Primarily, in patients with a (documented) history of a hypersensitivity reaction to a contrast medium, an alternative imaging modality should be considered. The more severe the reaction, the stronger omitting a contrast medium should be considered. For mild reactions in which alternative imaging modalities are of substantially inferior quality, the risk – benefit ratio may shift. In many cases, CT with iodine-based contrast media can be replaced by ultrasound, with or without contrast agents, or MRI, with or without contrast agents. When this is not possible, consider performing the examination without a contrast medium, but only if this has an acceptable degree of diagnostic quality. For this, close communication with the referring specialist is mandatory.

Use of premedication

In premedication, two types of drugs are used: H1-antihistamines and corticosteroids. Often, they are used concomitantly, making their individual effect difficult to assess, particularly since there are many variations in premedication schedules. H1-antihistamine monotherapy is not common practice in Europe and the US, but has been used successfully in milder HSRs, particularly by Korean research groups.

H1-antihistamines block histamine receptors on various effector cells, blocking the effect of one of the pivotal players in direct mast cell responses. However, mast cells and basophils secrete various other substances that are not blocked by these drugs. The main side effect of the older H1-antihistamines that are available for intravenous administration is drowsiness/sedation. For the newer nonsedating antihistamines this effect is usually mild, but these are mainly available for oral administration.

Corticosteroids have various effect on the immune system, including mast cells, and therefore can block both mast cell degranulation by upregulating inhibitory signalling receptors, and inhibit cytokine production through suppression of gene transcription. (Andrade, 2004; Park, 2009) These membrane stabilizing effects require that administration is started >6h before contrast media administration. Unfortunately, this comes with a less favourable side effect profile, particularly with higher doses and repeated exposure.

The old protocols for premedication shown below (Greenberger, 1981; Greenberger, 1986; Lasser, 1994) are still in widespread use. The Greenberger protocol is popular in the USA, while the Lasser protocol is more frequently used in Europe. There is no literature to establish an optimal indication or protocol. Recently, the Greenberger protocol has been modified into shorter options with intravenous administration for inpatients (Mervak, 2017).

Greenberger protocol (elective examinations 1981, 1984):

• Prednisolone 50 mg IV - 13h, 7h and 1h before the procedure.
• Diphenhydramine 50 mg IV - 1h before the procedure.

Greenberger protocol (emergency examinations 1986):

• Hydrocortisone 200 mg IV - immediately and every 4h until procedure is finished.
• Diphenhydramine 50 mg IV - 1h before the procedure Lasser protocol (elective examinations 1994).
• Methylprednisolone 32 mg IV - 12h and 2h before the procedure.

The evidence regarding the effectivity of corticosteroids and antihistamines for pharmacological prevention is very heterogeneous and of low quality; moreover, it stems from the time of use of high osmolar, ionic ICM (Delaney 2006; Tramer, 2006; Davenport, 2017). It seems that prophylactic premedication can prevent the number of hypersensitivity reactions after contrast administration, but premedication mainly reduces the number of mild reactions and therefore the total number of reactions (Lasser, 1994), and not the number of severe reactions (Jung, 2016). It has been shown that premedication can cause brief hyperglycaemia (Davenport, 2010), but may also be associated with longer hospital stay, increased costs, and worse clinical outcomes (Davenport, 2016).

Few studies have focused on H1-antihistamine monotherapy, and these are biased to patients with mild reactions (Lee, 2016; Park, 2018). In a large Korean multicentre study logistic regression analysis showed that changing the ICM (odds ratio 0.51; 95% CI: 0.36, 0.73) and premedication with H1-antihistamines (odds ratio 0.53; 95% CI: 0.33, 0.86) were protective against recurrent reactions (Cha, 2019).

Many studies report a use of antihistamine and corticosteroid combination premedication; often these regimens are stratified according to the severity of the previous HSR (antihistamines only in mild HSR; antihistamines + corticosteroids in moderate to severe HSR) (Lee, 2016; Park 2017; Park, 2018) or adapted based to the clinicals preference. Corticosteroid monotherapy has rarely been used in older studies (from the high osmolar, ionic ICM era) and their findings cannot reasonably be extrapolated to the current low osmolar, nonionic contrast media (Lasser, 1994) To our knowledge, there are no studies available in which prescription of premedication has been randomized. The currently discussed studies show no additional beneficial effect of corticosteroid premedication in preventing a recurrent HSR. (Park, 2018; Cha, 2019)

Not surprisingly, the Joint Task Force on Practice Parameters of the American Academy of Allergy, Asthma, and Immunology and the American College of Allergy, Asthma, and Immunology concluded in 2020 that “Evidence is lacking to support the role of glucocorticoid routine premedication in patients receiving low-osmolar or iso-osmolar ICMs to prevent recurrent radiocontrast media anaphylaxis” (Shaker, 2020).

