Understanding Fragrance Allergy Using An Exposure-based Risk Assessment Approach

Transcript

Copyright C Munksgaard 2001 Contact Dermatitis, 2001, 45, 333–340 Printed in Denmark . All rights reserved ISSN 0105-1873 Understanding fragrance allergy using an exposure-based risk assessment approach G. F G1, M K. R1, S P. F1, I R. W2  D A. B3 1 The Procter & Gamble Co., Miami Valley Laboratories, Cincinnati, OH, USA 2 St. John’s Institute of Dermatology, St. Thomas’ Hospital, London SE1 7EH, UK 3 SEAC Toxicology Unit, Unilever, Colworth House, Sharnbrook, UK Conducting a sound skin sensitization risk assessment prior to the introduction of new ingredients and products into the market place is essential. The process by which low-molecular-weight chemicals induce and elicit skin sensitization is dependent on many factors, including the ability of the chemical to penetrate the skin, react with protein, and trigger a cell-mediated immune response. Based on our chemical, cellular and molecular understanding of allergic contact dermatitis, it is possible to carry out a quantitative risk assessment. Specifically, by estimating the exposure to the allergen and its allergenic potency, it is feasible to assess quantitatively the sensitization risk of an ingredient in a particular product type. This paper focuses on applying exposure-based risk assessment tools to understanding fragrance allergy for 2 hypothetical products containing the fragrance allergen cinnamic aldehyde. The risk assessment process predicts that an eau de toilette leave-on product containing 1000 ppm or more cinnamic aldehyde would pose an unacceptable risk of induction of skin sensitization, while a shampoo, containing the same level of cinnamic aldehyde, would pose an acceptable risk of induction of skin sensitization, based on limited exposure to the ingredient from a rinse-off product application. Key words: skin; allergic contact sensitization; fragrances; exposure; potency; risk assessment; margin of safety. C Munksgaard, 2001. Accepted for publication 1 August 2001 For new products or product ingredients that contact the skin, it is necessary, prior to their introduction to the market, to conduct a thorough skin sensitization risk assessment to assure that the product will be well tolerated. The skin sensitization testing and risk assessment process for new ingredients and products generally follows a stepwise approach that may involve analytical assessments, preclinical skin sensitization testing, clinical testing, and benchmarking of resulting data against similar ingredients and product types. The details of these various elements and the overall process have been reviewed previously (1–5). Critical to conducting a sound skin sensitization risk assessment is having a thorough understanding of ingredient exposure, as well as allergenic potency and dose response. Clearly, the potential for an ingredient to cause a skin sensitization response is dependent on a variety of other factors including, for example, the application vehicle system (6, 7), the number (8) and duration (9) of exposures, underlying skin irritation (10, 11) and the use of occlusion (12). The importance of exposure and potency estimation in assessment of skin sensitization risk has recently been reviewed in a paper highlighting an exposure-based risk assessment process using methylchloroisothiazolinone/methylisothiazolinone (MCI/MI) as a case study (5). The paper shows how one can judge the sensitization risk for different products containing MCI/MI using an exposure-based risk assessment approach. This paper focuses on applying exposure-based skin sensitization tools to understanding the fragrance allergy potential for 2 distinct product types. Fragrance allergy is reported to be on the rise (13–16) and has received much attention recently regarding the skin sensitization risk of a number of specific perfume raw materials found in perfumes that are used in a variety of products. The goal of this paper is to illustrate how one can quantitatively evaluate the skin sensitization risk of a specific fragrance allergen by estimating the exposure and potency of the allergen under evaluation. It is important to note, however, that the 334 GERBERICK ET AL. approach described in this manuscript is specific to the induction of sensitization rather than elicitation in pre-sensitized individuals. In this manuscript, we have highlighted this approach, on a theoretical basis, by evaluating the skin sensitization risk of cinnamic aldehyde formulated in a leave-on product (eau de toilette) versus a rinse-off product (shampoo). Importance of Exposure and Potency Estimation in Skin Sensitization Risk Assessment While exposure for other endpoints is often expressed in units of mg/kg body weight, the relevant dose metric for skin sensitization potential is the amount of chemical per unit area of the allergen on the skin (5). In a series of elegant human skin sensitization studies, Friedmann and his colleagues proved that it is not the % (weight/volume) of material applied