, and dietary preference. No LTFU maintenance dose decrease was made for safety reasons or adverse symptoms. LTFU visits were conducted every 6–12 months up to month 98 after the start of OIT, which corresponds to a maximum of 72 months after reaching the 2 g maintenance dose for the first food in the original trial (median: 48.4 months, interquartile range 11.4 months) until May 2017. After completion of the original phase 1 trial, the participants switched from food flours/powders to food equivalents based on exact protein amounts for each specific food allergen. Some participants also took their maintenance dose every other day instead of every day. Adverse events were documented as per CTCAE v4.03 criteria [23] and in accordance with regulatory guidelines. An independent Data Safety and Monitoring Board (DSMB) met every 6–12 months to review participant safety data. All participants were required to carry reaction medications.

Standardized OFCs using validated and published techniques [24] were performed for each of the participant’s food allergens by a board-certified allergy and immunology specialist at each follow-up visit at the Sean N Parker Center for Allergy and Asthma Research at Stanford University, and in some cases, at the respective local allergist’s office. For each allergen, OFCs were done in a staged approach with approximately 150, 300, 600, 1200, and 2000 mg of each food allergen on separate days. Skin prick tests (SPTs) were performed as per the method in the original phase 1 trial [11].

Allergen-specific IgE and IgG₄ levels were measured at Johns Hopkins Allergy and Clinical Immunology Reference Laboratory by ImmunoCAP fluorescence enzyme immunoassay (Thermo Fisher scientific/Phadia, Waltham, MA).

The analysis was performed using the statistical language R (version 3.2.3). To evaluate the association between SPT wheal diameter and maintenance dose (low vs. high), a mixed effect linear model was fit controlling for months since maintenance was reached with random effects for patient and allergen type. A similar model was fit for the ratio of IgG₄ to IgE (log10).

Differences between Kaplan–Meier curves, depicting the timeline of participant’s continuing on a high vs. a low maintenance dose were assessed using a log-rank test (R package ‘Survival’ [25, 26]).

A total of 46 out of 48 participants from the original single-site, phase 1 OIT trial enrolled in this LTFU study (thus the enrollment rate = 95.8%). Of these 46 LTFU participants, 1 withdrew early secondary to issues related to inability to travel to the site. The characteristics of the participants at baseline are summarized in Table 1. Participants treated with OIT to 9 different food allergens (almond, cashew, egg, hazelnut, milk, peanut, pecan, sesame, walnut) were included in this study (Table 1). Most of the participants (21, 45.7%) were desensitized to only one food (i.e. peanut); while 6 participants (13%) ingested long-term maintenance doses of 5 different foods. The analysis of combination of foods administered to our multi-OIT participants suggested coexistence of allergies to egg and milk, and walnut and pecan (Additional file 1: Figure S1), which concurs with the previously published findings [27].

The long-term maintenance dose information per food per participant at the end of our follow-up study is summarized in Fig. 1.

Of the 25 participants with more than one food in their OIT, 40% (10 participants) were on a low and 40% (10 participants) were on a high long-term maintenance dose for all foods at the end of our follow-up study. Egg (1 of 9 participants on low dose, 11.1%) and milk (1 of 8 participants on low dose, 12.5%) were the foods that were ingested most often at a high dose since these allergens are common ingredients in many foods.

One participant stopped eating peanut after 25 months on a high maintenance dose. That participant however ate foods that contained small amounts (about 50 mg a day) of peanut at the conclusion of our study.

The percentage of participants, who continued the high long-term maintenance dose per allergen as opposed to low long-term maintenance dose over time is shown in Fig. 2. Six participants, who decreased their long-term maintenance dose to low, increased it back to high after varying times on the low dose. The time point of only the initial transition from the high into the low maintenance dose is shown in Fig. 2.

Adverse events related to allergic reactions were recorded for the ITT population during the LTFU and are summarized in Tables 2, 3 and Additional file 2: Table S1. In total 1207 reactions were recorded (2.29% of maintenance doses). Of those, 1073 reactions (88.9%) were mild and 129 (10.69%) were classified as moderate. There were 5 severe events (0.41% or all reactions) but these were severe based on nasal congestion and skin symptoms only. There were no fatal or serious adverse events. There were no anaphylactic reactions requiring the use of epinephrine. A key finding was that the frequency of allergic adverse events decreased over time (Tables 2, 3). Safety results did not differ based on low vs. high LTFU maintenance dose.

SPTs and assessment of plasma IgE and IgG₄ were performed for most participants at various time points over the course of LTFU.

The SPT wheal diameter for individual offending foods was significantly decreased during the active phase, and remained low over LTFU (p < 0.05). This phenomenon however was not observed with pecan (p = 0.074), almond (p = 0.71) and sesame (p = 0.13) (Fig. 3).

In concurrence with the previously published findings from the phase 1 study for a subset of peanut-allergic participants [11], a significant trend of increasing allergen-specific IgG₄/IgE ratios was observed at the end of active phase for all the food allergens. This trend persisted over LTFU (Fig. 4). The statistical significance for the serological readouts could be quantitated only for peanut (n = 20) and cashew (n = 10) (p < 0.01) owing to the limited number of participants with other foods (n < 6) for that the IgE/IgG₄ was obtained over LTFU.

The SPT wheal diameters (Fig. 3) and IgG₄/IgE ratios (Fig. 4) per participant and food allergen were analyzed for association with a low vs. a high maintenance dose in a mixed-effects linear model, including the months after reaching the maintenance dose, the food allergen and the participant as covariates. Neither SPT wheal diameter (p = 0.12) nor IgG₄/IgE ratio (p = 0.15) was significantly associated with a low vs. a high dose in the given dataset.

The outcome of a standardized, staged OFC performed by trained personnel was used to define the status of desensitization for the LTFU participants. The results of OFCs for each food allergen were analyzed to determine whether a reduced long-term maintenance dose was suitable to sustain desensitization for a given allergen. The time line of the long-term maintenance dose per participant per allergen as well as the time points and outcomes of OFCs are shown in Fig. 5. Each participant on long-term maintenance dosing was able to tolerate 2 g protein or more in a food challenge of their respective food allergens at the end of our follow-up phase, independent of high vs. low long-term maintenance dose. This finding thus demonstrates the feasibility of sustained desensitization even through a reduced long-term maintenance dose.