Penicillin allergy: a practical approach to assessment and prescribing

Penicillin allergies are not always lifelong. Approximately 50% are lost over five years. A reaction to penicillin during a childhood infection is unlikely to be a true allergy.

Only 1–2% of patients with a confirmed penicillin allergy have an allergy to cephalosporins. In patients with a low risk of severe allergic reactions, cephalosporins are a relatively safe treatment option.

Patients with a history of delayed non-severe reactions, such as mild childhood rashes that occurred over 10 years ago, may be suitable for an oral rechallenge with low-dose penicillin. This should be done in a supervised hospital environment.

In many cases, with appropriate assessment and allergy testing, it may be possible to remove the penicillin allergy label.

Most patients who say they have a penicillin allergy are not allergic to penicillins. While 10% of the population will report a penicillin allergy, less than 1% will be truly allergic.^1,2 They have been erroneously labelled as penicillin-allergic.

In the USA, penicillin allergies are the most commonly documented drug allergy, with up to 20% of hospitalised patients having a recorded penicillin allergy.^3,4 In Australian hospitals, national point prevalence data (2013–14) show that 8.9% of patients have a penicillin allergy label on their medical record.⁵ A high proportion of these labels are likely to be incorrect. The patient may have had a non-immune-mediated reaction such as nausea and vomiting, an exanthema (e.g. after taking amoxicillin during an Epstein-Barr virus infection) or an injection-site reaction.^6,7

Patient-reported penicillin allergies alter antibiotic management and may result in the use of suboptimal or broader spectrum drugs such as fluoroquinolones, macrolides, glycopeptides and cephalosporins.^6,8-11 Having a penicillin allergy label has been associated with an increased risk of Clostridium difficile, methicillin-resistant Staphylococcus aureus (MRSA), and vancomycin-resistant enterococci infections and colonisation.³ The increased use of broad-spectrum drugs in hospitalised patients with penicillin allergies also contributes to the growing global problem of antimicrobial resistance.^6,9,12,13 Antibiotic allergy labels are correlated with increases in length of hospital stay,³ hospital readmission rates,¹⁰ surgical site infections,¹⁴ and admissions to intensive care units.¹⁵ Similarly in general practice, penicillin allergy labels are associated with an increased risk of death and MRSA infection or colonisation.¹⁶

It has been demonstrated that more than 90% of patients labelled as having a penicillin allergy would be able to tolerate penicillins following appropriate assessment and allergy testing.^17-19 Even penicillin allergies confirmed by skin tests can wane over time. Half the patients who have a positive skin test for penicillins will lose that reactivity after five years.^13,20 There is therefore interest in penicillin allergy ‘de-labelling’. This is the removal of the allergy label following either allergy history reconciliation or testing (oral provocation or skin testing).

The classification of a patient-reported penicillin allergy label is the first important step in appropriate care (Table 1). Before prescribing, ask patients about their allergies, as not all allergies may have been documented in their medical records. Conversely, some reactions labelled as allergic may be other types of adverse events. Ask about the clinical features of suspected reactions.

The beta-lactam antibiotics include penicillins, cephalosporins, carbapenems and monobactams. Previously it was thought that patients with penicillin allergies had a 10% risk of cross-reactivity with cephalosporins and carbapenems.²¹ However, reviews have reported that the risk of cross-reactivity between cephalosporins, carbapenems and penicillins may be as low as 1%.^21-24

The cross-reactivity between beta-lactam antibiotics may be due to the beta-lactam ring itself, an adjacent thiazolidine or dihydrothiazine ring, or from the side chains (R1 in penicillins or R1 and R2 in cephalosporins) – see Fig. 1.. True cross-reactivity is largely due to the R1 side chains, with the highest risk being in beta-lactams with identical side chains.

Cross-reactivity is particularly seen with aminopenicillins (amoxicillin, ampicillin) and aminocephalosporins (cefalexin, cefaclor, cefadroxil, ceprozil).²⁴ The rate of cross-reactivity between aminopenicillins and aminocephalosporins has been reported to be as high as 30–40% in predominately European studies.^23,25-27 At the antibiotic allergy testing centres of Austin Health and the Peter MacCallum Cancer Centre in Melbourne, out of 15 patients reporting a severe-immediate cefalexin hypersensitivity, intradermal tests determined that six (40%) would not be able to tolerate ampicillin.^5,28

While the data regarding cross-reactivity have primarily been about immediate hypersensitivities, similar patterns have been reported in non-severe delayed penicillin allergies.^29,30 There are limited data regarding cross-reactivity in severe delayed reactions, such as Stevens-Johnson syndrome, toxic epidermal necrolysis, drug reaction with eosinophilia and systemic symptoms, and acute generalised exanthematous pustulosis. For these severe delayed reactions, information regarding cross-reactivity is not a reliable guide for empirical prescribing.

