Examining Cefazolin Utilization and Perioperative Anaphylaxis in Patients with and without a Penicillin Allergy Label: A Cross-Sectional Study

Abstract

Study Objective:

To compare the occurrence of cefazolin perioperative anaphylaxis (POA) in patients with and without a penicillin allergy label (PAL) to determine whether the prevalence of cefazolin POA differs based on the presence of a PAL.

Design:

Cross-sectional study.

Setting:

A large U.S. healthcare system in the Baltimore-D.C. region, July 2017 to July 2020.

Patients:

112,817 surgical encounters across inpatient and outpatient settings in various specialties, involving 90,089 patients. Of these, 4,876 (4.3%) encounters had a PAL.

Interventions:

Perioperative cefazolin administration within 4 hours before surgery to 4 hours after the procedure began.

Measurements:

The primary outcome was cefazolin POA in patients with and without PALs. Potential POA cases were identified based on tryptase orders or diphenhydramine administrations within the initial cefazolin administration to 6 hours postoperatively. Verification included two validation steps. The first checked for hypersensitivity reaction (HSR) documentation, and the second, led by Allergy specialists, identified POA and the probable culprit. The secondary outcome looked at cefazolin use trends in patients with a PAL, stratified by setting and specialty.

Main Results:

Of 112,817 encounters, 1,421 (1.3%) had possible cefazolin HSRs. Of these, 22 (1.5%) had POA, resulting in a 0.02% prevalence. Of these, 13 (59.1%) were linked to cefazolin and 9 (40.9%) attributed to other drugs. Only one cefazolin POA case had a PAL, indicating no significant difference in cefazolin POA prevalence between patients with and without PALs (p=0.437). Perioperative cefazolin use in patients with PALs steadily increased from 2.6% to 6.0% between 2017 and 2020, specifically in academic settings.

Conclusions:

The prevalence of cefazolin POA does not exhibit significant differences between patients with and without PALs, and notably, the incidence remains remarkably low. Based on these findings, it is advisable to view cefazolin as an acceptable choice for prophylaxis in patients carrying a PAL.

INTRODUCTION

Surgical-site infections (SSIs) are the most common nosocomial infections among surgical patients, with a significant impact on surgical outcomes. However, selecting the best perioperative antibiotic poses a notable challenge, especially for patients with a beta-lactam allergy [1,2]. Cefazolin is a first-generation cephalosporin that is highly effective against the most common pathogens responsible for SSIs [3]. As a result, it is widely used perioperatively as the first-line antibiotic for SSI prophylaxis due to its favorable safety and efficacy profile [4,5]. Despite the effectiveness of cefazolin at preventing SSIs, patients with a reported and/or documented beta-lactam allergy in the electronic health record (EHR) are frequently not prescribed cefazolin and are instead administered non–beta-lactam antibiotics [6].

In the past decade, studies have found that between 90 - 99% of patients with a penicillin allergy label (PAL) are not truly allergic and could safely receive a beta-lactam antibiotic, including cefazolin [7]. Further, the cross-reactivity degree among beta-lactams appears much less significant than previously reported [8]. The incidence of reactions (as determined through skin testing) to cefazolin among patients with an unverified penicillin allergy stands at 0.7% (95% CI: 0.1%-1.7%) [9]. In a study led by Romano et al., among 131 patients with confirmed penicillin allergies, only 0.8% experienced a cefazolin reaction (95% CI: 0.13%-4.1%) [10]. This cross-reactivity often originates from the shared beta-lactam ring or a dual allergy. Notably, immediate allergic responses to cephalosporins are predominantly linked to antigenic reactions involving the R1 or R2 group side chain groups rather than the central beta-lactam portion. Despite being a first-generation cephalosporin, cefazolin’s unique R1 and R2 side chain groups result in minimal cross-reactivity with other beta-lactam antibiotics [11–13].

