Vet Dermatol 2014; 25: 507–e85
DOI: 10.1111/vde.12159
Comparison of the results of intradermal test reactivity and serum allergen-specific IgE measurement for Malassezia pachydermatis in atopic dogs Willam E. Oldenhoff*, Glenn R. Frank† and Douglas J. DeBoer* *Department of Medical Sciences, School of Veterinary Medicine, University of Wisconsin, Madison, 2015 Linden Drive, Madison, WI 53706, USA †Heska Corporation, 3760 Rocky Mountain Avenue, Loveland, CO 80538, USA Correspondence: Douglas J DeBoer, Department of Medical Sciences, School of Veterinary Medicine, University of Wisconsin, Madison, 2015 Linden Drive, Madison, WI 53706, USA. E-mail: [email protected]
Background – Malassezia pachydermatis is part of the normal flora of canine skin. Malassezia hypersensitivity is recognized as a trigger for clinical signs of atopic dermatitis (AD) in some dogs. Determinations of Malassezia hypersensitivity are often made with intradermal testing (IDT), which may have limited availability in a first-opinion veterinary practice. Hypothesis/Objectives – The purpose of this study was to compare immediate IDT reactivity to M. pachydermatis with results of an enzyme-linked immunosorbent assay (ELISA) designed to detect anti-Malassezia IgE. Animals – Eighty-four dogs with a clinical diagnosis of AD. Methods – Multi-allergen IDT was performed on all dogs. Serum testing for allergen-specific IgE against a panel of common environmental allergens and M. pachydermatis was performed by ELISA using the FceRIa receptor fragment as a detection reagent, with results reported as adjusted optical density (OD). A receiver operating characteristic (ROC) curve was used to analyse the results of the two tests. Results – The median adjusted OD of the anti-Malassezia IgE ELISA for dogs reactive and nonreactive to M. pachydermatis on IDT was 0.137 and 0.024, respectively. Analysis of the ROC curve suggested a cut-off point for the anti-Malassezia ELISA that yielded a sensitivity of 77.0% and a specificity of 89% relative to IDT results. Conclusions and clinical importance – Substantial agreement was demonstrated between IDT reactivity and anti-Malassezia IgE as detected by the FceRIa receptor reagent. Although correlation with a clinical diagnosis of Malassezia dermatitis was not attempted in this study, the results indicate that the ELISA may be used to demonstrate the presence of immediate-type Malassezia hypersensitivity in dogs with AD.
Introduction The yeast Malassezia pachydermatis is a constituent of the normal skin flora of healthy dogs. It is typically present in low numbers, but can function as an opportunistic pathogen as a result of changes in the natural defenses of the epidermis.1–7 Conditions that may predispose a dog to M. pachydermatis overgrowth include, but are not limited to, atopic dermatitis (AD) and seborrhoea.1,8 In addition to acting as an opportunistic pathogen, M. pachydermatis can also act as a trigger for clinical signs in some dogs with AD if hypersensitivity to yeast allergens develops. Both immediate-type and delayed-type hypersensitivity
Accepted 29 May 2014 Sources of Funding: Allergen-specific IgE assays were performed gratis by the Heska Corporation; study otherwise self-funded. Conflict of Interest: Glenn Frank is an employee of Heska Corporation. Douglas DeBoer is a consultant to Heska Corporation and has received funding for postgraduate training activities from Heska. © 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
reactions to this yeast have been well documented in some dogs with AD.9–15 It is important for clinicians to understand that clinical signs of Malassezia dermatitis, the finding of yeast organisms on cytology and demonstration of potential Malassezia hypersensitivity, using serological or intradermal testing methods, are three separate (but often related) concepts. These three elements may or may not coexist in a given case. Likewise, it is recognized in humas that Malassezia species can trigger and exacerbate the clinical signs of head and neck dermatitis, a form of AD.16–20 Treatment with antifungal agents results in improvement in symptoms of human patients with AD.21–24 The diagnosis of AD is based on compatible history and clinical findings rather than through ‘allergy testing’. Tests of allergen reactivity, i.e. allergen-specific intradermal testing (IDT) and IgE serological testing, are performed in an attempt to identify potentially relevant hypersensitivity reactions in dogs assigned a clinical diagnosis of AD. Both IDT and serology test results are used to formulate allergen-specific immunotherapy (ASIT), an important treatment option for AD. Determinations of immediatetype hypersensitivity to M. pachydermatis are often 507
Oldenhoff et al.
made through IDT, which is often limited to veterinarians who specialize in dermatology. In contrast, most veterinary general practitioners perform IgE serology tests to formulate ASIT, yet Malassezia-specific IgE tests are not universally available from testing laboratories and may not be included on typical panels. There is a need for wider study of IgE serology tests to evaluate immediatetype hypersensitivity to M. pachydermatis in the dog. The purpose of this study was to compare the results of IDT reactivity with the results of an enzyme-linked immunosorbent assay (ELISA) test for IgE to M. pachydermatis using the FceRIa receptor detection reagent in a population of atopic dogs.
Materials and methods Case selection and sera Serum samples were obtained from privately owned dogs presented to our facility’s dermatology clinic between 2006 and 2013, in accordance with standards and approval from the School of Veterinary Medicine Animal Care and Use Committee. Samples had been stored continuously at 80°C. Sera (n = 84) were selected from dogs meeting the following criteria: (i) clinical diagnosis of AD according to the accepted criteria in use at the time of presentation (these varied according to the year of diagnosis);25,26 (ii) exclusion of other pruritic dermatoses, such as ectoparastic infestation, flea allergy dermatitis or dermatophytosis, using standard diagnostic techniques and lack of response to external parasite control; (iii) completion of a strict dietary restriction trial using a commercial novel protein or hydrolysed diet, as appropriate to the dog at the time; and (iv) completion of a multi-allergen IDT with concurrent (i.e. same day) serum sample obtained for serology testing. The minimal drug withdrawal time prior to IDT was as follows: 7 days for antihistamines; 14 days for oral glucocorticoids; and 28 days for injectable glucocorticoids.
Multi-allergen IDT Dogs from which sera were obtained had undergone the clinic’s standard multi-allergen IDT using appropriately diluted allergenic extracts (Greer Laboratories Inc., Lenoir, NC, USA). The IDT included M. pachydermatis extract at 1000 PNU/mL. This concentration was chosen based upon a previous report that supported its use in demonstrating type-1 Malassezia hypersensitivity in atopic dogs.15 Intradermal allergen tests were scored using a 0–5 point scale, with the positive (histamine) control assigned a score of 4. Reactions were assigned a score of 5 if the reaction was greater than the histamine control. For purposes of analysis, results were considered to be negative if they were scored 0, 1 or 2, and positive if they were scored 3, 4 or 5; see Discussion section for additional details. Based on this criterion, of the 84 dogs included in this study, 30 were positive to M. pachydermatis on IDT and 54 were negative.
Enzyme-linked immunosorbent assay for allergenspecific IgE All sera were assayed in a standard commercial allergen-specific IgE panel including seasonal and nonseasonal allergens (Allerceptâ Northeast Regional Panel; Heska Corporation, Loveland, CO, USA).27 In most instances, the panel had been performed when the serum was initially collected. Remaining sera that had not been tested at that time were retrieved from frozen storage and sent for testing with the identical panel. All sera were newly tested for the presence of IgE against Malassezia antigens by the same commercial laboratory, using identical proprietary assay protocols to those used in the standard commercial assay for IgE against other allergens.27 The optimal reagent
508
concentrations were determined by checkerboard titration as necessary. Assay steps in this microplate ELISA included (in sequence, with incubation and washing between steps): coating wells with M. pachydermatis extract (Greer Laboratories); addition of dog serum; addition of biotinylated FceRIa receptor reagent; addition of avidin–alkaline phosphatase; and addition of para-nitrophenylphosphate enzyme substrate. Following incubation and colour development, optical density at 405 nm was determined. Results for the anti-Malassezia ELISA were expressed as optical density adjusted for background reactivity in a ‘blank’ well (adjusted OD).