The key to both prescribing and de-labelling for patients with a history of penicillin allergy is an accurate assessment. This involves an understanding of the allergy particularly the severity, timing and tolerance. Therapeutic Guidelines: Antibiotic contains a guide for this assessment (Fig. 2).³¹

Severity

An understanding of the severity of an allergy includes obtaining a description of its ‘type’. Information can be obtained by asking about how the reaction was managed, for example, was the patient hospitalised? What treatments were given for the reaction (e.g. adrenaline (epinephrine), antihistamine, systemic steroids or no therapy)? Simply asking the patient if the reaction was ‘severe’ is unlikely to gather accurate information.

Timing

The timing of the reaction is important to determine if it was delayed (e.g. T-cell mediated reaction) or immediate (e.g. IgE-mediated reaction). Immediate reactions typically occur within a ‘few hours’ of the first or second dose of the antibiotic. A delayed reaction usually occurs after ‘days’ of taking the antibiotic and the reaction can be accelerated if the antibiotic is given again.

Ask how many years ago the reaction occurred. This is important for assessing the likelihood that the penicillin allergy has persisted. In patients with true immediate penicillin allergies, the response wanes over time, with 80% of patients becoming tolerant to penicillins after 10 years.³²

Tolerance

The patient should be questioned about antibiotics that they have tolerated since the reaction, particularly oral penicillins or cephalosporins. Antibiotics that have been tolerated following the reaction should be considered first. Being able to tolerate a specific antibiotic before the reaction does not predict tolerance following the reaction.

Risk assessment

The assessment of penicillin allergy enables classification of phenotypes as either severe versus non-severe and immediate versus delayed. This is helpful in stratifying the risk of using alternative beta- lactam antibiotics. Recommendations for prescribing based on the phenotypes appear in the Therapeutic Guidelines: Antibiotic (Fig. 3).³¹

There are tools that can be used to aid in the assessment of penicillin allergies.^32-34 An example is the Antibiotic Allergy Assessment Tool.³⁴ This underwent multidisciplinary validation by nursing staff, pharmacists, junior and senior medical staff with no training in allergy. It has subsequently been used by hospital pharmacists and nurses to assess beta-lactam allergy labels.³⁵ This tool classifies penicillin allergies into colour-coded risk groups and suggests an appropriate method for de-labelling:^34,36

• no risk – direct ‘de-label’
• low risk – potential direct oral rechallenge
• moderate risk – formal skin testing required before oral rechallenge
• high risk – formal allergy assessment (may include skin testing).

Table 2 shows an extract of the Antibiotic Allergy Assessment Tool.

Non-immune-mediated adverse drug reactions (type A) are not true allergic reactions. Common examples are gastrointestinal symptoms, such as nausea, vomiting and diarrhoea. If a patient has been labelled as penicillin-allergic because of a type A reaction, this should not stop the prescribing of beta-lactam antibiotics and patients do not need to undergo allergy testing. These labels should be directly removed from the patients’ medical records after a discussion about the nature of these reactions and the potential for treatment failure and adverse events if these antibiotics are avoided.

In severe type A reactions, the implicated drug should be avoided. However, it may be possible to use other drugs in the same class. 

Sometimes allergies are reported reflecting the history of a family member rather than the patient. These spurious cases of allergy can usually be de-labelled.

If there was a delayed, non-severe reaction (such as mild childhood rashes or a maculopapular rash that occurred over 10 years ago) an oral rechallenge with low-dose penicillin can be considered. Increasing evidence supports this in patients with a low risk of severe reactions, but the rechallenge should be in a supervised hospital environment.^37,38 At present, there is limited evidence for trying an oral rechallenge in general practice.

Considering which penicillin to use in an oral rechallenge is important as patients can retain hypersensitivity to one penicillin (e.g. amoxicillin) while tolerating another (e.g. penicillin VK) due to variations in the antibiotic R1 side chains. Before the widespread use of amoxicillin, most ‘penicillin allergies’ would be secondary to penicillin VK or G. This should guide the drug to be used for the rechallenge if the ‘penicillin’ is unspecified. For example, if the patient’s allergy dated back to the 1960s, it would be appropriate to use penicillin VK in the rechallenge.

Treatment options for patients with a penicillin allergy can be difficult. Prescribing should be guided by the information obtained from a thorough allergy assessment. Detailed advice regarding the use of cephalosporins and carbapenems is given in the Therapeutic Guidelines: Antibiotic (Fig. 3).³¹

While penicillin allergies can be life-threatening, it is important to ensure that all patients with a recorded penicillin allergy label undergo a thorough antibiotic allergy assessment. These labels should be removed if the patient did not have a true immune-mediated reaction. An assessment of the severity, timing and tolerance of allergic reactions will lead to more ‘de-labelling’ and improved prescribing.

If there has been a presumed immune-mediated reaction, formal antibiotic allergy testing should be considered. While the management of patients with a penicillin allergy can be challenging, the cross-reactivity between penicillins and other beta-lactams is lower than initially reported. In patients with a low risk of severe allergic reactions, cephalosporins can be considered as an appropriate treatment option to penicillins.

Conflicts of interest: none declared

Yuson CL, Katelaris CH, Smith WB. ‘Cephalosporin allergy’ label is misleading. Aust Prescr 2018;41:37-41.