Immunogenicity Aztreonam N. FRANKLIN
ADKINSON,
and Cross-Allergenicity
JR.,M.D. Ba/timore,
Mary/and
Beta-lactams are responsible for more than half of the allergic drug reactions encountered in the hospital setting. Although most such reactions are mild, the potential for acute and life-threatening reactions cannot be underestimated when considering readministration of beta-lactam agents, such as the penicillins or cephalosporins, to persons who have previously exhibited sensitivity. In addition to concerns about possible allergic reactions to the beta-la&am antibiotics individually, considerable cross-reactivity has been demonstrated among such classes as the penicillins, the cephalosporins, and the imipenems, although it cannot yet be predicted on an individual basis. Early studies of the unique monocyclic beta-lactam-or monobactam-aztreonam, indicated that the new class demonstrated negligible cross-reactivity with the standard beta-lactams both experimentally and clinically. Further, aztreonam was associated with an extremely low (2 percent) incidence of immunologic drug reactions. Aztreonam also has been found to be welltolerated by highly penicillin-allergic patients. Although further clinical study is indicated, data so far are encouraging. If it is confirmed that monobactams such as aztreonam are minimally cross-reactive, well-tolerated by subjects allergic to other beta-lactam antibiotics, and only weakly immunogenic, fewer allergic reactions may be associated with antimicrobial therapy in the future than have been seen with the other available beta-lactam antibiotic drugs.
From the Department of Medicine, Johns Hopkins University School of Medicine, Johns Hopkins Asthma and Allergy Clinic, and the Good Samaritan Hospital, BaltiJr., M.D., Johns Hopkins Asthma and Allergy Center, 301 Bayview Boulevard, Balti-
3C-12s
March 23, 1990
of
The American Journal of Medicine
eta-la&am allergies are the most common drugB allergy problem encountered in the practice of medicine. More than half of all allergic drug reactions that occur in hospitalized patients can be attributed to beta-lactams [l]. Most are relatively benign and easily handled clinically, but serious, life-threatening, acute allergic reactions occasionally develop that understandably produce a great deal of concern in both patients and physicians. It is the potential for such acute allergic reactions, largely mediated by drug-specific immunoglobulin (1g)E antibodies, that presents the greatest impediment to readministering beta-la&am agents to patients with any history of prior sensitivity
PI. IMMUNOCHEMISTRY OF BETA-LACTAM ALLERGY Beta-lactams are ideal drug allergens for two reasons: First, few people in this society escape exposure to beta-lactams, which are dispensed with such frequency in the United States that universal exposure is virtually guaranteed. Second, beta-lactams possess the sine qua non for any drug hapten-the facility to link dovalently with macromolecules to provide multivalent hapten-carrier complexes. This is readily accomplished via opening of the beta-la&am ring to form by acylation an amide bond with epsilon-amino groups of lysine residues in proteins, either in serum or on cell surfaces. This particular conformation, known as the penicilloyl group, represents the major antigenic determinant because of the frequency and intensity of drug-specific immune responses to this moiety [3,41. There are, however, other ways in which various beta-lactam molecules can form covalent interactions with macromolecules; these are collectively known as the minor antigenic determinants [5,6]. Once these beta-la&am protein conjugates have been formed, an immune response can be initiated, and the whole spectrum of immunopathology can become clinically evident. It is similarly necessary for beta-la&am drugs to form adducts with macromolecules in order to elicit an allergic reaction in sensitized hosts. In this case, polyvalent presentation of the drug hapten is necessary for cross-linking of IgE antibody molecules bound to high-affinity receptors on basophilic leukocytes and tissue mast cells. Upon cell activation by cross-linking, the release of preformed and newly synthesized inflammatory mediators produces the syndromes of acute urticaria and anaphylaxis, which, representing the type 1 reactions mediated by the IgE antibody, as described by Gel1 and Coombs 171, are clinically the most problematic. Cephalosporin Cross-Reactivity
With the introduction of cephalosporins in 1965, it was hoped that the second class of beta-lactams would
Volume 88 (suppl 3C)
SYMPOSIUM
avoid the problem of penicillin allergy frequently encountered in the 1940s and 1950s. The beta-la&am ring is common to both classes; the difference between the two resides in a six-membered dihydrothiazine ring that replaces the five-membered thalazolidine ring of the penicillin nucleus (Figure 1). There has been extensive study of the cross-reactivity of penicillin and cephalosporins. From the voluminous literature, it is clearly easy to demonstrate consistent cross-reactivity both among penicillins and between penicillins and cephalosporins in laboratory animals. Clinically, however, this cross-reactivity is considerably more variable and is unpredictable on an individual basis [8-lo]. Studies of the cross-reactivity of IgE antibodies, as classically determined by skin testing of the intradermal type, have shown an approximate 30 to 50 percent cross-reactivity between penicillins and cephalosporins [ill. This rate of crossreactivity for IgE antibodies does not translate directly into clinical risk, however. Even in patients with positive results from penicillin skin tests, only about 50 to 70 percent can be expected to develop an observable acute reaction if treated with penicillin [12,13]. This reaction rate is likely to be much lower if cephalosporins are administered, but again the risk for individual patients ,is currently unassessable. These observations have led to the general conclusion that if cephalosporins are to be administered to patients with documented penicillin allergy, they must be given with a great deal of caution and concern and with appropriate safeguards. MONOBACTAM IMMUNOLOGY When the novel monocyclic beta-lactams (monobactams) were identified from microbial growth taken from the New Jersey pine barrens in the early 1980s there was an obvious interest in studying the immunology of this new antimicrobial class, which has no adjoining cyclic structure but is comprised of a simple beta-lactam nucleus by itself. In cooperation with Squibb, an unprecedented effort was begun to study prospectively the immunology of an antibiotic during its preclinical development and early clinical trials. A number of interesting observations that emerged from these studies [14-161 have led to .conclusions that, first; aztreonam is only weakly cross-reactive with existing beta-lactam compounds, and second, this compound may be less immunogenic than many of its predecessors.
