Drug allergies: What to expect, what to do

November 1, 2006

Abstract: Allergic drug reactions are mediated by IgE. Other immunologic drug reactions are mediated by IgG and complement, T cells, systemic cytokine release, and mast cell activation. True drug allergies, which are IgE-mediated, account for less than 10% of all adverse drug reactions. IgE-mediated reactions are commonly seen with penicillins, but also occur with some other antibiotics, chemotherapeutic agents, therapeutic proteins, and neuromuscular blockers. They are often limited to urticaria but also include life-threatening anaphylaxis. Skin testing to identify patients who are allergic is available for many agents associated with IgE-mediated reactions. Protocols for desensitization permit the use of some culprit agents when indicated. (J Respir Dis. 2006;27(11):463-471)

Adverse drug reactions are a common occurrence. However, true drug allergies--those that are IgE-mediated--account for less than 10% of all adverse drug reactions.1 Even if all immune system-associated reactions are considered (IgG/complement-mediated, T-cell- mediated, mast cell-mediated, and cytokine-mediated--as well as IgE-mediated ones), these still account for a minority of adverse drug reactions.2 The most common causes of adverse drug reactions include overdoses, toxicities, age-related changes in drug metabolism, and expected but undesired pharmacologic effects. Although anaphylaxis can quickly lead to death, nonallergic adverse drug reactions kill many more people--just more slowly.

The seriousness of the problem posed by drug allergies is perhaps overblown in part because of the loose use of the word "allergy," especially in the European literature, to refer to all immunologically mediated reactions. Still, the problems posed by immunologically mediated drug reactions (and others that resemble such events) are considerable. In this article, I describe the more common types of allergic and "allergy-like" drug reactions that can occur (Table) and discuss how best to prevent or manage them.


IgE-mediated reactions

Only a minority of persons produce clinically significant amounts of IgE and, as a result, suffer from allergic disease. However, in susceptible persons, even cutaneous occupational exposures, such as when nurses or pharmacists accidentally spill antibiotic solutions on their skin, can efficiently sensitize them.

Many IgE-mediated reactions are limited to hives without symptomatic hypotension or dyspnea. Fortunately, anaphylaxis is rare even in people who produce allergic antibody.

Anaphylaxis, characterized by symptomatic hypotension with associated dyspnea, urticaria, and possibly GI symptoms, is the most severe manifestation of IgE-mediated drug allergy. It is most common after parenteral drug administration and is relatively rarer with oral or cutaneous exposure. Anaphylaxis results when antigen-specific IgE is present on mast cells and a systemic exposure to antigen occurs, cross-linking the IgE. This results in the simultaneous degranulation of large numbers of mast cells. Mast cell mediators increase vascular permeability; the sudden drop in blood pressure and tissue swelling can cause death.

Treatment for anaphylaxis includes intramuscular adrenaline and systemic antihistamines.1 The adrenaline reverses the hypotension, and the antihistamines block the histamine-mediated effects. Subcutaneous adrenaline is ineffective and should not be used. General support with fluid resuscitation and oxygen can keep a patient alive until his or her system reequilibrates.

Systemic cytokine release

Among the most serious systemic drug reactions are toxic epidermal necrolysis (TEN) and the slightly milder version of TEN known as Stevens-Johnson syndrome (SJS).3 In TEN, the skin is sloughed and symptoms like those seen with large surface area burns can lead to death. In SJS, mucosal lesions typically appear and there is a lesser amount of skin sloughing. In both TEN and SJS, symptoms develop over a period of days.

TEN appears to be mediated by systemic cytokine release. There is no evidence of antigen-specific events in which T or B cells respond specifically to a small molecule or a hapten in a classic immunologically mediated fashion. TEN can occur with lymphoma and may not require the presence of any drug. In TEN, Fas ligand, a member of the tumor necrosis factor family of cytokines, appears to be systemically overproduced.4 Fas ligand binds to CD95 on keratinocytes and causes apoptosis or cell death. There is very little local cellular infiltrate in the more severe cases of TEN; the soluble Fas ligand that mediates a significant portion of the reaction is produced systemically by mononuclear cells.