Statistical analysis A receiver operating characteristic (ROC) curve28 was used to analyse the results of the anti-Malassezia ELISA in comparison to the IDT result. Receiver operating characteristic curve analysis is a method to visualize the relationship between sensitivity and specificity given different assay interpretation criteria; it facilitates determination of an optimal cut-off point for a diagnostic test. On an ROC curve, the point of maximal inflection suggests the best cut-off point; the area under the curve can be used as a measure of test accuracy. Cohen’s kappa coefficient was used to compare test results for IDT (positive versus negative) with test results for anti-Malassezia ELISA (positive versus negative).29 In addition, we examined the proportions of animals serologically positive or negative to M. pachydermatis in relation to their ELISA results for other environmental allergens. To accomplish this, cases were deemed either positive or negative to each of the following test categories: grasses; weeds; trees; moulds; or indoor allergens. An animal was considered positive to a category if there was one or more individual positive test result within this category. Results for Malassezia (positive or negative) were compared with results for each allergen group (positive or negative) using Cohen’s kappa coefficient. Statistical analyses were performed using R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, Austria). The ROC curve was generated with the use of the ROCR package for R (www.r-project.org).30
Results Comparison of anti-Malassezia ELISA and antiMalassezia IDT The median adjusted OD of the anti-Malassezia ELISA for dogs positive to Malassezia on IDT was 0.137; the median adjusted OD for dogs negative to Malassezia on IDT was 0.024 (Figure 1). Dogs with higher IDT scores often had higher adjusted OD than those with lower IDT scores (Figure 2). An ROC curve for anti-Malassezia ELISA was constructed based on identification of immediate-type Malassezia hypersensitivity (Figure 3). In this analysis, an IDT score was considered to be positive if it was 3 or greater and negative if 2 or less. Analysis of the ROC curve revealed that a cut-off adjusted OD of 0.073 would be most appropriate and would result in a sensitivity of 77.0% and a specificity of 89%. When a cut-off adjusted OD of 0.073 was used, the anti-Malassezia ELISA and IDT had substantial agreement (j = 0.660). The kappa coefficient is a measure of agreement between two observations, with a range of 1 to 1. A value of 0 is equivalent to chance agreement, with 1 indicating complete disagreement and 1 indicating perfect agreement.29 The area under the ROC curve was 0.856, an indication that the anti-Malassezia ELISA had good diagnostic accuracy for identification of immediate-type Malassezia hypersensitivity.28 © 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
Comparison of tests for Malassezia hypersensitivity
Figure 1. Box and whisker plot showing anti-Malassezia ELISA adjusted optical density (OD; y-axis) for dogs that were negative to M. pachydermatis on intradermal testing (IDT; left) and dogs that were IDT-positive to M. pachydermatis (right). The middle horizontal line indicates the median. The box upper and lower boundaries indicate the 25th and 75th percentiles. The whisker length indicates the total range. Maximal OD for an IDT-negative dog, 0.941; maximal OD for an IDT-positive dog, 3.843.
Figure 2. Anti-Malassezia ELISA adjusted optical density (OD) values versus intradermal test (IDT) scores against Malassezia in the same dog. Outliers not shown: OD scores of 2.335 and 3.843 for IDT score of 4. The categorical nature of the IDT results precludes correlation analysis, but it is evident that there is a trend towards increasing OD scores as IDT score increases.
Analysis of anti-Malassezia ELISA versus other serology results Kappa coefficient results were calculated for comparison of anti-Malassezia ELISA results with ELISA results of other allergen groups. There was poor agreement between the anti-Malassezia ELISA and ELISA results for trees (j = 0.183), grasses (j = 0.314), weeds (j = 0.286), moulds (j = 0.155) and mites (j = 0.088).
Discussion In this study, we demonstrated substantial agreement between the results of an anti-Malassezia ELISA and IDT. The results of this study indicate that the anti-Malassezia ELISA described herein might be useful to demonstrate immediate-type hypersensitivity to Malassezia in the dog. © 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
Figure 3. Receiver operating characteristic curve for anti-Malassezia ELISA-based identification of immediate-type Malassezia hypersensitivity. True-positive rate (sensitivity) and false-positive rate (1 specificity) were calculated for all possible cut-off points. The arrow indicates the optimal anti-Malassezia ELISA OD cut-off point. AUC indicates area under the curve.
Agreement between these two tests of immediatetype hypersensitivity was not perfect; in some situations, one test was clearly positive but the other negative. This is not surprising, because historically intradermal test and ELISA results for other allergens do not always agree. Various hypotheses have been proposed to explain such discrepancies, including the possible existence of heterogeneous IgE subclasses and differences in the amounts of IgE in the circulation versus that bound to mast cells. It is also a fact that ELISA, by design, measures only quantities of allergen-specific IgE and not necessarily whether the IgE is able to function in a hypersensitivity reaction (as might be demonstrated by IDT). Our study was not meant to evaluate the clinical utility of these tests of hypersensitivity nor to correlate them with patient clinical signs or cytological findings. However, in evaluating a dog with clinical signs consistent with Malassezia dermatitis, both cytological evaluation of skin impression smears and tests for hypersensitivity may be useful. It is important to understand that clinical signs, cytological findings and IDT or ELISA results are independent elements of disease evaluation and that diagnosis should not rest on any single element. For example, failure to find yeast organisms on cytology does not rule out the possible contribution of yeast to clinical signs. What then, might be the clinical value of a positive IDT or ELISA to Malassezia antigen? As an example, a positive Malassezia hypersensitivity test in a dog might lead the clinician to consider M. pachydermatis overgrowth as a component of that animal’s disease and prompt cytological evaluation if the clinician had neglected to perform it previously. Conversely, a negative Malassezia IgE test does not rule out a pathological role for this organism in an atopic dog. The value of demonstrating immediate-type hypersensitivity to Malassezia yeast may have additional value in 509
Oldenhoff et al.