ON ATTREONAM
/ ADKINSON
0 F&-NH 0
m
N
COOH
Penicillins
W Ceftazidime
ktreonam
I CH,
tlgure 1. Structures ot penlctlllns, the cephalosporln bactam aztreonam. From [18] with permission.
ceftazldime,
and the mono.
weak cross-reactivity in clinical studies. In collaboration with Saxon et al [15] at UCLA, 41 penicillin-allergic subjects with documented positive penicillin skin test results and 40 control subjects with negative penicillin skin test results were studied. With the support of Squibb, skin-testing reagents were prepared using aztreonam in place of benzylpenicillin to form both major and minor determinant analogues. These reagents were then used to skin test the penicillin-allergic patients and control subjects. Of the 41 allergic subjects, all but four had entirely negative skin test results to all four of the aztreonam skin-test reagents, All of the four positive results were weak and none could be reproduced when the subjects were retested four to six months later. These observations sug gested that any existing cross-reactivity of human IgE antibodies is weak and tenuous.
Monobactam Cross-Reactivity
Clinical Tolerance of Aztreoham in Penicillin-Allergic Subjects
In the earliest studies, conjugates of benzylpenicillin, cephalothin, and aztreonam were prepared with bovine thyroglobulin as a carrier and injected into rabbits to study the specificity of the resulting drugspecific antibodies. From these studies, the expected appreciable cross-reactivity between benzylpenicillin and cephalothin, and vice versa, was confirmed. In contrast, aztreonam did not fit either cephalosporin or penicillin,antibodies to any significant degree and thus showed negligible cross-reactivity. When rabbits were given adjuvants that forced the production of antibodies that recognized atreonam, the resulting antibodies did not fit very well with protein conjugates of either benzylpenicillin or cephalothin [14]. This result encouraged the pursuit of this apparent
These skin-test results encouraged clinical studies to evaluate the consequences of administering full therapeutic doses of aztreonam to patients with positive penicillin skin test results, and thereby putatively at high risk for allergic reactions to penicillin, Again, subjects were patients with documented positive penicillin skin tests for major and/or minor determinants; some subjects had previously experienced serious allergic reactions to penicillin. Of 21 subjects entered in the study, 19 had entirely negative results from aztreonam skin tests and went on to receive l-g doses of intramuscular aztreonam without complication [17]. Two subjects had positive skin test results with aztreonyl-polylysine, the major determinant analogue. Interestingly, these two patients were ex-
March 23, 1990
The American Journal of Medicine
Volume 88 (suppl 3C)
3c13s
SYMPOSIUM
ON AZTREONAM /ADKINSON
z 2
drug reactions was about 2 percent. Although these studies included no concurrent control groups treated with other beta-lactams with similar route and dose, the impression left by the literature is that this 2 percent reaction rate is somewhat less than that found in similar earlier studies with penicillins and first- and second-generation cephalosporins. Some of the reactions observed in the clinical trials of aztreonam did appear clinically to be IgE-mediated (Table II). There was a single patient, for example, who developed rash, fever, eosinophilia, and hypotension’on Day 5. There also were six cases of urticarial rash, all but one of which occurred late in the course of therapy. Finally, there was a variety of other reactions that could be considered to have possible immunopathogeneses. Some of these allergic drug reactions may be attributable to cross-reactivity with other beta-lactams that had previously sensitized the patient. This is certainly the most likely explanation of the acute reactions that occurred in response to early doses. On the other hand, it is difficult to distinguish cross-reactivity from the possibility that the drug itself can induce a drugspecific reaction in its own right, particularly with prolonged treatment or repeated courses. In recognition of the potential for de vzovo aztreonam allergy, studies during the past two years have been conducted to address this possibility more definitively.
