Contact allergy to common ingredients in hair dyes

Background: p‐Phenylenediamine (PPD) is the primary patch test screening agent for hair dye contact allergy, and approximately 100 different hair dye chemicals are allowed. Objectives: To examine whether PPD is an optimal screening agent for diagnosing hair dye allergy or whether other clinically important sensitizers exist. Methods: Two thousand nine hundred and thirty‐nine consecutive patients in 12 dermatology clinics were patch tested with five hair dyes available from patch test suppliers. Furthermore, 22 frequently used hair dye ingredients not available from patch test suppliers were tested in subgroups of ∼ 500 patients each. Results: A positive reaction to PPD was found in 4.5% of patients, and 2.8% reacted to toluene‐2,5‐diamine (PTD), 1.8% to p‐aminophenol, 1% to m‐aminophenol, and 0.1% to resorcinol; all together, 5.3% (n = 156). Dying hair was the most frequently reported cause of the allergy (55.4%); so‐called ‘temporary henna’ tattoos were the cause in 8.5% of the cases. p‐Methylaminophenol gave a reaction in 20 patients (2.2%), 3 of them with clinical relevance, and no co‐reaction with the above five well‐known hair dyes. Conclusions: Hair dyes are the prime cause of PPD allergy. PPD identifies the majority of positive reactions to PTD, p‐aminophenol and m‐aminophenol, but not all, which justifies additional testing with hair dye ingredients from the used product.

2‐amino‐3‐hydroxypyridine; 3‐nitro‐p‐hydroxyethylaminophenol; hair dye allergy; m‐aminophenol; p‐aminophenol; p‐methylaminophenol; PPD; p‐phenylenediamine; p‐toluenediamine; toluene‐2; 5‐diamine

Hair dyes may cause allergic reactions [1] , [2] . For many years, the focus in this context has been on p‐phenylenediamine (PPD), as it is a hair dye ingredient and an extreme sensitizer, and is included in the European baseline series [3] , [4] , [5] , [6] , [7] , [8] . Currently, ∼ 100 ingredients/chemicals are permitted for use as hair dyes in Europe after review by the Scientific Committee on Consumer Safety; a number are extreme to strong sensitizers [9] , [10] , [11] .

A combination of local lymph node assay (LLNA) data and physical/chemical properties have previously been used in a quantitative structure–activity relationship (QSAR) analysis to predict sensitization potential for > 200 hair dye chemicals. The details of the analysis were described by Søsted et al. in 2004 [12] . Through QSAR analysis and a literature review of > 100 hair dye ingredients, PPD alone was challenged as an optimal screening agent for hair dye allergy. There are five substances in the European hairdressers series – PPD, toluene‐2,5‐diamine (PTD), resorcinol, m‐aminophenol, and p‐aminophenol – but it was suggested in Søsted's QSAR paper that other clinically important sensitizers may exist [12] . At the time of the former QSAR study, some 229 hair dye chemicals were available to industry, and these hair dyes were divided into clusters on the basis of their physical/chemical properties. It was expected that chemicals from the same clusters would show cross‐reactions in the patients. Substances that were used in amounts of > 2 tonnes per year in Europe and with a predicted potency of being moderate to strong sensitizers were selected. This resulted in a list of 22 hair dye ingredients that were often used in hair dye products but rarely patch tested in patients, and five hair dye substances from the European hairdressers series.

The aim of this multicentre evaluation was to identify substances suitable for inclusion in an extended patch test series for contact allergy to hair dye ingredients. The first aim was to study the frequency of allergic reactions to 22 not previously tested potential allergenic hair dye ingredients in eczema populations. The second aim was to study the relationship between reactions to one or more of the hair dye ingredients and a history of hair dye dermatitis. The third aim was to study the relationship between reactions to one or more of 22 rarely tested hair dye substances and a series of five well‐known hair dye screening agents.

