a b s t r a c t

Objective: To determine the impact of reported beta-lactam allergies on in-hospital mortality and other clinical outcomes in patients who presented with severe sepsis or septic shock.

Methods: This single-center, retrospective cohort study was performed at a 35-bed emergency department in central Kentucky. Patients presenting with sepsis, aged 18 years or older, were identified between October 2016 and June 2017.

Results: 438 patients with severe sepsis and/or septic shock were identified. Rates of the combined endpoint of in- hospital mortality or transfer to hospice were similar in patients with a beta-lactam allergy (7.2%) versus those with no reported beta-lactam allergy (10.4%) (p = 0.41). Time to initiation of antibiotic therapy was slightly lon- ger in the beta-lactam allergic group (2.2 h) versus those with no reported beta-lactam allergy (2.15 h), but the difference was not statistically significant (p = 0.993). Patients were 20.9% more likely to receive an appropriate empiric antibiotic, based off of retrospective culture review, if they did not report a beta-lactam allergy (p = 0.009). This led to a delay in effective therapy of 1.59 h in the reported beta-lactam allergy arm (p = 0.037).

Conclusions: Adequate documentation of beta-lactam allergies is vital to ensure timely and appropriate treatment in patients presenting with severe sepsis and septic shock. Choosing alternative treatment options results in in- creased time to effective antibiotics, reduced likelihood of covering cultures with first antibiotic, and increased total hospital and variable direct cost.

(C) 2020

1. Introduction

Sepsis is a medically urgent state that needs to be treated quickly and efficiently to ensure positive patient outcomes. Patients may pres- ent with hypotension, tachycardia, hyperthermia, hypothermia, tachypnea, leukocytosis, or evidence of organ dysfunction. Guidelines from the Centers for Medicare and Medicaid Services and the Surviving Sepsis Campaign recommend rapid antibiotic initiation, use of crystalloid fluid with or without vasopressors, and timely collection of blood cultures and other laboratory markers to appropriately assess patients’ needs [1]. Although systemic inflammatory response syn- drome (SIRS) criteria are currently utilized in the most updated version of the CMS guidelines, the Sequential [sepsis-related] Organ Failure As- sessment (SOFA) score is also recommended as a more robust tool used to screen patients [2]. Regardless of the tools used to identify sepsis, it is

* Corresponding author.

E-mail addresses: [email protected] (K.L. Komyathy), [email protected] (W.R. Judd), [email protected] (P.D. Ratliff), [email protected] (R.E. Hughes).

essential for the healthcare team to ensure there is no delay in treat- ment of septic patients which could lead to poor outcomes.

There is increasing evidence to support the claim that a delay in time to effective antibiotic initiation results in increased mortality in septic patients [3-7]. Up to an 8% increase in mortality has been shown with each hour appropriate antibiotics are delayed in this population [6,7]. broad spectrum antibiotics are chosen empirically to increase the likeli- hood of adequately treating the suspected infection. Alternative agents may be less reliable based on lower regional susceptibility rates. Other non-penicillin-beta-lactams (e.g., cephalosporins or carbapenems) may be selected for those with non-IgE mediated reactions or for pa- tients who are unsure of the nature of their reaction. This approach may result in increased rates of resistance or increased incidence of Clos- tridium difficile infection due to selective pressure on host bacteria.

Patients may report an allergy to a beta-lactam antibiotic without

experiencing a true IgE-mediated allergic reaction. Studies have demon- strated that 8-10% of the population report an allergy to beta-lactam an- tibiotics, while 90% of reported allergies are not IgE-mediated reactions which would preclude the use of some beta-lactam antibiotics [8]. The severity of the reaction can range from GI intolerance to anaphylaxis

https://doi.org/10.1016/j.ajem.2020.05.032

0735-6757/(C) 2020

K.L. Komyathy et al. / American Journal of Emergency Medicine 38 (2020) 1816–1819 1817

involving Laryngeal edema, hypotension, and shock. Penicillin skin test- ing has been used to “rule out” a beta-lactam allergy in patents who would benefit from therapy. In the inpatient setting, a care-team may choose to challenge the patient with a penicillin skin test in order to op- timize treatment with appropriate or first-line antibiotics. However, in critical situations, confirming a true allergy through this method may not be practical. The delay in time to antibiotic administration due to skin testing could greatly impact a septic patient’s likelihood of survival [9].