Agents that have been associated with TEN include various antibiotics, interleukin-2, and antiseizure medications. Do not perform patch testing or skin testing in patients who have a history of these reactions, since no useful information can be derived from the tests. Continued drug avoidance is prudent. Although there are only a few anecdotal reports on the risks of recurrent exposure, deaths have been documented. At present, persons who have a history of TEN or SJS cannot be "desensitized," and based on the current understanding of the pathophysiology of these reactions, desensitization should not be attempted.

Direct mast cell activation

Mast cells contain histamine and other vasoactive mediators. Their sudden release, due to either an IgE-mediated anaphylactic reaction or a similar non-IgE-mediated reaction (referred to as an "anaphylactoid" reaction), results in hypotension; flushing; itching; and potentially respiratory compromise, bowel edema, and death.

Acute reactions to intravenous contrast dye result from a direct mast cell effect precipitated by the hyperosmolarity of the contrast material. IgE plays no role in these reactions. (T-cell-mediated delayed-type hypersensitivity [DTH] reactions can occur with intravenous contrast dye, but these develop over a period of days following the exposure and can be confirmed by patch testing.) Similar direct mast cell- releasing effects are also seen with certain cancer chemotherapeutic agents and opioids.

The leading cause of morbidity in intravenous contrast reactions is their direct effect on kidney function. Hydration is important to prevent this side effect.

Patients with a history of an acute anaphylactoid contrast reaction are at a significant risk for having another. Premedicate such patients with systemic corticosteroids (prednisone, 1 mg/kg) 24 hours before administration of contrast and diphenhydramine, 50 mg, 1 hour before.1 The use of nonionic contrast, which contains iodine, is associated with a significantly lower rate of anaphylactoid reactions, although it can still cause T-cell- mediated DTH reactions.

T-cell-mediated reactions

Poison ivy or oak triggers the classic T-cell-mediated DTH reaction. A rash develops over a 2- to 5-day period after exposure to the resin, which binds to proteins in the skin. T cells reject the contaminated skin as they would a foreign-tissue graft; this results in a rash that often blisters and continues until new, uncontaminated skin grows out.

T-cell mediated DTH is the usual mechanism of contact dermatitis resulting from exposure to a local anesthetic. True IgE-mediated reactions to local anesthetics are extremely uncommon. Most acute reactions to local anesthetics are reactions to certain preservatives, such as methylparaben, that are commonly used in multidose vials of local anesthetics, or to coadministered epinephrine.

Patch tests for DTH reactions to local anesthetics are available. Skin testing for acute reactions to local anesthetics preserved with methylparaben can also be performed; this is usually done at the same sitting at which tolerance of the pure anesthetic is documented by means of a provocative dose test.5

If there is a question about which local anesthetics a patient can tolerate and there is not adequate time for testing, or if a patient has recently taken an antihistamine and skin testing is impossible, parenteral diphenhydramine can be infused locally as an anesthetic. It tends to burn during injection but is not associated with IgE-mediat-ed allergic reactions. (Although diphenhydramine can cause DTH reactions when used topically, it rarely does so when used systemically.)

Topical antibiotics, adhesive tape, and skin cleaners can also provoke DTH reactions.

IgG-mediated reactions and serum sickness

Many foreign materials, including antibiotics and proteins, can cause specific IgG to be produced over a period of several weeks.1 If the specific antibody then binds to residual antigen--or if the antigen is readministered--complement may be activated, leading to the development of such symptoms as drug fever, delayed-onset rash, and joint pain. When this type of reaction is associated with a second exposure to xeno-antisera (often horse sera, historically), it is called serum sickness. Similar reactions have been seen with cefaclor and many other antibiotics.


Most adverse reactions to antibiotics are not allergic in nature. When patients who have reported an adverse reaction are given any other antibiotic, another adverse reaction (usually a delayed-onset rash or a GI effect) occurs about 3% of the time6; only a small fraction, even of these repeated reactions, are IgE-mediated. Most rashes--even most episodes of urticaria--do not appear to be IgE-mediated.

