Penicillin antibiotics were first introduced to the market in the 1940s and have since become agents of choice for surgical prophylaxis and the treatment of numerous bacterial infections.1,2 The β-lactam drug class offers several advantages compared with other antibiotic classes, such as decreased cost, incidence of adverse effects, and emergence of antimicrobial resistance.3
Approximately 25% of patients in the United States self-report an allergy to at least 1 antimicrobial agent.4 Penicillin is the most commonly reported antimicrobial allergy, as well as the most commonly reported drug allergy. An estimated 10% of the population reports a penicillin allergy; however, approximately 80% to 90% of these patients have a negative penicillin skin test result for an immunologic (IgE)-mediated allergy.4–6 The incidence of anaphylaxis associated with the penicillin drug class is estimated to be as low as 0.015% to 0.004%, with an estimated fatality rate of 0.0015% to 0.002%.7,8 Studies have shown that most patient-reported penicillin allergies are not immunologically mediated. Instead, these reported allergies are common adverse drug reactions and side effects. This has been confirmed by the low incidence of positive penicillin skin tests and documented immediate type 1 allergic reactions for the past several years.1,9 Furthermore, studies have described a loss of antipenicillin IgE antibodies over time; therefore, as patients age, they could eventually become nonallergic.9
Historically, the reported risk of cross-reactivity between penicillin and cephalosporin antibiotics was estimated to be 10% and is partly due to the manufacturing of these products. This percentage is also confounded by nonimmunologic reactions and retrospective studies without control groups or confirmation of allergy by skin testing. Penicillins, cephalosporins, and carbapenems are all structurally similar because of the presence of a β-lactam ring. However, it has been posed that the cross-reactivity may be more highly correlated to the various side chains rather than the β-lactam ring. More recent studies estimate the rate of cross-reactivity between penicillins and first-generation cephalosporins is less than 5%, the cross-reactivity between penicillins and second-, third-, and fourth-generation cephalosporins is less than 2%, and the cross-reactivity between penicillins and carbapenems is less than 1%.6
Although minimal, this potential risk of cross-reactivity between penicillin and other β-lactam antibiotics has prohibited many penicillin-allergic patients from receiving first-line β-lactam agents. Instead, they are more likely to be treated with alternative antibiotics, such as fluoroquinolones, aztreonam, vancomycin, and clindamycin, compared with those patients without a reported antibiotic allergy.2,5 An estimated 30% of patients with antibiotic allergies receive antibiotic regimens that differ from the standard of care.10 Penicillin-allergic patients are twice as likely to receive vancomycin and 3 times as likely to receive fluoroquinolone antibiotics compared with nonallergic patients.11 These alternative antibiotics are also commonly associated with complications such as Clostridium difficile, renal failure, central nervous system reactions, cardiac and liver toxicity, increased cost, and emergence of multidrug-resistant organisms such as vancomycin-resistant enterococcus and extended-spectrum β-lactamase–producing Enterobacteriaceae.3,5 The estimated cost of antimicrobials increases 63% for inpatients and 38% for outpatients with penicillin allergies compared with nonallergic patients.12 Recent reports have also shown that penicillin allergies have been associated with increased length of hospital stay and increased mortality.1,3 It is evident that patients with a reported penicillin allergy are at a disadvantage compared with patients without an allergy because β-lactams are preferred for the treatment of many infectious diseases.
