Archives of Disease in Childhood 1997;76:513–517 513

Aminopenicillin allergy Arch Dis Child: first published as 10.1136/adc.76.6.513 on 1 June 1997. Downloaded from

A Romano, D Quaratino, G Papa, M Di Fonso, A Venuti

Abstract drug has been started, are not associated with Eighty two children with histories of IgE antibodies, and do not appear to predis- maculopapular or urticarial rashes dur- pose the patient to urticarial reactions.1 Over ing aminopenicillin treatment underwent the last decade, however, various investigators skin tests, patch tests, radioallergosorbent have come to the conclusion that these rashes assays and, in some cases, oral challenges. often represent type IV (cell mediated) hyper- Hypersensitivity was diagnosed in eight sensitivity, which is associated with patch test (9.8%): immediate in four with urticarial and/or delayed (that is six or more hours after reactions and delayed (that is cell medi- administration) intradermal test positivity.8–13 ated) in four with maculopapular rashes. In some cases, tested skin has been biopsied In 49 children (38 with maculopapular and found to present histological features simi- eruptions, 11 with urticarial/angiooede- lar to those of allergic contact dermatitis,9–13 matous reactions), negative allergologic and oral challenges with the suspected drug findings were confirmed using oral chal- have provoked delayed onset skin rashes.9101213 lenges with the suspected drug. Maculo- This approach has proved to be capable of papular rashes may reflect delayed detecting delayed forms of aminopenicillin hypersensitivity to aminopenicillins, hypersensitivity even when several years have which can be diagnosed on the basis of late elapsed since the last exposure to the drug.12 intradermal reactions and/or patch test Nevertheless, in many other cases of amino- positivity. The allergen panel must include penicillin associated maculopapular rashes the suspected aminopenicillin itself, as there is no evidence of immune mediated many cases are side chain specific. Most mechanisms of any type, and challenges with patients with urticarial reactions (more the suspected aminopenicillin are well typical of immediate hypersensitivity) will tolerated.12 14 also react to penicilloyl polylysine and The present study was conducted to deter- minor determinant mixture. The time mine the incidence of true allergy in a group of elapsed between the reaction and testing children with histories of cutaneous eruptions must be considered if negative results during aminopenicillin treatment. emerge, because IgE mediated sensitivity http://adc.bmj.com/ (unlike cell mediated forms) declines in Patients and methods the absence of antigen exposure. This study included all children with histories (Arch Dis Child 1997;76:513–517) of adverse cutaneous reactions to aminopeni- cillins seen by our staV between January 1990 Keywords: aminopenicillins; hypersensitivity; maculo- papular rash and December 1995. The following allergologic investigations

were performed. on September 23, 2021 by guest. Protected copyright. Many patients who develop rashes, urticaria, or other skin symptoms during â-lactam treat- (1) PRICK AND INTRADERMAL SKIN TESTS ment are automatically labelled ‘allergic to These were done using: penicillin’, and subsequent use of these antibi- (A) Penicilloyl polylysine (PPL) (Allergo- otics is avoided, even if they appear to be the pen; Reinbeck, Germany). The undiluted solu- drugs of choice for treatment of serious tion was obtained by reconstituting the infections. Urticaria (or hives), which is one of lyophilised contents of a vial containing 0.175 the classic manifestations of type I (or IgE mg PPL plus 20 mg mannitol in 5 ml buVer. A Institute of Internal and Geriatric mediated) hypersensitivity in general, occurs in 1:10 dilution of this solution in 0.9% sodium Medicine, Catholic 4.5% of patients treated with penicillin or its chloride (NaCl) was initially used. When University of St Cuore, analogues.1 However, many patients who have results were negative, testing was repeated with Rome, Italy experienced urticarial reactions to â-lactams the undiluted solution. A Romano prove to be skin test negative for the major and (B) Minor determinant mixture (MDM) D Quaratino minor penicillin determinants, and in many (Allergopen), containing 0.6 mg benzylpenicil- G Papa M Di Fonso cases subsequent administration of penicillin lin and 0. 5 mg benzylpenicilloate/ml, initially 2–4 A Venuti or the suspected drug is well tolerated. diluted 1:10 000 in 0.9% NaCl. Tests were Maculopapular or morbilliform rashes are repeated with a 1:10 dilution when initial Correspondence to: even more common during treatment with results were negative. Dr Antonino Romano, 5–7 Ambulatorio di Allergologia, these drugs, particularly ampicillin. The (C) Potassium penicillin G (Bristol Italiana; Complesso Integrato incidence of morbilliform rashes during Sermoneta, Italy), diluted in 0.9% NaCl and Columbus, Via della Pineta therapy with the latter drug has been estimated administered at increasing concentrations Sacchetti 506, 00168 Rome, at 9.5%.1 Until recently, these rashes were ranging from 0.1 IU/ml to 10 000 IU/ml. Italy. thought to be late, idiopathic drug reactions. (D) Ampicillin (Amplital, Farmitalia; Milan, Accepted 13 January 1997 They usually appear 2–3 days or more after the Italy) and amoxycillin (Velamox, Zambeletti; 514 Romano, Quaratino, Papa, Di Fonso, Venuti

