An Examination of Penicillin Allergy

  1. Introduction

As many as 10% of the general population claims to be allergic to penicillin. In reality only 1-3% of people have an IgE-mediated reaction to penicillin. In recent years, the number of available antibiotics has increased dramatically. This has led physicians to avoid penicillins if a patient reports any adverse reaction; no matter how mild. Consequently, there has been increased use of fluoroquinolones and vancomycin, in particular. The concomitant rise in antimicrobial resistance and health care costs has prompted the medical community to reexamine the issue of penicillin allergy. A thorough review of this topic should answer the following questions:

  • Are all patients that claim to be allergic to penicillin truly allergic?
  • How can we determine who is truly allergic?
  • Are there some infections for which penicillin is superior to other antibiotics?
  • How would the increased use of penicillin help with antimicrobial resistance and health care costs?
  • How does this apply to other β-lactams such as cephalosporins?
  • Are there certain populations of patients that would benefit from routine skin testing?
  • Is there a practical approach to patients with penicillin allergy?
  1. What is a true penicillin allergy?

Like many pharmacologic agents, penicillin is simple in structure and of low molecular weight. Low molecular weight substances that are able to produce an allergic response are known as haptens. By themselves they are unable to induce antibody formation. In order to induce an immune response they must be attached to carrier molecules that form a strong covalent bond. Patients do not exhibit an immune response to penicillin itself but rather to the breakdown products of penicillin (isomers) after they bind with tissue and plasma proteins to form hapten-carrier complexes. Many, if not all, of the breakdown products of penicillin are created in the vial, tablet, or powder prior to administration to the patient.

The b -lactam ring in the degraded penicillin is unstable and binds with lysine residues in the tissue and plasma proteins resulting in a penicilloylepitope known as benzylpenicilloyl (BSO) or the "major determinant". This major determinant is produced in the largest amount and therefore is immunodominant in penicillin specific immune responses. Although less common, the b -lactam rings can also make molecular rearrangements with carboxyl and thiol groups and form less dominant or "minor determinants" (Figure 1). The terms major and minor refer only to the amount of hapten available for binding and not to the importance of each hapten in an immunologic response.

These complexes of penicillin break down products and native proteins are recognized as foreign antigens by mast cells and basophils in some individuals. Subsequently, when a sufficient density of drug epitopes is formed, a drug-specific immune response ensues. Minor determinants appear to cause 10-20% of immediate IgE-mediated reactions, and major determinants appear to cause 80-90% of immediate IgE-mediated reactions. Major determinants can cause an immediate IgE-mediated reaction but more often cause accelerated or delayed reactions caused by IgG and IgM. This is a very important clinical distinction because patients can have IgE-mediated reactions to the minor determinant alone and if skin tests only include the major determinant those patients (who are more likely to have an anaphylactic reaction) will not be identified.

Note that prior to 1970, penicillin preparations were often contaminated with trace quantities of macromolecules and drug polymers (ampicillin), which increased the frequency of allergic reactions. The more recent "purified" penicillin preparations reduced but did not eliminate these reactions. A variety of infections create an inflammatory response that can enhance drug allergy. Rashes are common in patients with HIV, hepatitis B, coxsackie virus, echovirus and numerous bacterial infections. Therefore, patients with infections who take penicillin and develop a rash should not automatically be labeled with penicillin allergy.

Figure 1. Immunochemistry of penicillin molecules. (from Adkinson et al.)

Immediate Reactions (Type I)

  • The penicillin-protein antigen complex (often the minor determinant) binds to fixed IgE on mast cells and basophils, triggering the release of mediators of inflammation such as histamine, serotonin, and prostaglandins. This results in increased capillary permeability (edema), changes in smooth muscle tone, and increased production of viscid mucous in target tissues.
  • This usually occurs within one hour and clinically is manifested by any combination of diffuse erythema, pruritus, urticaria, angioedema, bronchospasm, laryngeal edema, hyperperistalsis, hypotension, and cardiac arrhythmias.
  • These reactions occur in 0.004% - 0.015% of penicillin courses - most often in adults between the ages of 20 and 49 years. A history of atopic reactions increases the risk for an anaphylactic reaction.
  • Immediate reactions are more common in parenteral administration.
  • Fatal outcomes occur in 1 per every 50,000 to 100,000 treatment courses.
  • Of note, it is possible for IgE-mediated reactions to occur from 1 to 72 hours after administration of penicillin. These are called "accelerated reactions" and are more often associated with the major determinant. They can result in urticaria, angioedema, laryngeal edema, and/or wheezing.
  • Life threatening reactions occurring beyond 1 hour of penicillin administration are rare.

