Peanut immunotherapy raises the dose children tolerate — and their anaphylaxis risk

Peanut oral immunotherapy reliably increases the amount of peanut a child can eat before reacting, and it reliably increases how often that child has an allergic reaction while on treatment. Both findings carry high-certainty evidence in the meta-analysis that assessed them, and they are not in conflict — they are answers to two different questions [s2]. Which question matters more is not something the trials settle.

The trial that got the treatment approved

PALISADE was a phase 3 trial of AR101, a peanut-derived oral immunotherapy product [s1]. Participants aged 4 to 55 were screened for dose-limiting symptoms at a challenge dose of 100 mg or less of peanut protein — roughly one third of a peanut kernel — in a double-blind, placebo-controlled food challenge [s1]. Those who reacted were randomised 3:1 to AR101 or placebo on an escalating-dose schedule, and those who completed the regimen, meaning about 24 weeks at 300 mg per day of maintenance, took a second double-blind food challenge on exit [s1].

Of 551 participants who received AR101 or placebo, 496 were aged 4 to 17 [s1]. In that age group, 250 of 372 on active treatment (67.2%) could ingest 600 mg or more of peanut protein without dose-limiting symptoms at the exit challenge, against 5 of 124 on placebo (4.0%) — a difference of 63.2 percentage points, 95% confidence interval 53.0 to 73.3, P<0.001 [s1]. Symptom severity at the exit challenge was lower on treatment too: maximum severity was moderate in 25% of the active group against 59% of placebo, and severe in 5% against 11% [s1].

The trial reports one more result that is easy to skip past: efficacy was not shown in participants aged 18 or older [s1].

The safety numbers inside the same trial

Adverse events during the intervention period affected more than 95% of participants aged 4 to 17 [s1]. In the active-drug group, 34.7% had mild events, 59.7% moderate and 4.3% severe; in the placebo group, 50.0%, 44.4% and 0.8% [s1]. In other words the balance of event severity shifted upward on treatment, which is the expected consequence of deliberately feeding an allergen to an allergic child every day.

The meta-analysis that put a number on it

The PACE systematic review searched eight databases plus the FDA and EMA to December 2018 for randomised trials comparing peanut oral immunotherapy against avoidance or placebo, and found 12 trials with 1,041 participants, median age across trials 8.7 years [s2].

Pooled, oral immunotherapy increased anaphylaxis risk (risk ratio 3.12, 95% CI 1.76-5.55, risk difference 15.1%, high certainty), anaphylaxis frequency (incidence rate ratio 2.72, 1.57-4.72, risk difference 12.2%, high certainty) and epinephrine use (RR 2.21, 1.27-3.83, risk difference 4.5%, high certainty) [s2]. The increase was similar during build-up and maintenance, with an interaction P of 0.92 — meaning the risk does not resolve once the dose stops climbing [s2]. Serious adverse events rose (RR 1.92, 1.00-3.66, moderate certainty), as did vomiting (1.79, 1.35-2.38) and angioedema (2.25, 1.13-4.47), both high certainty [s2].

On the efficacy side the review found the same thing PALISADE did, larger: passing a supervised challenge was far more likely with oral immunotherapy (RR 12.42, 6.82-22.61, risk difference 36.5%, high certainty) [s2]. And on the outcome patients tend to care most about, quality of life, there was no difference between groups (RR 1.21, 0.87-1.69, low certainty) [s2].

The review's own interpretation is that available regimens "considerably increase allergic and anaphylactic reactions over avoidance or placebo, despite effectively inducing desensitisation" [s2].

Why the two conclusions are both correct

Passing a supervised challenge is a surrogate endpoint. PACE says so explicitly: it is a stand-in for preventing out-of-clinic reactions, not a measurement of them [s2]. The trials were not long enough or large enough to measure whether the higher tolerated dose translates into fewer accidental reactions in real life, and the quality-of-life data, such as it is, did not show a benefit [s2].

So the case for treatment rests on the assumption that raising the reaction threshold reduces the consequences of accidental exposure, and the case against rests on the measured fact that treatment causes reactions in the meantime. Both are defensible readings of the same dataset, which is why guidance in this area is stratified and conditional rather than uniform.

The base rate this is set against

The size of the risk being managed is worth stating, because the intuitive estimate is usually far too high. A systematic review pooling 13 studies, covering 240 fatal food anaphylaxis episodes over an estimated 165 million food-allergic person-years, put the incidence of fatal food anaphylaxis in people with food allergy at 1.81 per million person-years (95% CI 0.94-3.45) [s3]. In people aged 0 to 19 the rate was 3.25 per million person-years (1.73-6.10) [s3]. Study quality was mixed and heterogeneity between studies was high, so these are estimates with wide margins [s3]. The authors' summary comparison: fatal food anaphylaxis for a food-allergic person is rarer than accidental death in the general European population [s3].

That base rate does not make the decision. It changes what the decision is about — not the prevention of a probable death, but the management of an improbable one, against a treatment whose harms are frequent and measured.

What is not known

None of these trials followed participants long enough to say whether desensitisation persists after treatment stops, or whether it prevents real-world reactions [s2]. None showed efficacy in adults [s1]. And the meta-analysis's authors close by calling for safer approaches and for trials that measure outcomes patients care about rather than challenge thresholds [s2] — an unusually direct admission that the field has been measuring the wrong thing.

Oral immunotherapy is a supervised medical treatment with a defined dosing schedule and a documented rate of anaphylaxis. Nothing in it is attemptable at home, and decisions about it belong with an allergy specialist.

Sources

• AR101 Oral Immunotherapy for Peanut Allergy — New England Journal of Medicine, 2018-11-18
• Oral immunotherapy for peanut allergy (PACE): a systematic review and meta-analysis of efficacy and safety — The Lancet, 2019-04-25
• Incidence of fatal food anaphylaxis in people with food allergy: a systematic review and meta-analysis — Clinical and Experimental Allergy, 2013-10-12