Anaphylaxis: assessment and referral after emergency treatment

New evidence

Section 7 of the RCUK's guideline recommends that:

'Mast cell tryptase should be measured in all patients with suspected anaphylaxis where the diagnosis is uncertain.'

With regards to sample timing, it recommends:

'a) Minimum: one sample, ideally within 2 h (when peak tryptase levels generally occur) and no later than 4 h after onset of symptoms.

b) Ideally: take three timed samples:

1) An initial sample as soon as feasible – but do not delay treatment to take sample.

2) A second sample 1 – 2 h (but no later than 4 h) after onset of symptoms.

3) A third sample at least 24 h after complete resolution, or in convalescence.'

These recommendations are largely consistent with NICE's guideline on anaphylaxis which recommends:

'Record the time of onset of the reaction.

After a suspected anaphylactic reaction in adults or young people aged 16 years or older, take timed blood samples for mast cell tryptase testing as follows:

NICE's guideline on anaphylaxis makes the same recommendations for children aged less than 16 years but they are consider recommendations, due to extrapolation from studies of adult populations about mast cell tryptase utility in diagnosing anaphylaxis.

NICE's guideline differs from the RCUK's guideline because it is less directive and explicit.

It recommends:

'Inform the person (or, as appropriate, their parent and/or carer) that a blood sample may be required at follow-up with the specialist allergy service to measure baseline mast cell tryptase.'

Impact assessment

The RCUK recommend a third sample at 24-h 'because it provides a baseline tryptase value (as)…some individuals have an elevated baseline level and may be at greater risk of anaphylaxis in response to some triggers.' This is therefore a risk management recommendation that acts to identify people who may be at higher risk of a subsequent anaphylactic episode or who may have raised levels due to another condition.

The NICE guideline recommendations about the timing of blood samples are based on 13 studies that investigated mast cell tryptase's utility to confirm anaphylaxis in adults, no studies in children were identified. No information on the optimal timing of baseline measurement of mast cell tryptase was identified. Very low quality evidence from 7 observational studies (n=178 patients) reported that the timing of peak levels ranged from 1 minute to 6 hours (median 30 minutes).

The committee noted that 'some patients had unexplained high levels of mast cell tryptase (for example, in mastocytosis), and therefore in order to interpret the results correctly it was important for a baseline sample to be taken. To aid the interpretation of the results it was agreed that this baseline sample would need to be taken at least 24 hours after the onset of symptoms, probably during specialist follow-up (full guideline pages 25 to 26)' This is the origin of the recommendation, which accommodates a third sample after at least 24 hours, although it is far less specific and directive than the RCUK's recommendation.

The RCUK's recommendation is based on that from the World allergy organization anaphylaxis guidance 2020 which states: 'It is recommended to evaluate baseline serum tryptase at least 24 h after resolution of anaphylaxis symptoms, even when tryptase concentration during episode remains within normal range'. This is a consensus recommendation that is taken from another consensus recommendation from the 2010 Mast cell activation working conference (Valent et al. 2011). This attempted to define mast cell activation (MCA) using 'accepted, objective, easily measurable, and commonly applicable parameters and criteria.' This is a goal predicated on knowing a person's baseline mast cell tryptase levels. The report states that 'the following criteria were regarded as indicative of systemic MCA: (1) typical clinical signs and symptoms, (2) substantial and transient increase in an MC-derived mediator in biological fluids…during or shortly after the acute event compared to a baseline level recorded either before the acute event or at least 24 h after all clinical signs and symptoms of the event have completely resolved.'

NICE's recommendation, accommodates taking a third post-anaphylaxis blood sample in a specialist allergy service. Feedback from a clinical colleague at NICE with experience in this area suggests this remains standard NHS practice. Considering this and that the evidence underpinning the RCUK's guideline is based on consensus, we do not propose to update the NICE guideline recommendations about blood sampling at this time.

April 2026: update to the timing and frequency of blood sampling for measuring mast cell tryptase

New evidence

To mitigate BRs theRCUK's Emergency treatment of anaphylaxis guideline (page 43) recommends observation periods stratified by risk of 2 hours, at least 6 hours and at least 12 hours. The latter duration is longer than the NICE guideline currently recommends, and NICE's guideline does not explicitly stratify by defined risk criteria other than suggesting that the response to treatment may indicate risk of BR.

NICE's guideline recommends that 'Adults and young people aged 16 years or older who have had emergency treatment for suspected anaphylaxis should be observed for 6 to 12 hours from the onset of symptoms, depending on their response to emergency treatment. In people with reactions that are controlled promptly and easily, a shorter observation period may be considered provided that they receive appropriate post-reaction care prior to discharge.'

NICE's guideline also recommends that 'Children younger than 16 years who have had emergency treatment for suspected anaphylaxis should be admitted to hospital under the care of a paediatric medical team.'

NICE's guideline recommendations about post-anaphylaxis observation periods are based on expert opinion as no studies were identified investigating an optimal observation period (full guideline page 34 section 3.2.3). As a result, the committee made a recommendation for research on length of observation period following emergency treatment for anaphylaxis. This recommends an RCT that compares differing observation durations.

