INTRODUCTION — Penicillin allergy is the most common drug allergy and is reported in up to 15 percent of hospitalized patients . A frequent clinical question is whether these patients can safely receive the structurally related cephalosporins or carbapenems (monobactams [ie, aztreonam] are not cross-reactive with penicillins and can be safely given to penicillin-allergic patients). This topic will present an approach to determining whether a patient with reported penicillin allergy can be treated with penicillins or related beta-lactam antibiotics, with or without access to allergy consultation. It is focused on the hospitalized patient but applies to ambulatory patients as well. The detailed allergy evaluation of a patient with past penicillin reaction based upon penicillin skin testing and studies of cross-reactivity among beta-lactam antibiotics are reviewed elsewhere.
Penicillin allergy (immediate and delayed), penicillin skin testing techniques, and rapid drug desensitization are reviewed in more detail separately:
●(See "Penicillin skin testing".)
IMPACT OF PENICILLIN ALLERGY ON CARE — There is morbidity, mortality, and economic cost associated with the unnecessary withholding of penicillins in patients who are labeled as allergic on the basis of history alone. Patients with a history of penicillin allergy are more likely to be treated with broad-spectrum antibiotics, such as quinolones or vancomycin [1-9]. There are distinct disadvantages to broad-spectrum agents, which are often more expensive, associated with more side effects (such as Clostridioides [formerly Clostridium] difficile infection), and less effective for some infections [6,10-13]. In addition, overuse of vancomycin and quinolones contributes to the development and spread of certain types of drug-resistant bacteria [7,14-17] (see "Vancomycin-resistant enterococci: Epidemiology, prevention, and control", section on 'Risk factors'):
●One study assessed various risk factors for the development of vancomycin-resistant enterococcus (VRE) in a medical intensive care unit (MICU) . Among the various classes of antibiotics used prior to transfer to the MICU, quinolones had the strongest association with subsequent development of VRE (odds ratio [OR] = 8.6), whereas treatment with penicillins/beta-lactamase inhibitors was not associated with developing VRE .
●In a study of over 51,000 hospitalized patients, patients labeled as penicillin-allergic had 23 percent more C. difficile infections, 30 percent more VRE infections than expected, 14 percent more methicillin-resistant Staphylococcus aureus (MRSA) infections, and longer hospital stays, compared with control subjects . The mechanism by which patients with a listed penicillin allergy would be at higher risk for MRSA is unknown, although greater time in the hospital is one possibility .
●In a multisite, prospective cohort analysis of inpatients who received an infectious diseases consultation, patients with a history of beta-lactam allergy who were not treated with a beta-lactam had more "adverse events" (composite outcome of readmissions for the same infection, acute kidney injury, C. difficile infection, and drug-related adverse events) than patients with beta-lactam allergy who were treated with a beta-lactam (adjusted OR 3.2, 95% CI 1.28-7.89) .
The implementation of an institutional program for addressing penicillin allergy as part of antimicrobial stewardship is a multifaceted process involving administrators, generalists, allergists, infectious diseases specialists, pharmacists, nursing personnel, and information technology support . (See "Antimicrobial stewardship in hospital settings" and "Antimicrobial stewardship in outpatient settings".)
OUR APPROACH — Many hospitals, including all United States facilities, are mandated to establish an antimicrobial stewardship program . However, access to an allergy and immunology consult service is often lacking. This topic presents an evidence-based, pragmatic approach, which permits a large majority of patients labeled as penicillin-allergic to receive related beta-lactam antibiotics with minimal risk of acute allergic reactions, without the necessity of allergy consultations. The approach is based on categorization of the past reaction and for those patients with possible immediate (eg, immunoglobulin E [IgE]-mediated) allergy, an assessment of the risk of a recurrent immediate reaction (algorithm 1). However, if an allergy specialist is available, consultation is strongly encouraged, especially when the algorithm suggests penicillin skin testing. (See 'Role of the allergy specialist' below.)