Antibiotic overuse is the leading cause of antibiotic-associated adverse reactions (both allergic and nonallergic). Antibiotics are often given inappropriately for viral upper respiratory tract infections. Rashes are common with viral infections. Consequently, many patients are erroneously labeled as being allergic to an antibiotic that was prescribed for a viral infection. Coordinated efforts to reduce inappropriate antibiotic use, such as have been made in Scandinavian countries, can curtail adverse drug reactions, reduce antibiotic resistance, and lead to more appropriate use of medical resources.

Penicillin allergy

IgE-mediated drug allergy has been best studied with penicillin-class antibiotics. About 1 in 10 persons seeking health care in the United States reports a history of penicillin "allergy." However, when persons with a history of a penicillin-class antibiotic allergy are skin tested with a complete panel of reagents, only about 7.5% of them have immediate positive responses--and thus a true penicillin allergy. The rate is higher in younger patient populations and in those who receive care at allergy clinics. (One reason the rate is higher in children is that skin testing is more likely to yield a positive result when performed closer to the time of the index reaction; all immunologically mediated reactions tend to decay with time.) In any event, a history of penicillin "allergy" is at least 10 times more common than actual, skin test-documented penicillin allergy.

Persons who have a history of adverse reactions to penicillins are also more likely to report adverse reactions to sulfamethoxazole; however, there is no evidence that this results from an immunologically mediated mechanism.

Cephalosporin allergy

Anaphylactic reactions to cephalosporins are much less common than anaphylaxis associated with penicillins. Persons who make IgE in response to cephalosporins seem to produce it only in response to a particular cephalosporin, whereas persons who make clinically significant IgE in response to penicillins tend to react to core penicillin break-down products. In case reports dating back to the 1980s, almost all of the patients with documented cephalosporin-associated anaphylaxis who had penicillin and cephalosporin skin testing showed IgE directed only against the specific implicated cephalosporin and rarely against other penicillins or other unrelated cephalosporins.7

These findings gave rise to the current theory that cephalosporin allergies involve mainly side-chain reactions, whereas a penicillin allergy is a reaction to the core ß-lactam ring of penicillin. Thus, in a patient with a history of a serious, potentially IgE-mediated reaction to a cephalosporin, it is critical to avoid reexposure to the same cephalosporin, to a cephalosporin that shares the same side chain, and even to other ß-lactams that share the same side chain (such as ceftazidime and aztreonam).

Penicillin-cephalosporin cross-reactions are extremely rare. Reports of clinical rates that are based on histories rather than on penicillin skin testing dramatically overstate the true risk of cross-reaction. One reason for the distortion is that a history of penicillin "allergy" is far more common than skin test-documented allergy. In fact, numerous publications support the relative safety of using cephalosporins in patients with a history of reaction to penicillin--even in those who have had positive results on penicillin skin tests.8


Avoidance is the mainstay of the management of drug allergy. However, needless avoidance of an agent or use of an alternative therapy that is more toxic or less effective can cause unnecessary morbidity. Thus, accurate diagnosis of drug allergies is essential.

Penicillin skin testing

Penicillin-class antibiotics are still widely used and are generally effective. When one of these agents is indicated, skin testing can help confirm its safety and qualm fears of a dangerous reaction.

Only about 1 in 300 persons who have a history of an adverse reaction associated with a penicillin-class antibiotic, whose results on a penicillin skin test are initially negative, and who then are given penicillin therapy will have an adverse reaction associated with a positive result on a repeated skin test.9 Persons with a history of penicillin "allergy" in whom penicillin skin testing yields a negative result have no greater risk of rashes associated with penicillin-class antibiotics than they do of rashes associated with any other class of antibiotics.6 Thus, use of skin testing can increase the number of instances in which penicillin--rather than a broad-spectrum antibiotic--can be safely used, thereby helping to reduce antibiotic resistance.

Ideally, penicillin skin testing would be done in all persons with a history of penicillin allergy. Unfortunately, because of the lack of commercial penicillin skin test reagents, this is not possible.

