Anaphylaxis guidelines rest on very low quality evidence, by their authors' account

The European Academy of Allergy and Clinical Immunology's 2021 anaphylaxis guideline recommends prompt intramuscular adrenaline as first-line management and the availability of adrenaline autoinjectors to patients in the community [s1]. It also states, in its own summary, that the evidence for the management of anaphylaxis remains mostly at a very low level, and calls the need for clinical trials that could improve care urgent [s1]. Those two facts sit in the same document, and the second one is the part that never gets reported.

This article describes what the epidemiology and the guideline documents say. It is not guidance on recognising or responding to an emergency, which is a matter for emergency services and a clinician.

What the guideline is and how it was built

The EAACI task force updated its 2014 guideline using the AGREE II framework and the GRADE approach, systematically reviewing the evidence and forming recommendations by weighing benefits against harms, with external peer review and a public consultation [s1].

Its content is procedural more than pharmacological. It suggests using clinical criteria to identify anaphylaxis, with blood sampling for later measurement of tryptase [s1]. It recommends intramuscular adrenaline first-line, with autoinjectors available in the community [s1]. It calls for pharmacokinetic data to be provided for autoinjector devices — a request that reveals how little comparative device data exists [s1]. It recommends structured, comprehensive training for people at risk, suggests simulation training and visual prompts for healthcare professionals, and suggests that school policies reflect anaphylaxis guidelines [s1].

The word "suggests" rather than "recommends" is doing work throughout, and it is doing it because of the evidence rating the guideline assigns itself [s1].

Why the evidence is thin

Anaphylaxis is the hardest thing in allergy medicine to study. A randomised trial of adrenaline against placebo in anaphylaxis would be unethical, so the central recommendation cannot be tested the way a drug for a chronic condition is tested. The events are rare, sudden, and occur outside clinical settings. What exists instead is registry data, case series, pharmacokinetic studies in healthy volunteers, and mechanistic reasoning — a body of evidence that GRADE, applied honestly, rates very low [s1].

That is not an argument against the recommendation. It is a statement about what kind of confidence is available, and the guideline's authors chose to state it plainly rather than let the recommendation's strength imply an evidence base that is not there [s1].

What the national data show

The best long-run picture comes from the United Kingdom, where researchers analysed national hospital admission, death and autoinjector prescription data from 1998 to 2018 [s2]. Over that period 101,891 people were admitted to hospital for anaphylaxis, of which 30,700 admissions (30.1%) were coded as food-triggered [s2].

Food anaphylaxis admissions rose from 1.23 to 4.04 per 100,000 population per year, an annual increase of 5.7% (95% CI 5.5% to 5.9%, P<0.001) [s2]. The steepest rise was in children under 15, from 2.1 to 9.2 admissions per 100,000 per year, an annual increase of 6.6% (6.3% to 7.0%), against 5.9% in people aged 15-59 and 2.1% in those 60 and over [s2].

Deaths moved the other way. The analysis identified 152 deaths where the fatal event was probably caused by food-induced anaphylaxis, and the case fatality rate fell from 0.7% to 0.19% for confirmed fatal food anaphylaxis (rate ratio 0.931, 0.904 to 0.959, P<0.001) [s2]. Over the same period, adrenaline autoinjector prescriptions increased by 336%, about 11% a year [s2].

The study does not establish that the prescriptions caused the falling fatality rate — more admissions coded for milder reactions would lower a case fatality rate on their own, and the authors present the two trends alongside each other rather than as cause and effect [s2].

Two findings about triggers are worth recording. At least 46% of the deaths analysed were triggered by peanut or tree nut [s2]. And in school-aged children, cow's milk accounted for 17 of 66 deaths, making it the most common single cause of fatal anaphylaxis in that group — a fact that runs against the public association of fatal food allergy with nuts [s2].

How rare fatal anaphylaxis is

A systematic review pooled 13 studies describing 240 fatal food anaphylaxis episodes across an estimated 165 million food-allergic person-years [s3]. In people with food allergy, the incidence of fatal food anaphylaxis was 1.81 per million person-years (95% CI 0.94 to 3.45, range across studies 0.63 to 6.68) [s3]. At ages 0-19 it was 3.25 per million person-years (1.73 to 6.10) [s3]. Sensitivity analyses using different assumed food allergy prevalences moved the overall figure between 1.35 and 2.71 [s3].

Study quality was mixed and heterogeneity was high, which the reviewers attribute partly to variation in food allergy prevalence and in data collection [s3]. Their headline comparison is that fatal food anaphylaxis for a food-allergic person is rarer than accidental death in the general European population [s3].

The access gap

Rarity does not make preparedness pointless, and the data suggest preparedness is incomplete. In a nationally representative US survey, among adults with a convincing food allergy, 51.1% had experienced a severe reaction and 38.3% reported at least one lifetime emergency department visit related to food allergy — while only 24.0% reported a current epinephrine prescription [s4].

That is the gap the EAACI guideline's community-availability recommendation addresses [s1], and it is a distribution problem rather than an evidence problem: on that survey, 24.0% of adults with symptom-consistent food allergy held a current prescription, against 51.1% who had already had a severe reaction [s4].

What to watch

The guideline's own request is the thing to follow: pharmacokinetic data for autoinjector devices, and prioritised clinical trials with the potential to improve management [s1]. Until those exist, anaphylaxis will remain a field where the recommendations are strong, the underlying evidence is graded very low, and the honest description of the situation is the one the guideline authors wrote themselves.

Sources

• EAACI guidelines: Anaphylaxis (2021 update) — Allergy, 2021-08-03
• Food anaphylaxis in the United Kingdom: analysis of national data, 1998-2018 — BMJ, 2021-02-17
• Incidence of fatal food anaphylaxis in people with food allergy: a systematic review and meta-analysis — Clinical and Experimental Allergy, 2013-10-12
• Prevalence and Severity of Food Allergies Among US Adults — JAMA Network Open, 2019-01-04