The use of penicillin skin testing as part of an antimicrobial stewardship program has gained popularity for the past few years. This is due to the significant number of patients who report penicillin allergies each year in combination with increased antimicrobial resistance and limited treatment options. Penicillin skin testing has proven to be a safe and reliable antimicrobial stewardship initiative for distinguishing true penicillin allergies to guide antibiotic therapy. The negative predictive value of penicillin skin testing is approximately 97% to 99% and therefore allows for the safe transition to β-lactam therapy when appropriate.1,2
A formal allergy assessment is an essential antimicrobial stewardship tool for the prevention of antibiotic treatment failure and emergence of antimicrobial resistance, especially with the high rate of inaccurate allergy labeling.4,7,13 At Morton Plant Hospital, patients with a reported antibiotic allergy undergo an allergy assessment to determine the severity of the reported reaction and screen for appropriateness of penicillin skin testing per the “Penicillin Skin Test (Pre-Pen) Protocol.” The objective of the Penicillin Skin Test (Pre-Pen) Protocol is to provide a standardized assessment questionnaire to be used as a stewardship tool and to assess the appropriateness of penicillin skin testing. The pharmacists are educated to recommend direct transition to β-lactam therapy if the patient describes an intolerance or side effect, graded challenge if the patient does not describe an IgE type reaction, and penicillin skin testing for a remote or historical (>5 years) IgE or unknown reaction. The pharmacists recommend alternative antibiotics if patients describe life-threatening reactions, Stevens-Johnson syndrome, or toxic epidermal necrolysis. The ultimate goal of this protocol was to guide antimicrobial therapy recommendations and assist in antimicrobial stewardship efforts.
Novel antimicrobial stewardship initiatives are crucial to decrease the prevalence of multidrug-resistant pathogens and also decrease morbidity and mortality at a time when the development of new antibiotic agents is slow-moving and antimicrobial resistance is increasing. Although information has been published on the utility of penicillin skin testing, there are no antimicrobial stewardship studies that evaluate the impact of an allergy assessment program. Thus, the primary purpose of this study is to compare the percentage of patients who received β-lactam therapy for the treatment of bacterial infections before and after the implementation of a formal allergy assessment protocol. This study also compared the clinical outcomes and incidence of adverse events in patients who underwent an allergy assessment to guide antibiotic therapy compared with patients who did not receive an allergy assessment during their hospitalization.
MATERIALS AND METHODS
Study Design and Patient Population
This was a prospective study design, which compared hospitalized adult inpatients with a reported β-lactam allergy at Morton Plant Hospital with those at a sister hospital within the Morton Plant Mease HealthCare System between October 2014 and April 2015. The frequency of conversion to β-lactam therapy between an intervention group of patients who received an allergy assessment was compared with a control group of patients who were not treated under a formal allergy assessment protocol. Secondary outcomes evaluated differences in all-cause mortality, hospital length of stay, 30-day readmission rates, and incidence of adverse events such as C. difficile, acute kidney injury, gastrointestinal upset, and infusion-related reactions between the 2 groups. Duration of therapy was further broken down into deep-seated and nondeep-seated infections. Bacteremia, osteomyelitis, and endocarditis were considered deep-seated infections, whereas pneumonia, urinary tract infections, intra-abdominal infections, and cellulitis were considered nondeep-seated infections.
The intervention group consisted of adult inpatients 18 years or older admitted to Morton Plant Hospital, a 687-bed community hospital in Clearwater, Florida, with a documented β-lactam allergy who were receiving at least 1 alternative antibiotic. Patients were excluded if they were unable to provide an accurate allergy history, only receiving antibiotics perioperatively, or if they were infected with a multidrug-resistant organism. Patients who met inclusion criteria underwent a detailed allergy assessment from a pharmacist within 24 hours of the antibiotic order. All pharmacists used a standard allergy assessment form, which was developed at Morton Plant Hospital (see Supplemental Material, http://links.lww.com/IDCP/A19). The pharmacist then provided antibiotic therapy recommendations to the provider on the basis of the results of the allergy assessment. Pharmacist recommendations could include continuation of current therapy, immediate switch to a β-lactam agent, penicillin skin testing, desensitization, or graded challenge. All clinical pharmacists receive education on the hospital-wide protocol and available literature regarding penicillin allergies and are asked to use their clinical judgment to make recommendations with the protocol and educational materials as a guide. Patient-specific empiric antimicrobial therapy recommendations are based on local susceptibility patterns and published guidelines. The primary investigators of the study are heavily involved in allergy assessments and are available for questions as well. The pharmacist was also responsible for updating and clarifying allergies in the electronic medical record after patient discussions. The pharmacists did not remove any allergies from the patient chart unless the patient had a negative penicillin skin test result.