long-term case management. A positive test raises the possibility that longer-term prophylactic topical antifungal treatment may be beneficial or that an M. pachydermatis extract could be used as part of an ASIT regimen. It must be emphasized, however, that the efficacy of yeast extracts in ASIT has not yet been reported. It is important to remember that dogs can have either immediate-type hypersensitivity or delayed-type hypersensitivity (or both concurrently) to Malassezia and that, in this study, we evaluated only the former type. It is possible that a dog may have negative results on the anti-Malassezia ELISA and/or negative immediate IDT reactivity, but have delayed-type hypersensitivity to Malassezia that could be clinically relevant. The latter could be detected, for example, by patch testing or by observing delayed (24–48 h) reactivity on IDT. In human patients with AD, correlation has been found between skin-prick testing and Malassezia-specific serum IgE, i.e. immediate hypersensitivity can be demonstrated by either method.16 Atopy patch testing has been used to document delayed-type hypersensitivity in patients with low Malassezia-specific IgE levels and negative skin-prick test results.16 To date, there are no published reports investigating patch test results to M. pachydermatis in atopic dogs, although one study found that positive patch test reactivity to yeast was more common in basset hounds with Malassezia dermatitis compared with healthy control dogs.14 Considering the perceived importance of yeast hypersensitivity in some atopic dogs, further studies investigating delayed IDT reactivity and atopy patch test reactivity to M. pachydermatis in atopic dogs are needed. We speculated that dogs positive to Malassezia might be more likely to have IgE against other environmental substances, such as fungi, perhaps even due to fungal antigen cross-reactivity. We therefore compared the results of the anti-Malassezia ELISA with those from the complete allergen-specific IgE panel. Comparison of the anti-Malassezia ELISA and the complete allergen-specific IgE panel indicated that dogs with positive results to Malassezia were somewhat more likely (than expected by chance) to have positive results to trees, grasses, weeds or moulds (j > 0). Nevertheless, the low kappa coefficients observed for these relationships indicated poor agreement. The fact that the agreement was greater than chance may simply reflect the fact that atopic dogs typically sensitize to a range of common allergens, rather than a single one. Statistical comparisons of results of these two tests mandated that we define in advance what constituted a ‘positive’ versus ‘negative’ IDT. This definition is by its nature somewhat arbitrary and subject to debate. In our analysis, we chose to compare the ELISA results with an IDT result that was in our view unequivocal, i.e. 3 or greater on our scale, as this typically prompts inclusion in ASIT in our practice, whereas scores of 2 or less are not included. Instead, we could have chosen to compare the ELISA results with a less strict definition of a positive IDT result, i.e. considering an IDT reaction of 2 also to be positive, though weakly so. Indeed, some authorities advocate defining an IDT of 2 or greater as positive. In making this alternative comparison, one might expect that the agreement between the two tests would be lower with 510
addition of the less certain, ‘weaker’ positive IDT results to the data set. To explore this, we recalculated the statistical analysis, defining a positive IDT as ‘2 or greater’ rather than ‘3 or greater.’ In this case (ROC curve not shown), the kappa statistic was 0.527, the area under the curve 0.76, and the adjusted OD cut-off with the highest kappa was 0.073. At this cut-off and including an IDT of 2 as positive, the sensitivity of the ELISA would then be 63.9% and the specificity 87.5%. Thus, if we reduced the positive threshold of the IDT to 2, as expected, the agreement of the two tests was lower. Another way of stating this is that the ELISA result has better agreement with a ‘strong positive’ IDT than with ‘any positive’ IDT. It is useful to consider the concept of ‘cut-off point’ in diagnostic assays, because the term can mean different things in different situations. Where a disease state is known to be present or absent with certainty, it is possible to calculate sensitivity and specificity of a diagnostic test for that disease and to set up a ‘cut-off point’ for the test based on evaluation of a large number of known negative and positive samples. The ‘cut-off point’ or positive threshold is often set at two-to-three standard deviations above the mean of the negative sample group.31 Samples above this ‘cut-off point’ are deemed positive and therefore diagnostic of the disease. In contrast, when comparing two different diagnostic tests with each other, in some cases the disease state may not be known with certainty. In this situation, for analytical purposes one test can be defined as the ‘gold standard’. The second test can then be evaluated for its ability to reach the same conclusion as the ‘gold standard’. In this situation, a ‘cut-off point’ is determined by ROC analysis and describes that optimal point at which the new test mirrors the results of the ‘gold standard’. 32 In the present study, it is important to understand that our calculated ‘cut-off point’ refers only to this agreement between tests, and not to the diagnostic efficiency of either test. Defining assay ‘cut-off values’ for the purpose of diagnosis would require further assay of very large collections of healthy dog sera and was not the aim of the present study. In brief, our analyses indicate that ELISA and IDT are conveying the same information, without specifying that either test is valid for diagnosis. This problem is a continuing plague in the use of allergen tests of all types; because we do not know ‘the truth’ of whether a patient is allergic to a substance without provocative testing, we are forced to define ‘gold standards’ (such as IDT) that may be of uncertain diagnostic efficiency. In summary, in this study we demonstrate that an antiMalassezia ELISA using the FceRIa receptor detection reagent can be used to demonstrate immediate-type hypersensitivity to Malassezia yeast in the dog, with results that agree substantially with those of IDT. This finding may have particular value for those veterinary practitioners who primarily use serological testing in the evaluation of a dog with allergic skin disease.
Acknowledgements The authors thank Nicholas S. Keuler (Statistical Consulting Service, College of Agricultural & Life Sciences, Uni© 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
Comparison of tests for Malassezia hypersensitivity
versity of Wisconsin-Madison) for assistance in statistical analysis.
References 1. Plant J, Rosenkrantz W, Griffin C. Factors associated with and prevalence of high Malassezia pachydermatis numbers on dog skin. J Am Vet Med Assoc 1992; 201: 879–882. 2. Mason IS, Mason KV, Lloyd DH. A review of the biology of canine skin with respect to the commensals Staphylococcus intermedius, Demodex canis and Malassezia pachydermatis. Vet Dermatol 1996; 7: 119–132. 3. Chen TA, Hill PB. The biology of Malassezia organisms and their ability to induce immune responses and skin disease. Vet Dermatol 2005; 16: 4–26. 4. Kennis R, Rosser E Jr, Olivier N et al. Quantity and distribution of Malassezia organisms on the skin of clinically normal dogs. J Am Vet Med Assoc 1996; 208: 1048–1051. 5. Nardoni S, Mancianti F, Corazza M et al. Occurrence of Malassezia species in healthy and dermatologically diseased dogs. Mycopathologia 2004; 157: 383–388. 6. Nardoni S, Dini M, Taccini F et al. Occurrence, distribution and population size of Malassezia pachydermatis on skin and mucosae of atopic dogs. Vet Microbiol 2007; 122: 172–177. 7. Cafarchia C, Gallo S, Romito D et al. Frequency, body distribution, and population size of Malassezia species in healthy dogs and in dogs with localized cutaneous lesions. J Vet Diagn Invest 2005; 17: 316–322. 8. Bond R, Ferguson E, Curtis C et al. Factors associated with elevated cutaneous Malassezia pachydermatis populations in dogs with pruritic skin disease. J Small Anim Pract 1996; 37: 103– 107. 9. Morris D, Olivier N, Rosser E. Type-1 hypersensitivity reactions to Malassezia pachydermatis extracts in atopic dogs. Am J Vet Res 1998; 59: 836–841. 10. Morris DO, DeBoer DJ. Evaluation of serum obtained from atopic dogs with dermatitis attributable to Malassezia pachydermatis for passive transfer of immediate hypersensitivity to that organism. Am J Vet Res 2003; 64: 262–266. 11. Morris DO, Clayton DJ, Drobatz KJ et al. Response to Malassezia pachydermatis by peripheral blood mononuclear cells from clinically normal and atopic dogs. Am J Vet Res 2002; 63: 358–362. 12. Nuttall TJ, Halliwell RE. Serum antibodies to Malassezia yeasts in canine atopic dermatitis. Vet Dermatol 2001; 12: 327–332. 13. Bond R, Hendricks A, Ferguson E et al. Intradermal test reactivity to Malassezia pachydermatis in atopic dogs. Vet Rec 2002; 150: 448–449. 14. Bond R, Patterson-Kane J, Perrins N et al. Patch test responses to Malassezia pachydermatis in healthy basset hounds and in basset hounds with Malassezia dermatitis. Med Mycol 2006; 44: 419–427.
15. Farver K, Morris DO, Shofer F et al. Humoral measurement of type-1 hypersensitivity reactions to a commercial Malassezia allergen. Vet Dermatol 2005; 16: 261–268. €m M, Bartosik J et al. Atopy patch test 16. Johansson C, Sandstro reactions to Malassezia allergens differentiate subgroups of atopic dermatitis patients. Br J Dermatol 2003; 148: 479–488. 17. Sonesson A, Bartosik J, Christiansen J et al. Sensitization to skin-associated microorganisms in adult patients with atopic dermatitis is of importance for disease severity. Acta Derm Venereol 2013; 93: 340–346. 18. Darabi K, Hostetler SG, Bechtel MA et al. The role of Malassezia in atopic dermatitis affecting the head and neck of adults. J Am Acad Dermatol 2009; 60: 125–136. 19. Wessels M, Dolkes G, Dijk A et al. IgE antibodies to Pityrosporum ovale in atopic dermatitis. Br J Dermatol 1991; 125: 227– 232. 20. Bayrou O, Pecquet C, Flahault A et al. Head and neck atopic dermatitis and Malassezia furfur-specific IgE antibodies. Dermatology 2005; 211: 107–113. 21. B€ ack O, Bartosik J. Systemic ketoconazole for yeast allergic patients with atopic dermatitis. J Eur Acad Dermatol Venereol 2001; 15: 34–38. 22. Svejgaard E, Ølholm Larsen P, Deleuran M et al. Treatment of head and neck dermatitis comparing itraconazole 200 mg and 400 mg daily for 1 week with placebo. J Eur Acad Dermatol Venereol 2004; 18: 445–449. 23. Ikezawa Z, Kondo M, Okajima M et al. Clinical usefulness of oral itraconazole, an antimycotic drug, for refractory atopic dermatitis. Eur J Dermatol 2004; 14: 400–406. 24. B€ ack O, Scheynius A, Johansson S. Ketoconazole in atopic dermatitis: therapeutic response is correlated with decrease in serum IgE. Arch Dermatol Res 1995; 287: 448–451. laud P, Guague re E, Alhaidari Z et al. Reevaluation of diag25. Pre nostic criteria of canine atopic dermatitis. Rev Vet Med 1998; 149: 1057–1064. 26. Favrot C, Steffan J, Seewald W et al. A prospective study on the clinical features of chronic canine atopic dermatitis and its diagnosis. Vet Dermatol 2010; 21: 23–31. 27. Stedman K, Lee K, Hunter S et al. Measurement of canine IgE using the alpha chain of the human high affinity IgE receptor. Vet Immunol Immunopathol 2001; 78: 349–355. 28. Akobeng AK. Understanding diagnostic tests 3: receiver operating characteristic curves. Acta Paediatr 2007; 96: 644–647. 29. Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. Fam Med 2005; 37: 360–363. 30. Sing T, Sander O, Beerenwinkel N et al. ROCR: visualizing classifier performance in R. Bioinformatics 2005; 21: 3940–3941. 31. Malvano R, Boniolo A, Dovis M et al. ELISA for antibody measurement: aspects related to data expression. J Immunol Methods 1982; 48: 51–60. 32. Bewick V, Cheek L, Ball J. Statistics review 13: receiver operating characteristic curves. Crit Care 2004; 8: 508–512.
