Allopurinol hypersensitivity is primarily mediated by dose-dependent oxypurinol-specific T cell response.

BACKGROUND: Allopurinol is a main cause of severe cutaneous adverse reactions (SCAR). How allopurinol induces hypersensitivity remains unknown. Pre-disposing factors are the presence of the HLA-B*58:01 allele, renal failure and possibly the dose taken. OBJECTIVE: Using an in vitro model, we sought to decipher the relationship among allopurinol metabolism, HLA-B*58:01 phenotype and drug concentrations in stimulating drug-specific T cells. METHODS: Lymphocyte transformation test (LTT) results of patients who had developed allopurinol hypersensitivity were analysed. We generated allopurinol or oxypurinol-specific T cell lines (ALP/OXP-TCLs) from allopurinol naïve HLA-B*58:01(+) and HLA-B*58:01(-) individuals using various drug concentrations. Their reactivity patterns were analysed by flow cytometry and (51) Cr release assay. RESULTS: Allopurinol allergic patients are primarily sensitized to oxypurinol in a dose-dependent manner. TCL induction data show that both the presence of HLA-B*58:01 allele and high concentration of drug are important for the generation of drug-specific T cells. The predominance of oxypurinol-specific lymphocyte response in allopurinol allergic patients can be explained by the rapid conversion of allopurinol to oxypurinol in vivo rather than to its intrinsic immunogenicity. OXP-TCLs do not recognize allopurinol and vice versa. Finally, functional avidity of ALP/OXP-TCL is dependent on both the induction dose and HLA-B*58:01 status. CONCLUSIONS AND CLINICAL RELEVANCE: This study establishes the important synergistic role of drug concentration and HLA-B*58:01 allele in the allopurinol or oxypurinol-specific T cell responses. Despite the prevailing dogma that Type B adverse drug reactions are dose independent, allopurinol hypersensitivity is primarily driven by oxypurinol-specific T cell response in a dose-dependent manner, particular in the presence of HLA-B*58:01 allele.

In vitro drug causality assessment in Stevens-Johnson syndrome - alternatives for lymphocyte transformation test.

BACKGROUND: Patients with Stevens-Johnson syndrome (SJS) or toxic epidermal necrolysis (TEN) are often exposed simultaneously to a few potentially culprit drugs. However, both the standard lymphocyte transformation tests (LTT) with proliferation as the assay end-point as well as skin tests, if done, are often negative. OBJECTIVE: As provocation tests are considered too dangerous, there is an urgent need to identify the relevant drug in SJS/TEN and to improve sensitivity of tests able to identify the causative drug. METHODS: Fifteen patients with SJS/TEN with the ALDEN score ≥ 6 and 18 drug-exposed controls were included. Peripheral blood mononuclear cells (PBMC) were isolated and cultured under defined conditions with drugs. LTT was compared to the following end-points: cytokine levels in cell culture supernatant, number of granzyme B secreting cells by ELISpot and intracellular staining for granulysin and IFNγ in CD3(+) CD4(+), CD3(+) CD8(+) and NKp46(+) cells. To further enhance sensitivity, the effect of IL-7/IL-15 pre-incubation of PBMC was evaluated. RESULTS: Lymphocyte transformation tests was positive in only 4/15 patients (sensitivity 27%, CI: 8-55%). Similarly, with granzyme B-ELISpot culprit drugs were positive in 5/15 patients (sensitivity 33%, CI: 12-62%). The expression of granulysin was significantly induced in NKp46(+) and CD3(+) CD4(+) cells (sensitivity 40%, CI: 16-68% and 53%, CI: 27-79% respectively). Cytokine production could be demonstrated in 38%, CI: 14-68% and 43%, CI: 18-71% of patients for IL-2 and IL-5, respectively, and in 55%, CI: 23-83% for IFNγ. Pre-incubation with IL-7/IL-15 enhanced drug-specific response only in a few patients. Specificities of tested assays were in the range of 95 (CI: 80-99%)-100% (CI: 90-100%). CONCLUSIONS AND CLINICAL RELEVANCE: Granulysin expression in CD3(+) CD4(+) , Granzyme B-ELISpot and IFNγ production considered together provided a sensitivity of 80% (CI: 52-96%) and specificity of 95% (80-99%). Thus, this study demonstrated that combining different assays may be a feasible approach to identify the causative drug of SJS/TEN reactions; however, confirmation on another group of patients is necessary.

Human leukocyte antigens (HLA) associated drug hypersensitivity: consequences of drug binding to HLA.