Table 1 Clinical manifestations described by subjects studied and results of allergologic adverse cutaneous reactions to aminopenicil- tests lins. In over 90% of the cases, the antibiotic had Arch Dis Child: first published as 10.1136/adc.76.6.513 on 1 June 1997. Downloaded from been prescribed to treat upper respiratory tract No of patients No with immediate No with delayed Symptoms studied hypersensitivity* hypersensitivity† infections. Our investigation was performed at a mean (SD) 30.6 (28.4) months (range Maculopapular rash 42 – 4 Urticaria 29 2 – 1–120) after the most recent drug exposure. Angio-oedema 4 1 – Sixty one of the 82 children (74.4%) had Urticaria/angio-oedema 2 – – reportedly reacted to amoxycillin (in three Urticaria/angio-oedema and dyspnoea 1 1 – cases administered with clavulanic acid), 17 Erythema 4 – – (20.7%) had reacted to ampicillin (bacampicil- * Immediate hypersensitivity diagnosed on the basis of skin tests and RAST. lin in one case, ampicillin plus sulbactam in † Delayed hypersensitivity diagnosed on the basis of delayed intradermal reactivity and patch tests. another), and four had experienced reactions to more than one aminopenicillin (ampicillin Milan, Italy), both used at concentrations of 1 and amoxycillin in three cases, ampicillin and mg/ml and 20 mg/ml, after dilution in 0.9% bacampicillin in one). Thirteen (15.85%) had NaCl. also reacted to cephalosporins. All haptens were initially tested on the volar The reactive manifestations (based on par- forearm skin using the prick technique. Reac- ents’ descriptions and, in some cases, paediatri- tions were considered positive when a weal >3 cians’ reports) are summarised in table 1. In mm in diameter appeared within 15–20 81/82 cases, the symptoms had been exclu- minutes. When negative results were observed, sively cutaneous; the remaining child had 0.02 ml of the hapten solution was injected experienced both cutaneous and respiratory intradermally, and readings were made at 20 symptoms. Forty two of the children had minutes and 48 hours. Results were defined as suVered maculopapular reactions, 36 had positive when weals of >5 mm appeared. Posi- experienced urticaria/angio-oedema, and four tive controls for prick and intradermal tests had developed diVuse erythematous reactions. were performed with histamine (10 mg/ml and On the basis of these tests, four of the 82 1 mg/ml, respectively); normal saline was used children (4.9%) were diagnosed as having as a negative control. immediate hypersensitivity. All four had expe- rienced their most recent reaction less than one (2) PATCH TESTS year before testing. Two had experienced urti- These were administered with (a) penicillin G carial reactions, one had developed angio- (5000 IU/g Vaseline), (b) ampicillin (5% in oedema, and the fourth had reacted with urti- Vaseline), and (c) amoxycillin (5% in Vaseline). caria, angio-oedema, and dyspnoea. In all four The patches were applied to the interscapular cases, the onset of symptoms had occurred region of the back and evaluated according to within one hour of the initial administration of the criteria of Wilkinson et al after 48 and 72 the drug and the rash had persisted for hours.15