i
Relative lmmunogenicity
TABLE I Adverse Reactions in 3,360 Patients Receiving Multiple Doses of Aztreonam* Reaction
Number of Patients (percent)
Adverse drug reactions Immunologic drug reactions
231(7.0)
I ‘: It;/ 64 (1.9)
XI’
*Data provided by E.R. Squibb as summarized from phase l/II IND clinical trials,
TABLE II Immunologic Drug Reactions in 3,360 Patients Receiving Aztreonam Number of Patients Type I (total) Rash, fever, hypotension, eosinophilra Urbcarial rash First dose Day 2 Day 3 Other, possibly immunologic (total) Nonurticarial rash and/or pruritus Purpuric rash Eosinophilia Urticaria, late onset Exfoliative dermatitis with eosinophilia Thrombocytopenia and hematuria Pyrexia with rash
:* : 3 64 53
1
\I1 in one patient on Day 5.
tremely sensitive to penicillin! with positive results from penicilloyl-polylysine skm tests at 1:lOO and l:l,OOO dilutions of the normal skin-testing concentration; they were not challenged with aztreonam. Finally, one patient initially had a weakly positive skin test reaction to aztreonam that was negative when repeated six months later; he was given aztreonam without incident. The important conclusion from this clinical study was that none of 20 penicillin-allergic subjects given a full therapeutic dose of aztreonam had an IgE-mediated reaction despite strikingly positive penicillin skin tests. Furthermore? there were no boosts in penicilloyl IgG or IgE antibodies in the serum of these penicillin-allergic patients as a result of aztreonam treatment [17]. This finding is compatible with the notion that there is very little immunologic cross-reactivity between aztreonam and penicillin. In the two years that aztreonam has been on the market, there has been further favorable experience with its use in penicillin-allergic subjects, including one published case report of a patient who tolerated aztreonam after unsuccessful attempts at penicillin desensitization [18]. CLINICAL EXPERIENCE WITH AZTREONAM Given the weak or negligible cross-reactivity associated with monobactams such as aztreonam, a low rate of allergic reactions would be expected from clinical experience with aztreonam. Table I shows the catalogue of adverse reactions reported from early clinical trials with aztreonam, involving 3,360 patients who received multiple doses and usually full clinical courses of aztreonam. The incidence of immunologic 3c-14s
March 23, 1990
The American Journal of Medicine
The study of relative immunogenicity of betalactam molecules is not an easy undertaking in clinical populations because virtually everyone has been previously exposed to penicillins and cephalosporins. The fact that monobactams such as aztreonam induce drug-specific antibody responses less frequently than antecedent compounds may quite simply reflect differential cumulative exposure. It was therefore decided to study an animal model by which a relatively naive set of animals could be evaluated and those with serologic evidence of prior exposure to beta-lactams eliminated. Outbred rabbits were chosen and screened for pre-existing penicilloyl IgG antibody, which was found in 27 percent of the rabbits; these were eliminated from further study. A large series of the naive rabbits were subsequently randomly assigned into treatment groups and administered benzylpenicillin, cephalothin, aztreonam, or ceftazidime. The initial treatment course consisted of 100 mg intramuscularly five days per week for three weeks. Following a twoweek rest period, a second course comprising two weeks of additional antibiotic treatment was given. The principal finding of the study is depicted in Figure 2. The cumulative response rate, defined as the percentage of rabbits in each group that responded with a detectable drug-specific IgG antibody response, is shown as a function of duration of treatment. Associated with benzylpenicillin and cephalothin was a rapid and progressive increase in the number of responding rabbits such that after a three-week course of therapy, about 60 to 70 percent had made detectable major-determinant antibody. Incidentally, this is the same rate of response that has been observed in clinical populations of inpatients treated with at least 2 g per day of penicillin for 10 or more days [19]. In contrast, both the ceftazidime and the aztreonam groups had a much lower frequency of re-
Volume 88 (suppl 3C)
SYMPOSIUM
sponse. Some rabbits responded to aztreonam during the second course of therapy, but no significant IgG antibody responses were seen in the ceftazidime group, This preliminary study strongly suggests that monobactams such as aztreonam and third-generation cephalosporins such as ceftazidime are much weaker immunogens in naive animals than are their antecedents, benzylpenicillin and cefalothin. The explanation that comes immediately to mind is that such differences in immunogenicity might reflect the facility with which acylation to protein occurs. In fact, however, although there are differences in the in vitro acylation rates with which various beta-lactam molecules couple to proteins, they var only within about one order of magnitude. Since at Peast two to three orders of magnitude of differences in immunogenicity have already been defined, it is unlikely that differing acylation rates represent the complete explanation. Perhaps the persistence of haptenized serum protein is important; studies to investigate this possibility are now under way. The influence of side chains versus the nuclear structure is also an important consideration: the bulky side chain of the later-generation cephalosporins and aztreonam may confer some protection from a drugspecific immune response against the nucleus. Pertinent here is the observation that all of the aztreonam antibody artificially raised ih rabbits, as well as all aztreonam-binding activity observed in the few human subjects in whom it was exhibited, have specificity for the side chain and not for the nuclear configuration or the protein carrier. Clinical studies will be needed to confirm this initial suggestion in rabbits that monobactams, such as aztreoham, will be significantly less immunogenic than some other beta-lactam molecules. Immunogenicity studies are now under way in populations of cystic fibrosis patients in Northern Eurbpe, where it is common practice to give prophylactic beta-lactam antibiotics every three or four months. It is hoped that with multiple courses of therapy in this patient population, prior antibiotic exposure will be normalized to some extent, which would allow a more direct estimate of relative immunogenicity in humans.
