Anaphylaxis observation periods: tradition, fear, or data?
Freddie, a 9-year-old boy, presents to the ED after accidental exposure to peanuts. He had widespread urticaria, vomiting, persistent cough, and increasing work of breathing. Mum gave his adrenaline auto‑injector at home, with good effect.
In ED, he looks settled. Observations are normal. You give an antihistamine for symptomatic relief and monitor him. He remains totally well.
Two hours later, Mum and Freddie are bored and tired. The waiting room is packed, there are no bays left, and the observed play area looks like a war zone.
Until now, guidance has been inconsistent, conservative, and based on limited data. A new large multicentre study by Dribin et al. may finally help us individualise observation periods after paediatric anaphylaxis.
Dribin, T.E., Sampson, H.A., Zhang, Y., Boyd, S., Zhang, N., Michelson, K.A., Neuman, M.I., Brousseau, D.C., Mistry, R.D., Freedman, S.B. and Aronson, P.L., 2025. Timing of repeat epinephrine to inform paediatric anaphylaxis observation periods: a retrospective cohort study. The Lancet Child & Adolescent Health, 9(7), pp.484-496.
Why does this matter?
Anaphylaxis presentations to paediatric EDs are increasing. While epinephrine works quickly and effectively for most children, fear of biphasic reactions has led to prolonged observation periods and higher admission rates.
Current guidance varies markedly:
- NICE (2021) recommends hospital admission for all children (CG134 Full guideline)
- RCUK suggests risk‑stratified observation (RCUK Anaphylaxis DRAFT.pdf)
- Many departments default to 4–6 hours or longer, just to be safe
But what are we gaining with long observation periods? Safety or just reassurance?
What did this study ask?
When does the risk of needing more epinephrine become so small that continued observation adds little value?
Rather than focusing solely on biphasic reactions, the authors used a pragmatic, clinically relevant outcome: repeat epinephrine administration.
How was the study done?
This was a large, multicentre retrospective cohort study conducted across 31 North American emergency departments, including children aged 6 months to 17 years who were treated with epinephrine for anaphylaxis between 2016 and 2019.
Population
Children aged 6 months to 17 years presenting to the ED with an acute allergic reaction or anaphylaxis treated with adrenaline/epinephrine, across 30 US EDs and one Canadian ED
Intervention
Observation after the first dose of adrenaline/epinephrine, stratified by initial reaction severity: no respiratory or cardiovascular involvement, respiratory involvement only, or cardiovascular involvement
Comparison
Different potential observation durations after the first adrenaline dose, particularly around 2 hours, 4 hours, 6 hours and 8 hours, and comparison between severity groups.
Children were categorised by initial reaction severity
– NO respiratory or cardiovascular involvement
– Respiratory involvement with no cardiovascular features
– Cardiovascular involvement
Outcome
Need for repeat adrenaline/epinephrine, measured as time from first to last dose. The study aimed to identify the observation time at which the increase in the cumulative incidence of repeat adrenaline administration was less than 2% per additional hour.
The clock started at the first epinephrine dose, and the “event” was a subsequent epinephrine dose, whether for biphasic, persistent, or refractory reactions.
The authors predefined an acceptable risk threshold:
An observation period after which the cumulative risk of repeat epinephrine increased by less than 2% per additional hour
This threshold is pragmatic, not magic – a key point the authors are very clear about.
What did they find?
Overall risk of repeat epinephrine, out of 5641 children:
- 4.7% received repeat epinephrine after 2 hours
- 1.9% after 4 hours
- 1.1% after 6 hours
- 0.8% after 8 hours
The cumulative risk curve flattened sharply after the early hours.
Observation thresholds by severity
| Group | Time when hourly risk fell below 2% |
| All children | 115 minutes |
| No respiratory or cardiovascular involvement | 105 minutes |
| Respiratory but no cardiovascular involvement | 109 minutes |
| Cardiovascular involvement | 161 minutes |
Applied to the cohort:
- 95% could have been safely discharged at 2 hours
- 98% at 4 hours
The clock starts at the time the EPIPen was given, NOT the time they booked into ED.
Cardiovascular involvement changes everything
Children with cardiovascular features (hypotension, syncope, altered mental state) had:
- Later repeat epinephrine use
- Higher rates of biphasic, persistent, and refractory anaphylaxis
In this group:
- 7.5% needed repeat epinephrine after 2 hours
- 2.8% after 4 hours
- 1.2% after 6 hours
This aligns with clinical instincts – these are the children we already “don’t quite trust”.
What about biphasic reactions?
True biphasic anaphylaxis occurred in:
- 1.5% overall
- 3.6% of those with cardiovascular involvement
Importantly:
- Biphasic reactions were rarely severe
- Fatal biphasic reactions were not observed
- Most repeat epinephrine doses occurred early
This challenges the dogma that prolonged observation is needed to “catch” rare late reactions in everyone.