that is critical, but the total dose/ area of exposed skin (17–20). They used a potent contact allergen, dinitrochlorobenzene (DNCB) and repeatedly exposed subjects to varying doses per unit area of skin to observe the incidence of sensitization upon challenge. When they applied increasing total doses to proportionately increased skin surface areas (keeping the dose/unit area the same), the sensitization incidence was equal. When they kept the total dose constant, but varied the area of application, those subjects exposed within smaller areas of skin (and hence larger dose/unit area exposures) had the greater incidence of sensitization. This is consistent with our understanding of the mechanism of sensitization, which requires that antigens are recognized and processed by epidermal Langerhans cells (LC) and are subsequently transported to the lymph nodes where the antigen is presented to responsive T lymphocytes. Thus, it is critical to express the sensitization dose as a dose per unit area measurement (e.g., mg/cm2) when conducting a quantitative skin sensitization risk assessment. The concept of dose per unit area provides a ‘‘common currency’’ enabling comparison of sensitization incidences across studies and facilitating comparison of the potencies of different chemicals. For example, if one compared exposure to an ingredient at 0.1% in a laundry product versus a facial skin cream there would be a greater than 100¿ difference in exposure when compared on a dose per unit basis (5). One also needs to consider dose per unit area when reviewing or conducting patch testing since different patch types come in different sizes and require different volumes to load the patches (5). While dose per unit area is an important parameter to use in comparative skin sensitization as- sessment, it is by no means the only factor to consider. Allergenic potency is another essential factor for consideration when conducting skin sensitization risk assessments. Sensitizing potency in this context is best described as a function of the amount of chemical that is required to induce contact sensitization in a previously naı¨ve subject or animal, and on this basis it has been estimated that chemicals vary significantly in terms of their intrinsic allergenic activity (21, 22). Despite the importance of potency estimation in the development of accurate risk assessments, there has been relatively modest progress in the definition of appropriate experimental models. Although the standard guinea pig tests such as the maximization test have been very successful at hazard identification (23, 24), there has been some interest in the use of a modified guinea pig maximization test for consideration of relative potency. Of particular note have been the efforts of Andersen et al. (25), who have manipulated the guinea pig maximization test in order to obtain dose-response data. However, the development of a novel predictive assay in the mouse, the local lymph node assay (LLNA) (22, 26–28), provides new opportunities for the objective and quantitative estimation of skin sensitization potency (29, 30). For the purposes of hazard identification, activity in the LLNA is measured as a function of proliferative responses induced in draining lymph nodes by test chemicals, with those chemicals that provoke a 3¿ or greater increase in lymph node cell proliferation compared with vehicle controls being classified as potential contact allergens. This method has more recently been applied to determine relative potency, based on the estimated concentration of chemical necessary to cause a 3¿ increase in proliferative activity (the ‘‘EC3’’). Experience to date with this approach has been encouraging; clear differences between skin sensitizing chemicals can be discerned and such differences appear to correlate with the ability of the materials to induce contact allergy in experimental models and with what is known of their sensitizing activity in humans (31– 34). For ethical reasons, there are no well-defined or widely applied methods for the determination of relative skin sensitization potency in humans. However, review of the published literature for reports of dose-response induction studies in humans reveals valuable information on the sensitizing potency of a variety of chemicals. Using available human repeat patch test data, together with expert judgment, numerous compounds were classified as strong, moderate, weak, extremely weak or non-sensitizers (33). Additionally, it has been shown clearly that LLNA EC3 values for the EXPOSURE-BASED RISK ASSESSMENT OF FRAGRANCE same chemicals are very comparable with the clinical no observed effect levels (NOELs) calculated from the literature (34). These investigations demonstrate that the LLNA can be used to provide quantitative estimates of relative skin sensitizing potency (EC3 values) that correlate closely with NOELs established from human repeat patch testing and from clinical experience. Fragrance Allergy: Clinical Perspective Although a considerable number of chemicals have