Two thousand nine hundred and thirty‐nine consecutive eczema patients attending 12 dermatology clinics during a 6‐month period (in 2007/2008) were patch tested with an extended series of hair dye ingredients, consisting of five basic hair dye ingredients from the European baseline and hairdressers series: PPD, PTD, resorcinol, m‐aminophenol, and p‐aminophenol. Moreover, 22 hair dye substances identified as potential contact allergens on the basis of their chemical structure and the volume of use in Europe [12] were patch tested. The exact volumes used in hair dyes in Europe and CAS numbers are described in the previous study [12] . For all of the substances used in this study, see Table [NaN] . All 12 departments tested the five basic hair dye substances, and the departments were split into four groups, with each group of three departments each receiving five or six rarely tested hair dye substances from different clusters according to the QSAR analysis results. Substances from the same cluster were suspected to give cross‐reactions [12] . The five basic substances were supplied (same batch) by Trolab® (Almirall Hermal Gmbh, Reinbek, Germany), and the raw materials for the 22 additional substances were supplied by courtesy of different hair dye manufacturers together with certificates of purity (range 95.5–99.9%).

Baseline characteristics of the cohort of eczema patients

1 AD, atopic dermatitis; Face, face eczema; Hand, hand eczema; Leg, leg eczema; NG, not given; SF, San Francisco.

• 2 Of 2864 patients.
• 3 Of 1950 patients.

Results from patch testing of 27 hair dye ingredients in 2939 consecutive patients

• 4 PPD, p‐phenylenediamine; PTD, toluene‐2,5‐diamine. A: chemical cluster is an arbitrary number that refers to physical/chemical properties. Chemicals with the same cluster number are closely related concerning their physical/chemical properties. It is expected that substances from the same clusters will give patch test cross‐reactions in patients.
• 5 Odense numbers: 352 of 366 (96.2%) of the doubtful reactions were registered at one clinic, where discoloration of the skin was taken as doubtful.
• 6 Six of these 20 patients were also positive for PTD, and 5 of these 20 patients were also positive for PPD; see Table .

The patch test materials were prepared by the pharmacy of the Capital Region of Copenhagen, Denmark.

The test concentrations were 1% in petrolatum for all of the 22 substances, based on knowledge of use concentrations from the producers and experience of testing in one centre, which had patch tested individual hair dye cases with ingredients in patients using hair dye products. The characteristics of the participants in this study are shown in Table [NaN] , and Table [NaN] shows where each substance was patch tested.

The extended series were applied on the upper back of the patients with 8‐mm Finn Chambers® (Epitest Ltd Oy, Tuusla, Finland) on Scanpor® tape (Alpharma AS, Norway), using 20 mg of each test material [13] , in all centres. Readings were performed on D2, D3/4, and D7. Reactions were scored according to the scale of the International Contact Dermatitis Research Group [14] .

If a patient had had a doubtful or positive patch test reaction to any of the hair dye substances, they were asked if they had ever had symptoms from hair dyes and the nature of the symptoms. It was possible to choose: itch only, facial/scalp rash, spreading rash, facial oedema, and/or hand eczema. It was not specified whether the symptoms were immediate or delayed.

In 4.5% of patients a positive reaction to PPD was found, in 2.8% a positive reaction to PTD, in 1.8% a positive reaction to p‐aminophenol, in 1% a positive reaction to m‐aminophenol, and in 0.1% a positive reaction to resorcinol (Table [NaN] ). In total, 5.3% (156/2939) reacted positively to one or more of the five basic substances (Table [NaN] ).

Concomitant patch test reactions among five basic hair dye ingredients in 156 patients

• 7 PPD, p‐phenylenediamine; PTD, toluene‐2,5‐diamine.
• 8 The five basic hair dye ingredients were: PPD; PTD; m‐aminophenol; p‐aminophenol; and resorcinol.
• 9 One or more positive patch test reactions in 156 patients. N = 2939.

PPD was the substance with most single positive reactions [60/133 (45.1%)], and identified 80.7–86.8% of the positive reactions to PTD, m‐aminophenol, or p‐aminophenol. Concerning resorcinol, 3 positive reactions were seen, in 2 of 3 cases with a concomitant positive reaction to PPD.

In 55.4% of the cases positive for one or more of the five substances, there was a history of rash caused by hair dyes, eyebrow dye, or semi‐permanent tattoo, more often in women (59.8%) than in men (36.4%) (Table [NaN] ).

Characteristics of 155 patients with a positive patch test reaction to one or more of the five basic hair dye substances

10 Five basic hair dye substances: p‐Phenylenediamine; toluene‐2,5‐diamine; resorcinol; m‐aminophenol and p‐aminophenol.

Facial rash or scalp rash caused by hair dyes constituted by far the largest group of adverse reactions (Table [NaN] ). Facial/scalp rash and facial oedema were the most prominent symptoms among those having experienced adverse reactions to hair dyes.

