Recognizing Adverse Reactions to Antibiotics

Emerg Med 38(6):11-20, 2006

After reviewing the four different immunologic mechanisms underlying adverse antibiotic reactions, the authors discuss the prevalence and clinical characteristics of beta-lactam and sulfa drug allergies and the more serious cutaneous, cardiac, and metabolic disturbances that antibiotics may cause.

By Stephen J. Playe, MD, and Gregory Murphy, MD

Many patients seek emergency care for adverse reactions to medications. Up to 7% of hospital admissions are for adverse drug reactions, and up to 16% of these reactions involve antibiotics. Some reactions caused by antibiotics occur soon after the drug is administered, while the patient is still in the emergency department. Adverse events include anaphylaxis, cardiac arrhythmias, endocrine abnormalities, drug-drug interactions, and cutaneous eruptions. It is important to recognize all such reactions and initiate appropriate treatment.

In this article, we will discuss the pathophysiology and presentation of severe reactions to antibiotics that may occur in the emergency department or may result in emergency department visits.

FOUR TYPES OF REACTIONS

Immediate reactions. Allergic reactions to antibiotics can be sudden and life-threatening. Type I hypersensitivity reactions most often occur within an hour after exposure to an antibiotic, although in rare cases they may occur after a day or more. The result of the interaction of an antigen with preformed IgE antibodies, these reactions cause the release of histamines and other inflammatory mediators, leading to urticaria, angioedema, and anaphylactic events. Urticaria manifests as blanching edematous papules or plaques approximately 1 to 2 centimeters in diameter, which are usually very pruritic. The lesions are generalized, bilateral, and symmetrical. Angioedema appears as swelling of underlying skin structures, most often occuring on the palms, soles, or the periorbital or perioral region; it is not pruritic. Both conditions are caused by capillary vein leakage; urticaria results from leakage in the superficial dermis and angioedema from leakage in the deep dermal tissue and subcutis. Often, urticaria and angioedema occur concomitantly.

Anaphylaxis accounts for approximately 500 deaths annually. It occurs when urticaria and angioedema advance to a stage that causes dysphagia, bronchoconstriction, and upper airway obstruction. Hypotension and cardiovascular collapse are the hallmarks of true anaphylactic shock. These systemic reactions are thought to be caused by increased capillary leakage and intravascular volume depletion. Antibiotics as a drug class cause anaphylaxis in about 1 in 5000 exposures, although these reactions are not uniformly fatal. Anaphylactic reactions are commonly reported after penicillin use, but they can also occur with cephalosporins, sulfonamides, and, more rarely, other antibiotics.

Cytotoxic reactions. These reactions are triggered when IgG or IgM antibodies become attached to red blood cells (RBCs) or renal interstitial cells that have an antigen bound to their surface. Activation of the complement cascade by this interaction causes cell lysis. Interstitial nephritis is a type II allergic reaction that is often associated with antibiotic use. Frequently, penicillins or sulfa drugs are found to be the inciting agent. Onset of symptoms can be delayed from days to weeks after starting the offending drug. Symptoms include eosinophilia, eosinophiluria, hematuria, and proteinuria. Patients may have declining renal function and a corresponding rise in serum blood urea nitrogen and creatinine. Rapid recovery will often occur after the drug is stopped. Thrombocytopenia and hemolytic anemia are other examples of cytotoxic reactions.

Delayed, immune complex reactions. This type of reaction occurs when IgG or IgM antibodies form circulating complexes with antigens, which then induce complement fixation and become lodged in small blood vessels of the skin, kidneys, and joints, causing inflammation. Serum sickness is an example of a type III reaction. It most often occurs 7 to 10 days after exposure and causes urticaria and angioedema, with fever, arthralgias, myalgias, and palpable purpura. Often, erythema will first occur on the sides of the fingers, toes, and hands before becoming more widespread. Penicillin, sulfonamides, and quinolones are most often the inciting agents in these reactions.