Penicillin skin testing should be done by specially trained professionals--board-certified allergists with access to a complete panel of penicillin skin test materials. These include the major determinant, penicilloyl-polylysine, and the minor determinants, native penicillin, penilloate, and penicilloate.10 IgE specific to the side chain of the most commonly used penicillin-class antibiotic, amoxicillin, can be identified using native amoxicillin.

Skin testing should be done using reagents with concentrations no greater than 0.01 mol/L, or about 2 mg/mL (although a number of European investigators have historically used much higher concentrations, often 20 mg/mL or greater). At a physiologic pH of 7.4, amoxicillin is soluble only to about 2 mg/mL. To achieve higher concentrations, the pH must be much higher--often as high as 10.

Because nonspecific skin irritations can result from the high pH and the high drug concentration of such solutions, the rate of false-positive results is increased. More concentrated solutions can also result in systemic exposure to pharmacologically active amounts of amoxicillin, which can lead to systemic allergic reactions. However, the occurrence of systemic symptoms with penicillin skin testing is rare--about 1 in 1000 persons tested.11

Since October 2004, the major penicillin skin test reagent, penicilloyl-polylysine, has not been commercially available in the United States, and it is currently unknown when it might be available again. Thus, for the foreseeable future, outside of certain academic medical centers that produce their own skin test reagents, penicillin skin testing is not widely available.

Desensitization to penicillin

When a patient has had a positive result on a penicillin skin test and has a life-threatening infection that can only be treated with a penicillin-class antibiotic, desensitization can allow the use of penicillin. If penicillin skin testing is unavailable and penicillin must be used in a patient who has a history of penicillin "allergy," then desensitization should be undertaken.

Desensitization should be performed with oral agents whenever possible, even if the final drug to be used is only available in a parenteral formulation. When a patient with a skin test-documented penicillin allergy needs a penicillin-class antibiotic (most commonly for CNS syphilis), the risks and benefits of the desensitization procedure should be explained, and written informed consent should be obtained.

The initial oral dose is 0.03 mg, or about twice as much penicillin as is used in an intradermal skin test.1 The oral dose is then doubled every 15 minutes until a therapeutic dose is reached. Once a therapeutic oral dose is tolerated, parenteral penicillin can be given if needed. Continued penicillin exposure is essential. If the systemic level of penicil-lin falls to zero and anti-penicillin IgE is once again freely present on mast cells, and a dose of penicillin is then given, systemic anaphylaxis can occur.

Similar protocols can be used for allergies to other medications.1

As with skin testing, desensitization should be handled by specially trained, board-certified allergists.


Consider neuromuscular blocking agents as a likely cause whenever anaphylaxis occurs during induction. These agents are divalent molecules that can directly cross-link IgE and cause true allergic reactions. Latex allergy still is a problem and should always be considered in the workup of intraoperative anaphylactic episodes--although it is much less common than it was in the late 1980s.1 Parenteral antibiotics are another common cause of anaphylaxis. Dyes, including isosulfan blue, cause anaphylactoid reactions in about 1% of patients.

Measuring tryptase within 3 hours of the onset of an anaphylactoid reaction can provide evidence that mast cell activation occurred during the event. A negative tryptase measurement essentially rules out a systemic IgE-mediated and/or mast cell-mediated mechanism.


IgE-mediated allergies are relatively common in patients who receive nonhuman or even modified human proteins; always consider the possibility of an allergic reaction when therapeutic proteins are administered. Severe IgE-mediated reactions, including anaphylaxis and death, occurred in the 1980s with autologous bone marrow transplantations in which fetal calf serum was used in the cell culture medium.12 Such reactions also occurred in patients undergoing artificial insemination in the 1990s13 and, in the first years of this decade, in patients who received cancer vaccines.14 Even humanized monoclonal antibodies have been associated with IgE-mediated reactions. Therapeutic nonhuman enzymes, such as streptokinase and chymopapain, have also been associated with high rates of anaphylaxis, particularly in patients with repeated exposures to these agents.1


Chemotherapy for cancer has been associated with both IgE- and di-rect mast cell-mediated reactions. A number of chemotherapeutic agents can either bind to proteins as haptens or are proteins, such as l-asparginase, and thus can directly cause allergic reactions. Skin testing for IgE to platins has been reported; positive results on such tests have a very high predictive value for allergy to all platins. Desensitization protocols have been published.15,16 Leucovorin and irinotecan may have a direct mast cell-releasing effect and can cause significant problems when given in combination.