Strengths of the study
One of the paper’s big strengths is its size. This was not a small, single-centre snapshot — it included a very large number of children across 31 emergency departments, making the findings feel much more relevant to day-to-day practice.
The authors also did something clinically useful: they looked at children by severity of presentation. That matters because a child with skin-only symptoms who has settled after adrenaline is not the same as a child who has had cardiovascular involvement. This severity-stratified approach makes the results easier to apply at the bedside.
The outcome they chose — whether the child needed more adrenaline — is also refreshingly practical. It is something clinicians care about, but it is also something families care about: “Is my child likely to get worse again?” or “Are they going to need more treatment?”
Finally, the authors tested their findings in many ways using sensitivity analyses. The results held up well, which makes the overall message more reassuring.
Limitations to keep in mind
The authors – rightly – urge caution.
There are some important caveats. The study uses repeat adrenaline as a marker of clinical deterioration. That is practical, but imperfect. A child may have worsened without receiving further adrenaline, or clinicians may have varied in their threshold for giving it.
The way discharge was handled also matters. Once a child left the ED, they were no longer observed in the dataset. In statistical terms, this was treated as censoring. But in real life, discharge is not random: children who look well go home, and children who look sicker tend to stay longer. This could mean the study slightly underestimates the risk of later deterioration in some groups.
The retrospective design brings the usual limitations. The researchers were working with data that had already been recorded, so they were dependent on the accuracy and completeness of clinical documentation.
Finally, these data came from large emergency departments in North America. That does not mean the findings are irrelevant elsewhere, but applying them to smaller units, rural settings, or healthcare systems with different access to follow-up and emergency care needs some clinical judgement.
Most importantly, this is not a discharge rule. It is a risk estimate to support clinical decision‑making.
What does this mean for practice?
This paper supports what many experienced clinicians already do – but now with numbers to back it up.
A reasonable, evidence‑based approach might be:
At 2 hours: No cardiovascular involvement
- Symptoms resolve completely.
- Child looks well
- Family has epinephrine and understands its use
At 4 hours: Cardiovascular involvement
- More severe respiratory features
- Clinician or parental concern
>6 hours: Ongoing symptoms
- Repeated epinephrine
- Refractory or persistent anaphylaxis
- Significant social or access concerns
Crucially, this also opens the door to shared decision‑making:
“The chance of needing more adrenaline after two hours is very small. Would you like to stay longer, or go home with clear safety‑netting?”
A final word from the author, Tim Dribin
His take-home was clear – a key concern was reader oversimplification
While most children without cardiovascular involvement can be observed for 2 hours or less, children with more severe respiratory or cardiovascular features clearly benefit from longer observation.
This nuance matters. The paper does not argue for “2 hours for everyone” – it argues for thinking, not ticking.
Bottom line
Most children treated with epinephrine for anaphylaxis do not need prolonged observation.
A risk‑stratified, 2–4 hour approach is supported by robust data.
Cardiovascular involvement is the key driver for longer observation.
Prolonged blanket observation captures very few additional events.
This study won’t replace clinical judgement – but it makes that judgement safer.
How good was the paper – CASP checklist
Does the study address a clearly focused issue?
Yes. It examines when the risk of requiring a repeat dose of epinephrine after paediatric anaphylaxis is low enough that prolonged observation may not be necessary.
Was the cohort recruited in an acceptable way?
Yes. A large, multicentre cohort from 31 EDs was included, capturing a broad and representative paediatric population treated for anaphylaxis.
Was the exposure accurately measured to minimise bias?
Partially. Exposure (initial anaphylaxis severity and epinephrine treatment) was clearly defined, but retrospective data collection may introduce variability.
Was the outcome accurately measured to minimise bias?
Reasonably. The primary outcome — repeat epinephrine administration — is clinically relevant and objectively recorded, though it is a proxy for deterioration.
Have the authors identified all important confounding factors?
Partially. Severity stratification was performed, but residual confounding is possible given the retrospective design.
Was the follow-up of subjects complete and accurate?
Likely adequate for the primary outcome, as repeat epinephrine would be captured during ED stay. However, late events after discharge may be underestimated.
What are the results?
The risk of requiring repeat epinephrine decreases rapidly after initial treatment, with most of the risk occurring within the first 2–4 hours.
Do you believe the results?
Yes. The large sample size, multicentre design, and consistent findings support the validity of the conclusions.
Can the results be applied to a local population?
Yes, with clinical judgement. While based on North American data, findings are broadly applicable to UK paediatric ED settings.
Do the results fit with other available evidence?
Yes. They align with emerging evidence supporting risk-stratified rather than fixed observation periods.
What are the implications for practice?
The study supports shorter, individualised observation periods (2–4 hours in most cases) and emphasises the importance of clinical judgement — particularly in children with cardiovascular involvement.
