First-line treatment with ceftriaxone for Neisseria gonorrhoeae infection less likely to be prescribed to patients with a penicillin allergy label: a retrospective audit of medical records

Introduction

The penicillin allergy label (AL) is designed to protect patients by alerting providers to avoid prescribing drugs that may induce anaphylaxis or other adverse effects. It is frequently encountered in Australian hospitals and community practice, occurring in 1 in 10 medical records.¹ However, allergy to penicillin is far less prevalent. Studies have found that up to 90% of patients with a penicillin AL are able to tolerate penicillin antibiotics.^2,3 This disparity occurs for a variety of reasons. For example, the original reaction may not have been a true allergic reaction (but intolerance or the result of a drug–infection interaction) or, if truly immunological, then it may not recur upon re-challenge or may wane over time.^3–5 Penicillin allergy mislabelling means that patients who are able to tolerate penicillins and other beta-lactams (e.g. cephalosporins) are less likely to be prescribed these antibiotics, to the detriment of care, healthcare costs and antimicrobial stewardship.⁶

First-line treatment for Neisseria gonorrhoeae infection in most areas of Australia includes ceftriaxone (a cephalosporin) plus azithromycin.^7,8 Although only 1–2% of patients with a true penicillin allergy also have a cephalosporin allergy, providers may avoid prescribing ceftriaxone to any patient with a penicillin AL.^9,10 National guidance to navigate antibiotic selection for these patients requires an accurate assessment of allergy/risk and accurate allergy labelling.^7,8

The aim of this study was to determine how the penicillin AL influences antibiotic selection for patients with gonorrhoea at Canberra Sexual Health Centre (CSHC). Differences in prescribed/administered antibiotics between patients with and without the label were identified and analysed. The influence of allergy severity was also assessed.

Methodology

A retrospective chart audit was performed on gonorrhoea cases treated at CSHC between January 2020 and October 2023. Power analysis was performed to determine required sample size and sampling timeframe. Cases were identified via a REDCap electronic notifiable disease database, and clinical information was accessed via EPIC Digital Health Record or CARDATA Specialised Health Information Program.

Data abstraction sheets were developed in Microsoft Excel to record patient demographics and case-related information. Patients were classified as having a penicillin AL if the allergy alert field on their medical record listed an allergy to any penicillin. Drawing on information from the records, allergy severities were classified using the previously validated Antibiotic Allergy Assessment Tool (AAAT).¹¹

Descriptive statistics were used to summarise demographics by individual patients. All further statistics were based on individual cases. Logistic regression was performed in STATA (Ver. 18) using the cluster post-estimation command to account for correlated observations (repeat cases). Statistical significance was set at a P-value <0.05.

Results

A total of 619 patients attended CSHC for treatment of gonorrhoea within the audit timeframe (totalling 728 cases). A total of 519 (83.8%) patients were male sex. The average age at first-case treatment was 33.5 years (penicillin AL 37.4 years, no penicillin AL 33.2 years). A total of 5.5% of patients and 5.6% of cases had a penicillin AL (Table 1).

Antibiotic selection was assessed according to penicillin AL status. Almost all cases without a penicillin AL received ceftriaxone, whereas those with the label received it significantly less frequently (penicillin AL 82.9%, no penicillin AL 98.8%, P < 0.000; (Table 2). Further, cases with a penicillin AL were less likely to receive ceftriaxone plus azithromycin (penicillin AL 53.6%, no penicillin AL 71.6%, P = 0.021), and more likely to receive antibiotics without ceftriaxone (penicillin AL 17.1%, no penicillin AL 1.2%, P < 0.000; Table 2). Similarly, cases with a penicillin AL and without co-infection or symptoms were more likely to receive antibiotics without ceftriaxone (penicillin AL 18.7%, no penicillin AL 1.2%, P < 0.000; Table 2).

Ceftriaxone selection was assessed according to penicillin allergy severity. Statistical tests of significance were not performed due to the small sample sizes. According to allergy severity reported in the medical records, all cases were more likely than not to receive ceftriaxone, although this likelihood was lowest in cases with moderate/high-severity allergy (low 95.5%, moderate/high 63.6%, not reported 75.0%; Table 3). According to allergy severity determined using the AAAT, cases in which the medical record had insufficient information about the allergy were most likely to receive ceftriaxone (92.9%; Table 3). A larger sample is necessary to assess ceftriaxone selection for other allergy severities.

Discussion

Existing and increasing levels of antimicrobial resistance in N. gonorrhoeae have discouraged alternatives to first-line treatment.^7,12 Still, providers may be reluctant to prescribe ceftriaxone to patients with penicillin allergy.^6,10 This study, for the first time in Australia, demonstrates that the penicillin AL adversely influences antibiotic selection for gonorrhoea treatment.

Penicillin AL status significantly influences whether or not a patient who attends CSHC for gonorrhoea treatment will be prescribed ceftriaxone, even if they may be able to tolerate the antibiotic. This finding is commensurate with McGuinness et al. (2021), who assessed antibiotic selection for suspected gonorrhoea infection in the emergency department of an urban US hospital over 5 years.⁶

Additionally, although most medical records included a non-specific allergy severity rating, there was insufficient information to more accurately classify the allergy using the AAAT. These were mostly patients with a low-severity AL. According to the electronic Therapeutic Guidelines, many of these patients should be able to tolerate ceftriaxone,⁸ which they indeed most likely received. Similarly, patients with non-severe immediate hypersensitivity should be able to tolerate ceftriaxone,⁸ which they generally received. Finally, the electronic Therapeutic Guidelines advise avoiding penicillins and cephalosporins in patients with severe immediate hypersensitivity (and assessing the allergy),⁸ and this study found ceftriaxone was prescribed haphazardly in these cases.

Limitations

This study is based on a single centre, which means that the findings may not be generalisable to other populations or care settings. Second, although sample sizes were sufficient based on prior power analysis for assessment of the primary outcome (prescribing by penicillin AL status), absolute numbers were too small to statistically analyse the secondary outcome of prescribing by allergy severity. Third, the retrospective nature of the audit confines data to that available in the medical records. Thus, any discussion of allergy or symptoms between patients and providers that was not documented, or not clearly outlined, may have been missed.

Future directions

There were >129,000 gonorrhoea notifications in Australia from 2020 to 2023,¹³ and this study audited 728 cases from one clinic. If the influence of penicillin AL on prescribing extends into other jurisdictions, populations and care settings, then the potential impact on Australian health care is far-reaching.

We need institutionally endorsed antibiotic allergy assessment tools and complementary allergy label templates for inpatient and outpatient care, as well as system-wide models of care and increased access to services for further allergy assessment. Further, more detailed advice within national sexually transmitted infection management guidelines about prescribing cephalosporins to patients with penicillin allergy would support accurate identification of patients for whom ceftriaxone is appropriate treatment for gonorrhoea.