Pre
Wkl
Wk2
Wk3
ON AZTREONAM /ADKINSON
Rest
Wk4
Wk5
Figure 2. Relative immunogenicity of selected beta+lactam antibiotics in naive rai bits. PEN = benzylpenicillin; CEPH = cephalothin; AZT = aztreonam; CEFT ceftazidime.
long run, this may be an important advantageous property of this new class of antitiicrobials that will lead to fewer allergic drug reactions with tnonobactams than have been seen with the preceding classes of beta-lactam antibiotics. REFERENCES 1. Amdt KA, Jick H: Rates of cutaneous reactions to drugs. A feport from the BostOn Collaborative Drug Surveillance Program. JAMA 1976; 235: 918-922. 2. Adkinson NF Jr: Diagnosis of immunologic drug reactions. N Engl Reg Allergy Proc 1984; 5: 104-110. 3. Parker CW: lmmunochemical mechanisms in penicillin allergy. Fed Proc 1965; 24: 51-54: 4. De Week AL, Schneider CH, Gutersohn J: The role of penicilloylated protein Impurities, penicillin polymers and dimers in penictllin allergy. Int Arch Allergy Appl lmmunol 1968; 33: 535-567. 5. Levine BB: lmmunologtc mechanisms of penlcillin allergy A haptenlc model system for the study of allergic diseases of man. N Engl J Med 1966 275: 1115-1125. 6. Levine BB, Redmond AP: fylnor haptenic determinant-specific reagins of penicillin hy persensitivity in man. Int Arch Allergy Appl lmmunol 1969; 35: 445-455. 7. Gell PGH, Coombs RRA: Classification of allergic reactions reasponsible for clinical hypersensitivity and disease. In: Gell PGH, Coombs RRA, Hachmann PJ, eds. Clinical as. pects of immunology. Oxford: Blackwell Scientific Publications, 1975; 761-781. 8. Levine BB: Antigenicity and cross-rehctivity of penicillins and cephalosporins. J Infect Dis 1973; 128: S364-S366. 9. Saxon A, Beall GN, Rohr AS, ef al: Immediate hypersensitivity reactions to beta-lactam antibiotics. Ann Intern Med 1987; 107: 204-215. 10. Weiss ME, Adkinson NF Jr: Immediate hypersensitivity reactions to penicillin and related antibiotics. Clin Allergy 1988; 18: 515-540. 11. Girard J? Common antigenic determinants of penicillin G, ampicillin and the cephalo-
sporins demonstiated in men. Int Arch Allergy Appl lmmunol 1968; 33: 428-438.
COMMENTS There is a substantial body of data supporting extensive cross-reactivity amonk penicillins; cephalosporins, and even imipenems [9]. Although clinical crosssensitivity is unpredictable in individual cases, the underlying immunologic cross-reactivity indicates a need for prudence and appropriate precautibns when any of these classes of antibiotics is given to a patient with known or suspected penicillin allergy. On the other hand, the monobactam prototype aztreonam appeatis to have ver”y weak, almost negligible, cross reactivity with other classes of beta-lactam antibiotics and has been well-tolerated by highly pefiicillin-allergic patients. Perhaps more importantly, the monobactams appear to be only weakly immunogehic. In the
12. drown BC, Price EV, Moore MB: Penicilloyl-polylyslne as an intradermal test of penicil. Iln sensitivity. JAMA 1964; 189: 599-604. 13. Levine BB, Zolov DM: Prediction of penicillin allergy by immunological tests. J Allergy 1969; 43: 231-244. 14. Adkinson NF JI, Swabb EA, Sugerman AA: Immunology of the monobactam aztre. onam. Antimicrob Agents Chemother 1983; 25: 93-97. 15. Saxon A, Hassner A, Swabb EA, et al Lack of cross-reactivity between aztreonam, il monqbactam antibiotic, and penicillin in penicillin.all&gic subjects. J Infect Dis 1984; 149: 16-22, 16. Adkinson NF Jr, Saxon A, Spence MR, et at Cross-allergenicity and immunogenicity of aztreonam. Rev Infect Dis 1985; 7: S613-S621. 17. Adkinson NF Jr, Wheeler B, Swabb EA: Clinical tolerance of the monobactam aztre. onam in penicillin allergic subjects (abstr WS-26-4). In: Proceedings of the 14th Internatlonal Congress of Chemothbrapy, Kyoto, Japan, June 23 to 28, 1985. 18. Loria RC, Finnerty N, Wedner HJ: Successful use of aztreonam in a patient wlio failed oral penicillin desensitization. J Allergy Clin lmmunol 1989; 83: 735-737. 19. Adkinson NF Jr, Wheeler B: Risk factors for IgE-dependent reactions to penicillin. In: Kerr JW, Ganderton MA, eds. Proceedings of the Xl International Congress of Allergology and Clinical Immunology, London: Macmillan Press Ltd., 1983; 55-59.