the capacity to cause allergic contact dermatitis, certain classes of materials appear to be more commonly associated with the condition. Amongst this group are the fragrance allergens; indeed allergic contact dermatitis from fragrance chemicals seems to be an increasing issue. Larsen highlighted the extent to which fragrances can cause skin allergy many years ago as a result of his work with the screening tray for fragrance allergens that he had developed (35, 36). Subsequently, this screening tray of 8 common fragrance allergens has been used widely to investigate perfume sensitivity and an exhaustive review article has been published (37). Most recently, evidence suggests that the incidence of allergy to fragrance chemicals is tending to increase (13–15, 38) and now represents approximately 10% of the patch test clinic population. From this work, it is also clear that certain fragrance allergens lead to a distinctly higher frequency of positive patch test reactions (e.g., isoeugenol) than do others (e.g., geraniol) (14). Cinnamic aldehyde (cinnamal) is one fragrance allergen that has been reported frequently in the dermatology literature (38–40). It is also 1 of the 8 constituents of the fragrance mix that is used for diagnosing fragrance contact allergy. Based on its well-known potential to cause contact allergy in humans, as well as information available on its inherent potency, cinnamic aldehyde was selected as a good candidate for use in our hypothetical risk assessment for induction of sensitization. Table 1. Data supporting potency classification of cinnamic aldehyde Supporting data O NOEL from Buehler guinea pig (unpublished data) O EC3 from LLNA (33) O NOEL from human RIPT (41) O LOEL from human RIPT (41) O NOEL from HMT (RIFM reference) Potency classification Default NOEL for use in risk assessment Ω30 mg/cm2 Ω500 mg/cm2 Ω591 mg/cm2 Ω1181 mg/cm2 Ω345 mg/cm2 Ωmoderate Ω100 mg/cm2 335 Potency Classification of Cinnamic Aldehyde The skin sensitization potency of cinnamic aldehyde has been evaluated in the murine LLNA, Buehler guinea pig and human patch tests in simple vehicles (Table 1). The NOEL value for cinnamic aldehyde calculated from a Buehler guinea pig dose response study was 30 mg/cm2 (unpublished data). The NOEL for cinnamic aldehyde was achieved when tested at 30 mg/cm2 (0.03%) in 80% EtOH for induction and 100 mg/cm2 (1%) for challenge; induction concentrations higher than 30 mg/cm2 yielded positive responses in the treated guinea pigs. Studies in the LLNA resulted in a threshold positive response (stimulation index of ⭓3, EC3) at an applied dose of 500 mg/cm2 when tested in acetone:olive oil (33). The NOEL reported in the literature for a human repeat patch test study is similar to the LLNA result giving a NOEL of 591 mg/cm2 when tested in ethanol (41). In the same study, a low effect level of 1181 mg/cm2 was reported. Finally, a NOEL of 345 mg/cm2 has been reported for cinnamic aldehyde when tested in petrolatum in a human maximization test (42). Overall, the threshold data on cinnamic aldehyde provide for a strong weight-of-evidence to classify cinnamic aldehyde as a moderate sensitizer. For quantitative risk assessment purposes, we have developed an arbitrary classification scheme to rank the potency of allergens based on weightof-evidence of data available in humans and/or animals (EC3 values from LLNA) (33) (Table 2). This scheme is comprised of 5 sensitization potency classes. It must be emphasized that this particular classification scheme has been adopted here solely for the purpose of facilitating identification of a default NOEL for risk assessment evaluations. Consequently it must be recognized that these classes, and the way in which they are defined with respect to EC3 values and NOEL data, do not necessarily represent the only or best approach to classification of skin sensitizing potency as a function of available animal and human data. The primary purpose here is to provide a mechanism whereby perfume raw materials may be ranked in terms of their potency using a weight-of-evidence of all available data. For each potency class, a ‘‘default NOEL’’ is assigned for use in quantitative risk assessments (Table 2). The main purpose of using a default NOEL is to recognize that the NOELs for most fragrance compounds available in the literature do not have a high degree of precision, sometimes being derived from a single human study. Thus, in the absence of good quantitative threshold data for sensitizers under evaluation, it is best to use conservative default NOELs for conducting skin sensitization risk assessments. 