Clinical history and record review — The evaluation of a patient with a history of penicillin allergy begins with the clinical history and review of available medical records. Penicillins include the following:
●The extended-spectrum penicillins – Carbenicillin, ticarcillin, and piperacillin
Hypersensitivity reactions to any of these drugs would correctly be considered a penicillin allergy. It is useful to name the various commonly used penicillins (ie, amoxicillin, ampicillin, specific brand names), as patients may recognize a common brand name but not realize it is a penicillin. It is also worthwhile to ask if the patient has avoided penicillins since the reaction, because in some cases, patients had been given a penicillin again despite a reported allergy, and if it was tolerated, then the patient is not allergic to penicillins.
Important questions — Important questions in a drug allergy history include the following (algorithm 1):
1. What exactly were the symptoms?
●Raised, red, itchy spots with each lesion lasting less than 24 hours (hives/urticaria)?
●Swelling of the mouth, eyes, lips, or tongue (angioedema)?
●Blisters or ulcers involving the lips, mouth, eyes, urethra, vagina, or peeling skin (seen in Stevens-Johnson syndrome [SJS], toxic epidermal necrolysis [TEN], other severe type IV reactions)?
●Respiratory or hemodynamic changes (anaphylaxis)?
●Joint pains (seen in serum sickness)?
●Did the reaction involve organs like the kidneys, lungs, liver, or lymph nodes (seen in drug rash eosinophilia systemic symptoms [DRESS, also known as, drug-induced hypersensitivity syndrome [DiHS]), other severe type IV reactions) or were there other features such as extensive rash (>50 percent body surface area), fever, and facial edema?
2. What was the timing of the reaction after taking penicillin: Minutes, hours, or days later? Did it begin after the first dose or after multiple doses? The rapid onset of symptoms (ie, within minutes to hours) after the first dose distinguishes immediate from delayed reactions.
3. How long ago did the reaction happen? This question is most important for immediate reactions, because patients may lose the sensitivity with time if they have no further penicillin exposures. Specifically, studies have shown that if patients with immediate penicillin allergy that was confirmed by skin testing are retested after 10 years of penicillin avoidance, only approximately 20 percent will still have positive skin tests, while the allergy will have resolved in 80 percent. The relevant studies demonstrating loss of immediate allergy are reviewed elsewhere (see "Penicillin allergy: Immediate reactions", section on 'Time elapsed since the reaction'). In contrast, the natural history of serious delayed reactions is not as well-studied.
4. How was the reaction treated? Was epinephrine (adrenaline) administered to suggest anaphylaxis? Was there a need for urgent care or hospitalization?
5. Has the patient tolerated similar medications, such as a cephalosporin since the initial penicillin reaction? This information can be very helpful. The information that can be deduced from various answers is reviewed separately. (See "Penicillin allergy: Immediate reactions", section on 'Exposure to related medications since the initial reaction'.)
A more detailed discussion of taking a drug allergy history is found separately. (See "An approach to the patient with drug allergy", section on 'Clinical history'.)
Categorize the past reaction — Based on the clinical history and review of the available medical record, we attempt to categorize the patient's penicillin reaction into one of the following groups:
●Reactions that are not allergic. (See 'Reactions that are not allergic' below.)
●Possibly serious forms of delayed reactions. (See 'Possible serious delayed reactions' below.)
●Mild reactions WITHOUT features of immediate allergy. (See 'Management of mild reactions WITHOUT history of features of immediate allergy (minimal risk of immediate allergy)' below.)
●Reactions WITH features of immediate allergy. (See 'Management of reactions WITH features of immediate allergy (some risk)' below.)