In the context of SSI prophylaxis, non-beta-lactam (e.g., clindamycin, vancomycin, and fluoroquinolones) options show compromised safety and efficacy compared to beta-lactam alternatives, with patients with a PAL facing a 50% higher SSI risk when receiving a non-beta-lactam antibiotic for surgical prophylaxis [12,14-16]. In addition to the human toll, SSIs entail substantial financial implications, with estimates indicating their cost exceeding USD 25,000 per case [17,18]. Recently, the 2022 Drug Allergy Practice Parameters were updated by the American Academy of Allergy, Asthma, and Immunology (AAAAI) and the American College of Allergy, Asthma, and Immunology (ACAAI), supporting cefazolin use for patients with a penicillin allergy history, even those with a history of anaphylaxis [19]. This shift aligns with recent research advocating cefazolin perioperative use in patients with a reported mild to moderate penicillin allergy [20–22]. Nevertheless, a disparity remains within the Clinical Practice Guidelines for Antimicrobial Prophylaxis in Surgery, as they continue to advise against beta-lactam usage for confirmed or suspected penicillin allergies, potentially stemming from an ongoing overestimation of the risk of cross-reactivity [23,24]. To enhance the understanding of cefazolin usage in patients with a PAL, we conducted a three-year analysis of perioperative data at a large U.S. healthcare system to determine the rate of PAL and POA. We hypothesized that there is no significant difference in the rate of cefazolin-induced POA when comparing patients with and without a PAL.

MATERIALS AND METHODS

This is an EHR-based cross-sectional study of patients who had received perioperative cefazolin from July 1st, 2017, to July 1st, 2020, at the Johns Hopkins Healthcare System (JHHS), a large northeastern U.S. healthcare system. The JHHS comprises five hospitals (academic and community hospitals) in the Baltimore-D.C. region that take care of more than 125,000 surgical cases per year.

The study included inpatient and outpatient (i.e., ambulatory) surgical encounters across all specialties where cefazolin had been administered between four hours before the start of surgery and up to four hours after the beginning of surgery, to account for delays in order entry. Surgical encounters were identified from the EHR, and details such as patient demographics, administered medications, comorbidities, drug allergies, laboratory results, and other procedures were extracted from the EHR. Comorbidities were identified using codes from the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10), which had been billed during the surgical procedure or subsequent continuous hospitalization [25–27]. Data was extracted at the level of surgical encounters; thus, multiple surgical encounters for the same patient were considered independently for the analysis.

All antibiotic drug allergy labels present at the beginning of surgery were grouped into a corresponding drug class using the American Hospital Formulary Service Pharmacologic-Therapeutic Classification, including penicillins, cephalosporins, carbapenems, and monobactams [28]. Surgical encounter characteristics were summarized and compared between encounters that had a PAL and those that did not. For this study’s scope, we defined potential POA cases as patients with a tryptase level ordered and/or received diphenhydramine within six hours after the initial cefazolin administration. These identified cases were flagged and subjected to EHR manual review [29,30]. The 6-hour time frame was selected in accordance with the International Consensus on Drug Allergy’s definition of immediate reactions, which comprises occurrences within 1 hour and up to 6 hours following drug exposure [29]. While epinephrine is the primary treatment for anaphylaxis, early results showed epinephrine administration to have very low sensitivity for expert-validated POA cases, missing every validated POA case flagged by tryptase order and/or diphenhydramine administration. This may be explained by the use of epinephrine to address hypotension not related to allergic reactions. Therefore, epinephrine was not included to flag POA cases. To refine our subset for manual review, we excluded patients who received cefazolin doses between the end of surgery and discharge, as those who experienced POA due to cefazolin are unlikely to be administered it again subsequently.

Surgical encounters that had a POA underwent two stages of manual validation by EHR review. The first validation assessed whether documentation of an adverse drug reaction was consistent with an immediate hypersensitivity reaction (HSR) phenotype (e.g., hives, angioedema, swelling, bronchospasm, hypotension, anaphylaxis) present in allergy lists, problem lists, diagnostic codes, and/or medical notes in the EHR. A secondary validation, performed by Allergy specialists (SAA and MCD), reviewed and diagnosed anaphylaxis based on the World Allergy Organization Anaphylaxis Guidance 2020 (Supplement 1) and assessed the likely culprit [31]. Manual annotation was performed to categorize characteristics such as the time interval between the cefazolin administration and the first clinical intervention, the addition of a beta-lactam allergy label following cefazolin administration, and whether an Allergy and Clinical Immunology consultation took place and consultation assessment.

To determine whether a POA was due to cefazolin or another agent, we considered several factors including the timing of cefazolin administration concerning the first diphenhydramine treatment or the order of tryptase level measurement, the involvement of an Allergy and Clinical Immunology consultation (when accessible), and the patient’s tolerance to cefazolin post-discharge.

Although an increase in total tryptase level from the baseline can offer heightened sensitivity, the baseline tryptase values were unavailable. Therefore, tryptase levels exceeding 11.4 ng/mL—the established threshold indicating clinically significant mast cell degranulation—were deemed positive [32]. The characteristics of the surgical encounters were then summarized and compared between instances where POA was likely attributed to cefazolin and those where it was not. The sensitivity analyses encompassed comparisons between patients with non-cefazolin-related POA and those without, as well as patients with any form of POA and those without.