approximately 24 hours after the antibiotic had http://adc.bmj.com/ been withdrawn and antihistaminic and steroid (3) RADIOALLERGOSORBENT TESTS (RASTS) treatment started. Two of these four children In vitro assays for antigen specific IgE to peni- presented skin test positivity for all of the aller- cilloyl G, penicilloyl V, ampicillin, and amoxy- gens tested, the third was positive for PPL and cillin were performed using the Phadebas penicillin G, and the fourth reacted only to RAST (Pharmacia; Uppsala, Sweden). penicillin G. RASTs were negative in 3/4; the

(4) ORAL CHALLENGES fourth was positive for penicilloyls G and V. These were performed only in negative cases, Thirty six other children with histories of on September 23, 2021 by guest. Protected copyright. on children with histories of either maculopa- urticarial/angio-oedematous or diVuse ery- pular reactions or erythematous or urticarial/ thematous reactions appearing within 24 hours angio-oedematous manifestations. The latter after the start of aminopenicillin treatment had been tested within one year of their last were negative in all tests. In 11 of these cases reaction. Informed consent was obtained from the most recent reaction had occurred less than all parents, and the children were admitted to one year before testing. Oral challenges were the day hospital for 24 hours for administration performed (nine with amoxycillin, two with of the challenge. The drugs—ampicillin (Am- ampicillin), none of which provoked symp- plital; Farmitalia) or amoxycillin (Velamox; toms. Zambeletti)—were administered orally at a Four other children (4.9%) presented results dose of 5 mg. If no reaction occurred within 72 indicative of cell mediated or delayed hypersen- hours, the challenge was repeated one week sitivity: all had patch test positivity and late later with 50 mg, and eventually with a intradermal reactions to both ampicillin and therapeutic dose (up to 500 mg, depending on amoxycillin; there were no delayed reactions to weight). PPL, penicillin G, or MDM. All four had experienced morbilliform rashes, in one case STATISTICAL ANALYSIS accompanied by angio-oedema. The symp- The diVerences between percentages were toms had appeared 24 hours or more after the evaluated with the ÷2 test. start of aminopenicillin therapy and had persisted for 7–15 days after the drug had been Results discontinued and steroids had been started. We examined a total of 82 children (39 boys, 43 The causes of previous reactions had been girls) ranging in age from 3 to 12 years (mean ampicillin in two cases (one of whom had (SD) age: 6.8 (2.9) years) with histories of experienced reactions during two separate Aminopenicillin allergy 515