Immediate reactions typically involve some combination of pruritus, urticaria, angioedema, bronchospasm, laryngeal edema, and/or hypotension, although an array of other symptoms may occur (table 1). These reactions usually occur within minutes to one hour or two of initial administration, beginning somewhat later for oral compared with intravenous medications. The most severe form of an immediate reaction is anaphylaxis. Patients with past anaphylaxis may report that they "could not breathe," "swelled up," or "almost died," and most patients would have required acute medical care. For the beta-lactam drugs, most immediate reactions are IgE-mediated, and the terms "IgE-mediated" and "immediate" are used interchangeably throughout this topic review. The clinical manifestations of immediate reactions to penicillins are reviewed in greater detail separately. (See "Penicillin allergy: Immediate reactions".)
The presence of IgE antibodies to penicillin can be reliably detected with penicillin skin testing. (See 'Penicillin skin testing' below.)
Management based upon category
Reactions that are not allergic — Some patients will describe reactions that are not allergic in nature, but instead are well-known adverse effects (eg, diarrhea, vomiting, yeast vaginitis). It should be explained to the patient that this is not an allergy, and penicillin allergy should be removed from the medical record or clarified to indicate that there was a side effect or intolerance. Patients with nonallergic reactions to penicillin can receive penicillins and related medications normally, although steps should be taken to avoid the specific agent that caused the previous adverse effect or to mitigate that adverse effect.
Family history of penicillin allergy — Other patients may never have actually taken penicillins, but are avoiding them because one or more family members has penicillin allergy. However, penicillin allergy is not a heritable trait, and family history is not a reason to avoid penicillin. Of note, there are drug reactions that are heavily influenced by genetics (eg, serious reactions to certain antiepileptics and antipsychotics).
Possible serious delayed reactions — A few patients may describe past reactions to penicillins that developed after days to weeks of therapy. Serious forms of delayed reactions to penicillins include:
●Toxic epidermal necrolysis (TEN) (see "Stevens-Johnson syndrome and toxic epidermal necrolysis: Pathogenesis, clinical manifestations, and diagnosis")
●Stevens-Johnson syndrome (SJS)
●Drug reaction with eosinophilia and systemic symptoms/drug-induced hypersensitivity syndrome (DRESS/DiHS) (see "Drug reaction with eosinophilia and systemic symptoms (DRESS)")
●Other exfoliating dermatoses/erythroderma (see "Erythroderma in adults")
●Serum sickness-like reactions (see "Serum sickness and serum sickness-like reactions")
●Drug-induced renal, hepatic, or other specific organ damage (see "Clinical manifestations and diagnosis of acute interstitial nephritis" and "Drug-induced liver injury")
●Acute generalized exanthematous pustulosis (AGEP)
These reactions are diagnosed by historical features and review of the patient's medical record to identify objective exam, laboratory, and/or biopsy results supportive of one of these diagnoses. Standardized diagnostic testing is not available for these types of reactions, although in vitro testing is an active area of drug allergy research. (See "Penicillin allergy: Delayed hypersensitivity reactions".)
Evaluation — For patients suspected of having one of the serious delayed types of hypersensitivity reactions to penicillin, all future use of penicillins should be avoided, since many of these reactions can recur (often more quickly) if the patient is exposed to the causative drug again.
Skin testing is not used in the diagnosis of these reactions, and there are no validated tests that can diagnose them retrospectively.
Future use of related antibiotics — There is very limited information in the published literature about whether patients with one of these reactions to a penicillin will experience the same reaction to a cephalosporin or carbapenems. Therefore, most clinicians also avoid related antibiotics in the future. However, if there is a strong indication for a cephalosporin or carbapenem, a drug allergy expert should be consulted if available to review the details of the case and assess the risk of restarting the same or similar medication. Examples of apparent lack of cross-reactivity with other beta-lactams in patients with delayed reactions to penicillins are reviewed elsewhere. (See "Penicillin allergy: Delayed hypersensitivity reactions".)