In total, 22 women and no men had worked as hairdressers, and 19 of these had current hand eczema. In 2 hairdressers, facial or scalp eczema was seen, and in 1 case there was no current dermatitis.

Among the additional substances that were tested in subgroups of patients, seven gave reactions in more than 1 patient (Tables [NaN] and [NaN] ). p‐Methylaminophenol gave a positive reaction in 20 patients (2.2%), and 4‐amino‐2‐hydroxytoluene gave a positive reaction in 6 patients (1.8%).

Concomitant positive reactions to one or more of the five basic hair dye allergens and each of the seven additional hair dye substances causing more than isolated positive patch test reactions

• 11 PPD, p‐phenylenediamine; PTD, toluene‐2,5‐diamine.
• 12 The five allergens are: PPD; PTD; m‐aminophenol; p‐aminophenol; and resorcinol.
• 13 Not all reactions to 4‐amino‐2‐hydroxytoluene, 2‐amino‐3‐hydroxypyridine, 3‐nitro‐p‐hydroxyamoniphenol and p‐methylaminophenol were identified by the five traditional commercially available hair dye patch test allergens.

In the case of 4‐amino‐2‐hydroxytoluene, 5 of the 6 positive patients had concomitant positive reactions to PPD, PTD, m‐aminophenol, and p‐aminophenol. In patients reacting positively to p‐methylaminophenol, 8 of 20 (40%) had concomitant positive reactions to one or more of the five basic substances, and among the 12 other patients negative for the five basic substances, 3 had presented a rash caused by a hair dye only, and not one caused by any other consumer product (25%). No brand names were given. In 2 of these cases, the physician had considered that the reaction to p‐methylaminophenol explained the current dermatitis.

3‐Nitro‐p‐hydroxyethylaminophenol was positive in 2 patients who gave negative reactions to all of the five basic substances. None of these gave a history of rash caused by hair dye, semi‐permanent tattoo, or eyebrow colour.

The patients with positive reactions to 2,4‐diaminophenoxyethanol‐HCl, 4‐amino‐m‐cresol or N,N′‐bis(2‐hydroxyethyl)‐p‐phenylenediamine were also positive for PPD in all cases.

None of the hairdressers showed positive reactions to any of the additional substances.

In 43 patients, a particular product with a brand name or producer was mentioned as the cause of the allergic reaction. In 2 cases, these were black henna/Turkish henna; the remainder were hair dyes most often produced by l'Oreal (12 cases), Wella (5 cases), Schwarzkopf (3 cases), and Clairol (4 cases). In all cases, the reactions concerned PPD, PTD, resorcinol, or the aminophenols, and these chemicals were labelled on the product.

Eleven of the 27 tested hair dye substances gave no positive patch test reactions in the patients (Table [NaN] ).

Contact allergy to PPD was a very frequent cause of contact allergy in European eczema patients in the present study, which is in line with previous experience over a century [15] .

In the current study, 4.5% of a cohort of almost 3000 consecutive eczema patients from different countries in Europe gave a positive patch test reaction to PPD 1% pet. The frequency of PPD allergy is compatible with previous retrospective studies in 21 515 patients tested between 2003 and 2007 in 10 European centres (4.6% reacted to PPD) [16] , and in 83 030 patch tested German patients (4% positive to PPD) [15] , whereas 7.1% of English patients reacted to PPD [4] .

Other hair dye ingredients, such as PTD, p‐aminophenol, m‐aminophenol, and resorcinol, are also well‐known hair dye allergens [12] , [15] , [18] , [19] , [20] , [21] , [22] , [23] , but have been much less studied.

In our study, PTD came second after PPD with 2.8%, followed by p‐aminophenol, m‐aminophenol, and resorcinol (Table [NaN] ). PTD is found more often than PPD in permanent hair dyes in Sweden [24] and Denmark [21] , but the reverse is true in Spain [25] . PPD and PTD have very similar chemical structures, so cross‐reactivity is expected, and concomitant reactions to PPD in those sensitized to PTD were found in 80.7% of cases (Table [NaN] ), similarly to what has been seen in previous investigations [23] , [26] . Also, a high degree of concordance with PPD allergy was seen in those with positive patch test reactions to p‐aminophenol (86.8%) or m‐aminophenol (82.7%). Resorcinol gave only 3 positive reactions in total (0.1%), even though it has been detected in ∼ 80% of permanent hair dyes [24] , [25] , probably because of its function as a coupler. Resorcinol has been predicted to be a moderately potent sensitizer by the LLNA [12] , but this is apparently not in accordance with experience in humans [27] . It is still recommended to test with resorcinol, even though it gives relatively few reactions as compared with, for example, PPD; it still detects some cases, it is often present in hair dyes, and can also be present in, for example, eye drops, so it is still relevant to test with it.

