Anal Bioanal Chem (2014) 406:6165–6178 DOI 10.1007/s00216-014-8066-3

PAPER IN FOREFRONT

Detection of potentially skin sensitizing hydroperoxides of linalool in fragranced products Susanne Kern & Hafida Dkhil & Prisca Hendarsa & Graham Ellis & Andreas Natsch

Received: 3 June 2014 / Revised: 18 July 2014 / Accepted: 24 July 2014 / Published online: 20 August 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract On prolonged exposure to air, linalool can form sensitizing hydroperoxides. Positive hydroperoxide patch tests in dermatitis patients have frequently been reported, but their relevance has not been established. Owing to a lack of analytical methods and data, it is unclear from which sources the public might be exposed to sufficient quantities of hydroperoxides for induction of sensitization to occur. To address this knowledge gap, we developed analytical methods and performed stability studies for fine fragrances and deodorants/antiperspirants. In parallel, products recalled from consumers were analysed to investigate exposure to products used in everyday life. Liquid chromatography–mass spectrometry with high mass resolution was found to be optimal for the selective and sensitive detection of the organic hydroperoxide in the complex product matrix. Linalool hydroperoxide was detected in natural linalool, but the amount was not elevated by storage in a perfume formulation exposed to air. No indication of hydroperoxide formation in fine fragrances was found in stability studies. Aged fine fragrances recalled from consumers contained a geometric mean linalool Electronic supplementary material The online version of this article (doi:10.1007/s00216-014-8066-3) contains supplementary material, which is available to authorized users. S. Kern Analytical Chemistry, Givaudan Schweiz AG, Ueberlandstrasse 138, 8600 Duebendorf, Switzerland H. Dkhil : P. Hendarsa Perfume Analysis, Givaudan France SAS, 19-23 Voie des Bans, 95102 Argenteuil, France G. Ellis RAPS Fragrance Toxicology, Givaudan International SA, 5 Chemin de la Parfumerie, 1214 Vernier, Switzerland A. Natsch (*) Biosciences, Givaudan Schweiz AG, Ueberlandstrasse 138, 8600 Duebendorf, Switzerland e-mail: [email protected]

concentration of 1,888 μg/g and, corrected for matrix effects, linalool hydroperoxide at a concentration of around 14 μg/g. In antiperspirants, we detected no oxidation products. In conclusion, very low levels of linalool hydroperoxide in fragranced products may originate from raw materials, but we found no evidence for oxidation during storage of products. The levels detected are orders of magnitude below the levels inducing sensitization in experimental animals, and these results therefore do not substantiate a causal link between potential hydroperoxide formation in cosmetics and positive results of patch tests. Keywords Skin sensitization . Allergy . Hydroperoxides . High-resolution mass spectrometry . Fragrance . Linalool Abbreviations GC Gas chromatography HR-MS High-resolution mass spectrometry LC Liquid chromatography LOD Limit of detection LOQ Limit of quantification MS Mass spectrometry

Introduction Monoterpenes are the main components of aromatic essential oils extracted from a wide variety of plants. The essential oils, terpenes extracted from these oils and synthetic products of these compounds are broadly found in many areas, including fragranced consumer products, aromatherapy oils, pharmaceutical preparations, over-the-counter topical treatments and flavour compositions, through occupational exposure and through exposure to plants and fruits. Owing to the presence of olefinic bonds, monoterpenes can undergo autoxidation on prolonged air exposure. This has been primarily studied in concentrated essential