Recent publications have shown that certain human leukocyte antigen (HLA) alleles are strongly associated with hypersensitivity to particular drugs. As HLA molecules are a critical element in T-cell stimulation, it is no surprise that particular HLA alleles have a direct functional role in the pathogenesis of drug hypersensitivity. In this context, a direct interaction of the relevant drug with HLA molecules as described by the p-i concept appears to be more relevant than presentation of hapten-modified peptides. In some HLA-associated drug hypersensitivity reactions, the presence of a risk allele is a necessary but incomplete factor for disease development. In carbamazepine and HLA-B*15:02, certain T-cell receptor (TCR) repertoires are required for immune activation. This additional requirement may be one of the 'missing links' in explaining why most individuals carrying this allele can tolerate the drug. In contrast, abacavir generates polyclonal T-cell response by interacting specifically with HLA-B*57:01 molecules. T cell stimulation may be due to presentation of abacavir or of altered peptides. While the presence of HLA-B*58:01 allele substantially increases the risk of allopurinol hypersensitivity, it is not an absolute requirement, suggesting that other factors also play an important role. In summary, drug hypersensitivity is the end result of a drug interaction with certain HLA molecules and TCRs, the sum of which determines whether the ensuing immune response is going to be harmful or not.

Etiology and pathogenesis of adverse drug reactions.

In clinical routine, adverse drug reactions (ADR) are common, and they should be included in the differential diagnosis in all patients undergoing drug treatment. Only part of those ADR are immune-mediated hypersensitivity reactions and thus true drug allergies. Far more common are non-immune-mediated ADR, e.g. due to the pharmacological properties of the drug or to the individual predisposition of the patient (enzymopathies, cytokine dysbalance, mast cell hyperreactivity). In true drug allergiesT cell- and immunoglobulin E (lgE)-mediated reactions dominate the clinical presentation. T cell-mediated ADR usually have a delayed appearance and include skin eruptions in most cases. Nevertheless, it should not be forgotten that they may involve systemic T cell activation and thus take a severe, sometimes lethal turn. Clinical danger signs are involvement of mucosal surfaces, blistering within the exanthematous skin areas and systemic symptoms, e.g. fever or malaise. Drug presentation via antigen-presenting cells to T cells can either involve the classical pathway of haptenization of endogenous proteins or be directly mediated via noncovalent binding to immune receptors (MHC molecules or T cell receptors), the so-called p-i concept. Flare-up reactions during the acute phase of T cell-mediated ADR should not be mistaken for true drug allergies, as they only occur in the setting of a highly activated T cell pool. IgE-mediated ADR are less frequent and involve mast cells and/or basophils as peripheral effector cells. Recent data suggest that certain patients with drug allergy have a preexistent sensitization although they have never been exposed to the culprit drug, probably due to cross-reactivity. Thus, allergic drug reactions on first encounter are possible. In general, the extent of cross-reactivity is higher in IgE-compared to T cell-mediated ADR. Based on a specific ethnic background and only for severe T cell-mediated ADR to certain drugs, a strong HLA association has been established recently.

Multiple drug hypersensitivity: normal Treg cell function but enhanced in vivo activation of drug-specific T cells.

BACKGROUND: Up to 10% of patients with severe immune-mediated drug hypersensitivity reactions have tendencies to develop multiple drug hypersensitivities (MDH). The reason why certain individuals develop MDH and the underlying pathomechanism are unclear. We investigated different T cell subpopulations in MDH patients and compared them with patients allergic to a single drug and with healthy controls (HC). METHODS: We analyzed the in vitro reactivity of peripheral blood mononuclear cells from MDH patients (n=7), patients with hypersensitivity to a single drug (monoallergic, n=6), and healthy controls (HD) (n=6) to various drugs (mainly antibiotics and antiepileptics). By depleting and selectively re-adding CD4(+) CD25(bright) T cells (T regulatory cells, Treg), their effect on drug-specific T cell reactivity was analyzed. The phenotype of reacting T cells was determined ex vivo by staining for markers of activation (CD38) and cell exhaustion (PD-1). RESULTS: No functional deficiency of Treg cells was observed in all drug-allergic patients. Drug-reactive T cells from MDH patients were found in the CD4(+) CD25(dim) T cell fraction and showed enhanced CD38 and PD-1 expression, while those from monoallergic patients reside in the resting CD4(+) CD25(neg) T cell fraction. CONCLUSION: In patients with MDH, the drug-reactive T cells are contained in an in vivo pre-activated T cell fraction. Therefore, they may show a lower threshold for activation by drugs. The reason for this in vivo T cell pre-activation needs further investigations.