Patients who met inclusion criteria at the sister hospital comprised the control group and were matched to patients in the intervention group on the basis of infectious disease diagnosis at discharge. Retrospective patient selection included screening of all patients at a sister hospital within the BayCare Health Systems who had a β-lactam allergy and were receiving at least 1 non–β-lactam antibiotic within the same time period of October 2014 through April 2015. The sister hospital served as a control because they did not have an approved allergy assessment protocol at the time of this study.
Patients in both groups were screened using the clinical surveillance decision support tool Theradoc (Premier, Inc, Salt Lake City, Utah). The pharmacy staff was educated on the allergy assessment procedure and appropriate recommendations for physicians. The patients were followed for clinical outcomes and adverse events. Outcome data were gathered through extensive chart reviews of the electronic medical record. Approval for this study was granted by the BayCare Institutional Review Board.
Sample Size Determination and Statistical Analysis
To the best of our knowledge, there are no published studies comparing outcomes after the implementation of a formal allergy assessment tool. Therefore, we assumed that an effect size of 20% would adequately power the study to detect a minimally important difference in proportion of patients receiving β-lactam therapy. Sample size calculations were made using MiniTab 16 Statistical Software (Minitab, Inc, State College, Pa).14 The MiniTab calculator for 2 proportions estimated a sample size of 82 patients in each arm (total sample size = 164) for this study to have 80% power for detecting a 20% increase in the primary end point at a 2-sided α level less than or equal to 0.05.
Statistical analyses of the primary and secondary outcomes were performed using the MiniTab 16 Statistical Software as well. Discrete and continuous variables were compared using a Student t test or descriptive statistics. Nominal data were compared using a Pearson χ2 test or a Fisher exact test.
A total of 148 adult inpatients with a documented β-lactam allergy on at least 1 non–β-lactam antibiotic were identified and screened for inclusion into the intervention (allergy assessment) group. Of these 148 patients, 63 patients were included and 85 patients were excluded from the study. The most common reason patients were excluded was perioperative antibiotic orders. These patients were excluded because they generally received antibiotics before an allergy assessment could be completed. A detailed flowchart of patient enrollment and exclusions can be found in Figure 1.
A total of 63 patients were included in the intervention group and matched on the basis of infectious disease diagnosis to 63 patients who served as the control group. The matched infection types included the following: pneumonia (n = 18), urinary tract infection (n = 10), abscess (n = 8), intra-abdominal infection (n = 7), bacteremia (n = 7), cellulitis (n = 3), osteomyelitis (n = 3), sepsis (n = 3), neutropenic fever (n = 2), and endocarditis (n = 2). A total of 126 patients between both groups were compared and analyzed. Detailed patient demographics and baseline characteristics are presented in Table 1. There were no significant clinical or statistical differences in baseline characteristics between groups.
It was found that 28 (44%) of 63 patients with a β-lactam allergy did not have a reaction listed in the electronic medical record or reported an unknown reaction. Other reactions recorded in the medical record included rash (n = 13), anaphylaxis (n = 8), hives (n = 7), itching (n = 2), and other reactions were reported by 5 patients. The percentage of patients transitioned to β-lactam therapy was 57% (n = 36) in the allergy assessment group and 22% (n = 14) in the control group (P = 0.019). A total of 7 (11%) of 63 pharmacist recommendations to switch to β-lactam therapy in the intervention group were rejected by physicians. The recommendations were recommended for various reasons including physician preference and improvement on current regimen.
In the 36 patients transitioned to β-lactam therapy, 19 (53%) were initially on aztreonam. Cefepime was the most common antibiotic that patients were transitioned to, followed by cefazolin (42% and 17%, respectively). Of the 36 patients transitioned to β-lactam therapy in the intervention group, 9 patients underwent penicillin skin testing (Pre-Pen, ALK, Round Rock, Tex). Penicillin skin testing was ordered by an infectious disease physician and administered and read by a trained infection control practitioner. The results of the penicillin skin test were documented in the medical record by the pharmacist who was present for the administration and patient education. Eight of the 9 patients had negative penicillin skin test results and were successfully transitioned to antibiotics within the penicillin drug class. One patient was on high-dose corticosteroids and did not respond to the histamine control; therefore, the test was indeterminate. No patients in the intervention group underwent desensitization or graded challenge.