Cell-mediated reactions. These reactions are induced when T-lymphocytes interact with an antigen and cytokines are released. Additional immune cells are attracted by the cytokines, causing local tissue inflammation. This type of reaction occurs with topical application; contact dermatitis is an example of a type IV hypersensitivity reaction. Type IV reactions are so common with topically applied beta-lactam antibiotics that those drugs are never used that way.

BETA-LACTAM ALLERGY

About 10% of patients report that they are allergic to penicillin, but only 1% to 2% of them have true allergic reactions documented by skin testing. How can emergency physicians appropriately prescribe antibiotics to patients with a reported penicillin allergy? Patients can experience a variety of reactions after taking an antibiotic, but it is the potentially life-threatening IgE-mediated type I hypersensitivity reactions that are of greatest concern. Some non-IgE-mediated reactions can also be debilitating, so much so that repeat use of the antibiotic or a drug that may crossreact with it is relatively contraindicated. Beta-lactam-induced Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are contraindications to future beta-lactam use. Maculopapular rashes that occur in isolation, however, especially more than 72 hours after antibiotic administration, are not IgE-mediated type I reactions and do not predict future life-threatening reactions.

To determine which patients are at high risk for IgE-mediated hypersensitivity reactions, skin testing must be performed. Skin testing costs approximately $17 and takes about 40 minutes to complete. In the United States, there have been no anaphylactic reactions to penicillin reported in patients with negative skin testing. Patients with a positive skin test have a 50% to 70% incidence of developing an acute allergic reaction to penicillin. Risk factors for developing severe IgE-mediated reactions include high dosage and concurrent beta-blocker therapy. Atopic individuals do not have an increased risk of having an adverse reaction; however, their reactions tend to be more severe.

The question of crossreactivity between penicillins and cephalosporins often arises. Hypersensitivity reactions to cephalosporins are much less common than reactions to penicillin, but they do occur. Crossreactions between penicillin-allergic individuals taking cephalosporins are much more likely to occur with first-generation preparations than with second- or third-generation cephalosporins. The rate of crossreactivity between patients with skin-test-positive penicillin allergy taking cephalosporins has been reported to be from 0% to 5.6%.

The carbapenem class of antibiotics (meropenem, ertapenen, and imipenem) has a very strong crossreactivity with the penicillins, approaching 50%, and should not be administered to penicillin-allergic patients. The monobactam class of antibiotics (aztreonam, for example) has a much lower rate of crossreactivity and may be administered safely to patients who have a penicillin allergy.

SULFA ALLERGIES

Approximately 3% of patients receiving sulfa antibiotics have an allergic reaction. In patients with AIDS, however, the rate may be as high as 60%. Antibiotics containing sulfa have been implicated in cases of fixed drug eruptions, urticaria, SJS, and TEN. Interestingly, recent studies have demonstrated that although hypersensitivity to sulfa antibiotics and to nonantibiotic drugs containing sulfa are associated, there is no direct crossreactivity between the two groups of medications. It is suggested that this association is due to a predisposition to allergic reactions rather than a direct crossreaction.

Approximately 10% to 15% of black men in the United States have a deficiency of glucose-6-phosphate dehydrogenase (G6PD) enzyme. This enzyme is required for production of nicotinamide adenine dinucleotide phosphate, which is necessary for maintaining glutathione in its reduced state. Red blood cells require glutathione to prevent oxidative damage; deficient cells are more susceptible to damage from oxidative stress. Sulfonamide drugs are commonly associated with increased oxidative stress and subsequent hemolysis in patients with a G6PD enzyme deficiency and should be avoided in these individuals. This type of reaction is not an allergic reaction but a direct interaction between the drug and enzymes within the RBCs. Patients with a G6PD enzyme deficiency can be diagnosed by a quantitative assay or visualization of Heinz bodies in the peripheral blood smear. Heinz bodies are precipitated hemoglobin within the RBCs caused by damage to the hemoglobin sulfhydryl groups.

CUTANEOUS REACTIONS

Cutaneous reactions to antibiotics cover the entire spectrum from benign to potentially life-threatening. Approximately 2% to 3% of hospitalized patients have been reported to have a skin reaction to antibiotic medications. The emergency physician must be able to readily recognize the manifestations of these reactions and accurately determine the cause and its associated morbidity and mortality.