In a recent study by McDonald (2021), published after our literature search, 1,973 high-risk patients with a history of HSR were retrospectively studied. Prophylactic measures consisted of changing the ICM and/or steroid premedication, with or without antihistamines. Only patients with a complete steroid premedication protocol (i.e., 2 doses of 32mg of methylprednisolone at 12 and 2 hours before) CT were include in the steroid group; patients with an incomplete protocol were put in the ‘not-steroid-premedicated’ group. In 4,360 examinations, 280 HSR occurred in 224 patients (11%), of which 19 (7%) were more severe than the previous HSR. Patients who received a different ICM with or without steroid premedication had a significantly lower rate of recurrent HSR than those who received the same ICM with steroid premedication (same ICM and steroid premedication: 80 of 423 examinations [19%]; different ICM and no steroid premedication: 10 of 322 examinations [3%]; odds ratio [OR], 0.14 [95% CI: 0.06, 0.33]; P , .001; different ICM and steroid premedication: five of 166 patients [3%]; OR, 0.12 [95% CI: 0.04, 0.36]; P < .001). A sub analysis of the first CT scans only revealed that patients who received the same ICM had a similar risk of recurrent HSR, regardless of whether they received steroid premedication. (Steroid premedication: 44 of 172 patients [26%] vs. no premedication: 73 of 298 patients [25%]; OR, 1.00 [95% CI: 0.64, 1.57]; P = .99).

Although there is less data on the effectivity of premedication in GBCA, the few studies available show comparable results. Premedication with antihistamines and corticosteroid did not eliminate moderate or severe reactions to gadobenate dimeglumine (Bhatti, 2018). Both premedication protocols employed by Ryoo (2019) (antihistamine, systemic steroid plus antihistamine) did not show a recurrence-lowering effect, compared with the non- premedicated cases (antihistamine administration [OR, 1.180; 95% CI, 0.647–2.154; P = 0.589] and systemic steroid plus antihistamine [OR, 1.668; 95% CI, 0.609–4.565; P = 0.316]).

Finally, there is a paucity of data on the benefits of premedication for non-severe nonimmediate hypersensitivity reactions. Most of these reactions are self-limiting or can be treated symptomatically. In the very recent large Korean analysis, changing the type of GBCA and premedication were preventive, but premedication was only preventive in nonimmediate reactions (Ahn YH, 2022). Major international guidelines suggest performing allergologic skin testing, but do not recommend the use of premedication for non-severe nonimmediate reactions (ACR, 2022; ESUR, 2018; Torres, 2021).

Changing of a specific contrast medium

In recent years, changing the culprit ICM has become a frequently employed prophylactic strategy that is used as an alternative or a complementary measure to premedication, the latter particularly if the change has been made empirically without performing skin tests.

A large comparative study with 771 patients showed that changing the CM was more effective than premedication in the prevention of adverse reactions (Abe, 2016). Similar results were achieved in patient cohorts with mild or moderate-severe HSR where changing the contrast medium led to fewer recurrent HSR (Park, 2017; Park, 2018).

A large retrospective study on 1,963 patients showed that changing the culprit ICM only led to significantly lower rates of recurrent HSR, odds ratio of 0.14 [95% CI: 0.06, 0.33]. Additional, corticosteroid premedication did not offer additional protection, odds ratio of 0.12 [95% CI: 0.04, 0.36] (McDonald, 2021). In severe HSR, skin testing is useful to provide a safe alternative ICM (Ahn, 2022; Sohn, 2021).

In a very recent meta-analysis (Umakoshi, 2022), published after our literature search, six retrospective observational studies at moderate to severe risk of bias assessed 4,329 patients in the ICM-change-group and 2,826 in the no-change group. Changing ICM was associated with a reduced risk of recurrent hypersensitivity reaction by 61% (risk ratio = 0.39; 95% credible interval [CrI]: 0.24, 0.58). Adverse events associated with ICM-change were not reported. It was concluded that in observational evidence of limited quality, ICM– change was associated with a reduced risk of recurrent immediate hypersensitivity reaction in patients with a prior ICM-induced hypersensitivity reaction.