In total, 72% of 155 patients reacting to one or more of the five basic hair dye ingredients from the hairdressing patch test series had ever dyed their hair; this applied to more women (82.8%) than men (33.3%) (Table [NaN] ). Among those who had ever dyed their hair, 71.4% had experienced symptoms related to hair dyeing, most frequently facial/scalp rash. In total, 27.5% had experienced more severe reactions in the form of facial swelling, and 20% had had a more generalized rash. Clinical manifestations in PPD patch test‐positive subjects who reacted to hair dyes have been described previously; 97% reported eczema of the scalp or neighbouring skin, 29% oedema of the eyelids, forehead, or neck, and 12% whole facial oedema [28] .

As shown in Table 4, a total of 14.8% had received a semi‐permanent tattoo, and 56% (13/23) of these had experienced a rash in the tattoo, suggesting the tattoo as the cause of contact allergy: this was equal to 8.3% (13/155) of those sensitized to one or more of the five basic hair dye patch test allergens. This is in line with the retrospective study of PPD allergy in Europe [15] , and underlines the fact that henna tattoos (semi‐permanent tattoos) are not the main culprits in PPD sensitization, but that they are not negligible and have to be avoided.

Hairdressers constituted 14% (22/156) of those with contact allergy to one or more of the five basic hair dye ingredients, and the majority had hand eczema of a probable occupational origin. No current or previous hairdressers reacted to any of the additional substances, which may indicate that the exposure to PPD is of sufficient magnitude to make it a prime cause of allergy in hairdressers [29] , and may force them to leave the trade [30] .

The selection of 22 other hair dye ingredients with different chemical composition (clusters) [12] was tested in a subset of patients. Eleven of the 22 substances were negative, four gave only an isolated reaction, and seven gave between 2 and 20 positive reactions; the prevalence was therefore between 0.2% and 2.2% (Table [NaN] ). Four substances had an almost total overlap with PPD‐positive reactions: 4‐amino‐2‐hydroxytoluene; 2,4‐diaminophenoxyethanol‐HCl; 4‐amino‐m‐cresol; and N,N′‐bis(2‐hydroxyethyl)‐p‐phenylenediamine; this means that the clinical relevance is indistinguishable from that of PPD, and the reactions may represent co‐reactivity or cross‐reactivity. Co‐reactivity of 4‐amino‐2‐hydroxytoluene with PPD has been described before in a patient with allergic contact dermatitis caused by a semi‐permanent hair dye [18] ; 70 tonnes of the substance were used in 2002, making it the fourth most used hair dye [12] .

For three substances – 2‐amino‐3‐hydroxypyridine (n = 2), 3‐nitro‐p‐hydroxyethylaminophenol (n = 2), and p‐methylaminophenol (n = 20) – no or very little overlap with PPD was seen. This is what we expected, owing to their different cluster numbers (Table [NaN] ). PPD is a cluster 8 substance, whereas 2‐amino‐3‐hydroxypyridine and 3‐nitro‐p‐hydroxyethylaminophenol are cluster 5 substances. We would have expected a cross‐reaction with m‐aminophenol or p‐aminophenol, which are also in cluster 5. A different cluster number is a sign of different physical/chemical properties, and less cross‐reactivity was therefore expected [12] , but this seems not to be the cases in this investigation. Contact allergy to 3‐nitro‐p‐hydroxyethylaminophenol was described previously in 2 consumers who had dyed their hair with a non‐permanent hair dye [22] , [31] ; in neither case was there a co‐reaction with PPD [31] .