The RCUK guideline recommends 12 hours or more in the following circumstances:

This is a 'weak' recommendation because it is based on what the RCUK assessed as very low certainty evidence. The new evidence for stratification of observation by risk identified by the RCUK that is in scope for NICE's guideline, includes 3 studies: Kraft et al. 2020, Lee et al. 2015, and Kim et al. 2019. The RCUK guideline's Evidence to Decision Framework (page 29) describes the rationale for this stratified approach (this is a supplement to Dodd et al. 2021).

Lee et al. 2015 is a meta-analysis of 27 observational studies (n=4,112, 192 BRs, including 10 studies of paediatric patients n=1,473, 54 BRs). The study reports that the median onset of a BR is 11 hours with a range of 0.2 to 11.0 hours. The ages of the participants included are not described in the abstract. The study reports that food as the allergen lowers the risk of BR (odds ratio [OR] 0.62, 95% confidence interval [CI], 0.4 to 0.94). Unknown triggers were associated with an increased risk of a BR (pooled OR 1.72, 95% CI, 1.0 to 2.95). Initial presentation with hypotension was also associated with the development of a BR (pooled OR 2.18, 95% CI, 1.14 to 4.15). Results for children are not reported separately. The study is included in the 2016 surveillance of NICE's guideline, which assessed it as not being enough on its own to warrant updating recommendations. The surveillance review reported that topic experts (TEs) said that new evidence was available about post-anaphylaxis observation periods but that they provided no references.

Kraft et al. 2020 is an analysis of registry data from n=9,171 cases from 11 countries, 10 European, not including the UK, and includes 2,187 children (<17 years). It reports a BR rate of 4.7% (n=435) with 60.5% (n=225) occurring within 12 hours after initial anaphylaxis; 23.9% (n=89) between 12- and 24-hours; and 15.6% (n=58) occurring after 24-hours. The rate of biphasic reactions did not significantly differ between children (5.1% [4.3% to 6%]) and adults (4.6% [4.1% to 5.1%]). The study reports that the following factors increased the risk of BR: reaction severity (grade III/IV versus grade II (OR 1.34; 95% CI: 1.1 to 1.62); multiorgan involvement; skin, gastrointestinal, severe respiratory, and cardiac symptoms; anaphylaxis caused by peanut/tree nut (OR 1.78; 95% CI: 1.38 to 2.23) or an unknown elicitor (OR 1.96; 95% CI: 1.41 to 2.72); and exercise as a cofactor (OR 1.44; 95% CI: 1.17 to 1.78).

Kim et al. 2019 is a meta-analysis of 12 studies (n=2,890 anaphylactic episodes, with 143 BRs) including adults and people of 'unidentifiable' age. It reports that 1-hour of observation achieved a 95% negative predictive value (NPV) and 6-hour or more, 97%. The NPV increased with longer observation periods with the trend of increasing NPV slowing after 6-hours. Incidence at 1- and 4-hours was 0.45 (95% CI 0.20 to 1.04) and 0.41 (95% CI 0.19 to 0.87) per 100 person-hours. After 8-hours an NPV > 98% was reported (incidence <0.10/100 person-hours). The authors conclude that although longer observation duration identified more people, a decreasing incidence is observed and a 6 to 12-hours observation window is probably practical.

Evidence provided by topic experts

A TE highlighted Dribin et al. 2025, a retrospective cohort study of children (n=5,641, median age 7.9 years, IQR 3.3 to 13.1) which investigated the incidence and timing of repeat adrenaline dosing based on initial reaction severity. The primary outcome was time from first dose to last dose of intramuscular or intravenous adrenaline administered before or during the emergency department stay. The study aimed to estimate the time threshold when there was <2% increase in the cumulative incidence in repeat adrenaline dosing as the observation time increased by 1-hour. The authors comment that <2% was 'a clinically acceptable risk' of receiving adrenaline after discharge.

This study suggests only a small proportion of children required repeat adrenaline doses, and that most received their last dose within a 2-hour window except for those with anaphylaxis with cardiovascular involvement. Risk factors for repeat adrenaline were severe respiratory (OR 1.47, 95% CI 1.16 to 1.84) or cardiovascular involvement (OR 3.18, 95% CI 2.35 to 4.17).

Following discharge 1.5% (n=83) had an emergency department (ED) revisit within 72-hours related to the original episode of which 33 of 83 received adrenaline. Of 2,562 patients who received pre-ED adrenaline 84.2% did not receive repeat adrenaline in the ED.

Some limitations of this study are that it uses final dose as a proxy for stability rather than the timing and rates of BR; it was conducted in the US healthcare system so it's relevance to a UK population may be questionable; and it excludes people transferred from other healthcare providers and those with MCA disorders. The latter criteria may both have led to the exclusion of more severe anaphylaxis cases that are more likely to require longer observation periods.

Further searches (Elicit, PubMed, Google Scholar) identified Short et al. (2024) a retrospective cohort study of children (n=292, mean age 8.8 years SD +/-5.7) investigating the frequency and timing of BRs. Participants were observed for an average of 233.1 minutes in line with NIAID (United States National Institute of Allergy and Infectious Diseases) recommendations of 4 to 6 hours. N=10 experienced a BR, 6 within an average of 96.3 minutes after initial symptom resolution and 4 between 10 and 33 hours after resolution. It reports no significant difference in the number of symptoms at anaphylaxis onset (cardiac, respiratory, GI, mucosal) or time to first adrenaline dose between groups. This suggests it is very difficult to definitively predict those at risk of a delayed BR, although n=10 is a very small number of events from which to make generalisations.

Impact assessment