Contact allergy to corticosteroids

Contact allergy to corticosteroids has been increasingly recognized worldwide as a problem of considerable clinical and therapeutic importance. The incidence of corticosteroid allergy observed (from 0.5% to 5%) varies in every report from one center to another ( 1–6) and depends on several factors such as awareness and testing for corticosteroid allergy, patient selection, preceding skin disease ( 3), prescribing habits and type of corticosteroids used in each country ( 3, 7, 8), and the test and reading methods used. In general, corticosteroid-sensitive patients present with long-standing dermatitis that fails to respond to locally applied corticosteroids. This fact may lead the physician to prescribe other, sometimes stronger preparations, often without improvement. Because of the aspecific clinical picture and the lack of alarming symptoms, most cases of contact allergies would be missed if corticosteroids were not routinely tested. Although not yet included in the European standard series for patch testing, two corticosteroids have been added in many countries to the routinely applied standard series: tixocortol pivalate (1% pet.) and budesonide (0.1% pet.). Patients with contact hypersensitivity to corticosteroids very often react upon patch testing to several corticosteroids. This might be partially explained by concomitant sensitization due to exposure to different types of substances during the long course of their skin disease. However, irrefutable proof of the existence of cross-reactions is provided by reactions to corticosteroids to which the patient has never been exposed. The cross-reactivity of different corticosteroid molecules has practical consequences for the identification of frequent sensitizers, and of screening agents for corticosteroid hypersensitivity. Corticosteroid allergy has very important therapeutic consequences, both locally and systemically, as has been demonstrated in clinical cases ( 9, 10). Because of the occurrence of cross-reactions between different corticosteroids, sensitive patients, in general, have to avoid many corticosteroids but can still be treated with others. Avoidance of all but only the possibly “dangerous” corticosteroids requires an accurate definition of groups of cross-reacting molecules. Corticosteroid sensitivity has long remained underdiagnosed, because the anti-inflammatory activity of corticosteroid preparations masks the contact-allergic reaction to the corticosteroid itself (or to another ingredient of the preparation), leading to an aspecific clinical picture. Indeed, the lesions of such patients are rarely spectacular, as they are generally chronic and just do not respond to local dermatologic treatment ( 11). Sometimes worsening of primary dermatologic symptoms can be observed ( 12). Occasionally, contact allergy to corticosteroids may express itself as acute eczema, acute local edema ( 13, 14), immediate-type reaction ( 15), or an id-like spread (including erythema multiforme-like symptoms) ( 16, 17) elsewhere on the body. Moreover, previously sensitized patients can present with systemic contact dermatitis, exanthema, purpura, and urticaria after the administration of systemic (orally or injected) corticosteroid preparations ( 18–22, Goossens, unpublished results), or, although this is infrequent relative to the very frequent use, allergic reactions can be observed to corticosteroids administered by inhalation preparations in the treatment of rhinitis or bronchial asthma ( 23, 24). Lesion localization mainly involves the hands, legs, and face. Indeed, patients suffering from atopic, allergic, and/or irritant contact dermatitis on the hands, as well as those suffering from stasis dermatitis/leg ulcers, often present with corticosteroid allergy ( 1). Occupational contact sensitization to topical corticosteroids may occur rarely in pharmacy work ( 25). The role of genetic predisposition in the development of contact sensitivity to corticosteroids cannot be ruled out ( 26). There are specific pitfalls in the detection of corticosteroid contact allergy related to the testing methods, such as the choice of the appropriate test concentrations and vehicles, as well as the reading of the reactions, which involves the necessity of late readings (on day 6 or 7) ( 27, 28), the edge effect, and false negative and false positive reactions. Difficulties with the determination of the appropriate test methods partly relate to the anti-inflammatory effect of the corticosteroid, as it can suppress the very allergic response it causes. The sensitivity level of the allergic individual also has a significant influence on the test results ( 29). Another problem is the bioavailability of the corticosteroid because it is applied when tested to the intact skin of the back, and not to the diseased and hence damaged skin. Moreover, a low concentration of