Continuous apple consumption induces oral tolerance in birch-pollen-associated apple allergy.

BACKGROUND: Patients with birch pollen allergy (major allergen: Bet v 1) have often an associated oral allergy syndrome (OAS) to apple, which contains the cross-reactive allergen Mal d 1. As successful birch pollen immunotherapy does not consistently improve apple related OAS symptoms, we evaluated whether regular apple consumption has an effect on OAS and immune parameters of Mal d 1 or Bet v 1 allergy. METHODS: A total of 40 patients with a clear history of birch pollen rhinoconjunctivitis and associated OAS to apple were included in an open, randomized, controlled clinical trial: 27 patients consumed daily defined amount of apple (1-128 g), doubling the amount every two to three weeks, while 13 patients remained untreated. Primary endpoint was the proportion of patients that achieved tolerance to at least 128 g of apple at the end of the study after 8 months. Exploratory endpoints were questionnaire about cross-reactive food and pollen allergy symptoms, conjunctival provocation test with birch pollen and Bet v 1, and in vitro tests (tIgE, sIgE, and IgG4 to Mal d 1 and Bet v 1; basophil activation test with both allergens). RESULTS: Seventeen of 27 patients in active group and none of 13 patients in control group (P = 0.0001) could tolerate a whole apple after the intervention. However, differences in endpoints reflecting systemic immune reactivity did not reach statistical significance. CONCLUSION: In patients with OAS to apple, tolerance can be safely induced with slowly, gradually increasing consumption of apple. However, the observation of a relapse after discounting of apple consumption and absence of immunologic changes suggest that induced tolerance is only transient.

NKp46+ cells express granulysin in multiple cutaneous adverse drug reactions.

BACKGROUND: The spectrum of cutaneous adverse drug reactions (cADRs) ranges from benign presentations to severe life-threatening forms such as toxic epidermal necrolysis (TEN). In TEN, granulysin has been shown to be the key cytotoxic molecule. Still, little is known about the expression of granulysin in other cADRs. As an important source of granulysin, natural killer (NK) cells are of major interest in cADRs. Recently, NKp46 has been identified as the most selective NK-cell marker. However, the role of NKp46(+) cells in cADRs and their contribution to granulysin expression remain to be elucidated. METHODS: Immunohistochemical and immunofluorescence staining of tissue sections from multiple cADRs were quantitatively and qualitatively evaluated. Further, in vivo and in vitro drug-stimulation tests were performed. RESULTS: Granulysin is expressed at different levels in multiple cADRs both by NKp46(+) cells and by CD8(+) T cells. Even in mild forms of cADRs, granulysin can be induced in vivo and in vitro in a drug-specific manner. NKp46(+) cells were found to infiltrate the dermal/epidermal junction particularly in TEN. CONCLUSION: The impressive clinical differences of cADRs may not be uniquely explained by the expression of granulysin. Additional factors such as drug-specific activation and recruitment of NKp46(+) cells to the epidermis may play a role in determining the severity of cADRs. Therefore, unraveling the effects of drugs on NK-cell activation and trafficking may help to better understand the cytotoxic mechanisms behind cADRs.

In vitro diagnosis of T cell-mediated drug allergy.

Diagnosis of drug hypersensitivity relies on history, skin tests, in vitro tests and provocation tests. In vitro tests are of great interest, due to possible reduction of drug provocation tests. In this review we focus on best investigated in vitro techniques for the diagnosis of T cell-mediated drug hypersensitivity reactions. As drug hypersensitivity relies on different pathomechanisms and as a single diagnostic test usually does not cover all possible reactions, it is advisable to combine different tests to increase the overall sensitivity. Recently, proliferation-based assays have been supplemented by a panel of novel in vitro tests including analysis of cytotoxic potential of effector cells (granzyme B, granulysin, CD107a), evaluation of cytokine secretion (IL-2, IL-5, IL-13, and IFN-γ) and up-regulation of cell surface activation markers (CD69). We discuss the latest findings and readout systems to identify causative drugs by detecting functional and phenotypic markers of drug-reacting cells, and their ability to enable a more conclusive diagnosis of drug allergy.

Robust expression of CCR3 as a single basophil selection marker in flow cytometry.