sume Re Contexte – Malassezia pachydermatis fait partie de la flore normale de la peau du chien. L’hypersensibilite a Malassezia est reconnue comme un facteur aggravant des signes cliniques de dermatite atopique (AD) termination de l’hypersensibilite chez certains chiens. La de a Malassezia est souvent faite par tests intra te rinaire ge ne rale. dermiques (IDT) qui sont peu disponibles en pratique ve ses/Objectifs – Le but de cette e tude e tait de comparer la re activite imme diate des IDT Hypothe a sultats d’un test ELISA (enzyme-linked immunosorbent assay) concßu pour M. pachydermatis avce les re tecter les IgE anti-Malassezia. de Sujets – Quatre vingt quatre chiens atopiques. thodes – Des IDT de plusieurs allerge nes ont e te re alise es sur tous les chiens. Des tests se riques pour Me cifiques d’allerge nes contre un panel d’allerge nes environnementaux fre quents et M. pachyderles IgE spe te re alise s par ELISA a l’aide d’un fragment de re cepteur FcƐRIa, les re sultats e tant rapporte s matis ont e optique ajuste e (OD). La courbe ROC (Receiver Operating Characteristic) a e te utilise e pour par densite sultats des deux tests. analyser les re © 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
511
Oldenhoff et al.
sultats – L’OD ajuste e moyenne des tests ELISA des IgE anti-Malassezia des chiens re actifs et non Re actifs a M. pachydermatis aux IDT e tait respectivement de 0.137 et 0.024. L’analyse de la courbe ROC re re un point de rupture pour l’ELISA anti-Malassezia qui produit une sensibilite de 77.0% et une sugge cificite de 89% relative aux re sultats d’IDT. spe lation substantielle a e te de montre e entre la re activite Conclusions et importance clinique – Une corre tecte s par le re actif de re cepteurs FcƐRIa. Bien que la corre lation avec des IDT et des IgE anti-Malassezia de te re alise e dans cette e tude, les re sultats indiun diagnostic clinique de dermatite a Malassezia n’ait pas e ^tre utilise pour de montrer la pre sence d’une hypersensibilite imme diate quent que le test ELISA peut e a Malassezia chez les chiens atopiques. Resumen n – Malassezia paquidermatis es parte de la flora normal de la piel canina. La hipersensibilidad Introduccio pica (AD) en alga Malassezia se reconoce como un factor iniciador de los signos clınicos de dermatitis ato n de la hipersensibilidad a Malassezia se hace mediante prueba intradermal unos perros. La determinacio n veterinaria. (ITT), que puede ser de acceso limitado en las pr acticas de primera opinio tesis/Objetivos – El propo sito de este estudio fue comparar la reactividad intrade rmica inmediata a Hipo M. pachidermatis con los resultados de un ensayo inmunoenzim atico designado para detectar la IgE frente a Malassezia. stico clınico de dermatitis ato pica. Animales – 84 perros con un diagno todos – Se desarrollo una prueba intrade rmica multi alerge nica en todos los perros. Las pruebas se ricas Me rgeno frente a un panel de alergenos ambientales comunes y frente a Malapara las IgE especıficas de ale mediante ELISA utilizando el fragmento del receptor FcƐRIa como el ressezia paquidermatis se desarrollo n. Los resultados se indicaron como densidad o ptica ajustada (OD). Una curva de activo de la deteccio para analizar los resultados de las dos pruebas. caracterıstica del operador receptor (ROC) se utilizo ptica ajustada de la prueba de ELISA para IgE frente a Malassezia Resultados – La mediana de densidad o en los perros reactivos y no reactivos frente a Malassezia pachidermatis tras la propuesta intradermal der un punto de corte para mal fue de 0,137 y 0,024, respectivamente. En el an alisis de la curva ROC se sugirio el ELISA frente a Malassezia que tiene una sensibilidad del 77% y una especificidad del 89% relativa a los rmica. estados la prueba intrade Conclusiones e importancia clınica – Hay un una concordancia sustancial demostrada entre la reactivirmica y la prueba de IgE frente Malassezia detectada mediante el uso del receptor FcƐRIa aundad intrade n con el diagno stico clınico de dermatitis por Malassezia no fue intentada en este estudio, que la correlacio los resultados indican que este ELISA puede utilizarse para demostrar la presencia de la hipersensibilidad pica inmediata a Malassezia en perros con dermatitis ato Zusammenfassung Hintergrund – Malassezia pachydermatis ist Teil der normalen Flora der Haut des Hundes. Eine Hypersen€ser fu €r die klinische Symptomatik der atopischen Dermatitis (AD) bei sibilit€at auf Malassezien ist als Auslo einigen Hunden bekannt. Eine Malassezien Hypersensibilit€ at wird oft mittels Intradermaltest (IDT) diagnos€glicherwiese nur limitiert zur Verfu €gung steht. tiziert, was in einer allgemeinen Praxis mo Hypothese/Ziele – Das Ziel dieser Studie war ein Vergleich der IDT Reaktivit€ at auf M. pachydermatis mit den Ergebnissen eines ELISA (Enzyme-linked Immunosorbent Assay), der zur Bestimmung von anti-Malassezien IgE entwickelt wurde. Tiere – Vierundachtzig Hunde mit der klinischen Diagnose von AD. €hrt. Ein serologischer Test zur AllerMethoden – Ein Multi-Allergen IDT wurde bei allen Hunden durchgefu gen-spezifischen IgE Bestimmung auf ein Panel von Umweltallergenen und auf M. pachydermatis wurde €hrt, wobei das FcƐRIa Rezeptorfragment als Reagens verwendet wurde. Die Ermittels ELISA durchgefu gebnisse wurden als „angepasste” optische Dichte (OD) angegeben. Eine ROC Kurve (Receiver Operating Characteristic Curve) wurde verwendet, um die Ergebnisse der beiden Tests zu analysieren. €r Hunde, die auf M. pachyderErgebnisse – Die Median angepasste OD des anti-Malassezien IgE ELISA fu matis im IDT reagierten bzw. nicht reagierten betrug 0,137 bzw. 0,024. Eine Analyse der ROC Kurve ergab €r den anti-Malassezien ELISA einen Cut Off, der in Relation zu den IDT Ergebnissen eine Sensibilit€ fu at von 77% und eine Spezifizit€at von 89% ergab. € Schlussfolgerungen und klinische Bedeutung – Es wurde eine deutliche Ubereinstimmung zwischen der IDT Reaktivit€at und dem anti-Malassezien IgE, welches mittels FcƐRIa Rezeptorreagens bestimmt wurde, festgestellt. Obwohl eine Korrelation mit der klinischen Diagnostik nicht Teil dieser Studie war, wei€glicherweise verwendet werden kann, um das Vorliegen sen die Ergebnisse darauf hin, dass der ELISA mo einer Malassezien Hypersensibilita€t vom Soforttyp bei Hunden mit AD festzustellen.