a corticosteroid in a commercial product can cause contact allergy on diseased skin but may produce a negative patch-test reaction on the intact skin of the back ( 30). Merely increasing the concentration of a corticosteroid does not increase its bioavailability. False negative reactions can also be caused by several other factors as well; e.g., physicochemical factors such as molecular and crystal size, and the lipophilicity of the molecule. Poor solubility of a given corticosteroid molecule in the test solution may result in inadequate penetration into the skin and subsequently in too low a concentration of the allergen in the epidermis to react with skin proteins. Besides all of these, there are even other pitfalls. One of them is the blanching or vasoconstriction effect (one of the intrinsic pharmacologic activities of corticosteroids) that is often observed at the first reading, particularly with stronger corticosteroids. On the contrary, another phenomenon is a reactive vasodilation, presenting as an impalpable erythema on the patch-test site; this reaction certainly has to be kept under observation because it could well be the beginning of an allergic reaction. Unlike the ordinary edge effect ( 31), which is regarded as irritation occurring under the edge of the patch-test chamber, the “edge effect” observed with corticosteroids is a rather unique phenomenon: at the first reading, there is often infiltration and erythema only on and around the edge of the patch-test chamber with a completely clear center. Later, this center also becomes eczematous as well. This effect is attributed to the anti-inflammatory effect of the concentrated corticosteroid in the middle that suppresses the allergic response that, in contrast, is appearing on the edge of the chamber, where the decrease in concentration due to the diffusion outside the patch-test site results in insufficient anti-inflammatory potency ( 11, 32). The correct concentration and vehicle for patch testing with corticosteroids has not yet been generally agreed upon. False negative reactions with corticosteroid patch tests have led some investigators ( 18, 33) to perform intradermal testing of hydrocortisone and other corticosteroids. However, patch testing with budesonide (0.1% pet. or eth.) and with the marker of hydrocortisone hypersensitivity, tixocortol pivalate (1% pet.), detected all patients with positive intradermal tests to budesonide and hydrocortisone succinate, respectively ( 34–36). This proved to be both sensitive and specific. Although some additional cases might be discovered by intradermal tests ( 36), the inevitable atrophy-inducing side-effect of the test itself limits its practical use. Most of the literature on corticosteroid testing is based on results obtained by patch testing corticosteroids at around 1% dilutions (except for those described with the commercial products). Our results ( 28) and also those of Isaksson et al. ( 29, 37) suggest that perhaps (much) lower concentrations should also be tested to avoid the pharmacologic effect of corticosteroids on a weak allergic reaction against the molecule. The choice of the appropriate test vehicle is critical for all patch-test materials because it strongly influences their skin penetration. The test substance has to be soluble or evenly dispersed in the vehicle, in order to guarantee optimal bioavailability, which is generally lower than in a commercial preparation that has been formulated to offer enhanced skin penetration. Besides, the vehicle should not interfere with the patch-test results. Moreover, the test preparation should be stable. Ethanol, petrolatum, physiological sodium chloride solution, dimethylsulfoxide, and the prepared pharmaceutic preparations have been used for patch testing of corticosteroid molecules. However, the ideal vehicle for testing corticosteroids has yet to be determined. Hermal and Chemotechnique Diagnostics provide commercial corticosteroid allergens at various concentrations diluted in petrolatum. In our studies ( 28), corticosteroid-allergic patients were tested with an extended series of corticosteroids, containing over 60 substances, diluted in ethanol. Statistical analysis (chi-square test) of the results obtained by testing tixocortol pivalate 1% ethanol and petrolatum did not reveal a significant difference between the two vehicles. Patients allergic to budesonide showed significantly (P=0.05) more positive reactions to budesonide 0.1% in ethanol than to the same allergen diluted 0.1% in petrolatum. There was also a significant difference (P=0.05) between the ethanol test vehicle and petrolatum when hydrocortisone-17-butyrate 1% was tested. When one tests substances diluted in ethanol, the possibility of positive reactions to the vehicle itself has to be considered; therefore, ethanol itself has to be tested separately ( 38). Moreover, it has been shown that