336 GERBERICK ET AL. Table 2. Sensitization potency classification: default NOELs for use in quantitative risk assessment (QRA) LLNA EC3 Sensitization potential Experimental human NOEL Default NOEL for use in QRA NCa ±10,000 mg/cm2 1000–10,000 mg/cm2 100–1000 mg/cm2 10–100 mg/cm2 Æ10 mg/cm2 non-sensitizing extremely weak weak moderate strong potent NCa ⬎10,000 mg/cm2 1000–10,000 mg/cm2 100–1000 mg/cm2 10–100 mg/cm2 Æ10 mg/cm2 NAb 10,000 mg/cm2 1000 mg/cm2 100 mg/cm2 10 mg/cm2 1 mg/cm2 a) NC, not calculated. No positive response is obtained at any concentration tested and, therefore, an EC3 value cannot be calculated. b) Not applicable. The material is a non-sensitizer and, thus, a default NOEL is not needed for risk assessment. However, if the reported NOEL value were on either side of a default NOEL, it would be the discretion of the risk assessor on what default NOEL level to use for the assessment. Of course, if the NOEL for a compound is robust (e.g., derived from a well-designed dose-response study), the calculated NOEL could be used for conducting a quantitative skin sensitization risk assessment. Cinnamic Aldehyde Exposure Calculations for Eau de Toilette and Shampoo The risk assessment process requires the ability to compare the chemical exposures used in animal or human patch test procedures with relevant or exaggerated exposures from the intended and foreseeable uses of a product. To determine the latter, one needs to know the concentration of the ingredient in the final product formulation, the amount of product applied, and the area of application. There will also be exposure differences based on the type of product. For example, for products that are applied to and left on the skin (leave-on products), the conservative assumption is that 100% of applied product (and all ingredients therein) is available. For products such as bath soaps or shampoos, that are applied and then rinsed off the skin (rinse-off products), there are calculations and assumptions made regarding residual product (and ingredients) remaining on the skin. Comparative exposure estimations for cinnamic aldehyde present at a hypothetical level of 0.1% (1000 PPM) in an eau de toilette or shampoo are shown in Table 3. It is important to note that the actual level of cinnamic aldehyde used in marketed eau de toilette and shampoo products varies significantly, with most products containing much lower values of cinnamic aldehyde. Many assumptions are used to calculate the exposures. These relate to skin surface area for the application site and other exposed skin surfaces, the amount of product used and number of applications per day and, for rinse off products, the residual material left on the skin. There are govern- ment and trade association guidelines that serve as resource material for most of the assumptions used in these calculations (24, 43–47), but these are currently being re-assessed by European authorities (EU Scientific Committee for Cosmetic and NonFood Products). The shampoo exposure determination is based on using body area parameters (44) and assuming 1% retention on the skin following use of the product. The eau de toilette exposure determination was based on measuring target ranges of product use (distances from product to skin); weight of product delivered per pump; and the skin area covered from product use at different distances. For the eau de toilette, we used the highest measured exposure dose value for our risk assessment (P&G unpublished data; presented at Colipa General Assembly, Athens 2000). Based on the parameters used (Table 3), the exposure to skin (in mg/cm2) for the ingredient of interest following use of these 2 products differs by greater than 30¿. Exposure-based Risk Assessment of Cinnamic Aldehyde Before a product is introduced into the market, all of the sensitization test results and exposure calculations are compiled and evaluated to determine a maximum allowable exposure limit (sensitization reference dose) for an ingredient in the intended product (48). As part of this process, sources of variability and uncertainty must be taken into account. These are generally acknowledged by the application of ‘‘uncertainty factors’’ that are sufficiently conservative so as to provide confidence that the allowable human exposure will, indeed, be safe for a heterogeneous population. Various considerations go into determining what is an appropriate uncertainty factor for a particular risk assessment. 3 important categories of uncertainty and variability include: O inter-individual response variability; O matrix differences between what was tested in EXPOSURE-BASED RISK ASSESSMENT OF FRAGRANCE 337 Table 3. Exposure calculations for eau de toilette and shampoo Eau de toilette Measured values (P&G unpublished data: presented at Colipa General Assembly) O Distance from target ranges from 5 to 20 cm O Dose per pump actuation ranges from 54 to 108 mg O Exposure area from product use ranges from 29 to 154 cm2 O Exposure dose per surface area ranges from 0.44 to 2.58 mg/cm2 Note: The exposure dose used for risk assessment is the highest measured value observed. Ω2.6 mg/cm2 Calculation for daily exposure to ingredient in eau de toilette: Ω[Amount in product (mg/g product) ¿ product exposure dose (mg/cm2)] ΩSkin exposure to ingredient (mg/cm2) Example: Cinnamic aldehyde at 1000 ppm (i.e 0.1%) in product Ω0.1% ¿ 2.6 mg/cm2 Ω2.6 mg/cm2 Shampoo Parameters (source: EPA exposure factor handbook; SCCNFP guidance document) O Median