March 23, 1990
The American Journal of Medicine
Volume 88 (suppl 3C)
3c-15s
ADKINSON,
and Cross-Allergenicity
JR.,M.D. Ba/timore,
Mary/and
Beta-lactams are responsible for more than half of the allergic drug reactions encountered in the hospital setting. Although most such reactions are mild, the potential for acute and life-threatening reactions cannot be underestimated when considering readministration of beta-lactam agents, such as the penicillins or cephalosporins, to persons who have previously exhibited sensitivity. In addition to concerns about possible allergic reactions to the beta-la&am antibiotics individually, considerable cross-reactivity has been demonstrated among such classes as the penicillins, the cephalosporins, and the imipenems, although it cannot yet be predicted on an individual basis. Early studies of the unique monocyclic beta-lactam-or monobactam-aztreonam, indicated that the new class demonstrated negligible cross-reactivity with the standard beta-lactams both experimentally and clinically. Further, aztreonam was associated with an extremely low (2 percent) incidence of immunologic drug reactions. Aztreonam also has been found to be welltolerated by highly penicillin-allergic patients. Although further clinical study is indicated, data so far are encouraging. If it is confirmed that monobactams such as aztreonam are minimally cross-reactive, well-tolerated by subjects allergic to other beta-lactam antibiotics, and only weakly immunogenic, fewer allergic reactions may be associated with antimicrobial therapy in the future than have been seen with the other available beta-lactam antibiotic drugs.
From the Department of Medicine, Johns Hopkins University School of Medicine, Johns Hopkins Asthma and Allergy Clinic, and the Good Samaritan Hospital, BaltiJr., M.D., Johns Hopkins Asthma and Allergy Center, 301 Bayview Boulevard, Balti-
3C-12s
March 23, 1990
of
The American Journal of Medicine
eta-la&am allergies are the most common drugB allergy problem encountered in the practice of medicine. More than half of all allergic drug reactions that occur in hospitalized patients can be attributed to beta-lactams [l]. Most are relatively benign and easily handled clinically, but serious, life-threatening, acute allergic reactions occasionally develop that understandably produce a great deal of concern in both patients and physicians. It is the potential for such acute allergic reactions, largely mediated by drug-specific immunoglobulin (1g)E antibodies, that presents the greatest impediment to readministering beta-la&am agents to patients with any history of prior sensitivity
PI. IMMUNOCHEMISTRY OF BETA-LACTAM ALLERGY Beta-lactams are ideal drug allergens for two reasons: First, few people in this society escape exposure to beta-lactams, which are dispensed with such frequency in the United States that universal exposure is virtually guaranteed. Second, beta-lactams possess the sine qua non for any drug hapten-the facility to link dovalently with macromolecules to provide multivalent hapten-carrier complexes. This is readily accomplished via opening of the beta-la&am ring to form by acylation an amide bond with epsilon-amino groups of lysine residues in proteins, either in serum or on cell surfaces. This particular conformation, known as the penicilloyl group, represents the major antigenic determinant because of the frequency and intensity of drug-specific immune responses to this moiety [3,41. There are, however, other ways in which various beta-lactam molecules can form covalent interactions with macromolecules; these are collectively known as the minor antigenic determinants [5,6]. Once these beta-la&am protein conjugates have been formed, an immune response can be initiated, and the whole spectrum of immunopathology can become clinically evident. It is similarly necessary for beta-la&am drugs to form adducts with macromolecules in order to elicit an allergic reaction in sensitized hosts. In this case, polyvalent presentation of the drug hapten is necessary for cross-linking of IgE antibody molecules bound to high-affinity receptors on basophilic leukocytes and tissue mast cells. Upon cell activation by cross-linking, the release of preformed and newly synthesized inflammatory mediators produces the syndromes of acute urticaria and anaphylaxis, which, representing the type 1 reactions mediated by the IgE antibody, as described by Gel1 and Coombs 171, are clinically the most problematic. Cephalosporin Cross-Reactivity
With the introduction of cephalosporins in 1965, it was hoped that the second class of beta-lactams would
Volume 88 (suppl 3C)
SYMPOSIUM
avoid the problem of penicillin allergy frequently encountered in the 1940s and 1950s. The beta-la&am ring is common to both classes; the difference between the two resides in a six-membered dihydrothiazine ring that replaces the five-membered thalazolidine ring of the penicillin nucleus (Figure 1). There has been extensive study of the cross-reactivity of penicillin and cephalosporins. From the voluminous literature, it is clearly easy to demonstrate consistent cross-reactivity both among penicillins and between penicillins and cephalosporins in laboratory animals. Clinically, however, this cross-reactivity is considerably more variable and is unpredictable on an individual basis [8-lo]. Studies of the cross-reactivity of IgE antibodies, as classically determined by skin testing of the intradermal type, have shown an approximate 30 to 50 percent cross-reactivity between penicillins and cephalosporins [ill. This rate of crossreactivity for IgE antibodies does not translate directly into clinical risk, however. Even in patients with positive results from penicillin skin tests, only about 50 to 70 percent can be expected to develop an observable acute reaction if treated with penicillin [12,13]. This reaction rate is likely to be much lower if cephalosporins are administered, but again the risk for individual patients ,is currently unassessable. These observations have led to the general conclusion that if cephalosporins are to be administered to patients with documented penicillin allergy, they must be given with a great deal of caution and concern and with appropriate safeguards. MONOBACTAM IMMUNOLOGY When the novel monocyclic beta-lactams (monobactams) were identified from microbial growth taken from the New Jersey pine barrens in the early 1980s there was an obvious interest in studying the immunology of this new antimicrobial class, which has no adjoining cyclic structure but is comprised of a simple beta-lactam nucleus by itself. In cooperation with Squibb, an unprecedented effort was begun to study prospectively the immunology of an antibiotic during its preclinical development and early clinical trials. A number of interesting observations that emerged from these studies [14-161 have led to .conclusions that, first; aztreonam is only weakly cross-reactive with existing beta-lactam compounds, and second, this compound may be less immunogenic than many of its predecessors.