Management of mild reactions WITHOUT history of features of immediate allergy (minimal risk of immediate allergy) — The majority of patients describe mild cutaneous reactions, such as mild drug eruptions, with or without pruritus, which lack other obvious features of IgE-mediated reactions (eg, urticaria, angioedema, bronchospasm, etc). Other symptoms, such as isolated sneezing or isolated pruritus of the eyes or nose may be occasionally reported. Note that some caution is warranted even with mild past reactions, because the clinical history can be inaccurate. Studies evaluating the accuracy of clinical history in penicillin allergy are reviewed elsewhere. (See "Penicillin allergy: Immediate reactions", section on 'Clinical history'.)
Options for treatment depend upon how urgently treatment is needed and whether consultation with an allergist is readily available (algorithm 1):
Penicillin skin testing NOT feasible — The risk of cross-reactivity between penicillins and cephalosporins is low, and the risk with carbapenems is even lower. The best designed studies suggest that 99 percent of patients reporting a penicillin allergy (without prior penicillin skin testing) tolerate a cephalosporin, partly because most of these patients do not have an immediate allergy to penicillin [22,23]. (See "Allergy evaluation for immediate penicillin allergy: Skin test-based diagnostic strategies and cross-reactivity with other beta-lactam antibiotics".)
Options for treating such patients include the following (if appropriate for the infection in question):
●Give a penicillin or first- or second-generation cephalosporin using a test dose procedure. If the test dose is tolerated without symptoms, it proves the patient does not have an immediate allergy to the antibiotic that was given. (See 'Test dose procedure (graded challenge)' below.)
●Give a third- or fourth-generation cephalosporin normally (as these are believed to be possibly less cross-reactive with penicillins compared with early-generation cephalosporins).
●Give a carbapenem normally.
●Give aztreonam (a monobactam) or an antibiotic that is unrelated to the beta-lactams normally.
Skin testing feasible — If an allergist is available, then consultation and penicillin skin testing and challenge can determine if the patient with a history of an IgE-mediated acute allergic reaction is currently allergic or not with certainty. Since testing is negative in the vast majority of patients and they may subsequently receive any beta-lactams safely, this greatly simplifies future antibiotic choices. (See 'Role of the allergy specialist' below.)
Management of reactions WITH features of immediate allergy (some risk) — Immediate reactions typically involve flushing, pruritus, urticaria, angioedema, bronchospasm, laryngeal edema, and/or hypotension (table 1). Patients reporting these symptoms are at some risk for future immediate allergic reactions. Patients reporting severe or recent reactions are at higher risk than those with milder, remote reactions.
Options for treatment depend upon how urgently treatment is needed, how severe the past reaction was, and whether consultation with an allergist is readily available (algorithm 1):
Penicillin skin testing NOT feasible — We suggest consultation with an allergist if available. However, if this is not possible, clinicians must attempt to identify those patients whose history suggests a past immediate reaction that was serious. The best designed cross-reactivity studies suggest that, among patients with skin test-confirmed penicillin allergy, 97 to 98 percent of patients tolerate a cephalosporin. However, an important limitation of these studies, which are nearly all retrospective, is the selection bias arising from the probable exclusion of patients with the most serious past reactions by the clinicians treating them at the time.
Options for treating such patients include the following (if appropriate for the infection in question):
●Give a third- or fourth-generation cephalosporin or carbapenems using a test dose procedure. (See 'Test dose procedure (graded challenge)' below.)
●Give aztreonam or a non-beta-lactam antibiotic normally.
Skin testing feasible — If an allergist is available, then consultation and penicillin skin testing and confirmatory challenge can determine if the patient is allergic to penicillin with certainty, which greatly simplifies future antibiotic choices. (See 'Role of the allergy specialist' below.)
An allergy evaluation should be pursed if a penicillin or a first- or second-generation cephalosporin is indicated or superior to other antibiotics. Either penicillin skin testing can be performed or the patient could be assumed to be allergic and receive a penicillin through an empiric (because the patient's allergy status is unknown) desensitization protocol or receive a first-generation cephalosporin via graded challenge or empiric desensitization. An infectious diseases consult may also be helpful to explore efficacy of different antibiotic options. (See "Rapid drug desensitization for immediate hypersensitivity reactions".)
