Sorry, you need to enable JavaScript to visit this website.
Skip to main content

Delayed Sepsis Management Due to Ambiguous Allergy

Kimberly G. Blumenthal, MD, MSc | June 1, 2019
View more articles from the same authors.

The Case

A 75-year-old man with a past medical history of hemorrhagic stroke, coronary artery disease, and severe aplastic anemia required immunosuppressive therapy (antithymocyte globulin, cyclosporine, and prednisone) and regular blood and platelet transfusions. On the day of presentation, the patient arrived at the infusion center for a scheduled platelet transfusion. Prior to starting the infusion, he became unresponsive and hypotensive (59/26 mm Hg). The patient was immediately transported to the emergency department (ED) where, after resuscitation, he described a 1-day history of lethargy, abdominal discomfort, diaphoresis, and bloody bowel movements. Additionally, the patient had an allergy to penicillin, with a reaction of hives documented in the electronic health record. However, it was unclear when this reaction had occurred or whether it had been witnessed by a health care provider.

In the ED, the patient's vital signs were initially stable: temperature 36.7°C, blood pressure 110/60 mm Hg, heart rate 57 beats per minute, respiratory rate 18 breaths per minute, and SpO2 94%. Physical examination was notable for a tired-appearing man with slight abdominal distention without significant tenderness or peritoneal findings. A complete blood count was notable for platelets of 8000/µL, white blood cell count of 2300/µL, and absolute neutrophil count of 1230/µL. Chemistry panel and liver function tests were within normal limits, and urinalysis was unremarkable.

Soon after the initial evaluation (during which he appeared hemodynamically stable), the patient developed hypothermia (35.3°C), hypotension (70/40 mm Hg), and worsening mental status. The patient promptly received 3 L of intravenous fluid. Given continued instability and concern for septic shock, vancomycin was administered. However, neither gram-negative nor anaerobic coverage were ordered due to concern for the penicillin allergy and unknown infection source.

The patient remained hypotensive, and a lactate level returned at 3.5 mg/dL. A CT scan of the abdomen/pelvis with contrast suggested colitis as a possible source of infection. The absence of gram-negative coverage went unrecognized until the primary admitting team took over 4 hours after the patient's initial presentation, at which point they ordered aztreonam. However, the aztreonam was not administered for another 2 hours. As the aztreonam was being given, anaerobic coverage with metronidazole was also ordered but not administered for another 9 hours. Thus, despite the patient's immunocompromised state (neutropenia and on active immunosuppressive therapy) and sepsis from a likely intraabdominal source, ordering and administration of antibiotics with gram-negative and anaerobic coverage was significantly delayed.

Following the patient's prolonged ED course, he was admitted to the intensive care unit where his sepsis worsened. Antimicrobial therapy was broadened to meropenem, caspofungin, azithromycin, and vancomycin. The patient developed anuric renal failure necessitating continuous renal replacement therapy. Ultimately, he was transitioned to comfort care and died. Autopsy revealed the likely cause of death to be septic shock due to neutropenic enterocolitis.

The cause of the patient's death was deemed to be multifactorial (i.e., high medical complexity, delayed sepsis recognition, unfamiliarity with antibiotic coverage, prolonged ED boarding time, and multiple involved services). However, given the data on relationship between time to antibiotic administration and sepsis mortality, the delayed antibiotic administration likely contributed to the death. Furthermore, the documented penicillin allergy may have contributed to this delay by preventing reflexive administration of empiric gram-negative antibiotic treatment with piperacillin-tazobactam or cefepime. Moreover, the lack of familiarity with the limited efficacy of aztreonam in anaerobic coverage led to a delay in the addition of appropriate anaerobic treatment.

The Commentary

by Kimberly G. Blumenthal, MD, MSc

Sepsis affects more than 1 million Americans per year and is a leading cause of death in the United States.(1) Sepsis requires swift recognition and management of infection, with beta-lactam antibiotics (often piperacillin-tazobactam or cefepime) indicated as empiric therapy. Additional coverage of resistant gram-positive bacteria may be required with vancomycin, and as in this case, anaerobe coverage may be needed for enteric organisms.

