Stop using steroids to pretreat contrast allergy! (CAR/CSACI Practice Guidance for Contrast Media Hypersensitivity)

I have spent a lot of time summarizing the evidence and trying to dispel myths when it comes to contrast and kidney injury. (Bottom line: we have massive amounts of evidence, and in controlled studies with modern contrast agents, we cannot find even the smallest hint of harm from contrast.) However, there is another persistent aggravation when it comes to ordering contrast CTs: allergies and arcane allergy protocols. There has never been any good evidence for these omnipresent and incredibly annoying protocols. I have been meaning to tackle the topic for a long time, but perhaps I no longer need to, because in their 2025 guideline, the Canadian Association of Radiologists and Canadian Society of Allergy and Clinical Immunology now tell us to STOP! Do not use steroids for pretreatment in patients with a prior history of allergy! Lets go through their recommendations.

The paper

Byrne A, Macdonald DB, Kirkpatrick IDC, Pham M, Green CR, Copaescu AM, McInnes MDF, Ling L, Ellis A, Costa AF. CAR/CSACI Practice Guidance for Contrast Media Hypersensitivity. Can Assoc Radiol J. 2025 Aug;76(3):400-416. doi: 10.1177/08465371241311253. Epub 2025 Jan 11. PMID: 39797723

The recommendations

1: Documenting

“The dose and precise name of any contrast media must be documented in the patient’s electronic health record, such that it can be easily retrieved in the event of an acute or delayed hypersensitivity reaction. If a patient experiences an acute contrast agent hypersensitivity reaction, the precise name of the offending contrast should also be documented in the patient’s radiology report, which is sent to the referring provider and stored on PACS. The onset and nature of the reaction as well as treatment should also be documented. Ideally, a reporting mechanism should also be in place for delayed hypersensitivity reactions.”

Comment: No argument here. Don’t just say “contrast allergy.” That is sort of like saying “antibiotic allergy”. Be specific. 

2: Allergies are to a specific contrast agent, not general

“Patients (adults and pediatric populations) should only be considered at risk for a hypersensitivity reaction if they had a previous hypersensitivity reaction to the same type of contrast agent. That is, HR to LOCM ICM is a risk factor for future HR to LOCM ICM, and HR to GBCA is a risk factor for future HR to GBCA. There is insufficient evidence to support routine screening of other risk factors. It is important to distinguish between a true hypersensitivity reaction and a physiological reaction.”

Comment: Again, this is very good advice. This is becoming less of an issue, as most places have been using low osmolar agents for a long time, but there is a big difference between low and high osmalar CT contrasts. It is also important to actually make a diagnosis. Too many documented “allergies” have no symptoms of allergy. (This is a broader issue in medicine. I get flagged all the time that a patient is allergic to morphine because it made them nauseous or constipated in the past. Please stop. Allergy means something medically. Let’s use the word correctly.)

3: The best protective measure is to use a different agent

“For patients with a history of a mild, moderate, or severe hypersensitivity reaction to ICM or GBCA, switching to a chemically different contrast agent is the single best protective measure to avoid a breakthrough reaction (Figure 1). For patients with a history of mild or moderate breakthrough reaction after contrast switching, a third agent should be administered. Contrast switching requires institutions and health authorities to have at least 2 different contrast agents. Single vendor contracts should allow for purchasing a second agent for this purpose.”

Comment: Unfortunately, financial conflicts of interest are rampant in hospital purchasing departments. Contracts often prevent you from buying from other companies. I have never worked in a hospital with multiple contrast types available (as far as I know), but the primary responsibility for preventing allergic reactions lies with the radiologist ordering the contrast material, not with the emergency department (because as we will see in a second, we need to abandon our current ridiculous pretreatment strategies.)

4: You can infer the prior CT contrast agent based on year

“For patients with a history of HR to ICM but the offending agent is not known, management is based on the date of the reaction (Figure 2):

• If the reaction occurred before the year 2000, it can be reasonably assumed that the offending agent was not a low-osmolality contrast media (LOCM) agent, and the patient is then cleared to be administered LOCM administered without premedication.
• If the reaction occurred on or after the year 2000, and the reaction was severe (acute or delayed), consider another imaging modality to answer the clinical question. Consider referral to an allergist if access and urgency of imaging allows. Premedication is not recommended. 
• If the reaction occurred on or after the year 2000, and the reaction was mild or moderate, consider premedicating with second generation antihistamines and monitoring the patient during and immediately after contrast media injection for any hypersensitivity reaction.”

Comment: The middle point will be difficult for emergency departments. Allergist referral is never going to be realistic, and we rarely have access to reasonable alternatives. (Sometimes an MRI would adequately answer the question we are asking on CT, but that would require having urgent access to an MRI.)

5: Steroids are out

“In patients with a history of hypersensitivity reaction to LOCM iodinated contrast media, premedication with steroids is not supported by high quality research. What evidence exists is methodologically flawed and pertains to HOCM. Due to lack of efficacy and likely harms associated with steroid pretreatment prophylaxis, the working group recommends against routine use of steroids in high-risk patients. This practice should be discontinued.” (Emphasis added.)