many corticosteroids degrade in ethanol solutions during storage ( 39); therefore, fresh solutions have to be prepared regularly. Since animals cannot be readily sensitized to corticosteroids, convincing cross-reaction studies are lacking. Some authors ( 40, 41) have conducted studies on sensitizing guinea pigs with corticosteroids, but so far have succeeded only with tixocortol pivalate. Statistical methods and three-dimensional chemical structure studies have to be used to analyze cross-reactivity. Clinical as well as molecular configuration studies have led to the classification of corticosteroids into four groups by the function of their allergenicity ( 42): • Group A: hydrocortisone type: no substitution on the D ring except a short-chain ester on C21 or a thioester on C21; i.e., tixocortol pivalate • Group B: triamcinolone acetonide type: C16, C17-cis-ketal or -diol structure • Group C: betamethasone type: C16-methyl substitution • Group D: hydrocortisone-17-butyrate type: long-chain ester at C17 or C17 and C21 with or without C16-methyl substitution. Members of a group cross-reacted with each other; corticosteroids belonging to different groups did not do so. Moreover, tixocortol pivalate was found to be a good screening agent for group A, a finding which was later confirmed by other authors ( 34, 35). Budesonide was found to be a marker for different groups of corticosteroids: not only for other acetonides (group B, to which it theoretically belongs) but also for certain esters (group D), such as hydrocortisone-17-butyrate and prednicarbate. These clinical observations were fully supported by conformational analysis of the electronic shape of the corticosteroids involved ( 43). Groups A, B, and D were found to be highly homogeneous within each group in terms of molecular structures, and significant differences were observed between the groups. The special behavior of budesonide can be attributed to its unique molecular structure; that is, its acetal function, which allows the molecule (actually the two evenly occurring isomers “R” and “S”) to mimic both the group B and the group D molecular shapes. Although conformational analysis did not detect group-specific characteristics of volume and shape for molecules of group C, it is still clinically important to maintain this grouping, as it seems that its members do not cross-react with other corticosteroid groups. Our studies on a large number of molecules tested in a big group fully support the classification of Coopman et al. ( 42). All these studies show that the epitope is located on or around the D ring of the corticosteroid structure. However, membership in a certain group does not indicate the sensitizing potential of a given corticosteroid. In general, many positive reactions are observed to group A (e.g., hydrocortisone, prednisolone, methylprednisolone, and their short-chain esters, cloprednol), group B (e.g., triamcinolon acetonide, amcinonide, and budesonide), and those group D corticosteroids that lack both a C16 methyl and halogen substitution (e.g., hydrocortisone-17-butyrate, -aceponate, -buteprate, methylprednisolone aceponate, and prednicarbate). However, group C and those group D corticosteroids that are C16 methyl and halogen substituted (e.g., betamethasone and its esters, such as valerate and propionate; clobetasone and its esters; and the new drugs mometasone furoate and fluticasone propionate) produce very few positive results. The distinctively different behavior of certain compounds of group D led us to propose subdivision of the ester steroids into two groups: group D1, containing C16 methyl substitution and halogenation on the B ring, and group D2, comprising the labile ester steroids that lack both substitutions. The latter are the so-called prodrug corticosteroids that, due to their high lipophilicity, easily penetrate into the skin, where they undergo rapid metabolization, giving rise to the formation of the corresponding structures with free hydroxyl groups at the C21 and C17 positions ( 44, 45). Hydrocortisone-17-butyrate is converted in the skin to its metabolite, C21 butyrate. The C21 ester is then enzymatically further converted to hydrocortisone. Methylprednisolone aceponate is metabolized via cutaneous esterases, and this enzymatic hydrolysis of the C21-acetate forms methylprednisolone-17-propionate. Via acylmigration, it is nonenzymatically converted to an intermediary component, methylprednisolone-21-propionate, which is then enzymatically hydrolyzed into the free corticosteroid methylprednisolone ( 46). Prednicarbate is rapidly hydrolyzed at position C21. The resulting intermediary product, prednisolone-17-ethylcarbonate, metabolizes slowly into prednisolone ( 47). The metabolization of corticosteroids in the skin influences not only the sensitizing potential of a given molecule but also the outcome of patch-test results as well as