Portal de Periódicos da CAPES

False-positive penicillin immunoassay: An unnoticed common problem

2012; Elsevier BV; Volume: 132; Issue: 1 Linguagem: Inglês

10.1016/j.jaci.2012.11.017

1097-6825

S. G. O. Johansson, Justus Adédoyin, Marianne van Hage, R. Grönneberg, Anna Nopp,

Food Allergy and Anaphylaxis Research

Penicillin is a small molecule that, most likely, is immunogenic after binding to macromolecules, forming a hapten-carrier complex. The major allergen is the benzylpenicilloyl determinant; however, minor determinants have also been reported.1Batchelor F.R. Dewdney J.M. Gazzard D. Penicillin allergy: the formation of the penicilloyl determinant.Nature. 1965; 206: 362-364Crossref PubMed Scopus (126) Google Scholar Adverse reactions to penicillin occur in up to 10% of the patients and anaphylaxis in 0.015% to 0.04%.2Vervloet D. Pradal M. Birnbaum Jl Koeppel M.-C. Drug allergy.4th ed. Phadia AB, Uppsala2009Google Scholar Correct diagnosis of allergy to antibiotics is of great importance because false-negative reports can lead to severe adverse reactions and false-positive reports to limited administration of the drugs or change to more expensive alternatives when they are indicated strongly.3Shapiro S. Siskind V. Slone D. Lewis G.P. Jick H. Drug rash with ampicillin and other penicillins.Lancet. 1969; 2: 969-972Abstract PubMed Google Scholar The diagnosis is based on case history, skin prick test (SPT), which has a good negative predictive value,4Salkind A.R. Cuddy P.G. Foxworth J.W. The rational clinical examination: is this patient allergic to penicillin? An evidence-based analysis of the likelihood of penicillin allergy.JAMA. 2001; 285: 2498-2505Crossref PubMed Scopus (231) Google Scholar and serum IgE antibodies (IgE-ab), for example, measured by ImmunoCAP (IDD Thermo Fisher Science, Uppsala, Sweden) sometimes followed up with intradermal and provocation testing. In April 2009, a serum sample with suspiciously high titers, that is, of the order of 1000-fold higher than the cutoff for a positive test, of IgE-ab to penicillin V (PcV) and penicillin G (PcG) was identified by using ImmunoCAP. In February, the patient, a 67-year-old woman with hay fever, was afflicted with mild, localized skin reactions the night after starting treatment with PcV (Kåvepenin; Meda, Solna, Sweden) for a gingival infection, for which she had been treated 6 months earlier without any adverse reactions. In April, the infection returned and her physician sent a blood sample for testing. It contained IgE-ab to PcG (78 kUA/L) and PcV (>100 kUA/L) (Table I) using ImmunoCAP. The patient was contacted for confirmation and follow-up at 6 occasions. The ethics committee in Stockholm, Sweden, approved the study.Table IIgE (kU/L) and IgE-ab (kUA/L) in serum samples from the patient as measured by ImmunoCAPVisitVisit datePcGPcVAmpicillinAmoxicillinPEA-PLLTimothyIgEIApril 21, 200978.0>100IIJune 1, 200940.794.3103.6IIIJune 30, 200938.985.05.90.4102.9968IVOctober 6, 200919.833.852.80.5654VDecember 1, 200913.136.634.7VIFebruary 20, 20108.328.224.2 Open table in a new tab In June, the level of IgE-ab to PcG was halved and the level of IgE-ab to PcV was less than 100 kUA/L, unchanged 4 weeks later, and further halved in September. Low levels of IgE-ab to ampicillin (5.9 kUA/L) and amoxicillin (0.4 kUA/L) were also obtained with commercially available ImmunoCAP. No reactivity was seen to ImmunoCAP with human serum albumin, or glycine excluding unspecific tests. Screening of IgE-ab to common inhalant allergens (n = 10) revealed only a low reactivity to timothy (0.5 kUA/L). Soluble PcV, that is, Kåvepenin, added