BACKGROUND: Basophil activation tests (BAT) rely on different combinations of basophil selection and activation markers. Whereas activation markers, especially CD63, are widely validated, the most suitable and robust marker for basophil selection is still a matter of debate. AIMS: Comparison of cell surface expression of two commonly used basophil selection markers (IgE, CD123/HLA-DR) with CCR3 in an unselected group of atopic and nonatopic donors in resting and activated basophils. METHODS: EDTA blood of 94 healthy adults, about half of them atopic by history, was analyzed using two different staining strategies: anti-CD123-PE/anti-HLA-DR-PerCP/anti-lin1-FITC and anti-IgE-FITC/anti-CD3-PerCP/anti-CCR3-PE. Additionally 40 pollen-allergic patients were recruited for the assessment of CCR3 expression after basophil activation. RESULTS: In resting basophils, cell surface expression of the three basophil selection markers was most constant for CCR3. IgE gating strategy showed the highest variation and up to 80% of nonbasophils in the selected gate in certain donors. During basophil activation, a shift of the mean fluorescence intensity for CCR3 toward the lower third of the CCR3-positive population could be demonstrated, but neither were CCR3-positive cells significantly lost for further analysis nor was differentiation between CCR3-positive and CCR3-negative cell populations hampered by this shift. CONCLUSIONS: CCR3 is a stable and highly expressed basophil selection marker, independent of the atopic background or basophil activation state and allows an accurate identification of basophils without need of a second marker. The basophil markers CD123/HLA-DR and IgE showed significantly higher interindividual variability in cell surface expression and are therefore less suited as selection markers.

General considerations on rapid desensitization for drug hypersensitivity - a consensus statement.

Drug hypersensitivity reactions can occur with most drugs, are unpredictable, may affect any organ or system, and range widely in clinical severity from mild pruritus to anaphylaxis. In most cases, the suspected drug is avoided in the future. However, for certain patients, the particular drug may be essential for optimal therapy. Under these circumstances, desensitization may be performed. Drug desensitization is defined as the induction of a temporary state of tolerance of a compound responsible for a hypersensitivity reaction. It is performed by administering increasing doses of the medication concerned over a short period of time (from several hours to a few days) until the total cumulative therapeutic dose is achieved and tolerated. It is a high-risk procedure used only in patients in whom alternatives are less effective or not available after a positive risk/benefit analysis. Desensitization protocols have been developed and are used in patients with allergic reactions to antibiotics (mainly penicillin), insulins, sulfonamides, chemotherapeutic and biologic agents, and many other drugs. Desensitization is mainly performed in IgE-mediated reactions, but also in reactions where drug-specific IgE have not been demonstrated. Desensitization induces a temporary tolerant state, which can only be maintained by continuous administration of the medication. Thus, for treatments like chemotherapy, which have an average interval of 4 weeks between cycles, the procedure must be repeated for every new course. In this paper, some background information on rapid desensitization procedures is provided. We define the drugs and drug reactions indicated for such procedures, describe the possible mechanism of action, and discuss the indications and contraindications. The data should serve as background information for a database (accessible via the EAACI-homepage) with standardized protocols for rapid desensitization for antibiotics, chemotherapeutic agents, monoclonal antibodies/fusion proteins, and other drugs.

In vitro detection of cytotoxic T and NK cells in peripheral blood of patients with various drug-induced skin diseases.

BACKGROUND: Cytotoxic cells are involved in most forms of drug-induced skin diseases. Till now, no in vitro test addressed this aspect of drug-allergic responses. Our report evaluates whether drug-induced cytotoxic cells can be detected in peripheral blood of nonacute patients with different forms of drug hypersensitivity, and also whether in vitro detection of these cells could be helpful in drug-allergy diagnosis. METHODS: GranzymeB enzyme-linked immunosorbent spot-forming (ELISPOT) and cell surface expression of the degranulation marker CD107a were evaluated on peripheral blood mononuclear cells from 12 drug-allergic patients in remission state and 16 drug-exposed healthy controls. RESULTS: In 10/12 allergic patients culprit but not irrelevant drug elicited granzymeB release after 48-72 h stimulation. It was clearly positive in patients with high proliferative response to the drug, measured in lymphocyte transformation tests. In patients, who showed moderate or low proliferation and low drug-response in granzymeB ELISPOT, overnight preincubation with interleukin (IL)-7/IL-15 enhanced drug-specific granzymeB release and allowed to clearly identify the offending agent. CD107a staining was positive on CD4+/CD3+, CD8+/CD3+ T cells as well as CD56+/CD3- natural killer cells. None of the drug-exposed healthy donors reacted to the tested drugs and allergic patients reacted only to the offending, but not to tolerated drugs. CONCLUSION: GranzymeB ELISPOT is a highly specific in vitro method to detect drug-reacting cytotoxic cells in peripheral blood of drug-allergic patients even several years after disease manifestation. Together with IL-7/IL-15 preincubation, it may be helpful in indentifying the offending drug even in some patients with weak proliferative drug-response.