cross-reactivity, because corticosteroid metabolites may behave very differently with regard to allergenicity from the corticosteroid applied to the skin surface. Our data have shown that D2, prodrug corticosteroid esters not only significantly cross-reacted within the group, but also showed a significant cross-reactivity pattern with budesonide and also with group A corticosteroids ( Fig. 1). This observation can be explained if these molecules behave as haptens before and after biometabolization as well. Typical members and cross-reaction patterns of corticosteroid groups classified by function of their allergenicity. Several mechanisms have been identified in the formation of hapten-protein bonds ( 48, 49), but these are still disputed for corticosteroids. Cortico-steroid alcohols are not likely to be the actual haptens. Monder and coworkers ( 50–52) have demonstrated the catabolization of hydrocortisone to an intermediary compound of the glyoxal type, 21- dehydrohydrocortisone, which is a minor but very reactive α-ketoaldehyde metabolite ( 53) ( Fig. 2). One may assume that, in a biologic environment, such molecules are apt to react with nucleophilic residues of proteins to form covalent bonds ( 54). While the fixation mechanism of the corticosteroid molecules to the skin proteins in contact allergy reactions has yet to be completely elucidated, it seems that the reaction site involved in the formation of the antigenic com-plex is not the α,β-unsaturated ketone (a “classic” reactive site) at position C3–C4 of the steroid moiety, which is much too “congested”, but the α-ketoaldehyde formed after oxidation of the alcohol function at C21. The presence of the reactive binding site on the lateral chain of corticosteroids is in good agree-ment with cross-reaction patterns, a major influence being modifications on cycle D ( 43). Oxidation of hydrocortisone to 21-dehydro-hydrocortisone. In 1980, Bundgaard ( 55) first formulated the hypothesis of the formation of steroid-glyoxal degradation products in commercial corticosteroid preparations and their possible involvement in contact-allergic reactions. This hypothesis was confirmed ( 56). Molecules of the 21-dehydrocorticosteroid type were found to react with the amino acids involved in the formation of the antigenic complex; i.e., guanidyl groups of arginine. These adducts, of which the chemical structure has recently been elucidated by Arbez-Gindre ( 57), are stable ( 58); therefore, they ought to be taken into account. It is difficult to confirm that the immunochemical reaction that results in delayed hypersensitivity is exclusively dependent on these adducts. Moreover, it has recently been shown that the sensitization potential of several corticosteroids correlates significantly with their ability to react with arginine residues ( 7). The frequency of positive patch tests to ethanol is raised in corticosteroid-allergic patients ( 28, 59). Since both alcohols and corticosteroids are metabolized by dehydrogenases via aldehydes, reactions to both corticosteroids and alcohol might be due to aldehyde formation, a defect in such enzymes perhaps allowing aldehyde accumulation. Our studies on testing corticosteroid aldehydes in corticosteroid-sensitive patients ( 38) supported the hypothesis that these aldehydes may be intermediates in contact sensitization and elicitation. In the case of enzyme dysfunction, possible accumulation of corticosteroid aldehydes might favor induction of delayed hypersensitivity to corticosteroids, and the parallel accumulation of acetaldehyde (the aldehyde formed from ethanol) might account for irritant reactions to ethanol. Contact allergy to corticosteroids is a frequent hypersensitivity reaction that usually lacks alarming clinical symptoms. All patients undergoing patch testing should be tested to screen for corticosteroid allergy; therefore, the screening markers tixocortol pivalate and budesonide have to be included in the standard test trays. From now on, the European Contact Dermatitis Society and the European Experimental Contact Dermatitis Research Group recommend the use of tixocortol pivalate 0.1% pet. and budesonide 0.01% pet. in the European standard series ( 60). Individual skin-metabolization characteristics and probably other mechanisms (such as other potential recognition sites on the hapten) certainly influence cross-reactivity patterns. Therefore, patients diagnosed as having corticosteroid sensitivity should be further tested with more extensive corticosteroid series. Further treatment and appropriate counseling of these patients regarding both local and systemic corticosteroid therapy for the future can be based on the outcome of this test and on the classification of corticosteroids based on the cross-reaction patterns proposed above. To avoid sensitization of the population, corticosteroids with low sensitizing capacity should be prescribed.