Management of patients with vague histories — Some patients with past drug eruptions often cannot recall enough detail for the interviewer to make a judgment about the nature of the skin lesions (hives versus other). It is important to ascertain that the skin eruption did not blister or peel or involve mucosal surfaces, and these are characteristics that patients generally remember. We consider reactions that are too vague to characterize as carrying some risk of a future immediate reaction. This risk is lower if the past reaction was (possible) isolated hives without other symptoms and occurred >10 years ago. Some experts would consider such a patient to be at minimal risk and manage them accordingly, but this requires clinical judgement.
Safety and impact of proposed approach — A guideline nearly identical to the approach advocated in this topic was implemented at one author's (KB) practice site (an academic tertiary care referral center) for hospitalized adults reporting a penicillin allergy who would benefit from treatment with a beta-lactam antibiotic [24,25]. Test doses, when indicated, were administered by the medical team, whereas prior to implementation of the guideline, all challenges had required allergy consultation. Allergy consultation was obtained if a penicillin or first- or second-generation cephalosporin was desired in a patient with possible immediate allergy, and the patient was skin tested to penicillin (with results guiding therapy).
The incidence of adverse drug reactions did not increase significantly after implementation, compared with the pre-guideline period, suggesting that test dosing can be safely performed by the general medical staff with proper guidance. After implementation, there were statistically significant reductions in the use of vancomycin, aztreonam, aminoglycosides, and fluoroquinolones. When replicated at another institution on internal medicine inpatients, the guideline resulted in an almost twofold increase in use of favorable (ie, not antibiotic stewardship-restricted) penicillins and cephalosporins (adjusted odds ratio [OR] 1.8, 95% CI 1.1-2.9) .
Based upon this initial success, the guideline was applied across a larger health care system, comprised of two academic tertiary care referral centers and three community teaching hospitals. Sites without access to allergy/immunology and penicillin skin testing either used empiric desensitization (ie, empiric because the patient's current sensitization status was not known) or an alternative antibiotic for inpatients with outpatient allergy/immunology follow-up for skin testing. Test dosing was performed similarly at all sites by nonallergist providers. Since guideline implementation, over 1000 test doses have been performed without any identified safety risks. Data collection on specific outcomes of this approach is in progress .
ROLE OF THE ALLERGY SPECIALIST — If an allergy specialist is available, consultation is strongly encouraged for suspected immediate allergy, with the urgency depending upon the clinical scenario. Allergists (or other specifically trained clinicians) can perform penicillin skin testing to determine with certainty if the patient has an immediate penicillin allergy, which simplifies future management and may further reduce the use of broader-spectrum agents.
Penicillin skin testing — Penicillin skin testing reagents are commercially available in many countries. The performance of skin testing is encouraged because it greatly simplifies future antibiotic choices for that patient and can reduce the use of broader-spectrum agents. The reason for this is that the vast majority of patients turn out to be penicillin skin test-negative, and by virtue of ruling out penicillin allergy, they are able to receive any and all beta-lactams antibiotics safely, provided they have not also reacted to cephalosporins. Penicillin skin testing is only validated in patients with past immediate reactions to a penicillin and should not be used to screen patients who have never reacted to a penicillin.
Skin testing takes between 30 minutes and one hour to perform, and the results are immediately available. It can be safely performed in pregnant women and critically-ill patients when necessary. Skin testing is discussed in greater detail elsewhere. (See "Penicillin skin testing".)