Mortality in sepsis increases even with very short delays in antimicrobial administration.(2) The patient in this case did not have gram-negative coverage for 6 hours, and the penicillin allergy history likely contributed to the delay to first dose of antibiotic. Among patients presenting to the ED with pneumonia, urinary tract infection, bacteremia, and sepsis, patients with a penicillin allergy had a longer mean time to first antibiotic dose than patients without a penicillin allergy history (236.1 minutes vs 186.6 minutes, p=0.03).(3) The antibiotic ultimately administered in this case was aztreonam, which is considered inferior to piperacillin-tazobactam and cefepime, especially for Pseudomonas species. Similarly, one prior study found patients with beta-lactam allergy histories experienced higher clinical failure for gram-negative bloodstream infections compared to patients without (39% vs 27%, p=0.03).(4)

Patients who have a history of penicillin allergy comprise approximately 10% of the US population.(5) However, many penicillin allergic patients actually are not allergic to the drug, yet the allergy label prevents them from being treated with beta-lactams. In fact, patients with historical penicillin reactions have their allergy disproved more than 95% of the time after undergoing formal testing.(5) Additionally, drugs on a patient's allergy list often include adverse effects and intolerances. These nonimmunologic reactions may be considered allergies by patients, but often do not warrant penicillin avoidance.(6) A history of "hives" or "rash" are among the most commonly reported reactions, but may have been from a childhood viral exanthem rather than a drug allergy. Even if there were an immunoglobulin E (IgE)-mediated penicillin reaction, penicillin allergy wanes over time. This means that years later, patients are unlikely to be allergic.(5)

Shared chemical structures between penicillins and cephalosporins can be the source for clinical cross-reactivity, including the beta-lactam ring and side chains (e.g., ampicillin shares a side chain with the cephalosporins cefaclor and cephalexin). However, the later generation cephalosporins, such as cefepime, are tolerated in the overwhelming majority of cases, even in patients who have anaphylactic penicillin allergy histories.(7) In patients with a history of nonsevere penicillin allergy, treatment with cephalosporins without preceding penicillin skin testing leads to a reaction rate lower than the expected new rate of reactions to cephalosporins.(7) Even in patients with proven penicillin allergy, only about 2% react to treatment with cephalosporins.(7) Higher reaction rates might be observed in acutely ill patients—one study found 4.4% reactions to cefepime and 1.2% reactions to ceftazidime in patients with IgE-mediated penicillin allergy histories not confirmed with testing.(8) Carbapenem use in patients with penicillin allergy histories is also safe; even among 56 acutely infected hospitalized patients with IgE-mediated penicillin allergy histories, none reacted to carbapenems.(8)

The choice to not give a penicillin, or beta-lactam antibiotic, due to a penicillin allergy history may have other immediate and longer term consequences. Use of non–beta-lactam alternatives in patients reporting penicillin allergies leads to more treatment failures in methicillin-sensitive Staphylococcus aureus bacteremia and more surgical site infections.(9,10) Infected hospitalized patients not treated with a beta-lactam had a threefold increased odds of adverse events that included drug toxicities and Clostridium difficile colitis.(11) Penicillin allergy reporters also had a 26% increased risk of Clostridium difficile and a 69% increased incidence of methicillin-resistant Staphylococcus aureus colonization or infection over a mean follow-up time of 6 years.(12)

To make better antibiotic choices, it is crucial to do a systematic evaluation of the history of penicillin response, beginning with a thorough allergy history. Many patients report symptoms that were highly unlikely to be a true allergy, such as family history or gastrointestinal intolerance. Patients whose allergy history is an adverse effect or intolerance should have the "allergy" deleted or the "reaction type" changed to indicate there was no real allergy. When time and resources permit, penicillin allergies that may be IgE-mediated (such as hives, angioedema, shortness of breath, anaphylaxis) can be evaluated with a penicillin skin test. While historically performed by allergy specialists, penicillin allergy testing in hospital-based patients has been performed by nurses, pharmacists, and physicians from other specialties (medicine, infectious diseases). Penicillin skin testing takes about 1 hour to perform. If tests are negative, one full dose of a penicillin (usually amoxicillin) is given under observation. If there is no reaction, then the penicillin allergy is deleted from the record. When penicillin skin testing is not possible (or not necessary because of low risk), drug challenges or test doses can be considered, either with amoxicillin to disprove the underlying penicillin allergy or with the indicated therapeutic beta-lactam. One example test dose protocol gives 1/10th drug dose followed by 100% dose 1 hour later under observation.