Descriptive statistics were reported in the study. Frequencies were compared using chi-square or Fisher’s exact tests, while continuous variables were compared using the two-sample t-test. P-values below 0.05 were regarded as statistically significant. The statistical analysis was conducted using the tableone package in Python version 3.8 [33]. The Johns Hopkins Medicine Institutional Review Board (IRB00258577) granted approval for this study.

RESULTS

Between July 1st, 2017, to July 1st, 2020, there were a total of 112,817 encounters representing 90,089 unique patients (Figure 1). Of the 112,817 encounters, 4,876 (4.3%) had a PAL. Patients with PALs were older (56.4, standard deviation [SD] 20 vs. 52.6, SD 21.9 years), more likely to be female (64.4% vs. 48.5%), more likely to be White (74.2% vs. 62.7%), and less likely to be Hispanic or Latino (3.1% vs. 6%, Table 1). Patients with PALs were also more likely to have comorbidities, including diabetes, hypertension, kidney disease, liver disease, asthma, metastatic cancer, and human immunodeficiency virus. Their surgeries were more likely to occur in academic settings (78.7% vs. 74.2%) and with an inpatient status (90.1% vs. 84.2%). Overall, the most common surgical services were orthopedics (22.4%), general surgery (13.1%), and urology (9.6%). Among all patients studied, sulfonamide antibiotics were the most prevalent allergy label in the EHR at 7.1% of surgical encounters, followed by penicillin (4.3%), macrolides (2.5%), and fluoroquinolones (2.5%). Encounters with a PAL were more likely to have allergy labels for all other antibiotic categories examined (Table 1).

Figure 1.

Retrospective cohort of 112,817 surgical encounters from July 1st, 2017, to July 1st, 2020, at the Johns Hopkins Healthcare System (JHHS). Abbreviations: POA, perioperative anaphylaxis.

Table 1.