e84
© 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
Comparison of tests for Malassezia hypersensitivity
© 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
e85
DOI: 10.1111/vde.12159
Comparison of the results of intradermal test reactivity and serum allergen-specific IgE measurement for Malassezia pachydermatis in atopic dogs Willam E. Oldenhoff*, Glenn R. Frank† and Douglas J. DeBoer* *Department of Medical Sciences, School of Veterinary Medicine, University of Wisconsin, Madison, 2015 Linden Drive, Madison, WI 53706, USA †Heska Corporation, 3760 Rocky Mountain Avenue, Loveland, CO 80538, USA Correspondence: Douglas J DeBoer, Department of Medical Sciences, School of Veterinary Medicine, University of Wisconsin, Madison, 2015 Linden Drive, Madison, WI 53706, USA. E-mail: [email protected]
Background – Malassezia pachydermatis is part of the normal flora of canine skin. Malassezia hypersensitivity is recognized as a trigger for clinical signs of atopic dermatitis (AD) in some dogs. Determinations of Malassezia hypersensitivity are often made with intradermal testing (IDT), which may have limited availability in a first-opinion veterinary practice. Hypothesis/Objectives – The purpose of this study was to compare immediate IDT reactivity to M. pachydermatis with results of an enzyme-linked immunosorbent assay (ELISA) designed to detect anti-Malassezia IgE. Animals – Eighty-four dogs with a clinical diagnosis of AD. Methods – Multi-allergen IDT was performed on all dogs. Serum testing for allergen-specific IgE against a panel of common environmental allergens and M. pachydermatis was performed by ELISA using the FceRIa receptor fragment as a detection reagent, with results reported as adjusted optical density (OD). A receiver operating characteristic (ROC) curve was used to analyse the results of the two tests. Results – The median adjusted OD of the anti-Malassezia IgE ELISA for dogs reactive and nonreactive to M. pachydermatis on IDT was 0.137 and 0.024, respectively. Analysis of the ROC curve suggested a cut-off point for the anti-Malassezia ELISA that yielded a sensitivity of 77.0% and a specificity of 89% relative to IDT results. Conclusions and clinical importance – Substantial agreement was demonstrated between IDT reactivity and anti-Malassezia IgE as detected by the FceRIa receptor reagent. Although correlation with a clinical diagnosis of Malassezia dermatitis was not attempted in this study, the results indicate that the ELISA may be used to demonstrate the presence of immediate-type Malassezia hypersensitivity in dogs with AD.
Introduction The yeast Malassezia pachydermatis is a constituent of the normal skin flora of healthy dogs. It is typically present in low numbers, but can function as an opportunistic pathogen as a result of changes in the natural defenses of the epidermis.1–7 Conditions that may predispose a dog to M. pachydermatis overgrowth include, but are not limited to, atopic dermatitis (AD) and seborrhoea.1,8 In addition to acting as an opportunistic pathogen, M. pachydermatis can also act as a trigger for clinical signs in some dogs with AD if hypersensitivity to yeast allergens develops. Both immediate-type and delayed-type hypersensitivity
Accepted 29 May 2014 Sources of Funding: Allergen-specific IgE assays were performed gratis by the Heska Corporation; study otherwise self-funded. Conflict of Interest: Glenn Frank is an employee of Heska Corporation. Douglas DeBoer is a consultant to Heska Corporation and has received funding for postgraduate training activities from Heska. © 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
reactions to this yeast have been well documented in some dogs with AD.9–15 It is important for clinicians to understand that clinical signs of Malassezia dermatitis, the finding of yeast organisms on cytology and demonstration of potential Malassezia hypersensitivity, using serological or intradermal testing methods, are three separate (but often related) concepts. These three elements may or may not coexist in a given case. Likewise, it is recognized in humas that Malassezia species can trigger and exacerbate the clinical signs of head and neck dermatitis, a form of AD.16–20 Treatment with antifungal agents results in improvement in symptoms of human patients with AD.21–24 The diagnosis of AD is based on compatible history and clinical findings rather than through ‘allergy testing’. Tests of allergen reactivity, i.e. allergen-specific intradermal testing (IDT) and IgE serological testing, are performed in an attempt to identify potentially relevant hypersensitivity reactions in dogs assigned a clinical diagnosis of AD. Both IDT and serology test results are used to formulate allergen-specific immunotherapy (ASIT), an important treatment option for AD. Determinations of immediatetype hypersensitivity to M. pachydermatis are often 507
Oldenhoff et al.
made through IDT, which is often limited to veterinarians who specialize in dermatology. In contrast, most veterinary general practitioners perform IgE serology tests to formulate ASIT, yet Malassezia-specific IgE tests are not universally available from testing laboratories and may not be included on typical panels. There is a need for wider study of IgE serology tests to evaluate immediatetype hypersensitivity to M. pachydermatis in the dog. The purpose of this study was to compare the results of IDT reactivity with the results of an enzyme-linked immunosorbent assay (ELISA) test for IgE to M. pachydermatis using the FceRIa receptor detection reagent in a population of atopic dogs.
Materials and methods Case selection and sera Serum samples were obtained from privately owned dogs presented to our facility’s dermatology clinic between 2006 and 2013, in accordance with standards and approval from the School of Veterinary Medicine Animal Care and Use Committee. Samples had been stored continuously at 80°C. Sera (n = 84) were selected from dogs meeting the following criteria: (i) clinical diagnosis of AD according to the accepted criteria in use at the time of presentation (these varied according to the year of diagnosis);25,26 (ii) exclusion of other pruritic dermatoses, such as ectoparastic infestation, flea allergy dermatitis or dermatophytosis, using standard diagnostic techniques and lack of response to external parasite control; (iii) completion of a strict dietary restriction trial using a commercial novel protein or hydrolysed diet, as appropriate to the dog at the time; and (iv) completion of a multi-allergen IDT with concurrent (i.e. same day) serum sample obtained for serology testing. The minimal drug withdrawal time prior to IDT was as follows: 7 days for antihistamines; 14 days for oral glucocorticoids; and 28 days for injectable glucocorticoids.
Multi-allergen IDT Dogs from which sera were obtained had undergone the clinic’s standard multi-allergen IDT using appropriately diluted allergenic extracts (Greer Laboratories Inc., Lenoir, NC, USA). The IDT included M. pachydermatis extract at 1000 PNU/mL. This concentration was chosen based upon a previous report that supported its use in demonstrating type-1 Malassezia hypersensitivity in atopic dogs.15 Intradermal allergen tests were scored using a 0–5 point scale, with the positive (histamine) control assigned a score of 4. Reactions were assigned a score of 5 if the reaction was greater than the histamine control. For purposes of analysis, results were considered to be negative if they were scored 0, 1 or 2, and positive if they were scored 3, 4 or 5; see Discussion section for additional details. Based on this criterion, of the 84 dogs included in this study, 30 were positive to M. pachydermatis on IDT and 54 were negative.
Enzyme-linked immunosorbent assay for allergenspecific IgE All sera were assayed in a standard commercial allergen-specific IgE panel including seasonal and nonseasonal allergens (Allerceptâ Northeast Regional Panel; Heska Corporation, Loveland, CO, USA).27 In most instances, the panel had been performed when the serum was initially collected. Remaining sera that had not been tested at that time were retrieved from frozen storage and sent for testing with the identical panel. All sera were newly tested for the presence of IgE against Malassezia antigens by the same commercial laboratory, using identical proprietary assay protocols to those used in the standard commercial assay for IgE against other allergens.27 The optimal reagent
508
concentrations were determined by checkerboard titration as necessary. Assay steps in this microplate ELISA included (in sequence, with incubation and washing between steps): coating wells with M. pachydermatis extract (Greer Laboratories); addition of dog serum; addition of biotinylated FceRIa receptor reagent; addition of avidin–alkaline phosphatase; and addition of para-nitrophenylphosphate enzyme substrate. Following incubation and colour development, optical density at 405 nm was determined. Results for the anti-Malassezia ELISA were expressed as optical density adjusted for background reactivity in a ‘blank’ well (adjusted OD).