cycles of treatment), bacampicillin in another, other viruses that provoke non-specific upper and amoxycillin in the fourth. The interval that respiratory tract symptoms can also produce Arch Dis Child: first published as 10.1136/adc.76.6.513 on 1 June 1997. Downloaded from had elapsed between the most recent reaction morbilliform rashes.19 The underlying mecha- and our work-up was two years in two cases, nisms are unclear in many cases: viral exan- one month in the third, and six years in the thems have been attributed to the destruction fourth. of infected epidermal cells by cytotoxic CD8 T The remaining 38 children who had experi- cells,20 but they may also be the result of inter- enced morbilliform rashes were negative in all action between the virus and the antibiotic, as allergologic tests. Oral challenges were carried described in Epstein-Barr viral infections.21 On out (34 with amoxycillin, four with ampicillin), the other hand, some authors have reported none of which provoked symptoms. positive reactions to oral challenges in cases of delayed rashes associated with â-lactam Discussion therapy in which skin tests were completely Our findings indicate that maculopapular negative, even at the 48 hour reading.913In the rashes are common during aminopenicillin light of these experiences (which diVer from treatment, particularly among children. our own), it would appear that mechanisms Roughly one third of the adults (60/195, unrelated to lymphocytes cannot be excluded. 30.8%) we examined recently in an analogous It should be emphasised that the RAST was study12 developed them as opposed to a signifi- negative for all 42 patients with maculopapular cantly higher percentage (51.2%) among reactions. This result confirms our findings the 82 children described here (÷2 = 9.52; with adults and lead us to believe that this test p <0.01). Although maculopapular rashes are could be omitted from the study protocol of common in both allergic conditions and such reactions. systemic infections, those occurring during Urticarial episodes during aminopenicillin antibiotic therapy are almost always attributed therapy for an acute febrile illness may well be to the drug rather than to the infectious agent. a direct eVect of the infectious organism (for Moreover, late (that is six hours or more) posi- example Epstein-Barr virus, coxsackie viruses tive reactions to intradermal tests with the sus- A9, A16, B4 and B5, echovirus 11, etc).19 They pected drug (particularly those associated with may, of course, also be allergic manifestations, patch test positivity for the same drug) are now and in these cases appear to be typical of type I considered indicative of delayed hypersensi- reactions. tivity.9–13 While this finding was quite common Our experience indicates that skin test posi- among our adults who reported morbilliform tivity for PPL, MDM, and/or penicillin G is ‘reactions’ to aminopenicillins (33/60,55.0%), much more common in patients with immedi- only 9.5% of such reactions experienced by the ate hypersensitivity to â-lactams than in those children described here seemed to be cell with the cell mediated form. In particular, all mediated (÷2 = 20.18; p <0.001). It would thus four children with IgE mediated sensitivity

appear that maculopapular rashes developing were penicillin G positive, while none of those http://adc.bmj.com/ in a child after several days of aminopenicillin with delayed hypersensitivity was. However, as treatment are much less likely to be manifesta- we observed in adults with immediate tions of drug allergy than those reported by reactions,12 there may be cases of side chain adults. specific hypersensitivity. Thus, until we have None the less, the possibility of an immune studies based on larger samples, we recom- mediated reaction should be excluded by mend using the entire panel of reagents. means of skin and patch testing, even in Although the RAST proved to be less sensi-

children. These tests must be performed with tive than the skin tests, we still recommend that on September 23, 2021 by guest. Protected copyright. the aminopenicillin suspected of having caused it be performed in immediate reactions be- the reaction, rather than with the major and cause, as we have observed with adults, there minor penicillin determinants alone. Voorhorst may be rare skin test negative, RAST positive and Sparreboom, in fact, found that the amino cases.12 group of the aminopenicillin side chain is The vast majority of the children in our essential for a delayed allergic reaction to this study with urticarial reactions, however, pre- group of drugs.16 Moreover, recent data sented negative skin test and RAST results. In indicate that T cell recognition of penicillins such cases, some allergologists use oral chal- can be influenced by the side chain structure of lenges to confirm the absence of allergy. In a the drug molecule.17 Failure to include the sus- study of 346 skin test negative patients with pected aminopenicillin in the panel of test rea- histories of reactions to â-lactams, 3% reacted gents might thus explain the negative allergo- to penicillin challenges.2 Mendelson et al logic findings reported in some patients who administered a 10 day oral penicillin challenge have reacted to these drugs.18 and repeated skin tests after four weeks or Our experience indicates that delayed hyper- more.4 In a group of 219 patients (aged 4 sensitivity to aminopenicillins is a persistent months–20 years) subjected to this protocol, condition. Therefore, negative results in patch three developed self limiting skin symptoms and intradermal tests using an appropriate during the challenge, but when repeated, skin antigen panel appear to be a reliable indicator tests were again negative. None of the other that the morbilliform rashes experienced by the 216 experienced symptoms during the chal- patient are not immunological reactions to the lenge, but two had positive skin tests after chal- drug, but rather manifestations of the underly- lenge. ing disease. In addition to those associated with In patients who have reacted to penicillins the classic childhood exanthems, a number of with urticaria and/or angio-oedema, however, 516 Romano, Quaratino, Papa, Di Fonso, Venuti