to the serum could not inhibit the PcV ImmunoCAP reaction. However, because penicillin on the PcV ImmunoCAP is conjugated to polylysine (PLL) as a linker, PcV-PLL was used and gave complete inhibition. An allergenic structure related to penicillin, but different from classical allergen and potentially present on the ImmunoCAP, is phenylethylamine with a benzyl group (PEA). PEA and its derivatives exist as neurotransmitters, in food, in many classes of drugs and are also produced by bacteria.5Sabelli H.C. Javaid J.I. Phenylethylamine modulation of affect: therapeutic and diagnostic implications.J Neuropsychiatry Clin Neurosci. 1995; 7: 6-14Crossref PubMed Scopus (65) Google Scholar Complete inhibition of the PcV ImmunoCAP reactivity was obtained with PEA conjugated to PLL (PEA-PLL) and also with PPS (PEA with open ring structure), which in contrast to PEA has a phenoxy group, conjugated to PLL (PPS-PLL). Our assumption is that the allergenic epitopes consist of one part shared by PcG, PcV, and PEA and another part present on PLL and succinat, thus creating a hapten-carrier epitope. Three new ImmunoCAP were prepared: PEA-PLL, PLL, and PLL succinate (PLL-succ). No IgE reactivity was seen toward PLL or PLL-succ, while the reactivity to PEA-PLL (Table I) was higher than to the PcV ImmunoCAP in all samples. The PEA-PLL/PcV ratio ranged from 1.1 to 1.6. Heating the serum for 30 minutes at 56°C lowered reactivity to the PcV ImmunoCAP by 95%, indicating a true IgE-mediated reaction. To document the patient's clinical sensitivity, a basophil allergen threshold sensitivity test, CD-sens,6Johansson S.G.O. Nopp A. van Hage M. Olofsson N. Lundahl J. Wehlin L. et al.Passive IgE-sensitization by blood transfusion.Allergy. 2005; 60: 1192-1199Crossref PubMed Scopus (102) Google Scholar, 7Nopp A. Johansson S.G.O. Ankerst J. Bylin G. Cardell L.O. Gronneberg R. et al.Basophil allergen threshold sensitivity: a useful approach to anti-IgE treatment efficacy evaluation.Allergy. 2006; 61: 298-302Crossref PubMed Scopus (118) Google Scholar was performed at 4 occasions. No reactivity was detected toward any of the several penicillins and penicillin-related preparations (n = 12) tested on her basophils with or without endogenous plasma present7Nopp A. Johansson S.G.O. Ankerst J. Bylin G. Cardell L.O. Gronneberg R. et al.Basophil allergen threshold sensitivity: a useful approach to anti-IgE treatment efficacy evaluation.Allergy. 2006; 61: 298-302Crossref PubMed Scopus (118) Google Scholar nor to PEA or PEA-related preparations (n = 7). CD-sens was positive to timothy both on basophils with and on basophils without endogenous plasma as were the positive controls, indicating functional basophils. The result is in contrast to Patent Blue V–induced anaphylaxis where only basophils in the patient's plasma are positive in CD-sens, indicating that the allergenic epitope is dependent on a carrier-hapten complex, the carrier being present in the plasma.8Johansson S.G.O. Nopp A. Oman H. Stahl-Skov P. Hunting A.S. Guttormsen A.B. Anaphylaxis to Patent Blue V, II: a unique IgE-mediated reaction.Allergy. 2010; 65: 124-129Crossref PubMed Scopus (26) Google Scholar There is a good correlation of CD-sens with allergen sensitivity in allergic rhinitis7Nopp A. Johansson S.G.O. Ankerst J. Bylin G. Cardell L.O. Gronneberg R. et al.Basophil allergen threshold sensitivity: a useful approach to anti-IgE treatment efficacy evaluation.Allergy. 2006; 61: 298-302Crossref PubMed Scopus (118) Google Scholar and asthma9Dahlen B. Nopp A. Johansson S.G.O. Eduards M. Skedinger M. Adedoyin J. Basophil allergen threshold sensitivity, CD-sens, is a measure of allergen sensitivity in asthma.Clin Exp Allergy. 