Characteristics stratified by the presence of penicillin allergy labels

Characteristics	Overall n = 112,817	No PAL n = 107,941	PAL n = 4,876	P-Value
Demographics
Age, mean (SD)	52.8 (21.8)	52.6 (21.9)	56.4 (20.0)	<0.001*
Female, n (%)	55,526 (49.2)	52,387 (48.5)	3,139 (64.4)	<0.001†
Race, n (%)				<0.001†
White	71,325 (63.2)	67,706 (62.7)	3,619 (74.2)
Black	26,663 (23.6)	25,804 (23.9)	859 (17.6)
Asian	5,361 (4.8)	5,215 (4.8)	146 (3.0)
Other	9,468 (8.4)	9,216 (8.5)	252 (5.2)
Ethnicity, n (%)				<0.001†
Non-Hispanic or Latino	105,077 (93.1)	100,410 (93.0)	4,667 (95.7)
Hispanic or Latino	6,585 (5.8)	6,432 (6.0)	153 (3.1)
Other/Unknown	1,153 (1.0)	1,097 (1.0)	56 (1.1)
Comorbidities, n (%)
Diabetes	8,690 (7.7)	8,178 (7.6)	512 (10.5)	<0.001†
Hypertension	23,344 (20.7)	21,955 (20.3)	1,389 (28.5)	<0.001†
Kidney disease	4,850 (4.3)	4,502 (4.2)	348 (7.1)	<0.001†
Liver disease	2,263 (2.0)	2,094 (1.9)	169 (3.5)	<0.001†
Asthma	4,515 (4.0)	4,159 (3.9)	356 (7.3)	<0.001†
Metastatic cancer	2,692 (2.4)	2,495 (2.3)	197 (4.0)	<0.001†
HIV	251 (0.2)	233 (0.2)	18 (0.4)	0.039†
Antibiotic DALs, n (%)
Aminoglycoside	131 (0.1)	108 (0.1)	23 (0.5)	<0.001†
Cefazolin	14 (<0.1)	11 (<0.1)	3 (0.1)	0.021‡
Cephalosporin	684 (0.6)	527 (0.5)	157 (3.2)	<0.001†
Carbapenem	46 (<0.1)	32 (<0.1)	14 (0.3)	<0.001‡
Fluoroquinolone	2,766 (2.5)	2,397 (2.2)	369 (7.6)	<0.001†
Macrolide	2,841 (2.5)	2,319 (2.1)	522 (10.7)	<0.001†
Monobactam	11 (<0.1)	6 (<0.1)	5 (0.1)	<0.001‡
Nitroimidazole	417 (0.4)	374 (0.3)	43 (0.9)	<0.001‡
Sulfonamide antibiotics	7,987 (7.1)	7,128 (6.6)	859 (17.6)	<0.001†
Tetracycline	1,459 (1.3)	1,254 (1.2)	205 (4.2)	<0.001†
Vancomycin	984 (0.9)	848 (0.8)	136 (2.8)	<0.001†
Other beta-lactam antibiotics	6 (<0.1)	6 (<0.1)		1.000‡
Other non-beta-lactam antibiotics	722 (0.6)	628 (0.6)	94 (1.9)	<0.001†
Number of antibiotic DALs, n (%)				<0.001†
0	94,818 (84.0)	94,818 (87.8)
1	13,388 (11.9)	10,347 (9.6)	3,041 (62.4)
2-3	4,117 (3.6)	2,549 (2.4)	1,568 (32.2)
>=4	494 (0.4)	227 (0.2)	267 (5.5)
Surgery Characteristics
Setting, n (%)				<0.001†
Inpatient	92,254 (84.4)	87,907 (84.2)	4,347 (90.1)
Outpatient	16,124 (14.8)	15,666 (15.0)	458 (9.5)
Unknown	868 (0.8)	851 (0.8)	17 (0.4)
Practice, n (%)				<0.001†
Academic	83,854 (74.4)	80,021 (74.2)	3,833 (78.7)
Community	27,936 (24.8)	26,913 (25.0)	1,023 (21.0)
Unknown	868 (0.8)	851 (0.8)	17 (0.3)
Service, n (%)				<0.001†
Breast	531 (0.5)	495 (0.5)	36 (0.7)
Cardiac	2,895 (2.6)	2,683 (2.5)	212 (4.3)
Cardiology	4,906 (4.3)	4,549 (4.2)	357 (7.3)
Dental	292 (0.3)	281 (0.3)	11 (0.2)
ENT	5,218 (4.6)	5,004 (4.6)	214 (4.4)
General	14,811 (13.1)	14,216 (13.2)	595 (12.2)
Neurosurgery	9,506 (8.4)	9,031 (8.4)	475 (9.7)
OB-GYN	4,499 (4.0)	4,197 (3.9)	302 (6.2)
Oncology	2,580 (2.3)	2,439 (2.3)	141 (2.9)
Ophthalmology	6,215 (5.5)	6,125 (5.7)	90 (1.8)
Orthopedics	25,281 (22.4)	24,222 (22.4)	1,059 (21.7)
Pain Management	139 (0.1)	134 (0.1)	5 (0.1)
Pediatrics	10,304 (9.1)	9,990 (9.3)	314 (6.4)
Plastics	6,173 (5.5)	5,933 (5.5)	240 (4.9)
Podiatry	311 (0.3)	295 (0.3)	16 (0.3)
Thoracic	2,652 (2.4)	2,487 (2.3)	165 (3.4)
Transplant	1,520 (1.3)	1,388 (1.3)	132 (2.7)
Urology	10,836 (9.6)	10,516 (9.7)	320 (6.6)
Vascular	4,143 (3.7)	3,951 (3.7)	192 (3.9)
Other	5 (<0.1)	5 (<0.1)

^*Two sample t-test

^†Chi-square test

^‡Fisher’s exact test

Abbreviations: PCN, penicillin, DAL, drug allergy label, HIV, human immunodeficiency virus, ENT, ear, nose and throat, OB-GYN, obstetrics and gynecology, POA, perioperative anaphylaxis, SD, standard deviation.

Patients with a PAL had higher rates of tryptase orders (p=0.009); however, the tryptase levels were not significantly different among those with a PAL (19.3, SD 18.6 ng/mL) compared to those without a PAL (18.1, SD 42.5 ng/mL, p=0.92). Elevated tryptase levels (>11.4 ng/mL) were more common in patients with PAL (p=0.034). Additionally, encounters with PALs had higher rates of diphenhydramine administration (2.4% vs. 2.0%, p=0.034, Table 2).

Table 2.