ON ATTREONAM
/ ADKINSON
0 F&-NH 0
m
N
COOH
Penicillins
W Ceftazidime
ktreonam
I CH,
tlgure 1. Structures ot penlctlllns, the cephalosporln bactam aztreonam. From [18] with permission.
ceftazldime,
and the mono.
weak cross-reactivity in clinical studies. In collaboration with Saxon et al [15] at UCLA, 41 penicillin-allergic subjects with documented positive penicillin skin test results and 40 control subjects with negative penicillin skin test results were studied. With the support of Squibb, skin-testing reagents were prepared using aztreonam in place of benzylpenicillin to form both major and minor determinant analogues. These reagents were then used to skin test the penicillin-allergic patients and control subjects. Of the 41 allergic subjects, all but four had entirely negative skin test results to all four of the aztreonam skin-test reagents, All of the four positive results were weak and none could be reproduced when the subjects were retested four to six months later. These observations sug gested that any existing cross-reactivity of human IgE antibodies is weak and tenuous.
Monobactam Cross-Reactivity
Clinical Tolerance of Aztreoham in Penicillin-Allergic Subjects
In the earliest studies, conjugates of benzylpenicillin, cephalothin, and aztreonam were prepared with bovine thyroglobulin as a carrier and injected into rabbits to study the specificity of the resulting drugspecific antibodies. From these studies, the expected appreciable cross-reactivity between benzylpenicillin and cephalothin, and vice versa, was confirmed. In contrast, aztreonam did not fit either cephalosporin or penicillin,antibodies to any significant degree and thus showed negligible cross-reactivity. When rabbits were given adjuvants that forced the production of antibodies that recognized atreonam, the resulting antibodies did not fit very well with protein conjugates of either benzylpenicillin or cephalothin [14]. This result encouraged the pursuit of this apparent
These skin-test results encouraged clinical studies to evaluate the consequences of administering full therapeutic doses of aztreonam to patients with positive penicillin skin test results, and thereby putatively at high risk for allergic reactions to penicillin, Again, subjects were patients with documented positive penicillin skin tests for major and/or minor determinants; some subjects had previously experienced serious allergic reactions to penicillin. Of 21 subjects entered in the study, 19 had entirely negative results from aztreonam skin tests and went on to receive l-g doses of intramuscular aztreonam without complication [17]. Two subjects had positive skin test results with aztreonyl-polylysine, the major determinant analogue. Interestingly, these two patients were ex-
March 23, 1990
The American Journal of Medicine
Volume 88 (suppl 3C)
3c13s
SYMPOSIUM
ON AZTREONAM /ADKINSON
z 2
drug reactions was about 2 percent. Although these studies included no concurrent control groups treated with other beta-lactams with similar route and dose, the impression left by the literature is that this 2 percent reaction rate is somewhat less than that found in similar earlier studies with penicillins and first- and second-generation cephalosporins. Some of the reactions observed in the clinical trials of aztreonam did appear clinically to be IgE-mediated (Table II). There was a single patient, for example, who developed rash, fever, eosinophilia, and hypotension’on Day 5. There also were six cases of urticarial rash, all but one of which occurred late in the course of therapy. Finally, there was a variety of other reactions that could be considered to have possible immunopathogeneses. Some of these allergic drug reactions may be attributable to cross-reactivity with other beta-lactams that had previously sensitized the patient. This is certainly the most likely explanation of the acute reactions that occurred in response to early doses. On the other hand, it is difficult to distinguish cross-reactivity from the possibility that the drug itself can induce a drugspecific reaction in its own right, particularly with prolonged treatment or repeated courses. In recognition of the potential for de vzovo aztreonam allergy, studies during the past two years have been conducted to address this possibility more definitively.
i
Relative lmmunogenicity
TABLE I Adverse Reactions in 3,360 Patients Receiving Multiple Doses of Aztreonam* Reaction
Number of Patients (percent)
Adverse drug reactions Immunologic drug reactions
231(7.0)
I ‘: It;/ 64 (1.9)
XI’
*Data provided by E.R. Squibb as summarized from phase l/II IND clinical trials,
TABLE II Immunologic Drug Reactions in 3,360 Patients Receiving Aztreonam Number of Patients Type I (total) Rash, fever, hypotension, eosinophilra Urbcarial rash First dose Day 2 Day 3 Other, possibly immunologic (total) Nonurticarial rash and/or pruritus Purpuric rash Eosinophilia Urticaria, late onset Exfoliative dermatitis with eosinophilia Thrombocytopenia and hematuria Pyrexia with rash
:* : 3 64 53
1
\I1 in one patient on Day 5.