2011; 41: 1091-1097Crossref PubMed Scopus (47) Google Scholar and thus the negative CD-sens to penicillin excludes clinical relevance of the detected IgE-ab. One plausible explanation of the inability of penicillin to cause basophil and mast cell activation could be that the exposed allergenic determinant is monovalent, prohibiting cross-linking of IgE molecules on basophils and mast cells. This was further confirmed by negative SPT results with PcG (Benzylpenicillin, AstraZeneca, Södertälje, Sweden) (300 mg/mL), PcV (Kåvepenin, Meda) (250 mg/mL), and amoxicillin (Amoxicillin, Sandoz, Switzerland) (50 mg/mL). However, in a standard panel of 10 classical allergens (SoluPrick; ALK Laboratories, Copenhagen, Denmark), timothy grass was positive as was histamine, indicating functional mast cells. To estimate the prevalence of this special IgE sensitization, 90 sera from patients with suspected penicillin allergy obtained from 3 laboratories were screened for IgE-ab to PEA by using PEA-PLL ImmunoCAP. Positive test results to PEA were obtained in 12 of the 90 sera (13%). Of the 46 sera positive to PcV and/or PcG, the same 12 sera (26%) were positive (>0.20 kUA/L was used, which is a compromise between recommended technical, 0.1 kUA/L, and clinical, >0.35 kUA/L, cut-off) to PEA and the geometric mean of IgE-ab to PcV was 1.8 kUA/L (range, 0.4-20.2). No penicillin-negative serum was PEA positive. Six of the 12 PEA-positive sera had a PEA/PcV reactivity ratio above 0.9 (range, 0.9-18.4) and 6 had a ratio under 0.4 (range, 0.005-0.3). In addition, of the 5 PcV-positive sera from patients without clinical manifestations to penicillin, 2 were positive to PEA. In a recent "Directions for use" (March 2012), it is recommended that "very low levels of allergen specific IgE antibodies should be evaluated with caution when total IgE values are above 500 kU/L when testing for specific IgE antibodies to beta-lactames." This statement is probably due to concerns for nonspecific reactions and requires the health professional to determine a total IgE as well. Although some of our sera had an IgE value above 500 kU/L the PEA IgE antibody values were not "very low" and there was no correlation of the PEA IgE with the total IgE concentration. We suggest that patients with a PEA/PcV ratio of more than 0.9, in this study comprising 13% of the 46 penicillin-positive patients, mainly have IgE-ab directed to the nonclinically relevant epitope present on the PcV and PcG ImmunoCAP. In contrast, patients with a low PEA/PcV ratio could, in addition, have IgE-ab directed to true, bivalent epitopes on the penicillin molecule and be at risk of adverse reactions. In summary, we have in this study found that 26% of the patients with a suspected IgE-mediated reaction to penicillin and a positive penicillin ImmunoCAP actually might have IgE-ab to PEA, which are not clinically relevant. Only those sera that were positive to penicillin were positive to PEA. A possible allergy in such a patient must be confirmed by clinical tests including penicillin provocation. Other commercial tests for IgE-ab to penicillin could also have this structure and should be analyzed. These false-positive tests make the ImmunoCAP a poor choice for the diagnosis of penicillin allergy. This flaw should be addressed by the manufacturer to avoid unnecessary false penicillin allergy diagnoses. We thank Clinical Immunology Laboratory at Sahlgrenska University Hospital (Gothenburg, Sweden) and Karolinska University Hospital and Phadia AB (Uppsala, Sweden) for donation of sera.