Impact of proactive allergy evaluation in advance of acute need — Several groups have evaluated the impact of allergy evaluation with penicillin skin testing or guidelines/algorithms to address the use of related antibiotics in patients reporting penicillin allergy (table 2) [24,28-49]. Studies performed testing in the hospital either when the need for penicillin arose or when less appropriate antibiotics were being used. None of these studies are randomized, many are subject to selection bias, and few have a control group or control for baseline patient differences between groups. Longer-term cost savings are more difficult to assess, as complications, suboptimal care, and antibiotic resistance would need to be considered.
TEST DOSE PROCEDURE (GRADED CHALLENGE)
Indications and precautions — Test doses (also called graded challenges) are indicated to exclude immediate allergic reactions. Test dosing is appropriate only when immediate allergy to the tested antibiotic is judged to be unlikely after careful consideration of the details of the history. Test doses may be administered by an allergy specialist (preferable) or a member of the general medical team . The purpose of a test dose is to expose the patient to a small amount of drug, followed by a period of close observation, in case there is a reaction.
Because graded challenges will not prevent or circumvent an immediate allergic reaction, the clinician must be prepared to recognize and treat such a reaction if one occurs. Diphenhydramine (oral and intravenous) and epinephrine (intramuscular) should be at the bedside. If a fixed dose epinephrine autoinjector is not being used, the appropriate concentration and dose of intramuscular epinephrine for that patient should be calculated in advance so that there are no delays or dosing errors if epinephrine is needed. Proper dosing of epinephrine in anaphylaxis is described in the tables for adults (table 3) and children (table 4).
When test doses are given, patients are usually not pretreated with antihistamines or glucocorticoids, because antihistamines may mask early signs of an allergic reaction. However, if a patient is already receiving a daily antihistamine for allergic rhinitis or another disorder, it is not necessary to withhold it. Treatment with beta-adrenergic blocking medications should be withheld for 24 hours before the first dose, if feasible, as these medications can interfere with treatment of anaphylaxis, should a reaction occur. Patients with asthma, chronic obstructive lung disease, or other chronic pulmonary diseases should be optimally controlled prior to undergoing challenge.
At one author's (KB) institution, test doses of beta-lactam antibiotics in hospitalized patients are ordered by any clinical provider using a protocol and administered by a nurse with a one-to-one nursing assignment for the challenge duration . The patient is assessed for symptoms, and vital signs are recorded every 30 minutes. More information about graded challenges, including interpretation of subjective symptoms, is reviewed elsewhere. (See "Penicillin allergy: Immediate reactions", section on 'Graded challenge (test dosing)'.)
The starting dose is usually 1/10th of the full dose, but could be higher or lower, depending on the route of administration, clinical stability of the patient, and level of certainty that the patient is not allergic to the drug.
Intravenous — In most cases, an intravenous test dose can be 1/10th (10 percent) of the final dose. If no reaction occurs, the full therapeutic dose is administered 30 to 60 minutes later. In a study including both hospitalized patients and outpatients in an allergy/immunology clinic, a simple two-step graded challenge in low-risk patients was as safe as three- or four-step challenges .
●An example of an intravenous test dose to ceftazidime, starting with 1/10th of a dose would be the following: Give 100 mg, followed by one hour of observation. If no symptoms, give 1000 mg, followed by one hour of observation. The next dose after that is given normally when it is due.
The test dose solution should not be significantly more dilute than the normal preparation, because patients may not react if the concentration is dramatically altered [51,52]. Thus, the 1/10th (10 percent) dose can be withdrawn from a full-strength bag with a syringe and administered by slow intravenous push (usually about 1 mL /min), followed in one hour (if tolerated) by the remaining 90 percent. Other methods of administering test doses and additional safety issues surrounding test dosing policies have been described .
Oral — Test doses can also be given orally. An oral test dose is usually 1/4th or 1/10th of the full dose.
●An example of an oral test dose to cephalexin, starting with 1/10th of a dose would be the following: Give 25 mg (ie, 0.5 mL of the 250 mg/5 mL suspension, given with a glass of water), followed by one hour of observation. If no symptoms, give 250 mg, followed by one hour of observation.