Patients with a reported beta-lactam allergy may pose a challenge to clinicians during initial antibiotic selection. Recent literature has also made claims that reported allergies may increase patient adverse events (e.g., infection progression, higher Healthcare costs) and mortality [3,10,11]. This study investigates the impact of reported beta-lactam al- lergies on all-cause, in-hospital mortality among patients who pre- sented to the emergency department (ED) with sepsis (formerly severe sepsis) or septic shock. It was hypothesized that patients with re- ported beta-lactam allergies have an increased risk of mortality due to delayed time to initiation of effective antibiotic therapy.

1. Methods

This retrospective, single-center cohort study was performed at a 35-bed emergency department in Lexington, Kentucky. The study was approved by the Catholic Health Initiatives Institute for Research and In- novation Institutional Review Board. Patients were identified after pre- senting to the hospital with a suspected or documented infection along with two or more SIRS criteria plus evidence of acute organ dysfunction and/or refractory hypotension requiring vasopressors. Patient data were collected through electronic health records.

The study population was determined via chart search of patients who presented to the emergency department between October 1, 2016 and June 30, 2017 and who met inclusion criteria of age greater than or equal to 18 years and diagnosis of severe sepsis or septic shock based on the CMS case definition. Patients were excluded if there was evidence of a confirmed fungal, viral, or parasitic infection; if they were receiving chemotherapy at the time of ED presentation; or if death, transfer to hospice, or comfort care measures occurred within the first 24 h of presentation.

Patients were considered to have severe sepsis or septic shock based on a documented or suspected source of infection, the presence of two or more SIRS criteria, and evidence of acute organ dysfunction and/or refractory hypotension requiring vasopressors, as defined by CMS criteria. Organ dysfunction was defined as one or more of the following: hypotension (i.e., systolic blood pressure b 90 mm Hg, mean arterial pressure b 65 mm Hg, or a 40 mm Hg drop in systolic blood pressure over 1 h); acute kidney injury, as evidenced by a serum creatinine N 2 mg/dL or sustained urine output b 0.5 mL/kg/h in the absence of chronic kidney disease; coagulopathy [i.e., International normalized ratio N 1.5 or activated partial thromboplastin time N 60 s in the ab- sence of chronic anticoagulation]; total bilirubin N 2 mg/dL; platelet count b 100,000/uL; or a lactic acid level N 2 mmol/L [2]. Patients were evaluated using Charlson Comorbidity Index (CCI) and SOFA scores to determine if groups were evenly matched based on severity of illness. intensive care unit admission was also used as a surrogate for dis- ease severity and risk of mortality.

The primary endpoint was a combination of in-hospital mortality or transfer to hospice. Secondary outcomes included time to first antibiotic administration, time to effective antibiotic administration in a cohort of patients with positive cultures as per retrospective review, readmission within 30 days of discharge, incidence of Clostridium Difficile infection developing 48 h or more after admission, ICU length of stay, total hospi- tal length of stay, and total and direct variable cost of hospitalization. Time to antibiotic administration was calculated using the ED Triage time as time zero and the date and time of antibiotic administration, as documented on the medication administration record (MAR). If

there were insufficient data to determine correct timing, patients were excluded from the final analysis. Readmission within 30 days of discharge was calculated using information from several local hospitals in the health system. Data regarding admissions to outside facilities could not be obtained. Culture data and susceptibilities followed an intention-to-treat method and were only compared to the cohort with positive cultures.