Late Reactions (Type II, Type III, Type IV, idiopathic)

  • These reactions occur after 72 hours of drug administration and are classified as Type II, III, or IV based on the underlying immune response. Because none of these reactions is IgE dependent, skin testing has no role in the evaluation of these patients. (Refer to Table I for a full summary).

Table 1: Classification of Penicillin Reactions

Classification / Time of Onset / Mediators / Clinical Signs / Skin Testing Useful? / Comments
Type I (immediate) / < 1 hour / IgE mediated cross-linkage with mast cells and basophils with release of vasoactivemediators. / Anaphylaxis and/or hypotension, laryngeal edema, wheezing, angioedema, urticaria / Yes / More likely if parenteral admin. Fatal in 1 per 50k-100k treatment
Late reactions / Variable
Type II / IgG, complement / Blood transfusion rxn., autoimmune hemolytic anemia / No / IgE not involved
Type III / IgG, IgM immune complexes / Serum sickness, Tissue injury (glomerulonephritis, RA, SLE) / No / Tissue lodging of immune complexes; drug fever
Type IV / Cytokines activate Tc cells causing direct cellular damage / Contact dermatitis, TB lesions, Graft rejection. / No
Idiopathic / Usually > 72 h / Maculopapular or morbilliform rash / No / 1%-4% of all patients receiving penicillin
  1. How can we determine who is truly allergic?

The role and utility of skin testing in various clinical settings.

Two basic classes of skin tests are used in clinical allergy practice – epicutaneous and intracutaneous (intradermal). The scratch and the prick tests are epicutaneous. The epicutaneous tests are more specific but less sensitive than the intradermal tests. The scratch test is performed by scratching the superficial epidermis with a needle or sharp instrument and applying a drop of antigen to the local abrasion. The prick test is performed by placing a drop of antigen on the skin and inserting a needle at a 45˚ angle through the drop to lift the epidermis. Intradermal testing is performed with a 26 to 30 gauge needle, creating a skin wheal of 0.02 cc of extract. Because the degree of cutaneous reactivity varies between patients, a diluent negative control and a positive control of histamine are used with the three methods.

The reactions are read in 15 minutes and graded on a scale of negative to 4 +. The wheal is measured at its largest diameter and at the perpendicular diameter, and an average of the two is calculated. One plus is erythema up to 2 cm in diameter, 2+ is greater than 2 cm in diameter, 3+ reactions are erythema and wheal formation, and 4+ is erythema, wheal formation, and pseudopod formation. Reactions of 2+ or greater are considered significant.

It is generally accepted that skin testing with both major and minor determinants is necessary in order to have the most sensitive and specific test possible because a minority of patients will have an adverse reaction to the minor determinant alone (often anaphylaxis). Currently, there is no commercial source for the minor penicillin skin test reagent in the US. Some use fresh penicillin G and its subsequent break down products as a minor determinant mixture. In recent months there have also been problems with the manufacture of major determinants. Centers that wish to skin test must custom make their materials or receive permission from the FDA to import them from Germany. Skin tests take approximately 40 minutes and the reagents and equipment cost approximately $17 per patient. The use of skin testing and an examination of its safety, cost, risks, and how predictive it is of penicillin allergy has been examined in a variety of settings. The following is a review of several of these studies.

  1. Outpatient STD clinic

Gadde J. et al. Clinical experience with penicillin skin testing in a large inner-city STD clinic. JAMA. (1993)

Purpose: To determine the prevalence of positive penicillin skin tests among outpatients with well-defined but variable history of penicillin allergy and to determine the reproducibility, safety, and negative predictive value of skin testing with major and minor determinant mixtures.

Design: Serial consenting outpatients with current indications for penicillin therapy were skin-tested with major and minor determinants. Those with negative skin tests received therapeutic benzylpenicillin or ampicillin. Negative predictive value of skin testing was established by 72 hour follow-up for adverse reactions to drug.

Methods: 5063 consecutive patients in a Baltimore, MD STD clinic. 66% male, 90% African-American. Follow-up was 94% complete.