Comment: We have known this for a long time. It is incredibly nice to see this laid out clearly in guideline format. There is really no evidence that steroids change anything except for minor rashes after contrast, but these authors are optimistic, and suggest that steroids might be a number needed to treat of 56,900 to prevent one death and an NNT of 569 to prevent one severe reaction. That might be fine if these steroid protocols were harm free, but in addition to the massive increases in length of stay (25 hours on average), steroids also have other side effects, such as infection. Given the NNT of 56,900, they estimate that in order to save on life from an allergic reaction you will end up causing 551 hospital acquired infections and 32 infection related deaths. We do this a lot in medicine. We become so myopic on a topic (in this case, trying to prevent allergic reactions at all cost), that we lose track of the bigger picture. This is the section of the paper that needs to be shared far and wide. The best evidence, as agreed on by the radiologists and allergists, is that steroid pretreatment protocols are killing patients. Harms outweigh benefits. This needs to stop now. 

6: Steroids are also out for gadolinium

“In patients with a history of hypersensitivity reaction to gadolinium-based contrast agents, premedication with steroids is not recommended. The evidence supporting steroid premedication to prevent a breakthrough reaction is mixed, but overall shows little benefit, with risk of potential harm.”

Comment: Probably not very relevant to most emergency department practices (although I did recently see a doctor successfully advocate for his pregnant patient with an equivocal ultrasound to get an MRI rule out appendicitis, so I guess you never know). 

7: Antihistamines might prevent a bit of itching

“Second-generation antihistamines can protect patients with history of mild hypersensitivity reaction from a breakthrough reaction. The added benefit of antihistamine after contrast switching is less clear and can be considered optional. Ideally, oral second-generation antihistamines should be given at least 1 hour prior to initiation of contrast.”

Comment: I haven’t had time to dig deeply into this literature, and given how minor an issue it is to give an anti-histamine I doubt I ever will. That being said, I would never delay a critical STAT scan by an hour just to give the antihistamine time to work. 

8 and 9: I am skipping

Basically, there is no reason to pick one specific CT contrast agent or gadolinium agent. Doesn’t really matter to us in emergency medicine.

10: For anaphylaxis, consider alternative tests. If there isn’t one, use a different agent and proceed.

“For patients with a history of anaphylaxis to contrast media, the risk of a breakthrough reaction can be limited by switching to a different imaging modality, if diagnostically acceptable. If not, switching to a different contrast agent is recommended. The patient should be monitored during and after the contrast media injection. Premedication is not recommended.”

Comment: This is basically a repeat of the above. The reminder to consider alternative testing is reasonable, as long as it doesn’t result in silly suggestions. (“I have this crashing patient that I think might have a PE… why don’t you order a VQ scan tomorrow?”) For emergency medicine, the key point of these guidelines is that if there is not a viable alternative, prior allergy is not a reason to avoid or delay a test. Just proceed, and monitor the patient appropriately.

11: Collect serum tryptase?

“For patients who experience an anaphylaxis reaction, once the patient has stabilized, collection of serum tryptase may help to determine if the reaction was an immediate hypersensitivity reaction. Serum tryptase should be collected 30 to 120 minutes after the onset of symptoms. Consider referring to allergy, when possible.”

Comment: I don’t know that this needs to be done routinely. Anaphylaxis is probably best left as a clinical diagnosis. Indeed, the few patients I have seen where I questioned the diagnosis, and have had negative serum tryptase levels come back, the allergists still suggest treating as if it was anaphylaxis, so I am not sure this test is helping anyone clinically. That being said, if the patient was already labelled as “allergic” in the chart, but has atypical symptoms, I suppose this might help you to remove that label. (Although, the label will be far less detrimental if we can get radiologists to follow these guidelines and do away with prolonged pretreatment regimens.)

12: Nonvascular contrast is the same

“For patients with a history of hypersensitivity reaction to nonvascular contrast, contrast switching without premedication is recommended for future administrations of non-vascular or vascular contrast. Adverse reactions to non-vascular contrast media are limited to case reports and small series. Based on a lack of evidence and overall rarity of these events, the expert consensus recommends against a pre-medication protocol.”

Comment: No real comment. Not sure how many of the patients I see generically labelled as “contrast allergy” were actually allergies to non-vascular contrast, but it sounds like it doesn’t matter, so I will continue to live in the dark. 

13: Fasting is not required (obviously)

“It is recommended that patients not be instructed to fast prior to receiving intravascular contrast media, unless required for the specific imaging examination (eg, CT or MR enterography).”

Comment: Of course fasting is not required. In fact, I had never even known this was a radiology myth until recently. This had never been a thing in any of the hospitals I worked at in the past, but when trying to change our NPO culture at my current hospital, this myth was raised. It is a nice bonus that they addressed it here.