tremely sensitive to penicillin! with positive results from penicilloyl-polylysine skm tests at 1:lOO and l:l,OOO dilutions of the normal skin-testing concentration; they were not challenged with aztreonam. Finally, one patient initially had a weakly positive skin test reaction to aztreonam that was negative when repeated six months later; he was given aztreonam without incident. The important conclusion from this clinical study was that none of 20 penicillin-allergic subjects given a full therapeutic dose of aztreonam had an IgE-mediated reaction despite strikingly positive penicillin skin tests. Furthermore? there were no boosts in penicilloyl IgG or IgE antibodies in the serum of these penicillin-allergic patients as a result of aztreonam treatment [17]. This finding is compatible with the notion that there is very little immunologic cross-reactivity between aztreonam and penicillin. In the two years that aztreonam has been on the market, there has been further favorable experience with its use in penicillin-allergic subjects, including one published case report of a patient who tolerated aztreonam after unsuccessful attempts at penicillin desensitization [18]. CLINICAL EXPERIENCE WITH AZTREONAM Given the weak or negligible cross-reactivity associated with monobactams such as aztreonam, a low rate of allergic reactions would be expected from clinical experience with aztreonam. Table I shows the catalogue of adverse reactions reported from early clinical trials with aztreonam, involving 3,360 patients who received multiple doses and usually full clinical courses of aztreonam. The incidence of immunologic 3c-14s
March 23, 1990
The American Journal of Medicine
The study of relative immunogenicity of betalactam molecules is not an easy undertaking in clinical populations because virtually everyone has been previously exposed to penicillins and cephalosporins. The fact that monobactams such as aztreonam induce drug-specific antibody responses less frequently than antecedent compounds may quite simply reflect differential cumulative exposure. It was therefore decided to study an animal model by which a relatively naive set of animals could be evaluated and those with serologic evidence of prior exposure to beta-lactams eliminated. Outbred rabbits were chosen and screened for pre-existing penicilloyl IgG antibody, which was found in 27 percent of the rabbits; these were eliminated from further study. A large series of the naive rabbits were subsequently randomly assigned into treatment groups and administered benzylpenicillin, cephalothin, aztreonam, or ceftazidime. The initial treatment course consisted of 100 mg intramuscularly five days per week for three weeks. Following a twoweek rest period, a second course comprising two weeks of additional antibiotic treatment was given. The principal finding of the study is depicted in Figure 2. The cumulative response rate, defined as the percentage of rabbits in each group that responded with a detectable drug-specific IgG antibody response, is shown as a function of duration of treatment. Associated with benzylpenicillin and cephalothin was a rapid and progressive increase in the number of responding rabbits such that after a three-week course of therapy, about 60 to 70 percent had made detectable major-determinant antibody. Incidentally, this is the same rate of response that has been observed in clinical populations of inpatients treated with at least 2 g per day of penicillin for 10 or more days [19]. In contrast, both the ceftazidime and the aztreonam groups had a much lower frequency of re-
Volume 88 (suppl 3C)
SYMPOSIUM
sponse. Some rabbits responded to aztreonam during the second course of therapy, but no significant IgG antibody responses were seen in the ceftazidime group, This preliminary study strongly suggests that monobactams such as aztreonam and third-generation cephalosporins such as ceftazidime are much weaker immunogens in naive animals than are their antecedents, benzylpenicillin and cefalothin. The explanation that comes immediately to mind is that such differences in immunogenicity might reflect the facility with which acylation to protein occurs. In fact, however, although there are differences in the in vitro acylation rates with which various beta-lactam molecules couple to proteins, they var only within about one order of magnitude. Since at Peast two to three orders of magnitude of differences in immunogenicity have already been defined, it is unlikely that differing acylation rates represent the complete explanation. Perhaps the persistence of haptenized serum protein is important; studies to investigate this possibility are now under way. The influence of side chains versus the nuclear structure is also an important consideration: the bulky side chain of the later-generation cephalosporins and aztreonam may confer some protection from a drugspecific immune response against the nucleus. Pertinent here is the observation that all of the aztreonam antibody artificially raised ih rabbits, as well as all aztreonam-binding activity observed in the few human subjects in whom it was exhibited, have specificity for the side chain and not for the nuclear configuration or the protein carrier. Clinical studies will be needed to confirm this initial suggestion in rabbits that monobactams, such as aztreoham, will be significantly less immunogenic than some other beta-lactam molecules. Immunogenicity studies are now under way in populations of cystic fibrosis patients in Northern Eurbpe, where it is common practice to give prophylactic beta-lactam antibiotics every three or four months. It is hoped that with multiple courses of therapy in this patient population, prior antibiotic exposure will be normalized to some extent, which would allow a more direct estimate of relative immunogenicity in humans.