Decision support around penicillin allergies can be incorporated into antibiotic prescribing and sepsis bundles to ensure the quick and safe administration of a beta-lactam antibiotic. Decision support might indicate which patients can have piperacillin-tazobactam and cefepime. For patients with higher risk beta-lactam allergy histories, decision support can still indicate that full dose carbapenems and the monobactam aztreonam are safe for use. At Massachusetts General Hospital and other academic and community Partners Healthcare System hospitals in the greater Boston area, we use test dose challenges in the ED and inpatient areas to optimize beta-lactam antibiotic use in patients with penicillin and cephalosporin allergy histories.(13) Our data indicate that inpatient penicillin and cephalosporin use increased approximately twofold; moreover, among more than 1000 test doses analyzed, fewer than 4% of patients had reactions (three patients needed intramuscular epinephrine).(8,14) Our penicillin allergy algorithm modified based on our most recent data is shown in the Figure. Clinical decision support tools to encourage standardized approaches to penicillin allergy in inpatients might include alerts, order sets, or web-based tools.(13) Alternative approaches to skin testing and test dose methods include implementation of allergy consultations (even virtually with telehealth [15]) and/or drug desensitization protocols to facilitate safe drug administration in patients who are truly allergic (e.g., recurrent IgE-mediated reactions or skin test positive), or too acutely ill to find out (i.e., the patient would be unlikely to recover from anaphylaxis).

Allergy documentation in health records is often inaccurate and incomplete. Although all allergy entries warrant attention to improve quality and safety, penicillin allergy documentation is a priority.(6) Large-scale efforts to improve the characterization of penicillin allergies include a recent professional society–endorsed review (5) and Centers for Disease Control and Prevention support in proclaiming a National Penicillin Allergy Day (September 28).(16)

Take-Home Points

  • Unverified penicillin allergy leads to increased treatment failures and delay to first antibiotic dose in sepsis patients.
  • Most patients who have a history of penicillin allergy are not allergic when the allergy is formally evaluated.
  • Even patients with true, IgE-mediated penicillin allergy can safely receive most cephalosporins, and all carbapenems and monobactams.
  • Other harms of unverified penicillin allergy labels include increased risk of adverse events, Clostridium difficile infection, surgical site infection, and colonization with antibiotic resistant organisms.
  • Systemwide approaches to hospitalized patients with beta-lactam allergies are needed to improve the quality and safety of care.

Kimberly G. Blumenthal, MD, MSc
Assistant Professor of Medicine
Quality and Safety Officer for Allergy
Professor of Radiology and Neuroradiology


1. Howell MD, Davis AM. Management of sepsis and septic shock. JAMA. 2017;317:847-848. [go to PubMed]

2. Kim RY, Ng AM, Persaud AK, et al. Antibiotic timing and outcomes in sepsis. Am J Med Sci. 2018;355:524-529. [go to PubMed]

3. Conway EL, Lin K, Sellick JA, et al. Impact of penicillin allergy on time to first dose of antimicrobial therapy and clinical outcomes. Clin Ther. 2017;39:2276-2283. [go to PubMed]

4. Jeffres MN, Narayanan PP, Shuster JE, Schramm GE. Consequences of avoiding ß-lactams in patients with ß-lactam allergies. J Allergy Clin Immunol. 2016;137:1148-1153. [go to PubMed]

5. Shenoy ES, Macy E, Rowe T, Blumenthal KG. Evaluation and management of penicillin allergy: a review. JAMA. 2019;321:188-199. [go to PubMed]

6. Blumenthal KG, Park MA, Macy EM. Redesigning the allergy module of the electronic health record. Ann Allergy Asthma Immunol. 2016;117:126-131. [go to PubMed]

7. Macy E, Blumenthal KG. Are cephalosporins safe for use in penicillin allergy without prior allergy evaluation? J Allergy Clin Immunol Pract. 2018;6:82-89. [go to PubMed]