Data and statistical analyses were performed using SPSS Version

23.0 [12]. Data were analyzed for normality using the Shapiro-Wilk Test. Study outcomes were compared between beta-lactam-allergic and non-beta-lactam-allergic groups using Mann-Whitney U, Chi- square or Fisher’s exact test, and Kruskal-Wallis test as appropriate. p- Values b 0.05 were considered statistically significant.

1. Results

A total of 438 patients met the inclusion criteria of age greater than or equal to 18 years with a documented or suspected source of infection plus a diagnosis of severe sepsis or septic shock based on the presence of two or more SIRS criteria as well as evidence of acute organ dysfunction and/or refractory hypotension requiring vasopressors, as defined by CMS. After applying exclusion criteria, a final cohort of 375 patients was identified. Ninety-six patients (26%) reported a beta-lactam allergy, while two hundred seventy-nine patients (74%) did not report an al- lergy to this class of antibiotics. Of note, forty-nine patients (51%) with a beta-lactam allergy also reported an allergy to another antibiotic class. Fifty-eight patients (21%) who did not report a beta-lactam allergy reported an allergy to another antibiotic class such as fluoroquinolones, aminoglycosides, sulfonamide derivatives, or glycopeptides.

Patient demographic data are included in Table 1. Groups were evenly matched with regard to baseline demographics, severity of ill- ness, and empiric number of antibiotics administered. Individual com- ponents of the Charlson Comorbidity Index were not statistically significant between patients who did and did not report a beta-lactam allergy with the exception of hypertension (71.3% vs. 59.8%, p = 0.035); hemiplegia and paraplegia (3.2% vs. 8.2%, p = 0.040); end organ damage from diabetes (8.6% vs. 21.6%, p = 0.001); and skin ulcer- ation (7.2% vs. 15.5%, p = 0.015).

The most common first-line antibiotic used in the ED was piperacillin-tazobactam, followed by ceftriaxone, and levofloxacin. The median number of antibiotics administered within the first 24 h of pre- sentation was similar between groups (p = 0.304).

There were no significant differences with regard to the primary sources of infection (Table 2). Bacteremia and urinary tract infections were the two most common conditions reported. Additionally, there was no significant differences in the overall incidence of resistant bacte- ria that were detected prior to the index visit, including methicillin-

Table 1

IQR, interquartile range; CrCl, creatinine clearance; SOFA, sequential organ failure assess- ment; ICU, intensive care unit.

	No reported beta-lactam allergy (n = 279)	Beta-lactam allergy (n = 96)	p value
Age, years (IQR)	71 (59-82)	69 (55.5-79)	0.106
Height, cm (IQR)	170.18	167.64	0.368
	(160.0-177.8)	(160.0-175.3)
Weight, kg (IQR)	77.2 (64.7-93.0)	79.6 (63.6-93.9)	0.550
CrCl (IQR)	49.2 (33.3-83.7)	59.6 (34.9-87.6)	0.151
Male gender (%)	132 (47.3)	39 (40.2)	0.226
Caucasian race (%)	240 (86.0)	84 (86.6)	0.887
Charlson Comorbidity Index (IQR)	6 (5-8)	7 (5-9)	0.080
SOFA score (IQR)	3 (2-6)	4 (1-6)	0.887
ICU admission (%)	117 (41.9)	36 (37.5)	0.440
Total no. of antibiotics given in the	3 (2-4)	3 (2-4)	0.304

first 24 h

1818 K.L. Komyathy et al. / American Journal of Emergency Medicine 38 (2020) 1816–1819

Table 2

MRSA, methicillin-resistant S. aureus; ESBL, extended-spectrum beta-lactamase; VRE, van- comycin resistant Enterococcus.