Statistical analysis A receiver operating characteristic (ROC) curve28 was used to analyse the results of the anti-Malassezia ELISA in comparison to the IDT result. Receiver operating characteristic curve analysis is a method to visualize the relationship between sensitivity and specificity given different assay interpretation criteria; it facilitates determination of an optimal cut-off point for a diagnostic test. On an ROC curve, the point of maximal inflection suggests the best cut-off point; the area under the curve can be used as a measure of test accuracy. Cohen’s kappa coefficient was used to compare test results for IDT (positive versus negative) with test results for anti-Malassezia ELISA (positive versus negative).29 In addition, we examined the proportions of animals serologically positive or negative to M. pachydermatis in relation to their ELISA results for other environmental allergens. To accomplish this, cases were deemed either positive or negative to each of the following test categories: grasses; weeds; trees; moulds; or indoor allergens. An animal was considered positive to a category if there was one or more individual positive test result within this category. Results for Malassezia (positive or negative) were compared with results for each allergen group (positive or negative) using Cohen’s kappa coefficient. Statistical analyses were performed using R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, Austria). The ROC curve was generated with the use of the ROCR package for R (www.r-project.org).30
Results Comparison of anti-Malassezia ELISA and antiMalassezia IDT The median adjusted OD of the anti-Malassezia ELISA for dogs positive to Malassezia on IDT was 0.137; the median adjusted OD for dogs negative to Malassezia on IDT was 0.024 (Figure 1). Dogs with higher IDT scores often had higher adjusted OD than those with lower IDT scores (Figure 2). An ROC curve for anti-Malassezia ELISA was constructed based on identification of immediate-type Malassezia hypersensitivity (Figure 3). In this analysis, an IDT score was considered to be positive if it was 3 or greater and negative if 2 or less. Analysis of the ROC curve revealed that a cut-off adjusted OD of 0.073 would be most appropriate and would result in a sensitivity of 77.0% and a specificity of 89%. When a cut-off adjusted OD of 0.073 was used, the anti-Malassezia ELISA and IDT had substantial agreement (j = 0.660). The kappa coefficient is a measure of agreement between two observations, with a range of 1 to 1. A value of 0 is equivalent to chance agreement, with 1 indicating complete disagreement and 1 indicating perfect agreement.29 The area under the ROC curve was 0.856, an indication that the anti-Malassezia ELISA had good diagnostic accuracy for identification of immediate-type Malassezia hypersensitivity.28 © 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
Comparison of tests for Malassezia hypersensitivity
Figure 1. Box and whisker plot showing anti-Malassezia ELISA adjusted optical density (OD; y-axis) for dogs that were negative to M. pachydermatis on intradermal testing (IDT; left) and dogs that were IDT-positive to M. pachydermatis (right). The middle horizontal line indicates the median. The box upper and lower boundaries indicate the 25th and 75th percentiles. The whisker length indicates the total range. Maximal OD for an IDT-negative dog, 0.941; maximal OD for an IDT-positive dog, 3.843.
Figure 2. Anti-Malassezia ELISA adjusted optical density (OD) values versus intradermal test (IDT) scores against Malassezia in the same dog. Outliers not shown: OD scores of 2.335 and 3.843 for IDT score of 4. The categorical nature of the IDT results precludes correlation analysis, but it is evident that there is a trend towards increasing OD scores as IDT score increases.
Analysis of anti-Malassezia ELISA versus other serology results Kappa coefficient results were calculated for comparison of anti-Malassezia ELISA results with ELISA results of other allergen groups. There was poor agreement between the anti-Malassezia ELISA and ELISA results for trees (j = 0.183), grasses (j = 0.314), weeds (j = 0.286), moulds (j = 0.155) and mites (j = 0.088).
Discussion In this study, we demonstrated substantial agreement between the results of an anti-Malassezia ELISA and IDT. The results of this study indicate that the anti-Malassezia ELISA described herein might be useful to demonstrate immediate-type hypersensitivity to Malassezia in the dog. © 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
Figure 3. Receiver operating characteristic curve for anti-Malassezia ELISA-based identification of immediate-type Malassezia hypersensitivity. True-positive rate (sensitivity) and false-positive rate (1 specificity) were calculated for all possible cut-off points. The arrow indicates the optimal anti-Malassezia ELISA OD cut-off point. AUC indicates area under the curve.
Agreement between these two tests of immediatetype hypersensitivity was not perfect; in some situations, one test was clearly positive but the other negative. This is not surprising, because historically intradermal test and ELISA results for other allergens do not always agree. Various hypotheses have been proposed to explain such discrepancies, including the possible existence of heterogeneous IgE subclasses and differences in the amounts of IgE in the circulation versus that bound to mast cells. It is also a fact that ELISA, by design, measures only quantities of allergen-specific IgE and not necessarily whether the IgE is able to function in a hypersensitivity reaction (as might be demonstrated by IDT). Our study was not meant to evaluate the clinical utility of these tests of hypersensitivity nor to correlate them with patient clinical signs or cytological findings. However, in evaluating a dog with clinical signs consistent with Malassezia dermatitis, both cytological evaluation of skin impression smears and tests for hypersensitivity may be useful. It is important to understand that clinical signs, cytological findings and IDT or ELISA results are independent elements of disease evaluation and that diagnosis should not rest on any single element. For example, failure to find yeast organisms on cytology does not rule out the possible contribution of yeast to clinical signs. What then, might be the clinical value of a positive IDT or ELISA to Malassezia antigen? As an example, a positive Malassezia hypersensitivity test in a dog might lead the clinician to consider M. pachydermatis overgrowth as a component of that animal’s disease and prompt cytological evaluation if the clinician had neglected to perform it previously. Conversely, a negative Malassezia IgE test does not rule out a pathological role for this organism in an atopic dog. The value of demonstrating immediate-type hypersensitivity to Malassezia yeast may have additional value in 509
Oldenhoff et al.
long-term case management. A positive test raises the possibility that longer-term prophylactic topical antifungal treatment may be beneficial or that an M. pachydermatis extract could be used as part of an ASIT regimen. It must be emphasized, however, that the efficacy of yeast extracts in ASIT has not yet been reported. It is important to remember that dogs can have either immediate-type hypersensitivity or delayed-type hypersensitivity (or both concurrently) to Malassezia and that, in this study, we evaluated only the former type. It is possible that a dog may have negative results on the anti-Malassezia ELISA and/or negative immediate IDT reactivity, but have delayed-type hypersensitivity to Malassezia that could be clinically relevant. The latter could be detected, for example, by patch testing or by observing delayed (24–48 h) reactivity on IDT. In human patients with AD, correlation has been found between skin-prick testing and Malassezia-specific serum IgE, i.e. immediate hypersensitivity can be demonstrated by either method.16 Atopy patch testing has been used to document delayed-type hypersensitivity in patients with low Malassezia-specific IgE levels and negative skin-prick test results.16 To date, there are no published reports investigating patch test results to M. pachydermatis in atopic dogs, although one study found that positive patch test reactivity to yeast was more common in basset hounds with Malassezia dermatitis compared with healthy control dogs.14 Considering the perceived importance of yeast hypersensitivity in some atopic dogs, further studies investigating delayed IDT reactivity and atopy patch test reactivity to M. pachydermatis in atopic dogs are needed. We speculated that dogs positive to Malassezia might be more likely to have IgE against other environmental substances, such as fungi, perhaps even due to fungal antigen cross-reactivity. We therefore compared the results of the anti-Malassezia ELISA with those from the complete allergen-specific IgE panel. Comparison of the anti-Malassezia ELISA and the complete allergen-specific IgE panel indicated that dogs with positive results to Malassezia were somewhat more likely (than expected by chance) to have positive results to trees, grasses, weeds or moulds (j > 0). Nevertheless, the low kappa coefficients observed for these relationships indicated poor agreement. The fact that the agreement was greater than chance may simply reflect the fact that atopic dogs typically sensitize to a range of common allergens, rather than a single one. Statistical comparisons of results of these two tests mandated that we define in advance what constituted a ‘positive’ versus ‘negative’ IDT. This definition is by its nature somewhat arbitrary and subject to debate. In our analysis, we chose to compare the ELISA results with an IDT result that was in our view unequivocal, i.e. 3 or greater on our scale, as this typically prompts inclusion in ASIT in our practice, whereas scores of 2 or less are not included. Instead, we could have chosen to compare the ELISA results with a less strict definition of a positive IDT result, i.e. considering an IDT reaction of 2 also to be positive, though weakly so. Indeed, some authorities advocate defining an IDT of 2 or greater as positive. In making this alternative comparison, one might expect that the agreement between the two tests would be lower with 510