IMMEDIATE TYPE (within 1 hour) ACCELERATED TYPE (>1–24 hours) LATE TYPE (after 24 hours or more) Arch Dis Child: first published as 10.1136/adc.76.6.513 on 1 June 1997. Downloaded from

Erythema, urticaria and/or angio-oedema, Urticarial or anaphylactic shock maculopapular rash

Perform skin tests and RAST Perform patch tests and skin tests at 20 min–48 hour readings

Negative Positive Positive Negative

Immediate Delayed hypersensitivity hypersensitivity

Consider time elapsed since last adverse reaction <12 months >12 months (all cases) (selected cases)

Advise avoidance of Perform oral challenge β-lactam therapy Perform oral challenge

Negative Positive Positive Negative

Repeat tests—after > 4 weeks—Repeat tests

Negative Positive Positive Negative

Probable Immediate Undetermined Undetermined Probable coincidental hypersensitivity pathogenic pathogenic coincidental reactions mechanism mechanism reactions

β-lactam β-lactam therapy therapy may be may be advised advised Figure 1 Algorithm for allergologic work-up of adverse reactions to aminopenicillins.

negative skin test results must be interpreted in children who did not suVer severe reactions, for http://adc.bmj.com/ the light of the period that has elapsed since the whom there is a greater than normal likelihood last known exposure to the drug. In these cases, that penicillin therapy will be necessary, oral the frequency of positive skin test results challenge confirmation of skin test negativity declines as the interval between drug exposure may be more appropriate. Taking into account and testing increases.232223 In a follow up the experiences of others as well as our own, we study of 42 children with documented IgE propose the diagnostic algorithm displayed in mediated hypersensitivity to penicillins, Chan- fig 1.

dra et al observed the negativisation of skin In any case, the testing protocol used here on September 23, 2021 by guest. Protected copyright. tests in one third of the children who had pre- allowed us to clarify the nature of the sented positive reactions one year earlier.3 ‘reactions’ experienced by 57 (69.5%) of the 82 In the present study, of the 40 children who children studied. Hypersensitivity was con- had experienced erythematous or urticarial firmed in 9.8% of them. In 49 other children reactions, only 15 (37.5%) were tested by our (59.8%), the negative results of skin and patch staV within one year of their most recent reac- testing as well as oral challenges allowed us to tion, including the four who presented skin reasonably exclude the possibility of allergy to tests indicative of IgE mediated sensitivity. In penicillins. the remaining 11 cases, skin test negativity In conclusion, while the percentage of should be considered a fairly reliable indicator aminopenicillin associated skin rashes actually of the absence of immediate hypersensitivity, as representing allergic phenomena is signifi- confirmed in all cases by the absence of cantly lower in children than in adults, our reactions to oral challenges. study demonstrates that the possibility of both In contrast, the test results are much less immediate and delayed hypersensitivity does reliable for the 25 children who had negative exist. The diagnostic protocol described here skin tests more than one year after their can be helpful in clarifying the nature of many urticarial/angio-oedematous reactions, and the of these reactions. Any doubt can be eliminated symptoms experienced by a small percentage with controlled oral challenges and, if neces- of them may have actually been manifestations sary, retesting. of IgE mediated hypersensitivity. In these cases, considering parental resistance to these procedures and the risk of triggering another 1 Saxon A. Immediate hypersensitivity reactions to beta- lactam antibiotics. Ann Intern Med 1987;107:204-15. reaction, we generally defer oral challenges and 2 Green GR, Rosenblum AH, Sweet LC. Evaluation of peni- advise avoidance of â-lactam treatment. In cillin hypersensitivity: value of clinical history and skin test- Aminopenicillin allergy 517

ing with penicilloyl-polylysine and penicillin G. J Allergy 13 Terrados S, Blanca M, Garcia J, et al. Nonimmediate