	No reported beta-lactam	Beta-Lactam Allergy (n =	p value
	allergy (n = 136)	49)
Bacteremia (%)	75 (55.1)	30 (61.2)	0.462
Urinary tract infections (%)	55 (40.4)	25 (51.0)	0.200
Respiratory infections (%)	25 (18.4)	10 (20.4)	0.756
Other (%) History of resistant organisms	31 (22.8)	10 (20.4)	0.730
(%) MRSA	9 (6.6)	4 (8.2)	0.500
ESBL	7 (5.1)	3 (6.1)	0.714
VRE	4 (2.9)	3 (6.1)	0.370

This table represents the cohort with positive cultures, but did not require susceptibility testing, at the time of retrospective review.

resistant S. aureus (MRSA) (p = 0.5), extended-spectrum beta- lactamase (ESBL) producing organisms (p = 0.714), and vancomycin resistant Enterococcus (VRE) species (p = 0.37). Culture and susceptibil- ity data from all past visits were reviewed to determine if there was a history of resistant bacteria.

The difference in the primary endpoint of combined in-hospital mortality or transfer to hospice was not statistically significant between patients with a reported beta-lactam allergy and those without (7.2% vs. 10.4%, p = 0.410) (Table 3). Differences in other secondary endpoints, including time to first antibiotic (p = 0.993), 30-day readmission (p

= 0.667), incidence of hospital-acquired C. difficile infection (p = 1.000), ICU length of stay (p = 0.487), total hospital length of stay (p

= 0.849), and total hospital cost (p = 0.421) were not statistically sig- nificant; however, the percentage of cultures reported as susceptible to the first antibiotic administered was higher among patients who did not report a beta-lactam allergy (78.0% vs. 57.1%, p = 0.009), and time to ef- fective coverage of all organisms was significantly faster in the non- allergy group (2.93 vs. 4.52 h, p = 0.037) (Table 4).

A subgroup analysis was performed with regard to the number of antibiotic allergies reported. Patients were stratified into three groups based on the number of antibiotic class allergies reported, including 0, 1-2, or >=3. There was a statistically significant difference in time to ef- fective antibiotic coverage of at least one organism that was recovered from the site of infection (2.4 vs. 3.3 vs. 3.9 h, p = 0.037) as well as time to effective coverage of all reported isolates (2.8 vs. 4.3 vs. 5.3 h, p = 0.022). Median total hospital cost were numerically lower in those who reported fewer antibiotic class allergies, but the difference was not statistically significant ($8828 vs. $9498 vs. $12,329, p = 0.070) (Table 5).

1. Discussion

Several studies have investigated the impact of beta-lactam allergies on the time to initiation of antibiotics in patients with sepsis. This study adds to the growing body of literature describing the clinical and eco- nomic impacts of reported beta-lactam allergies on patients with severe-sepsis or septic shock. We found a higher than average popula- tion of patients with reported beta-lactam allergies. Approximately 26% of patients in this study reported an allergy whereas previous stud- ies have reported an average of 8-10% as the standard proportion of beta-lactam allergic patients [13,14]. While there is no feasible way to confirm if these reported allergies are severe enough to justify alterna- tive antibiotic usage, in emergent situations such as severe sepsis and septic shock, alternative agents may be chosen for timely initiation of antibiotics. It is prudent to review a patient’s past medical history via a comprehensive interview and chart review to determine past tolerabil- ity of other beta-lactam antibiotics.

Patients reporting an allergy to beta-lactam antibiotics were 21% less likely to receive appropriate empiric antibiotic therapy based on final culture and susceptibility reports (p = 0.009). This resulted in almost a two-hour delay in time to effective coverage of all isolates that were recovered from the site of infection (p = 0.037). Of note, 40% of patients reporting a beta-lactam allergy received a beta-lactam antibiotic (i.e. ceftriaxone or meropenem) as first line therapy. Although rational of empiric coverage cannot be ascertained via retrospective review, it is prudent to consider that patient interview and chart review may have contributed to this delay in treatment.

Although not statistically significant, patients who reported a beta- lactam allergy had a higher incidence of multidrug-resistant organism (MDRO) infections due to MRSA, ESBL-producing organisms, and VRE based on historical records. This could play a role in determining appro- priate Empiric therapy in patients who report and allergy. An increased likelihood of presenting with a more resistant infection may justify the use of more broad- spectrum agents until culture and susceptibility re- ports are available.