In MRI, the experience of changing the culprit GBCA is more limited. In patients with mild immediate HSR, changing the contrast agent could reduce the recurrence rate (Ryoo, 2019). In a small study with mild to moderate HSR to a variety of linear and macrocyclic GBCA, empiric switching to gadoterate reduced the rate of recurrent HSR, independent of premedication with either corticosteroids and H1-antihistamines or corticosteroids only (Walker, 2021).

These findings are in line with the pathogenetic concept that the allergic reactions are not directed against a ubiquitous part of all ICM or GBCA (i.e., not against iodide), but are directed against a specific allergen that is unique to one or more contrast media; switching to a contrast medium that does not contain this epitope will prevent a recurrent allergic reaction. Unfortunately, the exact allergens/epitopes have not been identified and since contrast media are structurally related, the allergen may be present in other contrast media as well, leading to cross-reactivity for those specific agents. As a result, empiric switching of contrast media does not fully prevent a recurrent HSR. For ICM, the presence of the N-(2,3 dihydroxypropyl)-carbamoyl side chain may play a role in the HSR; after a HSR to an ICM containing this side chain, it is advised to switch to an ICM lacking this side chain (iobitridol, iopamidol), preferably supported by a negative skin test (Lerondeau, 2016).

Evidence to decision

There is no evidence that premedication reduces the risk of life-threatening anaphylactic reactions. The evidence for its role in less severe (moderate to mild) HSR remains weak and conflicting. Therefore, the GDG has decided to not advice premedication in patients with an history of immediate HSR to CM.

Corticosteroids do not appear to prevent immediate HSR to GBCA. Contrary, corticosteroids have significant side effects, particularly with cumulative use and in susceptible patients.

Antihistamines reduce the recurrence risk in milder reactions, but it remains uncertain if they also reduce the risk or ameliorate symptoms in moderate to severe reactions, as they are usually given in combination with steroids. Also, antihistamines have side effects, especially sedating side effects can occur (e.g., preventing driving a car). Changing the culprit CM as sole or complementary prophylactic measure significantly lowered the HSR recurrence rate for both ICM and GBCA.

Preferably the CM change is based on negative skin tests; if these are not available, an empiric but educated change should be performed, in which the currently known risks for cross-reactivity are considered (Table 1. Cross-reactivity rates between pairs of ICM in skin positive patients with immediate hypersensitivity reactions to iodine-based contrast media and Table 2. Cross-reactivity rates between pairs of ICM in skin positive patients with immediate hypersensitivity reactions to iodine-based contrast media). In case of an unknown previous culprit CM a testing dose of 10% of the alternative CM can be considered, especially in case of a previous severe reaction.

Breakthrough hypersensitivity reactions to contrast media

It’s becoming increasingly clear that premedication is far from perfect. In premedicated patients so-called “breakthrough” hypersensitivity reactions can occur despite premedication. These are usually of similar severity as the original culprit reaction and are seldom severe (Davenport, 2017; Mervak 2015), but occasionally are of greater severity than the index reaction (Bhatti, 2018).

Iodine-based contrast media

A large study of antihistamine premedication in patients with mild HSR showed no benefit of premedication with a breakthrough reaction frequency of 11%, identical to using no premedication (Lee, 2016).

In a study using a stratified premedication protocol, the frequency of breakthrough reactions was 17%. Most of these reactions (89%) were mild and required no treatment. In severe HSR underdosage of premedication led to a significant increase in breakthrough reactions (Lee, 2017).

Kim (2018) studied the effect of the administration route on breakthrough reactions. Re- exposure to intravascular CM yielded a breakthrough frequency of 19,5%. The number of reactions after extravascular CM was negligible.

Gadolinium-based contrast agents

Walker (2019) showed a high rate (35%) of breakthrough reactions in patients with HSR to gadobutrol. Both culprit and breakthrough HSR were usually mild but may escalate in severity. This rate is very similar to the rate in a previous large prospective study on HSR after gadobutrol (Power, 2016).

In a meta-analysis of breakthrough reactions, a similar 39% rate of breakthrough HSR was found. The frequency was similar between macrocyclic and protein-binding linear GBCA (Walker, 2020).

Evidence to decision

The frequency of breakthrough reactions varies on the severity of the culprit reaction and the specific premedication protocol. Rates after ICM vary between 2-20%, but rates after GBCA administration are higher, in the order of 35-40%. Most of the reactions are of similar severity as the culprit reaction, but incidental escalation in severity may be found.