Processing-dependent and -independent pathways for recognition of iodinated contrast media by specific human T cells.

BACKGROUND: One to three percent of patients exposed to intravenously injected iodinated contrast media (CM) develop delayed hypersensitivity reactions. Positive patch test reactions, immunohistological findings, and CM-specific proliferation of T cells in vitro suggest a pathogenetic role for T cells. We have previously demonstrated that CM-specific T cell clones (TCCs) show a broad range of cross-reactivity to different CM. However, the mechanism of specific CM recognition by T cell receptors (TCRs) has not been analysed so far. OBJECTIVE: To determine how T cells specifically recognize CM. METHODS: CM-specific TCCs were generated from human blood of three CM-allergic patients and a specific TCR was transfected into a mouse T cell hybridoma. Functional analysis such as proliferation assays, IL-2 secretion assays, and calcium influx experiments were performed using irradiated, glutaraldehyde-fixed, CM-pre-incubated, human leucocyte antigen (HLA)-DR-matched or -mismatched antigen-presenting cells (APCs), and HLA-blocking antibodies. RESULTS: We identified two mechanisms of T cell stimulation: some TCCs and the transfectant reacted to CM independent of uptake by APCs because proliferation/IL-2 secretion occurred in the presence of glutaraldehyde-fixed APCs, and intracellular calcium increased within seconds after drug addition. Other TCCs required functional APCs, compatible with uptake and presentation of CM on MHC-class II molecules, as implied by three findings: (1) glutaraldehyde fixation of APCs abrogated presentation; (2) CM could not be washed away from CM-pre-incubated APCs; and (3) the optimal pulsing time was 10-20 h. Because allogeneic, MHC-matched, CM-pulsed APCs could induce proliferative responses as well, the ability of CM uptake and presentation is not unique to APCs from patients with CM-induced delayed hypersensitivity. CONCLUSION: Our data suggest that CM may be stimulatory for T cells either by direct binding to the MHC-TCR complex or by binding after uptake and processing by APCs. This questions the assumed inert nature of CM.

Fulminant liver failure after vancomycin in a sulfasalazine-induced DRESS syndrome: fatal recurrence after liver transplantation.

DRESS syndrome (drug rash with eosinophilia and systemic symptoms) is a rare drug hypersensitivity reaction with a significant mortality. We describe a 60-year-old man with polyarthritis treated with sulfasalazine who developed DRESS and fulminant liver failure after additional vancomycin treatment. Liver histology revealed infiltration of granzymeB+ CD3+ lymphocytes in close proximity to apoptotic hepatocytes. After a superurgent liver transplantation and initial recovery, the patient developed recurrent generalized exanthema and eosinophilia, but only moderate hepatitis. Histology showed infiltration of FasL+ lymphocytes and eosinophils in the transplanted liver. Treatment with high-dose methylprednisolone was unsuccessful. Postmortem examination revealed extensive necrosis of the liver transplant. This case report illustrates that patients with DRESS may develop fulminant liver failure and that DRESS recurrence can recur in the transplanted liver. Histological and immunological investigations suggest an important role of granzymeB and FasL mediated cell death in DRESS associated hepatitis.

Effect of natural seasonal pollen exposure and repeated nasal allergen provocations on elevation of exhaled nitric oxide.

BACKGROUND: Exhaled nitric oxide (FENO) is a marker for allergic airway inflammation. We wondered whether in patients with intermittent allergic rhinitis only (i) natural pollen exposure and (ii) artificial pollen exposure by repeated nasal allergen provocations may lead to an elevation of FENO. METHODS: In two prospective studies, we compared the FENO of nonatopic controls with the FENO of nonasthmatic individuals with mild intermittent rhinitis to tree and/or grass pollen. Study I: 13 atopic individuals and seven controls had measurements of FENO, blood eosinophils and eosinophilic cationic protein (ECP) before, during and after pollen season. Study II: 16 atopic individuals and 12 controls had nasal allergen provocations on four following days out of pollen season, with daily measurements of FENO before, 2 and 6 h after provocation, and determination of blood eosinophils, ECP and FEV1 at baseline, on days 5 and 10-12. RESULTS: Natural pollen exposure (study I) caused a significant elevation of FENO in allergic individuals. Nasal allergen provocations (study II) did not elicit a statistically significant rise neither of FENO nor of blood eosinophils between baseline and day 5. However, a subgroup of four individuals with a rise of blood eosinophils during nasal allergen provocations showed also a rise of FENO. CONCLUSIONS: We suppose that in allergic rhinitis a concomitant reaction of the bronchial system is dependent on a strong local inflammation leading to a generalized immune stimulation.

