The experimental part of the study was approved by the ethics committee of the RWTH Aachen University Hospital (study number EK 309/19) and the North Rhine medical council (study number: 2020098). All donors signed an appropriate informed consent form. The pharmacoepidemiological part of the study was approved by the ethics committee of the University Hospital Bonn (study number 458/20).

An increase in the number of spontaneous reports of cefuroxime-associated anaphylactic reactions was observed for Germany (Figure 1) and the EEA, especially between 2016 and 2019. In more than half of the reports with respective information (EEA, 60.8%; Germany, 66.9%), the anaphylactic reaction occurred after intravenous administration and in 34.8% (EEA) and 31.8% (Germany) after oral administration (see Table 2). While the number of outpatient prescriptions of cefuroxime in Germany decreased substantially, the reporting rates (number of spontaneous reports/number of outpatient prescriptions) increased slightly (except for 2021).

Almost all of these spontaneous reports were classified as serious (EEA, 99.3%; Germany, 98.8%), and slightly more than half of them were designated as life-threatening (EEA, 55.4%; Germany, 56.6%). Fatal outcome was mentioned in 5.5% (EEA) and 8.2% (Germany), respectively.

In those reports with respective information (EEA, 71.7%; Germany, 71.4%), the most frequently reported indication was antibiotic prophylaxis (EEA, 42.0%; Germany, 51.5%).

Moreover, 40.4% of the EEA and 53.8% of the reports from Germany reported symptoms of the system organ classes, which form the basis for the classification by Ring and Messmer (Table 2). Most of these reports included symptoms of one or two of these organ systems (EEA, 84.8%; Germany, 88.7%). Of these ADR reports, respiratory disorders were reported in 53.0% of the EEA and 52.0% of the reports from Germany. Cardiac disorders were mentioned in 36.3% (EEA) and 49.0% (Germany). The evaluation of the co-occurrence of several system organ classes is shown in Table 2.

Out of the 15 patients, 12 were female, and 3 were male. The mean age was 55.7. An LTT was performed in all patients, but a BAT was not performed in three patients.

The BAT was positive in 3/12 patients compared to 7/15 patients in the LTT (sensitivity BAT 25% and LTT 46.7%) (see Table 1). Four control persons with no reported clinical history (note: no additional skin tests were performed in control persons) of sensitization to cefuroxime gave a (false) positive result in the LTT (specificity 73.3%, PPV = 0.70, NPV = 0.60).

The LTT with IL-5 as read-out was negative in all 13 patients and controls. Notably, PBMC from all patients and controls showed positive IL-5 and IFN-y responses to the positive control tetanus toxoid, confirming the technical validity of both tests.

Skin tests (prick and/or intracutaneous) were performed in 13/15 patients and were positive (prick and/or intracutaneous) in 11/13 cases. In the two patients with negative skin tests, a cefuroxime-induced anaphylactic reaction was still considered confirmed due to the very short time of onset of the reaction following administration of cefuroxime (5 min and immediately, respectively).

The pharmacoepidemiological part of our study underlines the clinical relevance of cefuroxime-associated anaphylactic reactions since (i) almost all of the anaphylactic reactions were classified as serious and (ii) the number of reports increased in the past. Although anaphylactic reactions in general can take a serious course (3, 23), we consider the number of spontaneous reports classified as serious to be remarkably high (Germany, 98.9%; EEA, 99.3%). Accordingly, involvement of the respiratory tract and cardiovascular system was observed in about half of the reports from Germany with respective information. Differences in the reporting behaviors between several countries could be the reason for the difference in the proportion of patients with cardiovascular ADRs in the ADR reports from the EEA (36.3%) compared to Germany (49.0%).

The increase in the reporting rates of cefuroxime-associated anaphylactic reactions over time (2013–2019 and 2021–2022) may reflect an increased inpatient use, a higher awareness of ADR reporting in general, changes in reporting obligations (27), or other so far unknown factors (16). The lower number of reports in 2020 and 2021 may be related to a preferential reporting of COVID-19 vaccines instead of ADRs related to other drugs as observed in other studies (28). Our calculated outpatient reporting rate is subject to underreporting (29) concerning the numerator since not all anaphylactic reactions are reported, thereby underestimating the real incidence. On the other hand, the denominator does not include inpatient cefuroxime administrations [not (publicly) available], thereby overestimating the real incidence. The real incidences cannot be determined based on spontaneous reports (16, 29).