Results: Positive skin tests were found in 7.1% of 776 individuals with a previous history of penicillin allergy and in 1.7% of 4287 individuals negative by history. A previous history of anaphylaxis or urticaria was associated with a significantly higher rate of positive skin tests (17.3% and 12.4% respectively). The interval from last use of penicillin did not influence skin test results. Thirteen patients had an adverse reaction to the skin test itself (one had "mild anaphylaxis", 11 had urticaria, one had a large local reaction). After penicillin administration among patients with a negative skin test, acute allergic reactions occurred in only 0.5% (18/3997) of subjects negative by history. Three patients had an immediate reaction and 15 patients had an accelerated reaction. No patients had anaphylaxis. 2.9% (17/596) of patients with a negative skin test and a positive history had an acute allergic reaction. Reactions were typically mild with only 2 patients having "mild anaphylaxis" – both of these patients had a history of IgE-mediated reactions and responded to epinephrine. No deaths occurred. Of the patients with a positive skin test, 90% demonstrated major-determinant sensitivity but 10% reacted to minor determinants alone. Ninety-seven percent of those with a history of "penicillin allergy" were able to tolerate full dose penicillin therapy after a negative skin test.

Conclusions: Skin testing with both major and minor determinants is safe and both reagents are needed to maximize the identification of sensitized subjects since minor determinants are associated with anaphylaxis. Penicillin skin testing can facilitate the safe use of penicillin in 90% of those patients with a previous history of penicillin allergy.

  1. Hospitalized adults (non-ICU settings):

Sogn D.D. et al. Results of the National Institute of Allergy and Infectious Diseases collaborative clinical trial to test the predictive value of skin testing with major and minor penicillin derivatives in hospitalized adults. Archives of Internal Medicine. (1992)

Purpose: A history of penicillin allergy does not predict reactions to subsequent administration of penicillin. This study tests the usefulness of skin tests with a major determinant and a minor determinant mixture in the prediction of IgE-mediated reactions.

Methods: Eight centers cooperated in testing hospitalized adults with infectious diseases for which penicillin was the drug of choice. Infections were often the reason for admission but could have been hospital-acquired. Patients were tested with major and minor determinants followed by a therapeutic course of penicillin with observation for 48 hours.

Results: There were 1539 patients who underwent skin testing (825 had histories of penicillin reaction – an astounding 54%, 104 were unsure of their status, and 610 had negative histories). No adverse reactions to the test materials were noted. One thousand, four-hundred and twenty-two patients completed the protocol. Among 726 history-positive patients, 566 with negative skin tests received penicillin and only seven (1.2%) had "possibly" IgE-mediated reactions. Among 600 history-negative patients, 568 with negative skin tests received penicillin and none had a reaction. Only nine of the 167 positive skin test reactors received penicillin, usually by "cautious incremental dosing". Two of these nine patients had reactions compatible with IgE-mediated or accelerated penicillin allergy and both were positive to two determinants.

Conclusions: Negative skin tests continue to have a high negative predictive value and the reaction rate in skin-test positive patients was significantly higher.

  1. ICU setting

ArroligaM.E. et al. A pilot study of penicillin skin testing in patients with a history of penicillin allergy admitted to a medical ICU. Chest. (2000)

Design: Prospective study of all patients with histories of penicillin allergy who were admitted to a medical ICU during a 3-month period and who received antibiotics at the Cleveland Clinic.

Methods: Twenty-four patients were enrolled and 21 underwent skin testing (prick test and intradermal) with benzylpenicilloyl (major) and penicillin G (minor) along with histamine controls. Three patients were not skin tested due to recent history of Type I reaction to penicillin.

Results: The characteristics of previous allergy were unknown in 11 patients, 6 had a history of rash, and 4 had a history of urticaria. Twenty of 21 patients had a negative skin test and a positive histamine control. One patient had a negative skin test and a negative histamine control but tolerated a penicillin challenge without problems. No adverse reactions to the skin test were noted and antibiotic coverage was changed in 10 patients (48%) as a result of the skin tests.

Conclusions: A very small pilot study. The authors conclude that most patients without a history of Type I reactions have negative skin tests and are able to receive penicillin. Penicillin skin testing may be a way to alter antibiotic use patterns in ICU settings.

  1. Safety – How often do patients have a systematic reaction to penicillin skin tests?

Valyasevi M.A. et al. Frequency of systematic reactions to penicillin skin tests. Annals of Allergy, Asthma, and Immunology. (2000)

Purpose: To determine the safety of penicillin skin testing and the rate of systemic reactions to skin testing in patients who gave a history of penicillin allergy.