Pre
Wkl
Wk2
Wk3
ON AZTREONAM /ADKINSON
Rest
Wk4
Wk5
Figure 2. Relative immunogenicity of selected beta+lactam antibiotics in naive rai bits. PEN = benzylpenicillin; CEPH = cephalothin; AZT = aztreonam; CEFT ceftazidime.
long run, this may be an important advantageous property of this new class of antitiicrobials that will lead to fewer allergic drug reactions with tnonobactams than have been seen with the preceding classes of beta-lactam antibiotics. REFERENCES 1. Amdt KA, Jick H: Rates of cutaneous reactions to drugs. A feport from the BostOn Collaborative Drug Surveillance Program. JAMA 1976; 235: 918-922. 2. Adkinson NF Jr: Diagnosis of immunologic drug reactions. N Engl Reg Allergy Proc 1984; 5: 104-110. 3. Parker CW: lmmunochemical mechanisms in penicillin allergy. Fed Proc 1965; 24: 51-54: 4. De Week AL, Schneider CH, Gutersohn J: The role of penicilloylated protein Impurities, penicillin polymers and dimers in penictllin allergy. Int Arch Allergy Appl lmmunol 1968; 33: 535-567. 5. Levine BB: lmmunologtc mechanisms of penlcillin allergy A haptenlc model system for the study of allergic diseases of man. N Engl J Med 1966 275: 1115-1125. 6. Levine BB, Redmond AP: fylnor haptenic determinant-specific reagins of penicillin hy persensitivity in man. Int Arch Allergy Appl lmmunol 1969; 35: 445-455. 7. Gell PGH, Coombs RRA: Classification of allergic reactions reasponsible for clinical hypersensitivity and disease. In: Gell PGH, Coombs RRA, Hachmann PJ, eds. Clinical as. pects of immunology. Oxford: Blackwell Scientific Publications, 1975; 761-781. 8. Levine BB: Antigenicity and cross-rehctivity of penicillins and cephalosporins. J Infect Dis 1973; 128: S364-S366. 9. Saxon A, Beall GN, Rohr AS, ef al: Immediate hypersensitivity reactions to beta-lactam antibiotics. Ann Intern Med 1987; 107: 204-215. 10. Weiss ME, Adkinson NF Jr: Immediate hypersensitivity reactions to penicillin and related antibiotics. Clin Allergy 1988; 18: 515-540. 11. Girard J? Common antigenic determinants of penicillin G, ampicillin and the cephalo-
sporins demonstiated in men. Int Arch Allergy Appl lmmunol 1968; 33: 428-438.
COMMENTS There is a substantial body of data supporting extensive cross-reactivity amonk penicillins; cephalosporins, and even imipenems [9]. Although clinical crosssensitivity is unpredictable in individual cases, the underlying immunologic cross-reactivity indicates a need for prudence and appropriate precautibns when any of these classes of antibiotics is given to a patient with known or suspected penicillin allergy. On the other hand, the monobactam prototype aztreonam appeatis to have ver”y weak, almost negligible, cross reactivity with other classes of beta-lactam antibiotics and has been well-tolerated by highly pefiicillin-allergic patients. Perhaps more importantly, the monobactams appear to be only weakly immunogehic. In the
12. drown BC, Price EV, Moore MB: Penicilloyl-polylyslne as an intradermal test of penicil. Iln sensitivity. JAMA 1964; 189: 599-604. 13. Levine BB, Zolov DM: Prediction of penicillin allergy by immunological tests. J Allergy 1969; 43: 231-244. 14. Adkinson NF JI, Swabb EA, Sugerman AA: Immunology of the monobactam aztre. onam. Antimicrob Agents Chemother 1983; 25: 93-97. 15. Saxon A, Hassner A, Swabb EA, et al Lack of cross-reactivity between aztreonam, il monqbactam antibiotic, and penicillin in penicillin.all&gic subjects. J Infect Dis 1984; 149: 16-22, 16. Adkinson NF Jr, Saxon A, Spence MR, et at Cross-allergenicity and immunogenicity of aztreonam. Rev Infect Dis 1985; 7: S613-S621. 17. Adkinson NF Jr, Wheeler B, Swabb EA: Clinical tolerance of the monobactam aztre. onam in penicillin allergic subjects (abstr WS-26-4). In: Proceedings of the 14th Internatlonal Congress of Chemothbrapy, Kyoto, Japan, June 23 to 28, 1985. 18. Loria RC, Finnerty N, Wedner HJ: Successful use of aztreonam in a patient wlio failed oral penicillin desensitization. J Allergy Clin lmmunol 1989; 83: 735-737. 19. Adkinson NF Jr, Wheeler B: Risk factors for IgE-dependent reactions to penicillin. In: Kerr JW, Ganderton MA, eds. Proceedings of the Xl International Congress of Allergology and Clinical Immunology, London: Macmillan Press Ltd., 1983; 55-59.
March 23, 1990
The American Journal of Medicine
Volume 88 (suppl 3C)
3c-15s