●A 1/4th dose of a pill is prepared with a pill cutter.
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Drug allergy and hypersensitivity".)
SUMMARY AND RECOMMENDATIONS
●Patients with a history of penicillin allergy are more likely to be treated with broad-spectrum antibiotics, which have distinct disadvantages, including greater cost, more side effects, promotion of drug-resistance bacterial strains, and in some cases, reduced efficacy against certain infections. (See 'Impact of penicillin allergy on care' above.)
●Evaluation of a patient with a reported penicillin allergy begins with a detailed history of the past reaction and a review of any relevant medical records. The intent of these steps is categorization of the patient's past penicillin reaction and identification of serious past reactions, particularly immediate, immunoglobulin E (IgE)-mediated reactions and serious forms of non-immediate (delayed) reactions, such as toxic epidermal necrolysis (TEN) or drug-induced organ damage or cytopenias. (See 'Our approach' above.)
●Treatment options for each group of patients are shown in the algorithm (algorithm 1).
•In some patients, the past adverse reaction to a penicillin was clearly not allergic in nature, and beta-lactam antibiotics can be given normally. However, the adverse reaction may recur if the identical agent is used, so clinicians should be mindful of this. (See 'Reactions that are not allergic' above.)
•In other patients, there is a convincing history of a serious type of delayed reaction, and these individuals should continue to avoid all beta-lactam antibiotics unless the details of the case are evaluated by an allergist or other clinician with expertise in the suspected type of reaction and it is deemed safe to use a related antibiotic. (See 'Possible serious delayed reactions' above.)
•Many patients will report mild past reactions that lacked clear features of IgE-mediated reactions (table 1). These individuals can be considered at minimal risk for future serious immediate reactions to beta-lactams and can receive third- or higher-generation cephalosporins, carbapenems, or aztreonam normally (without prior penicillin skin testing). However, because the clinical history can be unreliable, some caution with penicillins and first- and second-generation cephalosporins is still warranted, and these should be given using a test dose procedure. Another approach is to refer such patients to an allergist for penicillin skin testing, which will conclusively determine if the patient has an immediate penicillin allergy and simplify future management. (See 'Management of mild reactions WITHOUT history of features of immediate allergy (minimal risk of immediate allergy)' above.)
•Patients who describe past reactions with features of IgE-mediated allergy are at some risk for similar reactions in the future, and the risk is greatest in patients with severe and recent reactions. We also consider patients to be at risk if the history is too vague to categorize (eg, the clinician cannot tell if skin lesions were hives or something else). A more cautious approach is warranted for this group of patients: Third- and higher-generation cephalosporins and carbapenems can be given using a test dose procedure. If a penicillin or a first- or second-generation cephalosporin is strongly indicated, infectious diseases and allergy specialists should be consulted to confirm the need for these antibiotics and safe administration, respectively. Penicillin skin testing can be performed to clarify the patient's status or the patient can be assumed to be allergic and receive a penicillin or early-generation cephalosporin using a rapid "desensitization" protocol. (See 'Management of reactions WITH features of immediate allergy (some risk)' above.)
●The performance of penicillin skin testing is encouraged because more than 90 percent have negative results (ie, no evidence of allergy), which greatly simplifies future antibiotic choices for that patient and can reduce the use of broader-spectrum agents. Skin testing takes one to three hours to perform and should be done by specifically trained personnel. (See 'Role of the allergy specialist' above.)
●Test doses (also called graded challenges) to exclude immediate allergic reactions are appropriate only when immediate allergy to the tested antibiotic is judged to be unlikely after careful consideration of the details of the history. The purpose of a test dose is to expose the patient to a small amount of drug initially, followed by a period of close observation, in case there is a reaction. Because test dosing will not prevent or circumvent symptoms, the clinician must be prepared to recognize and treat an allergic reaction if one occurs. (See 'Test dose procedure (graded challenge)' above.)