The mean (SD) duration of antibiotic therapy in the intervention and control groups was 13 (11.8) days and 14.6 (11.9) days (P = 0.45), respectively. Although the mean length of antibiotics was not significantly different when all infection types were compared, it was significantly different with regard to nondeep-seated infections. The mean (SD) durations of antibiotic therapy for nondeep-seated infections were 8.6 (5.5) and 11.2 (6.8) days in the intervention and control groups, respectively (P = 0.018). There were no significant statistical differences in the mean length of hospital stay, all-cause mortality, 30-day all-cause readmission rate, or adverse events between groups. Similar to all-cause 30-day readmission rates, readmissions secondary to recurrent infection were similar between groups at 5 (8%) of 63 in-patients in the intervention group and 3 (5%) of 63 patients in the control group. Although not statistically significant, there were fewer deaths in the intervention group at 1 (2%) of 63 in-patients compared with 4 (6%) of 63 patients in the control group (P = 0.168). Interestingly, there were no infusion-related or allergic reactions reported in the intervention group despite transition to β-lactam antibiotics in patients with a documented allergy. The results of the primary and secondary outcomes are shown in Table 2 and Figure 2.
The results of this study demonstrate the importance of incorporating allergy assessments into daily antimicrobial stewardship activities. Patients with allergy histories require special consideration because studies have shown increased morbidity and mortality associated with allergy-labeled patients.1,3 Although allergies are reviewed with every patient during the intake process, the pharmacists in our study gathered more detailed information from their allergy assessments, which gave physicians more comfort in transitioning patients to β-lactam therapy. It was discovered that the greatest reservation providers had with regard to challenging patient allergies was the risk of a serious allergic or anaphylactic reaction when the patients' allergy history was unknown or unclear. When physicians were presented with detailed allergy histories and patients were willing to challenge their own allergies, the providers were generally very willing to prescribe β-lactam antibiotics unless a severe anaphylactic reaction was reported. Our allergy assessments found that most patients who reported a β-lactam allergy described a reaction consistent with medication intolerances or adverse effects rather than a true IgE-mediated reaction. We also found that patients were very willing to challenge their historical allergies after they were presented with information from pharmacists regarding the low rates of β-lactam cross-reactivity, minimal risks of anaphylaxis, improved manufacturing processes of penicillin antibiotics, potential to outgrow allergies over time, and the positive aspects of the β-lactam drug class.
Patients who described historical IgE-mediated anaphylactic reactions to penicillin, which occurred greater than or equal to 5 years ago, were often screened for penicillin skin testing to ensure safe conversion to β-lactam therapy. Of the 9 patients who underwent the penicillin skin testing procedure, 8 successfully tolerated β-lactam antibiotic therapy and 1 had an indeterminate result to the scratch test and therefore did not complete the penicillin skin testing procedure. Interestingly, this patient was ultimately transitioned to β-lactam therapy for treatment of bacteremia despite the indeterminate penicillin skin test. Although our sample size of patients who underwent penicillin skin testing was small, these findings of a 100% negative predictive value and negative penicillin skin test result in 89% of patients tested are consistent with previous reports in the literature. A study by Rimawi et al13 reported that 99% of the 146 patients on which they performed penicillin skin tests had a negative result and reported that the negative predictive value of the test was 100%.10 Other studies by Park et al and Arroliga et al2 reported that 94% and 95% of patients had negative penicillin skin test results, respectively.2,15 Interestingly, our study found that there were no incidences of infusion-related or allergic reactions in patients who were transitioned to β-lactam therapy on the basis of recommendations from the allergy assessments whether or not the patients underwent penicillin skin testing. These findings further illustrate the inaccuracy of allergy labels within the medical records and highlight the low risk of allergic hypersensitivity reactions in patients with a history of a penicillin allergy.
The greatest success in accepted recommendations to transition patients to β-lactam antibiotics involved timely identification of appropriate patients and allergy evaluation. Physicians were also more willing to change therapy if a complete allergy history was provided and especially if the patient had tolerated a β-lactam antibiotic agent in the past. Finally, physicians were willing to transition to β-lactam therapy 100% of the time if the patient had a negative penicillin skin test result.