In addition to a delay in treatment, it is also evident that reporting an antibiotic-allergy increases hospital costs. With the overall rising cost of healthcare, this may be an unnecessary and avoidable expense. Patients with a reported beta-lactam allergy were almost 30% more likely to re- port an allergy to another class of antibiotics (p b 0.001). Not only do these additional allergies pose a challenge for providers when choosing appropriate empiric therapy, it is also evident that they delay time to ef- fective antibiotic coverage of at least one organism (p = 0.037) as well as time to all isolates recovered (p = 0.022), and they increase the me- dian total hospital cost (p = 0.070).

This study has several limitations which should be considered. First, it is a small study which only investigates patients at a single center. Second, there was a higher than average percentage of patients who re- ported an allergy to any beta-lactam antibiotic. Authors did not further investigate the allergy aside from documentation in the electronic med- ical record. An allergy was considered to preclude the use of penicillin derivatives if a reaction was not documented in the medical record. Lastly, the study center reported a low observed/expected (O/E) mortal- ity ratio of 0.26 per risk-adjusted data from the Premier benchmarking database. This may impact applicability of mortality and length of stay data to outside hospitals with higher O/E ratios.

1. Conclusions

This study adds to the growing body of literature highlighting the importance of thorough and Accurate documentation of patient aller- gies. It is essential to quickly determine the most appropriate empiric therapy to cover potential pathogens in patients with severe sepsis and septic shock to improve patient outcomes and to reduce the likeli- hood of mortality. When patients report an allergy to beta-lactam anti- biotics, there is a reduced likelihood of appropriate initial antibiotic coverage, a delay in time to appropriate antibiotic initiation, and in- creased total hospital and variable direct cost. When multiple allergies are listed in a patient’s medical record, interviewing patients and inves- tigating past medication use may allow for more appropriate Empiric antibiotics.

Funding

The authors received no financial support for the research, author- ship, and/or publication of this article.

CRediT authorship contribution statement

Kelsey L. Komyathy: Methodology, Investigation, Data curation, Writing – original draft, Visualization. William R. Judd: Methodology, Resources, Writing – review & editing, Supervision. Patrick D. Ratliff:

K.L. Komyathy et al. / American Journal of Emergency Medicine 38 (2020) 1816–1819 1819

Table 3

IQR, interquartile range; ICU, intensive care unit.

	No reported beta-lactam allergy (n = 279)	Beta-lactam allergy (n = 96)	p value
In-hospital mortality or transfer to hospice (%)	28 (10.4)	7 (7.2)	0.410
Hospice (%)	19 (6.8)	6 (6.2)	0.832
Mortality (%)	9 (3.2)	1 (1.0)	0.464
Time to first antibiotic administration, hours (IQR)	2.15 (1.32-3.15)	2.20 (1.06-3.49)	0.993
30-day readmission (%)	55 (19.7)	17 (17.7)	0.667
Incidence of C. difficile after 48 h of presentation (%)	3 (1.1)	1 (1.0)	1.000
ICU length of stay (IQR)	2.7 (1.9-4.4)	2.9 (1.9-8.0)	0.487
Total hospital length of stay (IQR)	5.6 (3.4-8.6)	5.6 (3.8-8.0)	0.849
Total hospital cost (IQR)	9109 (6203-14,892)	10,241 (5805-19,157)	0.421
Variable direct cost (IQR)	4800 (3356-8190)	5641 (3041-10,545)	0.429

Table 4 IQR, interquartile range.
	No reported beta-lactam allergy (n = 107)	Beta-lactam allergy (n = 40)	p value
Cultures susceptible to first antibiotic (%)	96 (78.0)	24 (57.1)	0.009
Time to effective antibiotic administration of at least one culture, hours (IQR)	2.72 (1.8-4.4)	3.72 (1.6-5.5)	0.175
Time to all cultures covered, hours (IQR)	2.93 (2.1-6.5)	4.52 (3.0-11.0)	0.037

This table represents the cohort with positive cultures and susceptibilities at the time of retrospective review. Effective coverage was defined as administration of an antibiotic to which the organism was reported as susceptible. In some instances, Combination therapy was required when a single agent would not be effective against multiple pathogens. In this case, time to effective coverage was based on the time when the last antibiotic of a multidrug regimen was administered, as reported on the MAR.