In line with the high proportion of intravenous administration, antibiotic prophylaxis, e.g., in the context of surgery, was the indication most frequently reported. However, cefuroxime-associated anaphylactic reactions might also be more often detected, and therefore reported, in the context of surgery than in other settings. Notably, in the spontaneous reports, cefuroxime was considered suspected or interacting and was the only drug reported as suspected/interacting in the majority of the reports. However, other drugs administered concomitantly may also have contributed to or caused the anaphylactic reaction. Since no individual assessment of the ADR reports, e.g., with regard to the causal relationship, has been performed, the number of spontaneous reports in which cefuroxime was considered the only causal drug (among other drugs given) cannot be determined.

The timing of the LTT and BAT could, to some extent, explain why the LTT performed better than the BAT (46.7% vs. 25% sensitivity, respectively). While the BAT was performed prior to the skin tests in 11/12 patients, the LTT was always conducted thereafter except for one patient. Therefore, an immunological booster effect may have occurred in these patients before the LTT except for Patient 8 who tested negative in the LTT. However, it does not explain why the BAT proved negative in 75% of the patients. The specificity and sensitivity of the BAT with cephalosporins or beta-lactam antibiotics in general vary over a broad range from a high specificity of 92.3% but a low overall sensitivity of 20.8% (30) to a sensitivity of around 60% (1). The sensitivity and specificity of BAT can vary depending on factors such as the specific drug, test protocol, and patient population. For cefuroxime specifically, more targeted research would be needed to determine its precise sensitivities and specificity values in BAT. None of the 12 patients tested in the BAT was a non-responder in accordance with the specifications in the product information of the manufacturer (see Section 2.2.2) (26). This absence of a non-responder in 12 patients is compatible with the information given in the product information of the Flow Cast. There it is stated that 6.1% (out of n = 98) were non-responders to anti-FceRI mAb and 4.9% (out of n = 61) to fMLP (26).

Notably, in Patient 8, BAT and LTT were performed 0.5 months after the reaction with negative results in both tests and 3.5 months prior to skin testing (positive). It may be discussed that the time to testing was too short for both tests in this particular patient. However, since the negative results accounted for both tests, the exclusion of this patient would not substantially favor one of the tests with regard to the sensitivity [BAT 27.3% (3 pos./11 pat. tested); LTT 50% (7 pos./14 pat. tested)].

The modest sensitivity of the LTT (46.7%) in these patients with immediate-type reactions complies with the respective findings in literature. As stated above, the LTT in principle may also be used for immediate-type reactions (8), albeit with a lower sensitivity (9, 10).

Interestingly, we did not observe any IL-5 secretion of the PBMC of the patients upon cefuroxime stimulation, although one could expect the secretion of this Th2 cytokine since all patients developed an immediate (Th2)-type reaction by the clinical phenotype. However, it has been described in the literature that the in vivo phenotype of the reaction does not need to correspond to the in vitro phenotype (31).

Physicians should be aware that cefuroxime, in addition to delayed-type reactions, also induces severe anaphylactic reactions, in particular following intravenous application. Hence, monitoring patients receiving cefuroxime intravenously is important to avoid severe courses of anaphylactic reactions by early interventions. In addition, taking a break between the administration of cefuroxime before surgery and the administration of other drugs (e.g., muscle-relaxing drugs) may facilitate the identification of the causative drug based on the timing of a potential reaction. Finally, neither a negative BAT nor LTT can rule out a sensitization in cefuroxime-induced anaphylactic reactions.

Our results may trigger further studies specifically designed to compare the performance of the LTT to the BAT in immediate-type reactions induced by other drugs. Likewise, the impact of skin testing with the suspected drug before the BAT and LTT should be evaluated. We would like to highlight this aspect since the existence of such a temporal association could be considered to improve the sensitivity of these in vitro tests thereby facilitating the detection of drug hypersensitivity.

Meanwhile, if a patient is scheduled for a skin test and BAT or LTT, it may be considered to perform the skin test first, if appropriate and not contraindicated.