Board-certified allergists can provide the skin testing and desensitization expertise necessary to deal with serious allergic and allergy-like drug reactions. The Adverse Reactions to Drugs and Biologicals Committee of the American Academy of Allergy, Asthma, and Immunology (http://www.aaaai.org/) is a good resource for help with difficult drug reaction problems.



1. Macy E, Mellon MH, Schatz M, Patterson R. Drug allergy. In: Adelman DC, Corren J, Casale TB, eds.

Manual of Allergy and Immunology.

4th ed. Philadelphia: Lippincott Williams & Wilkins; 2002:219-241.
2. Gruchalla RS. Drug allergy.

J Allergy Clin Immunol

. 2003;111:S548-S559.
3. Letko E, Papaliodis DN, Papaliodis GN, et al. Stevens-Johnson syndrome and toxic epiderman necrolysis: a review of the literature.

Ann Allergy Asthma Immunol

. 2005;94:419-436.
4. Abe R, Shimizu T, Shibaki A, et al. Toxic epidermal necrolysis and Stevens-Johnson syndrome are induced by soluble Fas ligand.

Am J Pathol

. 2003;162:1515-1520.
5. Macy E, Schatz M, Zeiger RS. Immediate hypersensitivity to methylparaben causing false-positive results of local anesthetic skin testing or provocative dose testing.

The Permanente Journal.

6. Macy E, Burchette RJ. Oral antibiotic adverse reactions after penicillin testing: multi-year follow-up.


. 2002;57:1151-1158.
7. Warrington RJ, McPhillips S. Independent anaphylaxis to cefazolin without allergy to other beta-lactam antibiotics.

J Allergy Clin Immunol

. 1996;98:460-462.
8. Perez-Inestrosa E, Suau R, Montanez MI, et al. Cephalosporin chemical reactivity and its immunological implications.

Curr Opin Allergy Clin Immunol

. 2005;5:323-330.
9. Macy E, Mangat R, Burchette RJ. Penicillin skin testing in advance of need: multiyear follow-up in 568 test result-negative subjects exposed to oral penicillins.

J Allergy Clin Immunol.

2003; 111:1111-1115.
10. Macy E, Richter PK, Falkoff R, Zeiger R. Skin testing with penicilloate and penilloate prepared by an improved method: amoxicillin oral challenge in patients with negative skin test responses to penicillin reagents.

J Allergy Clin Immunol.

11. Macy E. Risks of penicillin skin testing.

Ann Allergy Asthma Immunol.

12. Macy E, Bulpitt K, Champlin RE, Saxon A. Anaphylaxis to infusion of autologous bone marrow: an apparent reaction to self, mediated by IgE antibody to bovine serum albumin.

J Allergy Clin Immunol.

13. Sonenthal KR, McKnight T, Shaughnessy MA, et al. Anaphylaxis during intrauterine insemination secondary to bovine serum albumin.

Fertil Steril

. 1991;56:1188-1191.
14. Mackensen A, Drager R, Schlesier M, et al. Presence of IgE antibodies to bovine serum albumin in a patient developing anaphylaxis after vaccination with human peptide-pulsed dendritic cells.

Cancer Immunol Immunother.

2000; 49:152-156.
15. Lee CW, Matulonis UA, Castells MC. Rapid inpatient/outpatient desensitization for chemotherapy hypersensitivity: standard protocol effective in 57 patients for 255 courses.

Gynecol Oncol.

16. Lee CW, Matulonis UA, Castells MC. Carboplatin hypersensitivity: a 6-h 12-step protocol effective in 35 desensitizations in patients with gynecological malignancies and mast cell/IgE-mediated reactions.

Gynecol Oncol.

2004;95: 370-376.