body surface area for female O % of total body surface area for head O Surface area for head (i.e. 7.1/100¿1.69) O Surface area for half head i.e. scalp (0.12/2) O % of total body surface area for hands O Surface area of the hands (i.e. 5.1/100¿1.69) O Amount shampoo used per day O Assume 1% retention on scalp, i.e. multiply final exposure by Ω1.69 m2 Ω7.1% Ω0.12 m2 Ω600 cm2 Ω5.1% Ω860 cm2 Ω12 g 0.01 Calculation for daily exposure to ingredient in shampoo: Ω[Amount in product (mg/g product) ¿ amount product applied (g)]/surface area exposed (cm2) ΩSkin exposure to ingredient (mg/cm2) ¿ 0.01 ΩSkin exposure to ingredient (mg/cm2) Example: Cinnamic aldehyde at 1000 ppm, i.e. 1000 mg/g product Ω1000 (mg/g) ¿ 12 (g)/1460 (cm2) Ω8.2 mg/cm2 ¿ 0.01 Ω0.08 mg/cm2 the patch test versus the product formulation to which the consumer will be exposed; O variations in product usage patterns that are not addressed in the exposure assessment (e.g., part of the body exposed and skin integrity, occluded or non-occluded, etc.). Using weighting factors of 1–10 for each of the 3 categories and multiplying the factors together provides an overall sensitization uncertainty factor that can range from 1 to 1000, unless there is justification for going beyond the factors identified. It is emphasized that these are guidelines and not a prescriptive approach; each risk assessment must rely on professional judgment to evaluate the weight-of-evidence to determine the appropriate uncertainty factors to apply. Once an ingredient sensitization uncertainty factor for a given product is defined (and these are commonly defined on the basis of product category), it is used with the default NOEL for that ingredient to determine the maximum human exposure limit for the ingredient in the product of interest (called sensitization reference dose, SRfD). Simply stated, S-RfD for an ingredient in the product is equal to the NOEL divided by the sensitization uncertainty factor (SUF): S-RfDΩNOEL/SUF. Thus, for example, if the default NOEL for an ingredient is 1000 mg/cm2 and the sensitization uncertainty factor is 100, the S-RfD would be 10 mg/ cm2. This S-RfD, then, represents a level of exposure that may be acceptable for the general public. However, consistent with other risk assessment methods, it does not – and cannot – imply absolute safety; nor, however, does it imply that exposures above the S-RfD will automatically be associated with a defined risk. It is not a ‘‘line in the sand’’ denoting ‘‘safe’’ on one side and ‘‘unsafe’’ on the other side; rather, it is a number that is calculated using conservative risk assessment approaches to help us estimate acceptable exposures by incorporating scientific data and principles to the extent possible, with uncertainty and variability being an inherent part of this process. These caveats notwithstanding, quantitative risk assessment remains a valuable tool. To illustrate how this approach maybe used, we will apply it to 2 products (eau de toilette and shampoo) hypothetically containing 1000 ppm cinnamic aldehyde. Based on the available biological ‘‘threshold’’ data on cinnamic aldehyde (Table 1), we have classified it as a moderate sensitizer and assigned 338 GERBERICK ET AL. Table 4. Retrospective assessment of cinnamic aldehyde acceptable exposure estimates for shampoo and eau de toilette formulations Shampoo assessment Cinnamic aldehyde default no effect level: 100 mg/cm2 Sensitization uncertainty factor assumptions 10¿ for inter-individual variability 3¿ for product matrix (surfactant matrix) 3¿ for use exposure patterns (lifetime use; compromised skin) 100¿ uncertainty factor Sensitization-reference dose (S-RfD) for shampoo products: default NOEL/uncertainty factorΩ100/100Ω1.0 mg/cm2 Margin of safetyΩS-RfD/Consumer exposureΩ1.0/0.08Ω12.5 Eau de toilette Cinnamic aldehyde default no effect level: 100 mg/cm2 Sensitization uncertainty factor assumptions 10¿ for inter-individual variability 1¿ for product matrix 10¿ for use exposure patterns (heavy users; compromised skin) 100¿ uncertainty factor Sensitization-reference dose (S-RfD) for eau de toilette: default NOEL/uncertainty factorΩ100/100Ω1.0 mg/cm2 Margin of safetyΩS-RfD/Consumer exposureΩ1.0/2.6Ω0.4 it a default NOEL of 100 mg/cm2 (33, 34). The weighting factors used to determine an uncertainty factor for cinnamic aldehyde in a shampoo are shown in Table 4. In this instance weighting factors of 10, 3 (1/2 log of 10 but for simplicity 3), and 3 were derived for inter-individual variability, matrix effects, and use patterns, respectively. The use of 3 as the 1/2 log of 10 is consistent with current thinking in approaches to quantitative risk assessment. The value of 10 for inter-individual variability is a default value that assumes that volunteer subjects for HRIPT studies do not fully reflect the heterogeneity of the general population. The value of 3 was assigned to matrix effects. This number was derived from the fact that the HRIPT conducted on cinnamic aldehyde was in ethanol (41), whereas the shampoo formulation