Potential surgical encounters with hypersensitivity reactions and perioperative anaphylaxis

Potential HSR Criteria	Overall n = 112,817	No PAL n = 107,941	PAL n = 4,876	P-Value
Tryptase a
Tryptase ordered, n (%)	30 (<0.1)	25 (<0.1)	5 (0.1)	0.009‡
Tryptase level, ng/mL, mean (SD)	18.3 (39.2)	18.1 (42.5)	19.3 (18.6)	0.920*
Tryptase >11.4 ng/mL, n (%)	7 (<0.1)	5 (<0.1)	2 (<0.1)	0.034‡
Diphenhydramine b
Diphenhydramine given, n (%)	2,230 (2.0)	2,113 (2.0)	117 (2.4)	0.034†
Possible POA Cases e
POA, n (%)	22 (<0.1)	19 (<0.1)	3 (0.1)	0.067‡
Cefazolin POA, n (%)	13 (<0.1)	12 (<0.1)	1 (<0.1)	0.437‡
Non-cefazolin POA, n (%)	9 (<0.1)	7 (<0.1)	2 (<0.1)	0.055‡

^*Two sample t-test

^†Chi-square test

^‡Fisher’s exact test

^aTryptase level ordered within initial cefazolin administration to 6 hours postoperatively.

^bDiphenhydramine IV and/or IM administered within the initial cefazolin administration to 6 hours postoperatively.

^cFirst validation, documentation of an ADR consistent with an immediate HSR phenotype.

^dItching/Pruritus not explicitly attributed to opioids/narcotics.

^fSecondary validation, Allergy specialists diagnosed anaphylaxis and identified the likely cause.

Abbreviations: HSR, hypersensitivity reactions, POA, perioperative anaphylaxis, SD, standard deviation.

Between Q3-2017 and Q2-2020, there was a significant rise in the frequency of perioperative cefazolin administration among patients with PALs, increasing from 2.6% to 5.6% (Figure 2). Specifically, among patients in academic practice locations, the prevalence of PALs among patients who received perioperative cefazolin increased from 2.4% to 6.4% during the same period, while the prevalence remained unchanged in community practices (Figure 3A). When analyzing the most represented specialties, a distinct upward trend was observed in the prevalence of PALs among patients receiving perioperative cefazolin across all surgical specialties. Notably, ears, nose, and throat (ENT) demonstrated the most significant increase, from 1.7% to 7.6% within the same timeframe (Figure 3B).

Figure 2.

Perioperative cefazolin administration in patients with penicillin allergy labels between July 1st, 2017, to July 1st, 2020, at the Johns Hopkins Healthcare System. Bars represent the absolute number of encounters with PALs, the line represents the percentage of patients receiving perioperative cefazolin that had a PAL. Abbreviations: PALs, penicillin allergy labels.

Figure 3.

Percentage of patients receiving perioperative cefazolin with penicillin allergy labels among patients who receive perioperative cefazolin, comparing academic and community practices (Figure 3A), as well as different surgical specialties (Figure 3B). Abbreviations: POA, perioperative anaphylaxis, PALs, penicillin allergy labels, ENT, ears, nose, and throat.

In 2,246 (2.0%) surgical encounters with a tryptase level ordered and/or diphenhydramine administered between the initial cefazolin administration and 6 hours postoperatively, 30 had a tryptase level ordered, and diphenhydramine was given in 2,230. Among these, 825 (36.7%) surgical cases received a least one dose of postoperative cefazolin so were excluded, leaving 1,421 (1.3% of total) surgical encounters for EHR manual review (Figure 1). Patients with a PAL were overrepresented in this cohort compared to patients not flagged for review (5.4% vs. 4.3%, p=0.048).

The primary review identified 284 (20%) surgical encounters with documentation of an immediate HSR (e.g., hives, angioedema, swelling, bronchospasm, hypotension, anaphylaxis) for expert review. The Allergy specialists deemed 22 cases (<1%) as having POA. Among these, 13 cases (59.1%) were considered plausibly related to cefazolin, while 9 cases (40.9%) were likely due to a different agent, given that these individuals received and tolerated cefazolin post-discharge (Table 3). Two cases had a tryptase ordered for unclear reasons, so they were treated as non-POA cases. POA cases had an age range from <1 to 84 years old (median 56, interquartile range [IQR] 39.8 – 67.75), and the sex was majority female (63.6%). Diverse surgical specialties were represented; however, the most prevalent were pediatrics (13.6%), general surgery (13.6%), transplant surgery (13.6%), urology (13.6%), and vascular surgery (13.6%).

Table 3.