8. Blumenthal KG, Li Y, Hsu JT, et al. Outcomes from an inpatient beta-lactam allergy guideline across a large US health system. Infect Control Hosp Epidemiol. 2019;40:528-535. [go to PubMed]

9. Blumenthal KG, Parker RA, Shenoy ES, Walensky RP. Improving clinical outcomes in patients with methicillin-sensitive Staphylococcus aureus bacteremia and reported penicillin allergy. Clin Infect Dis. 2015;61:741-749. [go to PubMed]

10. Blumenthal KG, Ryan EE, Li Y, Lee H, Kuhlen JL, Shenoy ES. The impact of a reported penicillin allergy on surgical site infection risk. Clin Infect Dis. 2018;66:329-336. [go to PubMed]

11. MacFadden DR, LaDelfa A, Leen J, et al. Impact of reported beta-lactam allergy on inpatient outcomes: a multicenter prospective cohort study. Clin Infect Dis. 2016;63:904-910. [go to PubMed]

12. Blumenthal KG, Lu N, Zhang Y, Li Y, Walensky RP, Choi HK. Risk of methicillin resistant Staphylococcus aureus and Clostridium difficile in patients with a documented penicillin allergy: population based matched cohort study. BMJ. 2018;361:k2400. [go to PubMed]

13. Blumenthal KG, Shenoy ES, Wolfson AR, et al. Addressing inpatient beta-lactam allergies: a multihospital implementation. J Allergy Clin Immunol Pract. 2017;5:616-625.e7. [go to PubMed]

14. Blumenthal KG, Wickner PG, Hurwitz S, et al. Tackling inpatient penicillin allergies: assessing tools for antimicrobial stewardship. J Allergy Clin Immunol. 2017;140:154-161.e6. [go to PubMed]

15. Staicu ML, Holly AM, Conn KM, Ramsey A. The use of telemedicine for penicillin allergy skin testing. J Allergy Clin Immunol Pract. 2018;6:2033-2040. [go to PubMed]

16. National Penicillin Allergy Day. [Available at]

17. Wolfson AR, Huebner EM, Blumenthal KG. Acute care beta-lactam allergy pathways: approaches and outcomes. Ann Allergy Asthma Immunol. 2019 Apr 19; [Epub ahead of print]. [go to PubMed]


Figure. Partners HealthCare System Penicillin Allergy Pathway. Reprinted from (17) with permission from Copyright Clearance Center.

Figure of Partners HealthCare System Penicillin Allergy Pathway

Test dose procedures use 1/10th of the dose for parenteral beta-lactams and 25% dose for oral beta-lactams followed by 30–60 minutes of observation. If there is no reaction, 100% of the dose is administered with 60 minutes of observation. If there is no reaction, the next dose of the beta-lactam antibiotic is given per its usual therapeutic schedule. Unlike desensitization procedures, which typically require intensive care unit monitoring, test dose procedures can be performed on the medical floors. Common cephalosporins by generation: 1st cephalexin/cefazolin; 2nd cefoxitin/cefuroxime; 3rd ceftriaxone/cefixime/cefotaxime/cefpodoxime/ceftazidime*; 4th cefepime; 5th ceftaroline* Abbreviations: HSR, hypersensitivity reaction; PCN, penicillin * Antibiotics restricted by Partners antimicrobial stewardship committees. § Alternative agents by microbial coverage: Gram positive coverage: Vancomycin, linezolid*, daptomycin*, clindamycin, doxycycline, trimethoprim/sulfamethoxazole Gram negative coverage: Quinolones, sulfamethoxazole/trimethoprim, aminoglycosides, aztreonam*

This project was funded under contract number 75Q80119C00004 from the Agency for Healthcare Research and Quality (AHRQ), U.S. Department of Health and Human Services. The authors are solely responsible for this report’s contents, findings, and conclusions, which do not necessarily represent the views of AHRQ. Readers should not interpret any statement in this report as an official position of AHRQ or of the U.S. Department of Health and Human Services. None of the authors has any affiliation or financial involvement that conflicts with the material presented in this report. View AHRQ Disclaimers
Related Resources