Table 5

Antibiotic-classes were defined as beta-lactams, fluoroquinolones, tetracyclines, etc.

Acknowledgements

Time to effective antibiotic administration of at least one culture, hours (IQR)

Beta-lactam allergy

p value

0.037

The authors thank Philip L. Whitfield from the Saint Joseph hospital pharmacy department for his contribution to data collection.

References

0 reported allergies 2.4

(1.6-4.2)

1-2 reported allergies 3.3

(2.1-6.8)

N2 reported allergies 3.9

(2.5-5.9)

Time to all cultures covered, hours (IQR) 0.022

0 reported allergies 2.8

(2.0-5.7)

1-2 reported allergies 4.3

(2.5-7.2)

N2 reported allergies 5.3

(3.5-28.8)

Median total hospital cost (IQR) p value

Total reported antibiotic-class allergies

0 8828.19 (6220.46-14,653.48) 0.070

1-2 9497.66 (5723.87-15,495.76)

N2 12,328.62 (8487.94-20,784.52)

Methodology, Formal analysis, Writing – review & editing, Supervision.

Robert E. Hughes: Conceptualization, Writing – review & editing.

Declaration of competing interest

Authors of this presentation have nothing to disclose concerning possible financial or personal relationships with commercial entities that may have a direct or indirect interest in the subject matter of this presentation.

1. Rhodes A, Evans LE, Alhazzani W, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016. Crit Care Med. Mar 2017;45(3):486-552.
2. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus def- initions for sepsis and septic shock (Sepsis-3). JAMA. February 23 2016;315(8): 801-10.
3. Whiles BB, Deis AS, Simpson SQ. Increased time to initial antimicrobial administra- tion is associated with progression to septic shock in severe sepsis patients. Crit Care Med. Apr 2017;45(4):623-9.
4. Gaieski DF, Mikkelsen ME, Band RA, et al. Impact of time to antibiotics on survival in patients with severe sepsis or septic shock in whom early goal-directed therapy was initiated in the emergency department. Crit Care Med. 2010;38:1045-53.
5. Ferrer R, Martin-Loeches I, Phillips G, et al. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guide- line based performance improvement program. Crit Care Med. 2014;42:1749-55.
6. Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of ef- fective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med. 2006;34:1589-96.
7. Seymour CW, Gesten F, Prescott HC, et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med. 2017;376:2235-44.
8. Solensky R. hypersensitivity reactions to beta-lactam antibiotics. Clin Rev Allergy Immunol. 2003 Jun;24(3):201-19.
9. Macy E. Penicillin and beta-lactam allergy: epidemiology and diagnosis. Curr Allergy Asthma Rep. Nov 2014;14(11):476.
10. Huang KG, Cluzet V, Hamilton K, Fadugba O. The impact of reported beta-lactam al- lergy in hospitalized patients with hematologic malignancies requiring antibiotics. Clin Infect Dis. Jan 18 2018;67(1):27-33. https://doi.org/10.1093/cid/ciy037.
11. MacFadden DR, LaDelfa A, Leen J, et al. Impact of reported beta-lactam allergy on in- patient outcomes: a multicenter prospective cohort study. Clin Infect Dis. Oct 1 2016;63(7):904-10.
12. IBM Corp. Released. IBM SPSS Statistics for Windows, version 23.0. Armonk, NY: IBM

Corp; 2014.

1. Gadde J, Spence M, Wheeler B, Adkinson NF. Clinical experience with penicillin skin testing in a large inner-city STD clinic. JAMA. 1993;270:2456-63.
2. ACAAI Public Website. drug allergies. Available from: https://acaai.org/allergies/ types/drug-allergies; 2018. [Accessed 28 March 2018]. [online].