Laboratory studies performed in the 1960s and 1970s demonstrated immunologic cross reactivity between penicillins and cephalosporins, specifically to the R1 side chain off the β-lactam ring.[7–10] These studies show very little cross reactivity between the β-lactam rings themselves (reaction to the β-lactam rings was one of the original explanations for the link between penicillin and cephalosporin allergies). More recent laboratory studies and several cohort studies confirm the role of the R1 side chain in the cross reactivity.[11–17] Patients with an allergy to a specific cephalosporin also demonstrate cross reactivity to penicillin and other cephalosporins, as evidenced by an elevated immunoglobulin E (IgE) response when challenged.
Assem and Vickers challenged 24 penicillin-allergic patients with cephaloridine. Three (12.5%) of them had an adverse reaction. This study may be flawed in that the penicillin and cephaloridine were obtained from the same manufacturer, which increases the risk of cross contamination of the two medications and the chance that they were manufactured in the same Acremonium fungus.
Two studies specifically addressed the reaction rate of penicillin-allergic patients to cefamandole (second generation), which is no longer available in the United States. Blanca et al. challenged 19 patients with a confirmed penicillin allergy with cephaloridine (first generation) and cefamandole (second generation). Two patients (10.5%) had an adverse reaction to cefamandole; no reactions were noted with cephaloridine. Miranda et al. challenged 21 patients confirmed to be allergic to amoxicillin by skin testing to cefadroxil (first generation) and cefamandole. No reactions were seen to cefamandole, but 8 patients (38%) had a reaction to cefadroxil. The combined adverse reaction rate for cefamandole is 0.05%.
Sastre et al. specifically addressed cross reactions in patients with an allergy to amoxicillin and cefadroxil. The authors confirmed a penicillin allergy in 76 (13%) of the 576 patients with a reported history of penicillin allergy. The 16 patients who were specifically allergic to amoxicillin were challenged with cefadroxil. Two of them (12.5%) had an immediate allergic event.
Fonacier and associates sought to quantify the true risk of an allergic reaction to cephalosporins in patients with a documented penicillin allergy. The investigators sent a survey to 186 patients who had a history of penicillin allergy and received a cephalosporin during a hospital stay. Eighty-three patients responded, yielding a response rate of 44%. Seven of the 83 patients reported an allergic reaction (8.4%): 2 of the 44 who received a first-generation cephalosporin, 3 of the 10 who received a second-generation cephalosporin, and 2 of the 19 who received a third-generation cephalosporin. None of the patients who received a fourth-generation cephalosporin had a reaction. Limitations of this study are recall bias and the small sample size. Patients who had a reaction are more likely to respond, so true cross-reactivity rates in the larger group are likely to be lower.
In a cohort study by Romano and colleagues, 128 penicillin-allergic patients (those who sustained anaphylactic shock or urticaria) underwent skin testing with cephalosporins (cephalotin, cefamandole, cefuroxime, ceftazidime, ceftriaxone, and cefotaxime). Fourteen (10.9%) had a positive reaction, mostly to the first-generation cephalosporin, cephalothin. All of the patients with negative skin tests tolerated challenges with cefuroxime (second generation) and ceftriaxone (third generation).
Novalbos et al. challenged penicillin-allergic (confirmed by positive skin test or provocation test) patients with cephalosporins (cephazoline, cefuroxime, and ceftriaxone) that had side chains dissimilar from the one in the penicillin that caused the reaction. All of the patients tolerated therapeutic doses without adverse effects.
The study by Beam and Spooner questioned the reliability of a reported history of penicillin allergy. Only 2 of the 20 patients who gave a history of a type 1 hypersensitivity reaction to penicillin actually had a positive skin test. Similar results were reported by Solensky et al., who performed skin testing on 58 patients with a history of an IgE-mediated allergic response to penicillin. Fifty-three had a negative skin test and were then challenged with three 10-day courses of oral penicillin. Among the 46 patients who completed the protocol, there was no increased risk of resensitization. Of the 7 patients who dropped out of the study, none was known to have experienced an allergic reaction.
A cohort observational study of patients receiving cephalothin (first generation) found that 7 of 54 patients had an adverse reaction (i.e., rash [n = 3], urticaria [n = 2], or anaphylaxis [n = 2]). Five of these patients reported a history of a penicillin allergy, though only 3 had a positive penicillin skin test. Interestingly, 2 of the 7 had a positive skin test to cephalothin, and one patient had positive skin tests to both cephalothin and penicillin—this patient's adverse reaction was reported as anaphylaxis that occurred within 30 s after receiving the cephalothin.
In a 6-year cohort observation study, Macy and Burchette followed 249 patients and documented the number of adverse reactions they experienced after receiving antibiotics. Of the 83 patients who were confirmed to be penicillin allergic, 42 were given a cephalosporin, and one in that group had an adverse event. The reaction was attributed to cefixime (third generation). Interestingly, the authors found that the reaction rate in penicillin-allergic patients was actually lower with cephalosporins than with non-β-lactams (p = 0.005). Cephalosporins were also associated with fewer adverse reactions, independent of penicillin skin test results (p = 0.005).
Finally, three large retrospective cohort studies confirm that the cross-reactivity rate between penicillin and cephalosporins is very low. From the records of 606 patients who had a history of penicillin allergy and who received cephalosporins during their hospital stay, Daulat et al. found that 1 (0.17%) patient had an adverse reaction. These investigators recorded a total of 16 adverse reactions to cephalosporins out of the 27,230 charts they reviewed, yielding an overall adverse reaction rate to cephalosporins of 0.07%. These numbers may be biased by pharmacists who reviewed the orders and recommended alternative therapies and by incomplete coding of patients' medical records. Goodman et al. reviewed the charts of 2933 patients who received prophylactic cefazolin (a first-generation cephalosporin) before surgery. Of the 300 patients with a documented allergy to penicillin, 1 (0.3348%) had an adverse reaction to the cefazolin. The largest retrospective review was by Apter et al., who reviewed the records of 534,810 patients who received penicillin followed by a cephalosporin within 60 days. Of the 3920 patients who reported a reaction to penicillin, 43 (1.09%) had a reaction to a cephalosporin. Among the 530,890 patients who did not report a reaction to penicillin, 581 (0.11%) had a reaction to a cephalosporin. These three studies represent a total of 45 adverse events in 4826 documented penicillin-allergic patients, for a composite cross-reaction rate with cephalosporins of 0.93%.
The meta-analyses by Pichichero and Casey, and Anne and Reisman show that a cross allergy of penicillin with first-generation cephalosporins (odds ratio 4.8, confidence interval 3.7–6.2) does exist.[2,29] The estimated incidence is 1% to 10%. Their data show a negligible cross allergy of second-generation cephalosporins (odds ratio 1.1, confidence interval 0.6–2.1) with penicillin.
Solensky et al. addressed physicians' willingness to administer a cephalosporin to patients with penicillin allergy. For patients with a vague penicillin allergy, 58% and 59% of physician responders would choose a cephalosporin to treat mild or moderate disease, respectively, and 40% would choose vancomycin. For patients with a convincing penicillin-allergy history and severe disease, 55% of physicians would choose erythromycin, 44% would prescribe a quinolone for oral use, and 63% would choose vancomycin. The study is limited by a 16% response rate.
J Emerg Med. 2012;42(5):612-620. © 2012 Elsevier Science, Inc.