Methods: Retrospective chart review of 1710 patients with a history of penicillin allergy who underwent skin testing with major and minor (penicillin G) determinants (prick tests and intradermal).

Results: Eighty-six patients had a positive skin test and 2 patients had systemic reaction. Neither patient required hospitalization and responded well to treatment with epinephrine in one case and chlortrimeton and prednisone in the other. The systemic reaction rate was 0.12% for all patients and 2.3% for the penicillin skin test positive group. No fatalities occurred.

Conclusions: The incidence of systemic reaction to skin tests is low and skin prick tests should be performed. If patients have a history of a previous serious reaction, skin tests may be avoided and if done should be diluted. Practitioners who perform skin testing should be ready to treat a systemic reaction.

  1. What are the potential antimicrobial resistance and economic benefits of clarifying penicillin allergy?

In recent years, two very important issues have confronted medicine – antimicrobial resistance and the rising cost of health care. Penicillin skin testing may provide some relief for both of these problems. The following is a look at three recent studies that examine the issue.

  1. Macy E. Elective penicillin skin testing and amoxicillin challenge: effect on outpatient antibiotic use, cost, and clinical outcomes. Journal of Allergy and Clinical Immunology. (1998)

Purpose: To test the effect of penicillin skin testing on oral outpatient antibiotic use and to evaluate concerns over penicillin skin testing in advance of therapeutic penicillin use causing resensitization and increased drug allergy rates.

Design: A natural history study on 236 patients who underwent penicillin skin testing with or without oral amoxicillin challenge between November 1994 and November 1995 to include those with both positive and negative skin tests. This was a follow-up to a study by Macy et al. published in 1997. That study tested a purified minor determinant mixture as part of a skin test protocol to see if it was predictive of penicillin allergy. Patients received the custom made skin tests followed by an oral challenge with amoxicillin if their skin tests were negative. Macy found that his purified minor determinant was safe and predictive of penicillin allergy.

Methods: A review of computerized pharmacy records and outpatient medical records was undertaken with antibiotic prescriptions, clinic visit rates, and adverse reactions to antibiotics noted over a 2-year period to include the year prior to penicillin skin testing and the year after penicillin skin testing.

Results:

  1. In the follow-up period the use of penicillins increased from 13.7% to 39.8% of all antibiotics prescribed.
  2. The overall use of antibiotics decreased from 779 courses dispensed to 191 patients to 558 courses dispensed to 169 patients.
  3. Among patients with a negative skin test who received an antibiotic in the follow-up year, 47.4% received penicillin compared to 20.4% of those same patients who required antibiotics in the year prior to skin testing.
  4. The total cost of antibiotics fell 32% ($17,211.88 to $11,648.27) during the year after skin tests. The average cost fell 7.2% (from $22.21 to $20.61) for patients with a negative skin test but rose 3.3% (from $21.61 to $22.34) for patients with a positive skin test.
  5. The overall clinic visit rate did not change, a shift from primary care to specialty care visits was noted (51.9% to 39.1% to primary care).
  6. Fifteen patients with positive skin tests received at least 1 course of a cephalosporin in the year after skin testing (33 total courses) with one reaction (rash in a one-year-old) noted that required no treatment.
  7. Of the 93 patients with negative skin tests who subsequently received penicillin, 5 (5.4%) had an adverse reaction – all to amoxicillin. However, 2 of the 5 patients subsequently tolerated 5 additional courses of penicillin (including amoxicillin) and 1 patient reacted only after her 4th course to penicillin. All reactions were noted at least 7 days after the start of treatment. The adverse reaction rate in patients with a negative skin test and no amoxicillin rechallenge was 3.2% and the reaction rate to non-penicillins was 5.4%.

Conclusions: Penicillin skin testing continues to be safe and changes subsequent frequency of antibiotic use and type of antibiotic prescribed. Cost was also decreased. Patients who received oral amoxicillin challenge with their initial skin testing were not resensitized with therapeutic penicillin use. In addition, those who had skin testing without oral amoxicillin challenge did not become sensitized when receiving penicillin. Also, cephalosporin use in penicillin allergic patients was safe. The fact that patients saw allergists at skin testing and more specialists in follow-up may have contributed to decreased antibiotic use by more accurate diagnosis and symptom relief. This may have confounded the study results.