addition of the less certain, ‘weaker’ positive IDT results to the data set. To explore this, we recalculated the statistical analysis, defining a positive IDT as ‘2 or greater’ rather than ‘3 or greater.’ In this case (ROC curve not shown), the kappa statistic was 0.527, the area under the curve 0.76, and the adjusted OD cut-off with the highest kappa was 0.073. At this cut-off and including an IDT of 2 as positive, the sensitivity of the ELISA would then be 63.9% and the specificity 87.5%. Thus, if we reduced the positive threshold of the IDT to 2, as expected, the agreement of the two tests was lower. Another way of stating this is that the ELISA result has better agreement with a ‘strong positive’ IDT than with ‘any positive’ IDT. It is useful to consider the concept of ‘cut-off point’ in diagnostic assays, because the term can mean different things in different situations. Where a disease state is known to be present or absent with certainty, it is possible to calculate sensitivity and specificity of a diagnostic test for that disease and to set up a ‘cut-off point’ for the test based on evaluation of a large number of known negative and positive samples. The ‘cut-off point’ or positive threshold is often set at two-to-three standard deviations above the mean of the negative sample group.31 Samples above this ‘cut-off point’ are deemed positive and therefore diagnostic of the disease. In contrast, when comparing two different diagnostic tests with each other, in some cases the disease state may not be known with certainty. In this situation, for analytical purposes one test can be defined as the ‘gold standard’. The second test can then be evaluated for its ability to reach the same conclusion as the ‘gold standard’. In this situation, a ‘cut-off point’ is determined by ROC analysis and describes that optimal point at which the new test mirrors the results of the ‘gold standard’. 32 In the present study, it is important to understand that our calculated ‘cut-off point’ refers only to this agreement between tests, and not to the diagnostic efficiency of either test. Defining assay ‘cut-off values’ for the purpose of diagnosis would require further assay of very large collections of healthy dog sera and was not the aim of the present study. In brief, our analyses indicate that ELISA and IDT are conveying the same information, without specifying that either test is valid for diagnosis. This problem is a continuing plague in the use of allergen tests of all types; because we do not know ‘the truth’ of whether a patient is allergic to a substance without provocative testing, we are forced to define ‘gold standards’ (such as IDT) that may be of uncertain diagnostic efficiency. In summary, in this study we demonstrate that an antiMalassezia ELISA using the FceRIa receptor detection reagent can be used to demonstrate immediate-type hypersensitivity to Malassezia yeast in the dog, with results that agree substantially with those of IDT. This finding may have particular value for those veterinary practitioners who primarily use serological testing in the evaluation of a dog with allergic skin disease.
Acknowledgements The authors thank Nicholas S. Keuler (Statistical Consulting Service, College of Agricultural & Life Sciences, Uni© 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
Comparison of tests for Malassezia hypersensitivity
versity of Wisconsin-Madison) for assistance in statistical analysis.
References 1. Plant J, Rosenkrantz W, Griffin C. Factors associated with and prevalence of high Malassezia pachydermatis numbers on dog skin. J Am Vet Med Assoc 1992; 201: 879–882. 2. Mason IS, Mason KV, Lloyd DH. A review of the biology of canine skin with respect to the commensals Staphylococcus intermedius, Demodex canis and Malassezia pachydermatis. Vet Dermatol 1996; 7: 119–132. 3. Chen TA, Hill PB. The biology of Malassezia organisms and their ability to induce immune responses and skin disease. Vet Dermatol 2005; 16: 4–26. 4. Kennis R, Rosser E Jr, Olivier N et al. Quantity and distribution of Malassezia organisms on the skin of clinically normal dogs. J Am Vet Med Assoc 1996; 208: 1048–1051. 5. Nardoni S, Mancianti F, Corazza M et al. Occurrence of Malassezia species in healthy and dermatologically diseased dogs. Mycopathologia 2004; 157: 383–388. 6. Nardoni S, Dini M, Taccini F et al. Occurrence, distribution and population size of Malassezia pachydermatis on skin and mucosae of atopic dogs. Vet Microbiol 2007; 122: 172–177. 7. Cafarchia C, Gallo S, Romito D et al. Frequency, body distribution, and population size of Malassezia species in healthy dogs and in dogs with localized cutaneous lesions. J Vet Diagn Invest 2005; 17: 316–322. 8. Bond R, Ferguson E, Curtis C et al. Factors associated with elevated cutaneous Malassezia pachydermatis populations in dogs with pruritic skin disease. J Small Anim Pract 1996; 37: 103– 107. 9. Morris D, Olivier N, Rosser E. Type-1 hypersensitivity reactions to Malassezia pachydermatis extracts in atopic dogs. Am J Vet Res 1998; 59: 836–841. 10. Morris DO, DeBoer DJ. Evaluation of serum obtained from atopic dogs with dermatitis attributable to Malassezia pachydermatis for passive transfer of immediate hypersensitivity to that organism. Am J Vet Res 2003; 64: 262–266. 11. Morris DO, Clayton DJ, Drobatz KJ et al. Response to Malassezia pachydermatis by peripheral blood mononuclear cells from clinically normal and atopic dogs. Am J Vet Res 2002; 63: 358–362. 12. Nuttall TJ, Halliwell RE. Serum antibodies to Malassezia yeasts in canine atopic dermatitis. Vet Dermatol 2001; 12: 327–332. 13. Bond R, Hendricks A, Ferguson E et al. Intradermal test reactivity to Malassezia pachydermatis in atopic dogs. Vet Rec 2002; 150: 448–449. 14. Bond R, Patterson-Kane J, Perrins N et al. Patch test responses to Malassezia pachydermatis in healthy basset hounds and in basset hounds with Malassezia dermatitis. Med Mycol 2006; 44: 419–427.
15. Farver K, Morris DO, Shofer F et al. Humoral measurement of type-1 hypersensitivity reactions to a commercial Malassezia allergen. Vet Dermatol 2005; 16: 261–268. €m M, Bartosik J et al. Atopy patch test 16. Johansson C, Sandstro reactions to Malassezia allergens differentiate subgroups of atopic dermatitis patients. Br J Dermatol 2003; 148: 479–488. 17. Sonesson A, Bartosik J, Christiansen J et al. Sensitization to skin-associated microorganisms in adult patients with atopic dermatitis is of importance for disease severity. Acta Derm Venereol 2013; 93: 340–346. 18. Darabi K, Hostetler SG, Bechtel MA et al. The role of Malassezia in atopic dermatitis affecting the head and neck of adults. J Am Acad Dermatol 2009; 60: 125–136. 19. Wessels M, Dolkes G, Dijk A et al. IgE antibodies to Pityrosporum ovale in atopic dermatitis. Br J Dermatol 1991; 125: 227– 232. 20. Bayrou O, Pecquet C, Flahault A et al. Head and neck atopic dermatitis and Malassezia furfur-specific IgE antibodies. Dermatology 2005; 211: 107–113. 21. B€ ack O, Bartosik J. Systemic ketoconazole for yeast allergic patients with atopic dermatitis. J Eur Acad Dermatol Venereol 2001; 15: 34–38. 22. Svejgaard E, Ølholm Larsen P, Deleuran M et al. Treatment of head and neck dermatitis comparing itraconazole 200 mg and 400 mg daily for 1 week with placebo. J Eur Acad Dermatol Venereol 2004; 18: 445–449. 23. Ikezawa Z, Kondo M, Okajima M et al. Clinical usefulness of oral itraconazole, an antimycotic drug, for refractory atopic dermatitis. Eur J Dermatol 2004; 14: 400–406. 24. B€ ack O, Scheynius A, Johansson S. Ketoconazole in atopic dermatitis: therapeutic response is correlated with decrease in serum IgE. Arch Dermatol Res 1995; 287: 448–451. laud P, Guague re E, Alhaidari Z et al. Reevaluation of diag25. Pre nostic criteria of canine atopic dermatitis. Rev Vet Med 1998; 149: 1057–1064. 26. Favrot C, Steffan J, Seewald W et al. A prospective study on the clinical features of chronic canine atopic dermatitis and its diagnosis. Vet Dermatol 2010; 21: 23–31. 27. Stedman K, Lee K, Hunter S et al. Measurement of canine IgE using the alpha chain of the human high affinity IgE receptor. Vet Immunol Immunopathol 2001; 78: 349–355. 28. Akobeng AK. Understanding diagnostic tests 3: receiver operating characteristic curves. Acta Paediatr 2007; 96: 644–647. 29. Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. Fam Med 2005; 37: 360–363. 30. Sing T, Sander O, Beerenwinkel N et al. ROCR: visualizing classifier performance in R. Bioinformatics 2005; 21: 3940–3941. 31. Malvano R, Boniolo A, Dovis M et al. ELISA for antibody measurement: aspects related to data expression. J Immunol Methods 1982; 48: 51–60. 32. Bewick V, Cheek L, Ball J. Statistics review 13: receiver operating characteristic curves. Crit Care 2004; 8: 508–512.