The clinical relevance of the positive reactions among those not reacting to one or more of the five basic hair dye ingredients was 25% (3/12), and was thus less than for PPD. These results show that it is important to test patients with suspected allergy to hair dye with the ingredients in the products, to detect patients not allergic to PPD and to correctly inform patients about possible future problems and avoidance. In this study, none of the patients who only had reactions to 3‐nitro‐p‐hydroxyethylaminophenol gave a history of rash caused by hair dye, semi‐permanent tattoo, or eyebrow colour.

p‐Methylaminophenol was a frequent cause of reactions; based on experience from routine testing in one centre with 1% in pet., it was not expected to be irritant, and only a few irritant reactions were scored in this study, but further assessment of the optimal patch test concentration may be necessary.

The clinical data presented in this study show that we still have much to learn about cross‐reactions between the hair dye substances. The cross‐reaction hypothesis that chemicals with similar physical/chemical properties give cross‐reactions in humans [12] seems to be true sometimes, but not in all cases. Table [NaN] is a good starting point for testing hair dye allergy, but there is still work to be done concerning optimal patch test concentrations and knowledge about cross‐reactions.

A final patch test series for hair dyes cannot be based on the data from this investigation. A study is needed in which the hair dye substances are patch tested in different concentrations. Zero reactions of some substances might be false‐negative results, because of concentrations that are too low. Furthermore, the study is relatively small, as the 22 additional hair dye ingredients were only tested in three different centres. Therefore, on the basis of the current results, it is still recommended to test with those substances that, in this study, showed zero reactions.

In conclusion, PPD allergy is common among European eczema patients, and both the current and previous studies have shown that hair dyes constitute the prime cause of PPD allergy. PPD identified the majority of positive reactions to PTD, p‐aminophenol, and m‐aminophenol, but not all, which justifies continued screening with these basic ingredients of hair dyes. The testing with additional hair dye ingredients shows that it is important to test with the ingredients of the particular hair dye in question, to diagnose hair dye allergy and to correctly inform patients about possible future problems and avoidance.

1 Søsted H, Agner T, Andersen K E, Menné T. 55 cases of allergic reactions to hair dye: a descriptive, consumer complaint‐based study. Contact Dermatitis 2002 : 47 : 299 – 303.