One type of cutaneous reaction is a fixed drug eruption, which appears in the same location with each administration of the causative agent. The lesions usually appear one to two weeks after the first exposure but may appear within several days of subsequent exposures. They are round or oval, erythematous, edematous plaques that may contain vesicles; there is often a burning or pruritic sensation associated with the outbreak. Lesions often appear on the face, genitalia, and peripheral locations such as the hands and feet. Continued administration of the offending drug can cause lesions to appear in areas other than the original site. Antibiotics that are frequently cited as causes of fixed drug eruptions include tetracyclines, sulfa drugs, and penicillins.

Patients often experience a diffuse, erythematous, maculopapular rash after taking a new antibiotic. This reaction most often occurs more than 72 hours after the drug is started and is not a true IgE-mediated allergic reaction. Although concerning to the patient, this should not be considered a contraindication to the use of the causative agent in the future. A maculopapular rash that appears several days after amoxicillin is started is not a reason to withhold beta-lactam antibiotics in the future.

ERYTHEMA MULTIFORME, SJS, AND TEN

Antibiotics are among the more common causes of the continuum of cutaneous reactions that include erythema multiforme (EM), SJS, and TEN. There is much debate as to whether these conditions all represent the same disease process or are distinct entities. The incidence of EM is unknown, but it is estimated that the incidence of SJS is 1 to 6 per million person-years and the incidence of TEN is 0.4 to 1.2 cases per million person-years. Stevens-Johnson syndrome and TEN usually appear one to three weeks after initial exposure to the inciting agent but may manifest in hours to days after re-exposure.

Erythema multiforme is quite variable in presentation, although patients often complain of malaise, fever, myalgia, arthralgia, and pruritus prior to the onset of cutaneous manifestations. Most commonly, the lesions are maculopapular target lesions. The maculopapular regions enlarge over 24 to 48 hours, with their central areas becoming cyanotic or dusky-looking. The lesions can develop anywhere on the body, but usually the initial appearance is a symmetrical outbreak on the dorsa of the hands and feet and on the extensor surfaces of extremities. They then spread to proximal regions of the body. Mucosal involvement, most often limited to the mouth, occurs in 25% to 70% of cases.

Stevens-Johnson syndrome and TEN by definition involve multiple sites of mucosal involvement and are more severe than EM. Mucosal lesions can occur in the mouth, bronchial tree, eyes, urethra, and gastrointestinal tract. About 69% of patients with SJS and 50% of patients with TEN were found to have ocular involvement, compared to only 9% of patients with EM. The eruption of SJS and TEN, caused by epidermal detachment, consists of large flaccid bullae that often coalesce late in the disease, producing large denuded areas of dermis. A positive Nikolsky sign is elicited when a shear force applied to the skin causes epidermal separation. The cutaneous involvement usually reaches its peak within four days before beginning to resolve.

The distinction between SJS and TEN rests on the percentage of epidermal detachment. In SJS, there is less than 10% detachment, while in TEN there is more than 30%. Patients presenting with between 10% and 30% detachment are considered to have an overlap of the two conditions. Patients with these symptoms can have massive fluid loss and often need admission to a burn unit for care.

Case-controlled studies have retrospectively evaluated the risk of SJS and EM after various medications have been taken. Among the antibiotics, sulfa-containing drugs had the highest relative risk, followed by the imidazole antifungal agents, cephalosporins, quinolones, aminopenicillins, tetracyclines, and macrolides. Interestingly, in the same study, thiazide diuretics and sulfonylureas, both of which are structurally related to antibacterial sulfa drugs, had no increased risk.

QT INTERVAL PROLONGATION AND TORSADES DE POINTES

Both congenital and acquired prolonged QT intervals can predispose patients to polymorphic ventricular tachycardia (torsades de pointes). Drugs are a common cause of an acquired prolonged QT interval. The QT interval is affected by heart rate; the slower the heart rate, the greater the period of prolongation. To account for rate-related variations, the corrected QT interval (QTC) is determined by dividing the QT interval by the square root of the preceding RR interval (Bazette’s formula). This is automatically calculated and reported by ECG machines. A normal QTC is less than 430 ms in men and less than 450 ms in women.