sume Re Contexte – Malassezia pachydermatis fait partie de la flore normale de la peau du chien. L’hypersensibilite a Malassezia est reconnue comme un facteur aggravant des signes cliniques de dermatite atopique (AD) termination de l’hypersensibilite chez certains chiens. La de a Malassezia est souvent faite par tests intra te rinaire ge ne rale. dermiques (IDT) qui sont peu disponibles en pratique ve ses/Objectifs – Le but de cette e tude e tait de comparer la re activite imme diate des IDT Hypothe a sultats d’un test ELISA (enzyme-linked immunosorbent assay) concßu pour M. pachydermatis avce les re tecter les IgE anti-Malassezia. de Sujets – Quatre vingt quatre chiens atopiques. thodes – Des IDT de plusieurs allerge nes ont e te re alise es sur tous les chiens. Des tests se riques pour Me cifiques d’allerge nes contre un panel d’allerge nes environnementaux fre quents et M. pachyderles IgE spe te re alise s par ELISA a l’aide d’un fragment de re cepteur FcƐRIa, les re sultats e tant rapporte s matis ont e optique ajuste e (OD). La courbe ROC (Receiver Operating Characteristic) a e te utilise e pour par densite sultats des deux tests. analyser les re © 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
511
Oldenhoff et al.
sultats – L’OD ajuste e moyenne des tests ELISA des IgE anti-Malassezia des chiens re actifs et non Re actifs a M. pachydermatis aux IDT e tait respectivement de 0.137 et 0.024. L’analyse de la courbe ROC re re un point de rupture pour l’ELISA anti-Malassezia qui produit une sensibilite de 77.0% et une sugge cificite de 89% relative aux re sultats d’IDT. spe lation substantielle a e te de montre e entre la re activite Conclusions et importance clinique – Une corre tecte s par le re actif de re cepteurs FcƐRIa. Bien que la corre lation avec des IDT et des IgE anti-Malassezia de te re alise e dans cette e tude, les re sultats indiun diagnostic clinique de dermatite a Malassezia n’ait pas e ^tre utilise pour de montrer la pre sence d’une hypersensibilite imme diate quent que le test ELISA peut e a Malassezia chez les chiens atopiques. Resumen n – Malassezia paquidermatis es parte de la flora normal de la piel canina. La hipersensibilidad Introduccio pica (AD) en alga Malassezia se reconoce como un factor iniciador de los signos clınicos de dermatitis ato n de la hipersensibilidad a Malassezia se hace mediante prueba intradermal unos perros. La determinacio n veterinaria. (ITT), que puede ser de acceso limitado en las pr acticas de primera opinio tesis/Objetivos – El propo sito de este estudio fue comparar la reactividad intrade rmica inmediata a Hipo M. pachidermatis con los resultados de un ensayo inmunoenzim atico designado para detectar la IgE frente a Malassezia. stico clınico de dermatitis ato pica. Animales – 84 perros con un diagno todos – Se desarrollo una prueba intrade rmica multi alerge nica en todos los perros. Las pruebas se ricas Me rgeno frente a un panel de alergenos ambientales comunes y frente a Malapara las IgE especıficas de ale mediante ELISA utilizando el fragmento del receptor FcƐRIa como el ressezia paquidermatis se desarrollo n. Los resultados se indicaron como densidad o ptica ajustada (OD). Una curva de activo de la deteccio para analizar los resultados de las dos pruebas. caracterıstica del operador receptor (ROC) se utilizo ptica ajustada de la prueba de ELISA para IgE frente a Malassezia Resultados – La mediana de densidad o en los perros reactivos y no reactivos frente a Malassezia pachidermatis tras la propuesta intradermal der un punto de corte para mal fue de 0,137 y 0,024, respectivamente. En el an alisis de la curva ROC se sugirio el ELISA frente a Malassezia que tiene una sensibilidad del 77% y una especificidad del 89% relativa a los rmica. estados la prueba intrade Conclusiones e importancia clınica – Hay un una concordancia sustancial demostrada entre la reactivirmica y la prueba de IgE frente Malassezia detectada mediante el uso del receptor FcƐRIa aundad intrade n con el diagno stico clınico de dermatitis por Malassezia no fue intentada en este estudio, que la correlacio los resultados indican que este ELISA puede utilizarse para demostrar la presencia de la hipersensibilidad pica inmediata a Malassezia en perros con dermatitis ato Zusammenfassung Hintergrund – Malassezia pachydermatis ist Teil der normalen Flora der Haut des Hundes. Eine Hypersen€ser fu €r die klinische Symptomatik der atopischen Dermatitis (AD) bei sibilit€at auf Malassezien ist als Auslo einigen Hunden bekannt. Eine Malassezien Hypersensibilit€ at wird oft mittels Intradermaltest (IDT) diagnos€glicherwiese nur limitiert zur Verfu €gung steht. tiziert, was in einer allgemeinen Praxis mo Hypothese/Ziele – Das Ziel dieser Studie war ein Vergleich der IDT Reaktivit€ at auf M. pachydermatis mit den Ergebnissen eines ELISA (Enzyme-linked Immunosorbent Assay), der zur Bestimmung von anti-Malassezien IgE entwickelt wurde. Tiere – Vierundachtzig Hunde mit der klinischen Diagnose von AD. €hrt. Ein serologischer Test zur AllerMethoden – Ein Multi-Allergen IDT wurde bei allen Hunden durchgefu gen-spezifischen IgE Bestimmung auf ein Panel von Umweltallergenen und auf M. pachydermatis wurde €hrt, wobei das FcƐRIa Rezeptorfragment als Reagens verwendet wurde. Die Ermittels ELISA durchgefu gebnisse wurden als „angepasste” optische Dichte (OD) angegeben. Eine ROC Kurve (Receiver Operating Characteristic Curve) wurde verwendet, um die Ergebnisse der beiden Tests zu analysieren. €r Hunde, die auf M. pachyderErgebnisse – Die Median angepasste OD des anti-Malassezien IgE ELISA fu matis im IDT reagierten bzw. nicht reagierten betrug 0,137 bzw. 0,024. Eine Analyse der ROC Kurve ergab €r den anti-Malassezien ELISA einen Cut Off, der in Relation zu den IDT Ergebnissen eine Sensibilit€ fu at von 77% und eine Spezifizit€at von 89% ergab. € Schlussfolgerungen und klinische Bedeutung – Es wurde eine deutliche Ubereinstimmung zwischen der IDT Reaktivit€at und dem anti-Malassezien IgE, welches mittels FcƐRIa Rezeptorreagens bestimmt wurde, festgestellt. Obwohl eine Korrelation mit der klinischen Diagnostik nicht Teil dieser Studie war, wei€glicherweise verwendet werden kann, um das Vorliegen sen die Ergebnisse darauf hin, dass der ELISA mo einer Malassezien Hypersensibilita€t vom Soforttyp bei Hunden mit AD festzustellen.
e84
© 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
Comparison of tests for Malassezia hypersensitivity
© 2014 ESVD and ACVD, Veterinary Dermatology, 25, 507–e85.
e85