contains surfactant ingredients. This 3¿ assumption is appropriate given the known ability of surfactants to increase the skin penetration of chemicals (49, 50). Finally, the value of 3 is assigned to use pattern. Although a shampoo is a rinse-off product, the product is used daily throughout lifetime, might be used more than 1¿ per day, and may be applied to compromised skin. Multiplying the weighting factors (10¿3¿3) results in an uncertainty factor of 100. As Table 4 indicates, dividing the default NOEL of 100 mg/cm2 by the sensitization uncertainty factor of 100¿ gives a sensitization reference dose of 1.0 mg/cm2. The S-RfD represents the NOEL that has been adjusted to account for the areas of un- certainty and variability in extrapolating from a NOEL identified in one scenario to be applied in another. For example, the cinnamic aldehyde NOEL obtained from a single patch study in a simple vehicle under certain patch test conditions can now be applied to a shampoo product that has a complex matrix and will be used by consumers with widely varying sensitivities. The final step is to compare the S-RfD to the estimated consumer exposure to calculate a ‘‘margin of safety’’ (MOS): MOSΩS-RfD/consumer exposure. In the case of the cinnamic aldehyde-containing shampoo at 1000 ppm, the MOS is 12.5. The MOS is a unitless term since both the S-RfD and the consumer exposure are expressed in units of mg/ cm2. In this hypothetical situation, the MOS of 12.5 suggests that the amount of cinnamic aldehyde in a shampoo would pose little-to-no sensitization risk. Again, it is important to note that we are talking about induction of sensitization and not the elicitation of reactions in patients previously sensitized to cinnamic aldehyde (38). The uncertainty factor and exposure limit determinations for cinnamic aldehyde in a leave-on eau de toilette application are also summarized in Table 4. In this situation the weighting factors are slightly different. The inter-individual variability was the same as for the shampoo product (10¿) and a 10¿ factor was also used for use exposure pattern since the product use is long-term and can potentially be used on compromised skin. The product matrix received a factor of 1¿ since the product vehicle is similar to the vehicle used for establishing the default NOEL (41). Uncertainty factor multiplication in this instance gives a factor of 100 (10¿1¿10), and dividing this value into the default no observed effect level of 100 mg/cm2 gives a S-RfD dose for the ingredient of 1 mg/cm2. In the case of the cinnamic aldehyde-containing eau de toilette (1000 ppm), the MOS is 0.4 demonstrating that the exposure is in a range at which our confidence for safe use by the general population would not be sufficiently high without additional data. Summary Allergic contact dermatitis is a complex process that is influenced by a variety of factors related to chemical exposure and the eventual development of skin sensitization (induction) and its clinical manifestation. In producing and marketing consumer products, it is critical to understand the exposure characteristics for different EXPOSURE-BASED RISK ASSESSMENT OF FRAGRANCE product types and the ingredients contained therein, especially if any of those ingredients have the potential to be contact allergens. The importance of understanding exposure estimation and allergenic potency for conducting sound skin sensitization risk assessments has been addressed previously (5). Specifically, for skin sensitization (induction) risk assessments, it is essential to understand the threshold or no observed effect levels for potentially sensitizing ingredients, in the common currency of dose per unit area. Moreover, it is equally important to estimate the extent of exposure to the chemical when formulated in different product types. With this vital information, it is then possible, using producttype-specific uncertainty factors, to determine ‘‘safe’’ exposure limits for specific ingredients. In this paper, we have conducted a quantitative exposure-based skin sensitization (induction) risk assessment using 2 consumer products containing the known human allergen cinnamic aldehyde at a hypothetical level of 0.1%. The risk assessment process reveals that an eau de toilette product containing 1000 ppm or more cinnamic aldehyde could pose an unacceptable skin sensitization (induction) risk and therefore would require additional testing to support marketing. In contrast, a shampoo containing the same level of cinnamic aldehyde would probably pose an acceptable skin sensitization (induction) risk based on limited exposure to the ingredient from this rinse-off product application. Again, it is imperative to emphasize that we are discussing the risk of inducing skin sensitization and not elicitation of skin sensitization in individuals pre-sensitized to the ingredient. However, it is important to point out that these exposure-based risk assessment tools may be applied also to understanding elicitation of skin sensitization if the threshold for elicitation