Corticosteroids in Anaphylaxis: Rethinking the Automatic Approach

Anyone working in prehospital emergency care knows the routine. Anaphylaxis arrives, and the sequence begins: epinephrine, and then, almost without thinking, corticosteroids and antihistamines. In our practice, along with epinephrine, we do it automatically. Not because someone has established that it's necessary, but because it's always been done this way.

It's precisely this automatism that's worth questioning. Not to overturn it with a slogan, but for a simple and uncomfortable reason: the guidelines we should be following have stopped recommending that corticosteroid, and a good portion of us haven't noticed.

What the guidelines really say

On one point, recent literature is unusually unified. The Resuscitation Council UK, already in the 2021 update, removed corticosteroids from routine emergency treatment of anaphylaxis. The European Resuscitation Council, in the 2025 Guidelines, does not recommend routine use of corticosteroids in the management of anaphylactic reactions. The EAACI 2021 guidelines, the World Allergy Organization, the American AAAAI/ACAAI practice parameters (updated in 2020 and then in 2024), and the Australasian ASCIA all converge in the same direction:

This isn't a detail hidden in a footnote. It's an explicit position, shared by all major scientific societies. Yet the cocktail survives in our drug kits and in our local protocols. The gap between what the evidence says and what the field does is, in this case, particularly wide.

The myth that keeps corticosteroids alive

The historical justification for administering the steroid is almost always the same: to prevent the biphasic reaction, that recurrence of symptoms that can appear hours after apparent resolution (typically between 8 and 12 hours). The idea has its pathophysiological logic. The problem is that it doesn't hold up to verification.

A 2017 systematic review analyzed dozens of studies without finding convincing evidence that corticosteroids reduce the risk of biphasic reaction. The subsequent meta-analysis put numbers to the concept: no significant benefit (OR 0.87; 95% CI 0.74-1.02), with a number needed to treat of 161 to prevent a single biphasic reaction at a 5% prevalence. In children, the signal even reverses, with a possible increase in risk (OR 1.55; 95% CI 1.01-2.38). Cochrane, for its part, found not a single randomized trial supporting the practice.

In other words: we are administering a drug with known side effects to prevent an event that the drug does not prevent.

The real cost of the cocktail isn't the cocktail

Here's the point that, in my opinion, matters more than any odds ratio. The problem isn't the single vial of hydrocortisone. It's what that vial represents and what it brings with it.

Corticosteroids are not life-saving and have no immediate effect on anaphylaxis symptoms: their latency is measured in hours, while severe anaphylaxis plays out in minutes. Epinephrine, on the other hand, is the only intervention that reliably reverses both the respiratory and hemodynamic components. Every minute and every ounce of clinical attention spent preparing and justifying the rest of the cocktail is taken away from the only drug that changes the outcome.

And this is where observational data becomes interesting. In the Cross-Canada Anaphylaxis Registry, prehospital use of corticosteroids was associated with an increase in intensive care or ward admissions (adjusted OR 2.84; 95% CI 1.55-6.97), correcting for severity, epinephrine, antihistamines, asthma, sex, and age. In the same registry, prehospital epinephrine reduced the likelihood of needing repeat doses in the emergency department. The World Allergy Organization goes so far as to write that corticosteroids, in anaphylaxis, might even be harmful.

A point of honesty

It would be easy, at this point, to conclude with "corticosteroids are harmful, stop giving them." It would also be dishonest, and not very useful to colleagues who think critically.

The evidence against corticosteroids is also of low quality. It's almost all observational, and that data on admissions is very likely affected by confounding by indication: the sickest patients more often receive the steroid, and therefore the association with worse outcomes may reflect baseline severity rather than a drug effect. The authors themselves openly acknowledge this.

But this doesn't weaken the argument, it clarifies it. The thesis is not "steroids worsen anaphylaxis." The thesis is more sober and more solid: there is no demonstrated benefit that justifies automatic administration, there is a possible signal of harm, and above all there is a real opportunity cost with respect to epinephrine. For this, we don't need a large trial: pathophysiology and clinical common sense are sufficient.

Reasoning instead of automating

The alternative to the automatic cocktail is not an automatic prohibition. It's, simply, a clinical decision.

It means that IM epinephrine comes first, always, upon recognition. It means that antihistamines are considered, if at all, only after stabilization and only for cutaneous symptoms. And it means that corticosteroids are not the obligatory fourth step of a sequence, but a choice reserved for specific situations: refractory anaphylaxis after initial resuscitation, or concomitant uncontrolled asthma. The same guidelines that exclude it from routine use leave this door open, albeit without strong evidence for or against.

When I wanted to verify whether my field impression held up, I queried the evidence with the review engine we use in EMSy. The answer converged with international guidelines, but the interesting point is another: even a tool that synthesizes the literature must be questioned, not followed blindly, exactly as an inherited protocol should not be followed blindly. Automatism, in clinical practice, is always the real risk, whatever its source.

In closing

I'm not proposing to throw corticosteroids out of the drug kit. I'm proposing to stop giving them without thinking.

There's a huge difference between a deliberate action and an inherited action. The first can be defended in front of a difficult case; the second is just done, until someone stops to ask why.

I'm interested to know how it works in your systems. Is the epinephrine plus steroid plus antihistamine cocktail still an automatism in your protocols, or have you already started to reason it out case by case? And if you've changed it, what really made the difference: the evidence, training, or an audit that put the numbers on the table?

References

1. European Resuscitation Council Guidelines 2025: Special Circumstances in Resuscitation. Resuscitation. 2025.
2. Muraro A, et al. EAACI guidelines: Anaphylaxis (2021 update). Allergy. 2022;77:357-377.
3. Resuscitation Council UK. Emergency treatment of anaphylaxis: Guidelines for healthcare providers. 2021.
4. Cardona V, et al. World Allergy Organization Anaphylaxis Guidance 2020. World Allergy Organ J. 2020;13(10):100472.
5. Shaker MS, et al. Anaphylaxis: a 2020 practice parameter update, systematic review, and GRADE analysis. J Allergy Clin Immunol. 2020;145(4):1082-1123.
6. Alqurashi W, Ellis AK. Do Corticosteroids Prevent Biphasic Anaphylaxis? J Allergy Clin Immunol Pract. 2017;5(5):1194-1205.
7. Liyanage CK, Galappatthy P, Seneviratne SL. Corticosteroids in management of anaphylaxis: a systematic review of evidence. Eur Ann Allergy Clin Immunol. 2017;49(5):196-207.
8. Gabrielli S, et al. Evaluation of Prehospital Management in a Canadian Emergency Department Anaphylaxis Cohort. J Allergy Clin Immunol Pract. 2019;7(7):2232-2238.
9. ASCIA. Acute Management of Anaphylaxis Guidelines. 2024.