[Intolerance as consequence of hyperreactivity]. / Unverträglichkeit als Folge einer Hyperreaktivität.

Many patients complain about hyperreactivity of the skin or mucosal membranes, whereby rather harmless "triggers" lead to exaggerated reactions like running nose, bronchospasm, urticaria, etc. Most of these hyperreactivities have an underlying inflammation. It is important not to focus only on the triggers, but to look for the agent causing the inflammation and to avoid/treat it: elimination of the cause also eliminates the hyperreactivity. Four examples are presented where this cause--trigger model of hyperreactivity is illustrated (exercise induced asthma, pollen associated food allergy, flare-up reactions in drug allergy and hyperreactivity in chronic urticaria).

Drug-specific in vitro release of IL-2, IL-5, IL-13 and IFN-gamma in patients with delayed-type drug hypersensitivity.

BACKGROUND: The most prevalent drug hypersensitivity reactions are T-cell mediated. The only established in vitro test for detecting T-cell sensitization to drugs is the lymphocyte transformation test, which is of limited practicability. To find an alternative in vitro method to detect drug-sensitized T cells, we screened the in vitro secretion of 17 cytokines/chemokines by peripheral blood mononuclear cells (PBMC) of patients with well-documented drug allergies, in order to identify the most promising cytokines/chemokines for detection of T-cell sensitization to drugs. METHODS: Peripheral blood mononuclear cell of 10 patients, five allergic to beta-lactams and five to sulfanilamides, and of five healthy controls were incubated for 3 days with the drug antigen. Cytokine concentrations were measured in the supernatants using commercially available 17-plex bead-based immunoassay kits. RESULTS: Among the 17 cytokines/chemokines analysed, interleukin-2 (IL-2), IL-5, IL-13 and interferon-gamma (IFN-gamma) secretion in response to the drugs were significantly increased in patients when compared with healthy controls. No difference in cytokine secretion patterns between sulfonamide- and beta-lactam-reactive PBMC could be observed. The secretion of other cytokines/chemokines showed a high variability among patients. CONCLUSION: The measurement of IL-2, IL-5, IL-13 or IFN-gamma or a combination thereof might be a useful in vitro tool for detection of T-cell sensitization to drugs. Secretion of these cytokines seems independent of the type of drug antigen and the phenotype of the drug reaction. A study including a higher number of patients and controls will be needed to determine the exact sensitivity and specificity of this test.

Skin testing in patients with hypersensitivity reactions to iodinated contrast media - a European multicenter study.

BACKGROUND: Iodinated contrast media cause both immediate and nonimmediate hypersensitivity reactions. The aim of this prospective study was to determine the specificity and sensitivity of skin tests in patients who have experienced such reactions. METHODS: Skin prick, intradermal and patch tests with a series of contrast media were conducted in 220 patients with either immediate or nonimmediate reaction. Positive skin tests were defined according to internationally accepted guidelines. Seventy-one never-exposed subjects and 11 subjects who had tolerated contrast medium exposure, served as negative controls. RESULTS: Skin test specificity was 96-100%. For tests conducted within the time period from 2 to 6 months after the reaction, up to 50% of immediate reactors and up to 47% of nonimmediate reactors were skin test positive. For immediate reactors, the intradermal tests were the most sensitive, whereas delayed intradermal tests in combination with patch tests were needed for optimal sensitivity in nonimmediate reactors. Contrast medium cross-reactivity was more common in the nonimmediate than in the immediate group. Interestingly, 49% of immediate and 52% of nonimmediate symptoms occurred in previously unexposed patients. Many of these patients were skin test positive, indicating that they were already sensitized at the time of first contrast medium exposure. CONCLUSIONS: These data suggest that at least 50% of hypersensitivity reactions to contrast media are caused by an immunological mechanism. Skin testing appears to be a useful tool for diagnosis of contrast medium allergy and may play an important role in selection of a safe product in previous reactors.