is known or a factor for converting an induction threshold to a an elicitation threshold is used (e.g., 1000¿). Finally, in regard to improving or reducing the clinical cases of fragrance allergy, the risk assessment approach discussed in this paper should be used to establish improved guidelines for specific perfume raw materials known to cause sensitization and, equally important, be used for setting limits for newly developed perfume raw materials that might pose a sensitization (induction) hazard. They can also be used to help us incorporate better science such that evaluation of sensitization as a risk for the consumer is based not on hazard alone, but incorporates both potency and exposure. In this way, allowances may appropriately be made for higher levels of fragrance compounds in some product types than in others, depending on known or foreseeable consumer exposures. 339 Acknowledgements The authors thank Dr. Kim Rich and Ms Cindy Ryan for their critical review and comments on the manuscript. In addition, the authors acknowledge all members of the European Fragrance Think Tank (led by Dr. Ian White) for their critical discussions on this skin sensitization risk assessment approach. References 1. Robinson M K, Stotts J, Danneman P J, Nusair T L, Bay P H. A risk assessment process for allergic contact sensitization. Food Chem Toxicol 1989: 27: 479–489. 2. Gerberick G F, Robinson M K, Stotts J. An approach to allergic contact sensitization risk assessment of new chemicals and product ingredients. Am J Contact Dermatitis 1993: 4: 205–211. 3. Basketter D A, Gerberick G F, Robinson M K. Risk Assessment. In: Kimber I, Maurer T (eds.): Toxicology of contact hypersensitivity. London: Taylor & Francis, Ltd., 1996: 152–164. 4. Basketter D A. Skin sensitization: risk assessment. International Journal of Cosmetic Science 1998: 20: 141–150. 5. Robinson M K, Gerberick G F, Ryan C A, McNamee P, White I, Basketter D A. The importance of exposure estimation in the assessment of skin sensitization risk. Contact Dermatitis 2000: 42: 251–259. 6. Heylings J R, Clowes H M, Cumberbatch M, Dearman R J, Fielding I, Hilton J, Kimber I. Sensitization to 2,4dinitrochlorobenzene: Influence of vehicle on absorption and lymph node activation. Toxicology 1996: 109: 57–65. 7. Robinson M K, Cruze C A. Preclinical skin sensitization testing of antihistamines: Guinea pig and local lymph node assay responses. Food Chem Toxicol 1996: 34: 495–506. 8. Robinson M K, Parsell K W, Breneman D L, Cruze C A. Evaluation of the primary skin irritation and allergic contact sensitization potential of transdermal triprolidine. Fundam Appl Toxicol 1991: 17: 103–119. 9. McFadden J P, Wakelin S H, Holloway D B, Basketter D A. The effect of patch duration on the elicitation of paraphenylenediamine contact allergy. Contact Dermatitis 1998: 39: 79–81. 10. Kligman A M. The identification of contact allergens by human assay (3). The maximization test: a procedure for screening and rating contact sensitizers. J Invest Dermatol 1966: 47: 393–409. 11. Cumberbatch M, Scott R C, Basketter D A, Scholes E W, Hilton J, Dearman R J, Kimber I. Influence of sodium lauryl sulphate on 2,4-dinitrochlorobenzene-induced lymph node activation. Toxicology 1993: 77: 181–191. 12. Funk J O, Maibach H I. Propylene glycol dermatitis: Reevaluation of an old problem. Contact Dermatitis 1994: 31: 236–241. 13. Buckley D A, Wakelin S H, Seed P T, Holloway D, Rycroft R J, White I R, McFadden J P. The frequency of fragrance allergy in a patch-test population over a 17-year period. Br J Dermatol 2000: 142: 279–283. 14. Scientific Committee on Cosmetic Products and Non-Food Products Intended for Consumers. Opinion concerning fragrance allergy in consumers. SCCNFP/0017/98, 1999. 15. Lunder T, Kansky A. Increase in contact allergy to fragrances: patch test results 1989–1998. Contact Dermatitis 2000: 43: 107–109. 16. Johansen J D, Menne T, Christophersen J, Kaaber K, Veien N. Changes in the pattern of sensitization to common aller- 340 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. GERBERICK ET AL. gens between 1985–86 and 1997–98, with a special view of preventive strategies. Br J Cont Dermatitis 2000: 142: 490– 495. Friedmann P S, Moss C, Shuster S, Simpson J M. Quantitation of sensitization and responsiveness to dinitrochlorobenzene in normal subjects. Br J Dermatol 1983: 109 (25): 86–88. Friedmann P S, Moss C, Shuster S, Simpson J M. Quantitative relationships between sensitizing dose of DNCB and reactivity in normal subjects. Clin Exp Immunol 1983: 53: 709–715. White S I, Friedmann P S, Moss C, Simpson J M. The effect of altering area of application and dose per unit area on sensitization by DNCB. Br J Dermatol 1986: 115: 663– 668. Rees J L, Friedmann P S, Matthews J N. The influence of area of application on sensitization by dinitrochlorobenzene. Br J Dermatol 1990: 122: 29–31. Basketter D A, Rodford R, Kimber I, Smith I, Wahlberg J E. Skin sensitization risk assessment: a comparative evaluation of 3 isothiazolinone biocides Contact Dermatitis 1999: 