Characteristics of cases diagnosed with perioperative anaphylaxis

Pre-procedure information				Reaction description and treatment					Follow-up
Age/Gender Race	Procedure	DALs at surgery	PCN DAL	Reaction	Culprit	Time to Treatment (h:m)	Tryptase	Diphenhydramine	Cefazolin DAL added	AI
POA LIKELY DUE TO CEFAZOLIN
64F AI/AN	Academic Outpatient Lumpectomy	Penicillin	Rash	Hives, hypotension, wheezing	Cefazolina	0:08	6.3	Yes	Yes
26F White	Academic Inpatient I&D			Bronchospasm, hypoxia	Cefazolin	0:56		Yes
84M White	Academic Inpatient PCI			Hypotensionb	Cefazolin	1:21	17	Yes	Yes	Yes
55F Black	Academic Inpatient Laparotomy	Levofloxacin		Anaphylaxisc	Cefazolin	1:07	9.5	Yes
36F White	Academic Inpatient Ureteroscopy			SCA, rash	Cefazolin	1:56	46	Yes	Yes	Yesd
54F White	Academic Inpatient Kidney Tx			Hypotension, wheezing	Cefazolin	0:07	56.1		Yes
22F Black	Academic Outpatient Cyst resection			Anaphylaxis	Cefazolin	0:01		Yes	Yes	Yes
66F White	Academic Inpatient Hernia repair	Metronidazole		Hypotension	Cefazolin	0:16	4	Yes
57F White	Academic Inpatient Cystoscopy			Hypotension, SCA	Cefazolin	0:30	212		Yes
69M White	Academic Inpatient Prostatectomy			Hypotension	Cefazolin	2:48	3.4
76M White	Academic Inpatient Angiogram			Hypoxia, rash, tongue swelling	Cefazolin	3:47		Yes
67M Asian	Academic Inpatient Kidney Tx			Hypotension, hypoxia	Cefazolin	1:26		Yes
64M White	Academic Inpatient Liver Tx			Hypotension, rash	Sugammadexd	0:44		Yes
POA NOT LIKELY DUE TO CEFAZOLIN
55M White	Academic Inpatient Whipple			Hypotensionb	Unknowne	N/A	5
74F White	Community Inpatient IM nail			Anaphylaxisc	Unknowne	N/A		Yes
68M Black	Community Inpatient AVF creation			Hypotension	Protamine	N/A		Yes
19F Black	Academic Inpatient I&D	Amoxicillin Sulfa		Hypotension	Unknowne	N/A		Yes
51F Other	Academic Inpatient Angiography			Anaphylaxis, SCA	Aspirin	N/A		Yes		Yes
<1M White	Academic Inpatient Bladder exstrophy			Hypoxia	Opioids	N/A		Yes
1F White	Academic Inpatient Laparoscopy			Hypotension, rash	Unknowne	N/A	3
74F White	Academic Inpatient Spine decompression	Penicillins Sulfa		Hypotension	Sugammadex	N/A	47	Yes
54F Black	Academic Inpatient Fistulogram	Vancomycin		Hypotension	Contrast	N/A		Yes

^aIn cases where cefazolin cannot be ruled out as the culprit, the diagnosis was cefazolin-induced perioperative anaphylaxis

^bUnexplained hypotension.

^cNo further description of signs/symptoms related to the episode of POA

^dSuggamadex placed on allergy list after this incident, but cannot tell what cause was definitively.

^eTolerated cefazolin multiple times after.

Abbreviations: PCN, penicillin, DAL, drug allergy label, HSR, hypersensitivity reaction, AI/AN, American Indian and Alaska Native, I&D, Incision and drainage, PCI, percutaneous coronary intervention, SCA, sudden cardiac arrest, plmt, placement, Tx, transplant, AVF, arteriovenous fistula, perioperative anaphylaxis, SD, standard deviation.

Among patients with POA likely secondary to cefazolin, the ages ranged from 22 to 84 years old (median 64, IQR 54 – 67), and most were female (61.5%). The most prevalent surgical specialties were urology (23%), transplant (23%), and general surgery (15.4%). The median time between cefazolin administration and tryptase level ordered and/or diphenhydramine administered was 56 minutes (IQR 16 – 86). Six out of the 13 (46.2%) patients had a cefazolin allergy label placed in their EHR after the reaction, 8 (61.5%) had a tryptase level ordered, and 4 (30.8%) had a positive tryptase level (>=11.5 ng/mL). Allergy was consulted for 3 (23.1%) cases. One individual was tested by Allergy and displayed a positive skin test reaction to cefazolin, while the skin test for penicillin yielded a negative result. One individual had negative skin testing results for cefazolin as well as vecuronium, chlorhexidine, and penicillin. However, despite these negative results, the patient experienced another episode of POA, reportedly following the administration of cefazolin on a subsequent encounter.