Limitations of this study include the retrospective component of the comparator arm. This forces the investigators to rely on the accuracy of the electronic medical record and potentially introduces bias and confounding with regard to the selection of the controls. For instance, the severity of illness was not taken into consideration when matching the intervention group to the controls. There was also a lack of pretest in the 2 comparator institutions, which can lead to potential for selection bias. Although 63 patients were prospectively enrolled in this trial, we did not meet our initial sample size goal of 82 patients in each arm. Also, the study was not truly powered for the secondary endpoints. Prospectively, although pharmacists were trained on conduction of allergy assessments and appropriate therapeutic recommendations, interpharmacist variability in recommendations may be present on the basis of their individual clinical judgment. Other limitations include delays to allergy assessment completion, which may have caused interventions to be increasingly difficult. It is unknown whether pharmacists at the sister hospital conducted informal allergy assessments and recommended conversion to β-lactam therapy within their daily clinical responsibilities, which could have also confounded the data. Finally, pharmacoeconomic analyses were not preformed. Although these are limitations to the study, these inconsistencies will always play a role in a clinical environment.
Standardized, pharmacist-led allergy assessments allow for easier conversion to first-line β-lactam antibiotic therapy. The results of our study show that most patients who underwent an allergy assessment were converted to β-lactam antibiotic therapy without any complications and tolerated therapy well. No infusion-related or anaphylactic reactions occurred after the conversion to β-lactam therapy in previously labeled β-lactam allergic patients. This study indicates that formal allergy assessments may serve as an essential part of every antimicrobial stewardship program.
The authors thank the pharmacists at Morton Plant Hospital for all of their contributions in enrolling patients into this study.
1. Unger NR, Gauthier TP, Cheung LW. Penicillin skin testing: potential implications for antimicrobial stewardship. Pharmacotherapy
2. Arroliga ME, Wagner W, Bobek MB, et al. A pilot study of penicillin skin testing in patients with a history of penicillin allergy admitted to a medical ICU. Chest
3. Macy E, Contreras R. Health care use and serious infection prevalence associated with penicillin “allergy” in hospitalized patients: a cohort study. J Allergy Clin Immunol
4. Charneski L, Deshpande G, Weiss S. Impact of an antimicrobial allergy label in the medical record on clinical outcomes in hospitalized patients. Pharmacotherapy
5. Solensky R. Penicillin allergy as a public health measure. J Allergy Clin Immunol
6. Trubiano J, Phillips E. Antimicrobial stewardship’s new weapon? A review of antibiotic allergy and pathways to ‘delabeling’. Curr Opin Infect Dis
7. Forrest DM, Schellenberg R, Thien VV, et al. Introduction of a practice guideline for penicillin skin testing improves the appropriateness of antibiotic therapy. CID
8. Pichichero ME. A review of evidence supporting the American Academy of Pediatrics recommendation for prescribing cephalosporin antibiotics for penicillin-allergic patients. Pediatrics
9. Macy E, Ho NJ. Adverse reactions associated with therapeutic antibiotic use after penicillin skin testing. Perm J
10. Lutomoski Lafollette JA, Biaglow MA, et al. Antibiotic allergies in the medical record: effect on drug selection and assessment of validity. Pharmacotherapy
11. Lee CE, Zembower TR, Fotis MA, et al. Implications regarding prescribing patterns and emerging bacterial resistance. Arch Intern Med
12. Sade K, Holtzer I, Levo Y, et al. The economic burden of antibiotic treatment of penicillin-allergic patients in internal medicine wards of a general tertiary care hospital. Clin Exp Allergy
13. Rimawi RH, Cook PP, Gooch M, et al. The impact of penicillin skin testing on clinical practice and antimicrobial stewardship. J Hosp Med
14. Mini-tab Statistical Software, Version 16. Mini-tab.com. Accessed July 29, 2014.
15. Park MA, McClimon BJ, Ferguson B, et al. Collaboration between allergists and pharmacists increases beta-lactam antibiotic prescriptions in patients with a history of penicillin allergy. Int Arch Allergy Immunol