Cross-reactivity between specific contrast media (see also Introduction to chapter 3.5 Follow up strategieën na hypersensitiviteitsreacties na CM)

In most studies on contrast media hypersensitivity, the term cross-reactivity is used when patients have a HSR to two or more different contrast media, or if there are positive skin tests for two or more contrast media. In the latter case, it is not always entirely certain whether the skin test positivity is clinically relevant, as a drug provocation test is generally not performed. It has recently been suggested to discriminate polyvalent reactivity from cross-reactivity. Polyvalent reactivity comprises patients that have positive skin tests to multiple contrast media. It is argued that the term cross-reactivity should be reserved for polyvalent reactivity within a defined chemical group (e.g., with a N-(2,3 dihydroxypropyl)- carbamoyl side chain), and that multiple positive reactions against non-group CM should be defined as individual reactivity that is probably more prominent between contrast media (Schmid, 2021). However, this is a much stricter definition than has been used in most studies and for clarity we here stick to the broader definition of cross-reactivity.

Iodine -based contrast media

Schrijvers (2018) found most cross-reactivity between agents with a N-(2,3 dihydroxypropyl)-carbamoyl side chain. For immediate HSR, iomeprol and iopromide showed the highest test positivity (41%), while for nonimmediate HSR this was between ioversol and iomeprol (55%) (Table 1. Cross-reactivity rates between pairs of ICM in skin positive patients with immediate hypersensitivity reactions to iodine-based contrast media and Table 2. Cross-reactivity rates between pairs of ICM in skin positive patients with immediate hypersensitivity reactions to iodine-based contrast media).

Sohn (2021) showed in 250 patients with positive skin tests, polyvalent reactivity to at least 2 different ICM in 157 patients. The highest frequency was between iomeprol and iohexol (36%). The frequency was higher in pairs with common N-(2,3 dihydroxypropyl)-carbamoyl side chains than between CM with non-common side chains. This was significant for severe immediate HSR. In contrast, Gamboa (2021) found in IgE-mediated allergic reactions that cross-reactivity of iomeprol with iopamidol, iopromide, and iobitridol was low. In their study, iopamidol was a valid alternative in patients with IgE-mediated allergy to iomeprol and negative skin tests to iopamidol. The culprit ICM itself can be administered safely in patients having experienced nonallergic immediate hypersensitivity. In the CIRTACI study on immediate HSR it was also shown that cross-reactivity was predominantly present in allergic immediate reactions, but seldom in nonallergic immediate HSR (Clement, 2018).

In 43 patients with skin tests for nonimmediate HSR, Gaudin (2019) showed a high rate of cross-reactivity between ICM, that followed the Lerondeau classification (Lerondeau, 2016). Iobitridol was a well-tolerated alternative ICM in 77% of patients. Very similar findings have been found in a 19/142 patients with non-immediate HSR and positive intradermal tests (Gracia Bara, 2019).

In an older meta-analysis of 21 studies on skin testing, extensive data are presented on the frequency of cross-reactivity in immediate and nonimmediate reactions (Yoon, 2015). The percentage of cross-reactivity is in general lower than the percentages found in other studies (Schrijvers, 2018: Sohn, 2021). This may be related to the inclusion of older studies with a lower overall positive yield of the skin test.

Table 1. Cross-reactivity rates between pairs of ICM in skin positive patients with immediate hypersensitivity reactions to iodine-based contrast media 

_{Average percentages and [range] of findings by Yoon 2015, Schrijvers 2018 and Sohn 2021. ICM containing the common N-(2,3-dihydroxypropyl) carbamoyl side chain is grouped within the black line. Risk of cross-reactivity is marked as very low (dark green, <10%), low (green, 10-20%), medium (orange 20-30%), high (red, 30-50%) and very high (dark red, >50%).}

Table 2. Cross-reactivity rates between pairs of ICM in skin positive patients with immediate hypersensitivity reactions to iodine-based contrast media

 _{Average percentages [range] of findings by Yoon, 2015 and Schrijvers, 2018. ICM containing the common N-(2,3- dihydroxypropyl) carbamoyl side chain are grouped within the black lines. Risk of cross-reactivity is marked as very low (dark green, <10%), low (green, 10-20%), medium (orange 20-30%), high (red, 30-50%) and very high (dark red, >50%).}

Gadolinium-based contrast agents

The CIRTACI study showed that a high percentage of Ring-Mesmer type 3-4 reactions after contrast media administration were allergic. Cross-reactivity among GBCA was only shown in these allergic immediate HSR. The overall number of cross-reactivity reactions was higher for GBCA than for ICM, but the number of patients was low for GBCA (Clement, 2018).