Among the 13 patients diagnosed with POA likely secondary to cefazolin, only one case involved a patient with a previously documented PAL. The reported reaction in this instance was “rash” without any other specific details provided in the EHR. During the surgery, the patient had hives, hypotension, and wheezing. Diphenhydramine IV was given 8 minutes after cefazolin administration. Tryptase levels were ordered 22 minutes later, and the measured level was 6.3. The symptoms were promptly addressed and managed. Subsequently, an allergy label for cefazolin was documented following the surgical encounter. Notably, in the 13 cefazolin-induced POA patients, there was no difference in the rate of PALs between patients with and without cefazolin POA (7.7% vs. 4.3%, p=0.437). Sensitivity analyses showed no difference in the prevalence of PALs between patients with and without a non-cefazolin POA (p=0.055) and with and without any POA (p=0.067) compared to patients without.

DISCUSSION

In this multicenter cross-sectional study involving patients undergoing both outpatient and inpatient surgical procedures, our investigation found similar and notably low occurrences of POA secondary to cefazolin among patients with and without a PAL. During the study, we observed a consistent upward trend in the use of perioperative cefazolin among patients with PALs, with this trend being particularly prominent within academic hospital settings. This finding aligns with the national initiatives aimed at promoting appropriate antibiotic usage in surgical patients in recent years [34]. Out of 112,817 surgical cases, we detected a prevalence of POA events at 0.02%, aligning with the reported incidence range of 1 in 1,000 to 10,000 instances of anesthesia-related episodes [35,36].

Our research corresponds with U.S. data, highlighting that antibiotics, particularly beta-lactams, are the most frequently implicated drugs in POA [37]. In contrast, European studies, particularly those conducted in France, have identified neuromuscular blocking agents as the primary cause [38]. These regional differences have been attributed to variations in allergen sensitization and medical practices across different institutions [39]. Despite a comprehensive manual review of all clinical documentation by experts in our study, it was not always possible to establish a causative drug(s). Additionally, only a small percentage of these patients underwent allergy consultation, and cefazolin was confirmed as the causative agent in only one case. These findings highlight the rarity of POA and the challenges associated with subsequent allergy investigations [39,40].

The reactions observed in our study exhibited varying severity, ranging from hypotension and wheezing to cardiac arrest. This emphasizes the importance of stratifying reaction severity when evaluating HSRs. While no statistically significant contrast emerged in the incidence of cefazolin-induced POA between patients with and without PALs, it is worth acknowledging that the interpretation of this outcome is significantly constrained due to uncertain documentation of cefazolin HSR in a single patient with a PAL. Furthermore, our study revealed that patients with PALs had higher rates of tryptase orders, but there were no significant differences in tryptase levels compared to patients without PALs. This finding suggests a potential inclination to consider cefazolin HSRs in patients with a history of PALs. The approach to investigating these complex reactions is not standardized, and it is increasingly recognized that collaboration between experts in the fields of allergy and anesthesiology is necessary to provide optimal care for these patients [39].

Recent studies have explored the safety of cefazolin in PAL. One study involving 340 patients with a penicillin allergy found that no patients experienced symptoms related to cefazolin [24]. Another study focused on hip and knee arthroplasty patients and observed a self-limited allergic reaction to cefazolin in only one patient (0.3%) with a penicillin allergy [41]. A retrospective study comparing cefazolin, clindamycin, and vancomycin as preoperative treatments for patients with PALs found similar frequencies of HSRs among the three antibiotics (0.9%, 1.4%, and 1.1%, respectively) [42]. In a review of patients with penicillin or cephalosporin allergy labels who underwent arthroplasty, a lower incidence of postoperative SSIs was seen in the cefazolin group compared to the clindamycin and/or vancomycin group, without a significant difference in intraoperative HSRs [43]. Furthermore, an analysis categorizing PAL into IgE-mediated and non-IgE-mediated HSRs demonstrated lower reaction rates to cefazolin for IgE-mediated reactions compared to non-IgE-mediated HSRs (0.9% and 4%, respectively) [44]. Lastly, a multidisciplinary team raised cefazolin usage in children with mild PCN allergies for SSI prevention from 60% to 80% during a quality improvement initiative involving 400 children. This improvement was sustained for five months post-project completion, with no POA reported [45]. Our study, which represents the largest reported data on perioperative cefazolin administration in subjects with a PAL to date, is consistent with this previous research and shows a very low prevalence of POA to cefazolin among patients with PALs, with a rate of less than 0.1%.