The mechanism of QT interval prolongation is prolonged action potentials of the cells of the ventricular myocardium, which may be caused by a reduction of outward currents or increased inward currents during phase 2 and 3 of the action potential, or both. In animal studies, it has been found that after such a prolongation of the action potential the next action potential can become deformed (called early after-depolarization), which has been implicated in the genesis of torsades de pointes.

Many antibiotics can cause prolongation of the QT interval. These include ketoconazole, itraconazole, fluoroquinolones, macrolides, pentamidine, and trimethoprim/sulfamethoxazole. The macrolide antibiotics cause a proportionally greater increase in QTC compared to other antibiotics, most likely due to their intrinsic arrhythmogenic capability and inhibition of cytochrome P450 (CYP) enzymes. A recent study found that patients taking erythromycin had an adjusted rate of sudden death from cardiac causes that was twice as high as other patients. The adjusted rate of sudden death rose to five times when patients took erythromycin with another drug that was a CYP3A inhibitor.

The incidence of torsades de pointes with macrolide antibiotics has been found to be increased with higher doses and with intravenous administration. Azilides (such as azithromycin) and ketolides (such as telithromycin) have minimal effect on the QT interval. Trimethoprim/sulfamethoxazole is thought to cause prolongation of the QTC in persons with a specific single-nucleotide polymorphism in the MiRP1 gene. It has been estimated that 1% to 2% of the population has this polymorphism, which puts them at increased risk for prolonged QTC. The azoles cause prolongation of the QT interval by inhibiting the CYP3A4 enzyme. Ketoconazole and itraconazole are much stronger inhibitors of this enzyme than fluconazole and thus are more likely to incite torsades de pointes. There is a very low risk of QT prolongation and torsades in patients taking fluoroquinolone.

These drugs should be prescribed with caution in patients with risk factors predisposing them to developing a prolonged QT interval, especially patients already taking a drug that can cause QT prolongation. Optimally, an ECG should be done on all patients with risk factors for developing a prolonged QT interval prior to starting a drug known to prolong the QTC. In 1997, the European Committee for Proprietary Medicinal Products issued a statement that an increase in the QTC of 30 to 60 ms after a patient starts a new drug “may represent a drug effect” and an increase of more than 60 ms or a QTC that exceeds 500 ms “raises a clear concern about the potential risk of torsades.” Again, optimally, an ECG should be performed when the drug is at steady state but emergency physicians often do not have this luxury. For an individual patient, the risk of developing an arrhythmia is most strongly predicted by the baseline QTC interval, the QT interval after the first drug dose when at steady state, potassium concentration, medical illness (such as structural heart disease), and the use of additional QTC -prolonging drugs. Studies of cardiovascular drugs have demonstrated that women are at higher risk for developing drug-induced torsades.

HYPOGLYCEMIA AND OTHER DRUG-DRUG REACTIONS

Persistent hypoglycemia has been reported in patients taking both oral hypoglycemic agents and the fluoroquinolone gatifloxacin. The onset of hypoglycemia ranged from 45 minutes to 24 hours after the first dose of gatifloxacin. In all cases, the hypoglycemia was refractory to conventional therapy for 24 to 36 hours. While there are no clear-cut recommendations concerning the use of gatifloxacin in patients taking oral hypoglycemic agents, it seems prudent to choose a different fluoroquinolone or a different class of antibiotic for patients taking such medications.

Many antibiotics will crossreact with other medications that patients are taking. These reactions can have many unintended consequences ranging from improper action of the antibiotic to altered bioavailability and cardiac arrhythmias. Any time a new antibiotic is prescribed it should be checked against other medications for unintended interactions. The table below contains a list of common drug-drug reactions caused by antibiotics.

back to top

Printable version of this article

E-mail this article