In a 7-year retrospective analysis of patients with hypersensitivity to GBCA, 13,6% (18/132) had positive skin tests and were deemed allergic. Cross-reactivity occurred in 38% and was more frequent among the macrocyclic GBCA. Cross-reactivity between macrocyclic and linear GBCA also occurred (Mankouri, 2021).

In a small retrospective study, Grüber (2021) showed cross-reactivity among macrocyclic GBCA and between macrocyclic and linear GBCA, but not among linear GBCA.

In a small case-series of 5 patients with immediate HSR to gadobutrol, only cross-reactivity with gadoterate was demonstrated (Gallardo-Higueras, 2021).

Evidence to decision

In ICM cross-reactivity is common in allergic immediate and even more in nonimmediate HSR. It occurs most frequently among ICM with a common N-(2,3 dihydroxypropyl)- carbamoyl side chain such as iopromide, iohexol, ioversol, iomeprol and iodixanol.

In GBCA cross-reactivity in allergic HSR is more common than with ICM and is especially prevalent among macrocyclic GBCA.

Serum tryptase evaluation and skin testing are key in diagnosing allergic vs. nonallergic HSR and skin tests can identify safe alternative contrast media for future diagnostic studies.

Unknown severity of previous hypersensitivity reaction to contrast media

Unfortunately, there is a lack of data about the recurrence rate and severity of HSR to CM of patients in which there is no data about the severity of the initial HSR. Although in our daily practice this is a substantial part of the population, in studies these patients are not included. Therefore, we want to stress the importance of proper documentation (see below).

A practical guideline to assess the severity of the initial reaction can be adapted from the Hartwig’s Severity Assessment Scale (Hartwig, 1992):

• Did the hypersensitivity reaction to contrast media caused permanent harm to the patient?
• Was the hypersensitivity reaction to contrast media reason for admission to the hospital or reason for increasing of hospital stay?
• Was the hypersensitivity reaction to contrast media treated with an adrenaline auto- injector (Epipen)?

The GDG advice to treat patients in line with a previous mild reaction if these questions are answered with ‘no’. In case one of these questions is answered with ‘yes’ patient should be treated as having a previous severe reaction.

Documentation of hypersensitivity reactions to contrast media

With an increasing use of changing between specific contrast media and the use of skin testing for identifying possible safe alternatives to culprit contrast media causing hypersensitivity reactions, proper documentation in the electronic patient record (EPR) has become very important.

However, the practice is quite different. Documentation in the EPR is not well standardized, is often done by physicians without any experience in the administration of contrast media, and is therefore often insufficient and incomplete (Ananthakrishnan, 2021; Deng, 2019). Recommendations for standardization have recently been published (Böhm, 2020). In selected institutions semi-structured tools for documentation of adverse events have only just been developed and implemented (Lang, 2022).

We would like to re-iterate the recommendations from Safe Use of Contrast Media, part 2: It is mandatory that the physician responsible for the administration of the CM or (EPR only) the drug allergy specialist accurately records the following:

• The place, date, and time of CM administration in the imaging report and in the electronic patient record.
• The specific contrast medium name and dose (volume, concentration) in the imaging report and in the electronic patient record.
• The type of hypersensitivity reaction, immediate or non-immediate, in the imaging report and in the electronic patient record.
• All patient symptoms and vital signs (blood pressure, pulse, respiration rate, oxygen saturation) in the imaging report and in the electronic patient record.
• The treatment given, and the response of the patient to the treatment in the imaging report and in the electronic patient record.
• Any clinical follow-up and advice on need for future premedication in the imaging report and in the electronic patient record.
• Any results of the consultation with a drug allergy specialist on future CM administration in the electronic patient record.

In addition:

• The presence of a documented allergic or nonallergic hypersensitivity reaction in the electronic patient record allergy registry (“allergie registratie”). It is essential that this reporting should be based on the name of the specific contrast medium and be done by radiologists/cardiologists or drug allergy specialists with experience in the use of contrast media.
• If the adverse reaction to a contrast medium is severe or unusual, the physician responsible for the administration of the CM or the drug allergy specialist should report all details of the reaction to the National Pharmacovigilance Authority (LAREB).