Clin Immunol 1977;60:339-45. reactions to betalactams: prevalence and role of the diVer- Arch Dis Child: first published as 10.1136/adc.76.6.513 on 1 June 1997. Downloaded from 3 Chandra RK, Joglekar SA, Tomas E. Penicillin allergy: anti- ent penicillins. Allergy 1995;50:563-7. penicillin IgE antibodies and immediate hypersensitivity 14 Bierman CW, Pierson WE, Zeitz SJ, HoVman LS, skin reactions employing major and minor determinants of VanArsdel PP Jr. Reactions associated with ampicillin penicillin. Arch Dis Child 1980;55:857-60. therapy. JAMA 1972;220:1098-100. 4 Mendelson LM, Ressler C, Rosen JP, Selcow JE. Routine 15 Wilkinson DS, Fregert S, Magnusson B, et al. Terminology elective penicillin allergy skin testing in children and of contact dermatitis. 1970;50:287-92. adolescents: study of sensitisation. Acta Dermatovener J Allergy Clin Immunol 16 Voorhorst R, Sparreboom S. The use of stereoisomers in 1984;73:76-81. 5 Shapiro S, Siskind V, Slone D, Lewis GP, Jick H. Drug rash patch testing. Ann Allergy 1980;45:100-3. with ampicillin and other penicillins. Lancet 1969;ii:969- 17 Padovan E, Mauri-Hellweg D, Pichler WJ, Weltzien HU. T 72. cell recognition of penicillin G: structural features deter- 6 Collaborative Study Group. Prospective study of ampicillin mining antigenic specificity. Eur J Immunol 1996;26:42-8. rash. BMJ 1973;i:7-9. 18 GraV-Lonnevig V, Hedlin G, Lindfors A. Penicillin 7 Kerns DL, Shira JE, Go S, et al. Ampicillin rash in children. allergy—a rare paediatric condition ? Arch Dis Child 1988; Relationship to penicillin allergy and infectious mononu- 63:1342-6. cleosis. Am J Dis Child 1973;125:187-90. 19 Cherry JD. Cutaneous manifestations of systemic infections. 8 de Haan P, Bruynzeel DP, Van Ketel WG. Onset of penicil- In: Feigin RD, Cherry JD, eds. Textbook of pediatric infectious lin rashes: relation between type of penicillin administered diseases. Philadelphia: WB Saunders, 1987: 786-817. and type of immune reactivity. Allergy 1986;41:75-8. 20 Chosidow O, Bourgault I, Roujeau JC. Drug rashes. What 9 Aihara M, Ikezawa Z. Evaluation of the skin test reactions are the targets of cell-mediated cytotoxicity? Arch Dermatol with delayed type rash induced by penicillins and 1994;130:627-9. cephalosporins. J Dermatol 1987;14:440-8. 21 Haverkos HV, Amsel Z. Drotman P. Adverse virus-drug 10 Vega JM, Blanca M, Carmona MJ, Delayed allergic et al. interactions. 1991;13:697-704. reactions to betalactams. Allergy 1991;46:154-7. Rev Infect Dis 11 Warrington RJ, Silviu-Dan F, Magro C. Accelerated 22 Finke SR, Grieco MH, Connel JT, Smith EC, Sherman WB. cell-mediated immune reactions in penicillin allergy. J Results of comparative skin tests with penicilloyl-polylysine Allergy Clin Immunol 1993;92:626-8. and penicillin in patients with penicillin allergy. Am J Med 12 Romano A, Di Fonso M, Papa G, et al. Evaluation of 1965;38:71-82. adverse cutaneous reactions to aminopenicillins with 23 Sullivan TJ, Wedner HJ, Shatz JS, Yiecies LD, Parker CW. emphasis on those manifested by maculopapular rashes. Skin testing to detect penicillin allergy. J Allergy Clin Immu- Allergy 1995;50:113-8. nol 1981;68:171 -80. http://adc.bmj.com/ on September 23, 2021 by guest. Protected copyright.