This finding is reflected in the Drug Allergy 2022 Practice Parameter Update by the AAAAI and ACAAI [19]. This authoritative source provides valuable guidance on administering cephalosporins to patients with different phenotypes of penicillin allergy. These guidelines provide a systematic approach to categorize reactions based on anaphylactic and non-anaphylactic responses, as well as verified and non-verified penicillin allergies. For individuals with unverified non-anaphylactic penicillin allergies, cephalosporin antibiotics can be administered without testing. In rare cases of anaphylaxis to penicillin, a non-cross-reactive cephalosporin like cefazolin can be given routinely without prior testing [19]. In our study, we found that cefazolin was identified as the probable cause of a POA in just one out of 4,876 subjects. This highlights the infrequency of cross-reactivity and provides additional reinforcement for these recommendations.

Efforts to promote perioperative cefazolin usage in patients with PALs have been gradually gaining traction, initially driven by Blumenthal et al.’s study. Their research revealed that patients with PALs are at a higher risk of SSIs due to less effective surgical prophylaxis [46]. VanderVelde et al. conducted a qualitative improvement project to clarify PAL in pediatric patients preoperatively. Their intervention resulted in an increase in cefazolin use in penicillin-allergic patients from 50% to 74%, noting a four-fold increase in clarified PAL cases [47]. Pharmacists in a community-based center managed preoperative antibiotics for patients with a history of penicillin allergy. Over 2 years, they found that 356 patients, including 52 with a history of anaphylaxis, could safely tolerate preoperative cefazolin [48]. Moreover, the integration of allergy e-consultations resulted in a doubled utilization of cefazolin, whereas the introduction of allergy referrals for preoperative evaluation led to a 27-fold surge in cefazolin usage [49,50]. Our findings indicate increasing use of cefazolin in patients with PALs during the study period, despite the current recommendations against beta-lactam use in individuals with penicillin allergy according to the Clinical Practice Guidelines for Antimicrobial Prophylaxis in Surgery [23]. This indicates a positive advancement in advocating for sustained initiatives aimed at addressing this concern.

This work has limitations. One limitation is that HSR cases were not coded into structured clinical data and required manual review that depended on the clinical quality of documentation. We implemented risk mitigation strategies by employing expert review and using broad criteria to identify cases of POA. However, it is important to acknowledge the possibility that some cases may have been missed, especially if there were no orders for tryptase or diphenhydramine. Still, our detected 0.02% POA prevalence aligns with the reported 1 in 1,000 to 10,000 anesthesia-related episode incidence [35,36]. While we cautiously examined univariate tests, these low absolute counts preclude any adjusted analyses based on underlying patient differences or to take account for repeated measures. Furthermore, we only analyzed antibiotic allergy labels, which limits our ability to contextualize these results in concomitant drug allergy labels, notably including sugammadex and neuromuscular blocking agents. Given POH’s infrequency, it is crucial to forge collaborations within and across specialties, forming centers to nurture and disseminate expertise in this specialized domain [40]. Future studies may analyze a larger consortium of sites to find higher absolute numbers of these rare events, enabling more robust statistical analysis.

CONCLUSION

Our research shows no significant difference in cefazolin-related POA between patients with and without PALs. The low incidence reaffirms cefazolin’s safety for prophylaxis in PAL-carrying patients. This underscores the importance of updating surgical antimicrobial prophylaxis guidelines to improve antimicrobial utilization and reduce the occurrence of SSIs.

HIGHLIGHTS.

• POA rate remains low at 0.02%, aligning with 1 in 1,000 to 10,000 in prior reports

• Cefazolin POA rates are similar with/without PALs, echoing safety for PAL patients

• Rising cefazolin use in PAL patients signals a positive prescribing shift

• The complex POA study underlines the need for allergy-anesthesiology collaboration

• Promote guideline updates for cefazolin use in PAL patients

Abbreviations:

AAAAI

American Academy of Allergy, Asthma, and Immunology

ACAAI

American College of Allergy, Asthma, and Immunology

EHR

Electronic health record

ENT

Ears, nose, and throat

HSR

Hypersensitivity reaction

ICD-10

International Statistical Classification of Diseases and Related Health Problems, Tenth Revision

IV

Intravenous

IM

Intramuscular

IQR

Interquartile range

JHHS

Johns Hopkins Healthcare System

SSI

Surgical-site infection

NLP

Natural language processing

PAL

Penicillin allergy label

PMAP

Precision Medicine Analytics Platform

POA

Perioperative anaphylaxis