40: 150–154. Basketter D A, Gerberick G F, Kimber I, Loveless S E. The local lymph node assay – A viable alternative to currently accepted skin sensitisation tests. Food Chem Toxicol 1996: 34: 985–997. Andersen K E, Maibach H I. Guinea pig sensitisation assays: an overview. In: Andersen K E and Maibach H I (eds.): Contact allergy predictive tests in guinea pigs, Current Problems in Dermatology, vol. 14. New York: Karger, 1985: 59–106. Botham P A, Basketter D A, Maurer Th, Mueller D, Potokar M, Bontinck W J. Skin sensitization – a critical review of predictive test methods in animal and man. Food Chem Toxicol 1991: 29: 275–286. Andersen K E, Volund A, Frankild S. The guinea pig maximization test with a multiple dose design. Acta Dermatovenereologica 1995: 75: 463–469. Kimber I, Weisenberger C. A murine local lymph node assay for the identification of contact allergens. Archives of Toxicology 1989: 63: 274–282. Kimber I, Basketter D A. The murine local lymph node assay; collaborative studies and new directions: a commentary. Food Chem Toxicol 1992: 30: 165–169. Gerberick G F, Ryan C A, Kimber I, Dearman R J, Lea L J, Basketter D A. Local lymph node assay validation assessment for regulatory purposes. Am J Cont Derm 2000: 11: 3–18. Kimber I, Basketter D A. Contact sensitization: A new approach to risk assessment. Human and Ecological Risk Assessment 1997: 3: 385–395. Dearman R J, Basketter D A, Kimber I. The local lymph node assay: use in hazard and risk assessment. J Applied Toxicol 1999: 19: 299–306. Hilton J, Dearman R J, Harvey P, Evans P, Basketter D A, Kimber I. Estimation of relative skin sensitizing potency using the local lymph node assay: A comparison of formaldehyde and glutaraldehyde. Am J Contact Derm 1998: 9: 29–33. Basketter D A, Evans P, Gerberick G F, Kimber I. Chemical allergy: estimation of potency, thresholds and risk assessments. Comments in Toxicology 1999: 7: 79–88. Basketter D A, Blaikie L, Dearman R J, Kimber I, Ryan C A, Gerberick G F, Harvey P, Evans P, White I R, Rycroft R J G.. Use of the local lymph node assay for the estimation of relative contact allergenic potency. Contact Dermatitis 2000: 42: 344–348. 34. Gerberick G F, Robinson M K, Ryan C A, Dearman R J, Kimber I, Basketter D A, Wright Z, Marks J G. Contact allergenic potency: correlation of human and local lymph node assay data. Am J Contact Derm, in press. 35. Larsen W. Perfume dermatitis. A study of 20 patients. Arch Dermatol 1977: 12: 623–627. 36. Larsen W. Perfume dermatitis. J Am Acad Dermatol 1985: 12: 1–9. 37. De Groot A C, Frosch P J. Adverse reactions fragrances. A clinical review. Contact Dermatitis 1997: 36: 57–86. 38. Johansen J D, Andersen K E, Rastogi S C, Menne´ T. Threshold responses in cinnamic aldehyde-sensitive subjects: results and methodological aspects. Contact Dermatitis 1996: 34: 165–171. 39. Larsen W, Nakayama H, Lindberg M, Fischer T, Elsner P, Burrows D, Jordan W, Shaw S, Wilkinson J, Marks J, Sugawara M, Nethercott ?. Fragrance contact dermatitis. A worldwide multicentre investigation (Part 1). J Am Acad Dermatol 1996: 7: 77–83. 40. Malten K E, Van Ketel W G, Nater J P, Liem D H. Reactions in selected patients to 22 fragrance materials. Contact Dermatits 1984: 11: 1–10. 41. Danneman P J, Booman K A, Dorsky J, Kohrman K A, Rothenstein A S, Sedlak R I, Steltenkamp R J. Cinnamic aldehyde: A survey of consumer patch-test sensitization. Food Chem Toxicol 1983: 21: 721–725. 42. Research Institute for Fragrance Materials/Flavors and Extract Manufacturers, Association Database. Human Health Data, unpublished. 43. EPA. Dermal exposure assessment: principles and applications. US EPA publication no. 600/8-91-001, Exposure Assessment Group, Office of Health and Environmental Assessment, Office of Research and Development; Washington, D.C. 1992. 44. EPA. Exposure factors handbook. US EPA publication no. 600/P-95/002, Exposure Assessment Group, Office of Health and Environmental Assessment, Office of Research and Development, Washington, D.C. 1998. 45. EPA. Exposure factors handbook. Internet URL: Http:// www.epa.gov/ncea/exposfac.htm 1999. 46. ECETOC. Assessment of non-occupational exposure to chemicals. Technical Report no. 58, 1994. 47. Scientific Committee on Cosmetic products and non-food products intended for consumers. Notes for guidance for testing of cosmetic ingredients for their safety evaluation. SCCNFP/0119/99, 1999. 48. Gerberick G F, Robinson M K. Skin sensitization risk assessment of new products. Am J Contact Dermatitis 2000: 11: 65–73. 49. Seidenari S, Motolese A, Belletti B. Pre-treatment of nickel test areas with sodium lauryl sulfate detects nickel sensitivity in subjects reacting negatively to routinely performed patch tests. Contact Dermatitis 1996: 34: 88–92. 50. Effendy I, Weltfriend S, Kwangsukstith C, Singh P, Maibach H. Effects of all-trans retinoic acid and sodium lauryl sulphate on the permeability of human skin in vitro. Br J Dermatol 1996: 135: 428–432. Address: G. Frank Gerberick The Procter & Gamble Co. P.O. Box 538707 Cincinnati, OH 45253-8707 USA