• 2 Søsted H, Johansen J D, Andersen K E, Menné T. Severe allergic hair dye reactions in 8 children. Contact Dermatitis 2006 : 54 : 87 – 91.
• 3 Armstrong D K, Jones A B, Smith H R, Ross J S, White I R, Rycroft R J, McFadden J P. Occupational sensitization to p‐phenylenediamine: a 17‐year review. Contact Dermatitis 1999 : 41 : 348 – 349.
• 4 Patel S, Basketter D A, Jefferies D, White I R, Rycroft R J, McFadden J P, Ho S Y. Patch test frequency to p‐phenylenediamine: follow up over the last 6 years. Contact Dermatitis 2007 : 56 : 35 – 37.
• 5 Sertoli A, Francalanci S, Acciai M C, Gola M. Epidemiological survey of contact dermatitis in Italy (1984–1993) by GIRDCA (Gruppo Italiano Ricerca Dermatiti da Contatto e Ambientali). Am J Contact Dermatitis 1999 : 10 : 18 – 30.
• 6 Thyssen J P, White J M. Epidemiological data on consumer allergy to p‐phenylenediamine. Contact Dermatitis 2008 : 59 : 327 – 343.
• 7 Uter W, Lessmann H, Geier J, Schnuch A. Contact allergy to ingredients of hair cosmetics in female hairdressers and clients – an 8‐year analysis of IVDK data. Contact Dermatitis 2003 : 49 : 236 – 240.
• 8 Bruze M, Andersen K E, Goossens A. Recommendation to include fragrance mix 2 and hydroxyisohexyl 3‐cyclohexene carboxaldehyde (Lyral) in the European baseline patch test series. Contact Dermatitis 2008 : 58 : 129 – 133.
• 9 European Commission, Scientific Committee on Consumer Products. Opinion of p‐phenylenediamine, 2006.
• 10 European Commission, Scientific Committee on Consumer Products. Memorandum on hair dye substances and their skin sensitising properties. 2006.
• 11 European Commission, Scientific Committee on Consumer Products. Opinion on toluene‐2,5‐diamine. 2007 October 2.
• 12 Søsted H, Basketter D A, Estrada E, Johansen J D, Patlewicz G Y. Ranking of hair dye substances according to predicted sensitization potency – quantitative structure–activity relationships. Contact Dermatitis 2004 : 51 : 241 – 254.
• 13 Bruze M, Isaksson M, Gruvberger B, Frick‐Engfeldt M. Recommendation of appropriate amounts of petrolatum preparation to be applied at patch testing. Contact Dermatitis 2007 : 56 : 281 – 285.
• 14 Wilkinson D S, Fregert S, Magnusson B et al. Terminology of contact dermatitis. Acta Derm Venereol 1970 : 50 : 287 – 292.
• 15 Schnuch A, Lessmann H, Frosch P J, Uter W. para‐Phenylenediamine: the profile of an important allergen. Results of the IVDK. Br J Dermatol 2008 : 159 : 379 – 386.
• 16 Thyssen J P, Andersen K E, Bruze M et al. p ‐Phenylenediamine sensitization is more prevalent in central and southern European patch test centres than in Scandinavian: results from a multicentre study. Contact Dermatitis 2009 : 60 : 314 – 319.
• 17 Cathelineau H. Note sur 18 cases d'accidents provoques par teinture pour cheveaux a base de chlorhydrate de PPD. Bull Soc Franc Syph 1898 : 9 : 28 – 35.
• 18 Ellis R A, Wilkinson S M. Contact dermatitis to 4‐amino‐2‐hydroxytoluene in hair dye. Contact Dermatitis 2009 : 60 : 118 – 119.
• 19 Frosch P J, Kugler K, Geier J. Patch testing with hydroxyethyl‐p‐phenylenediamine sulfate – cross‐reactivity with p‐phenylenediamine. Contact Dermatitis 2011 : 65 : 96 – 100.
• 20 Laberge L, Pratt M, Fong B, Gavigan G. A 10‐year review of p‐phenylenediamine allergy and related para‐amino compounds at the Ottawa Patch Test Clinic. Dermatitis 2011 : 22 : 332 – 334.
• 21 Søsted H, Rastogi S C, Andersen K E, Johansen J D, Menné T. Hair dye contact allergy: quantitative exposure assessment of selected products and clinical cases. Contact Dermatitis 2004 : 50 : 344 – 348.
• 22 Søsted H, Menné T. Allergy to 3‐nitro‐p‐hydroxyethylaminophenol and 4‐amino‐3‐nitrophenol in a hair dye. Contact Dermatitis 2005 : 52 : 317 – 319.
• 23 Winhoven S M, Rutter K J, Beck M H. Toluene‐2,5‐diamine may be an isolated allergy in individuals sensitized by permanent hair dye. Contact Dermatitis 2007 : 57 : 193.
• 24 Yazar K, Boman A, Lidén C. Potent skin sensitizers in oxidative hair dye products on the Swedish market. Contact Dermatitis 2009 : 61 : 269 – 275.
• 25 Yazar K, Boman A, Lidén C. P‐phenylenediamine and other hair dye sensitizers in Spain. Contact Dermatitis 2011 : 66 : 27 – 32.
• 26 Basketter D A, English J. Cross‐reactions among hair dye allergens. Cutan Ocul Toxicol 2009 : 28 : 104 – 106.
• 27 Broeckx W, Blondeel A, Dooms‐Goossens A, Achten G. Cosmetic intolerance. Contact Dermatitis 1987 : 16 : 189 – 194.
• 28 Krasteva M, Cottin M, Cristaudo A et al. Sensitivity and specificity of the consumer open skin allergy test as a method of prediction of contact dermatitis to hair dyes. Eur J Dermatol 2005 : 15 : 18 – 25.
• 29 Warshaw E M, Wang M Z, Mathias C G et al. Occupational contact dermatitis in hairdressers/cosmetologists: retrospective analysis of North American Contact Dermatitis Group Data, 1994 to 2010. Dermatitis 2012 : 23 : 258 – 268.
• 30 Lysdal S H, Søsted H, Andersen K E, Johansen J D. Hand eczema in hairdressers: a Danish register‐based study of the prevalence of hand eczema and its career consequences. Contact Dermatitis 2011 : 65 : 151 – 158.
• 31 Dejobert Y, Piette F, Thomas P. Contact dermatitis to 2‐hydroxyethylamino‐5‐nitroanisole and 3‐nitro‐p‐hydroxyethyl aminophenol in a hair dye. Contact Dermatitis 2006 : 54 : 217 – 218.