6166

oils and in pure preparations of their constituents [1–4]. The primary oxidation products are hydroperoxides, which react further to form secondary oxidation products. Air-oxidized linalool (a monoterpenic allylic alcohol) was studied in great detail and was found to be skin sensitizing in animal tests [5, 6]; for a summary of the published studies, see Fig. 1. A series of elegant studies have elucidated the role of the hydroperoxides as key sensitizing components formed during oxidation and the chemical pathways leading to these products [3, 7]. Oxidized linalool is increasingly being used as a diagnostic tool in patch tests in dermatitis patients [6, 8, 9], with recent studies showing a consistently high frequency of positive reactions. In patch tests, dermatological patients with suspected allergies are challenged with skin patches of putative allergens for 48 h, and the development of clinical signs is assessed for 3–7 days. On the basis of the frequent skin reactions to the tested hydroperoxide in these tests, it was proposed that oxidized linalool is among the commonest causes of contact allergy [6] and among the commonest contact allergens [8]. A strong dose response was observed in clinical studies, with a particularly high proportion of patients reacting only to patch-test preparations with a high concentration of linalool hydroperoxide [6]. However, there remains a missing link between the detailed laboratory studies showing that pure linalool can undergo autoxidation to sensitizing components on prolonged air exposure, on the one hand, and the patch-test results in clinics, on the other hand: so far evidence for consumer exposure to significant levels of linalool hydroperoxide and the potential source of such exposure has not been established. Around 60 % of perfumed products contain linalool [10], and the prevalence is even higher in fine fragrances (see this work). Therefore, exposure is common, and the mere presence of the parent compound in products used by patients cannot be

Fig. 1 Main primary and secondary oxidation products for linalool with their reported local lymph node assay concentrations for threefold stimulation of thymidine incorporation (EC3) [3, 5, 22]. Linalool-7-OOH 2 is normally found at fourfold higher concentrations as compared with linalool-6-OOH 3 in air-exposed linalool [3]

S. Kern et al.

taken as evidence for the relevance of the patch-test reactions, so long as the presence of a significant amount of linalool hydroperoxide in the products used by patients has not been verified. This gap can be addressed only with analytical studies on the hydroperoxides in products used by consumers. However, analysis of these organic hydroperoxides poses significant analytical challenges. So far, method development work has been reported [11, 12], but the methods were applied only to neat essential oils and have never been used for consumer products. To overcome this limitation, a method with high selectivity is needed as consumer products and fragrances, in particular, are complex mixtures of multiple components. As indicated above, the laboratory studies on oxidation of terpenes were primarily performed starting either from pure preparations of the terpenes or from concentrated essential oils, which were exposed to air in an open Erlenmeyer flask and submitted to a standardized procedure of stirring four times a day for up to 80 weeks [3]. This worst-case scenario proved very useful to evaluate and compare the potential of different compounds to undergo autoxidation. However, terpenes used in consumer products may be protected from oxidation by several factors. Firstly, they are used in diluted form in the products (with the possible exception of some aromatherapy applications), and autoxidation may be slowed by competition for oxygen by other product ingredients. In addition, products are sold in closed containers, and any air contact comes mainly from the headspace air trapped in partly emptied containers. Finally, antioxidants are often used to protect cosmetic products against oxidation. Although fragranced consumer products are not routinely tested for the formation of terpene hydroperoxides, it is well known that the olfactory properties of fragrances may change on storage if products are not correctly stabilized. Therefore, industry is routinely performing stability studies on fragrances. These are normally done at elevated temperatures, and a usual procedure is to follow the stability for 2–3 months at 37–45 °C. Storage at 45 °C for 2 months is considered to mimic storage for 8 months at room temperature as lipid oxidation rates are known to increase by around twofold with an increase in temperature of 10 °C [13]. Fragrance levels in different consumer products differ widely. Whereas typical cosmetic creams and lotions contain 0.3–0.6 % fragrance, deodorants and antiperspirants typically contain 0.5– 2 %. The highest fragrance levels are used in fine fragrances, which are hydroalcoholic products containing a fragrance level which may exceed 20 %, but is usually around 10 % (products at the latter concentration are known as eau de toilette). Next to the fragrance level in the product, exposure is mainly driven by the use pattern of a product. For skin sensitization, the relevant metric is the dose applied per unit area and retained on the skin [14]. Deodorants/antiperspirants and fine fragrances are important contributors to exposure if expressed as the dose of the final product applied and retained per unit area [15]. Thus, these two product types are both applied at a significant dose per unit area

Detection of potentially skin sensitizing hydroperoxides

and they contain high levels of fragrance. Therefore, they are two of the most relevant sources of exposure to fragrance ingredients when skin sensitization is considered. Here we developed and compared methods and performed stability studies on fine fragrance formulations and antiperspirants containing defined amounts of linalool. We monitored the stability of the parent component over time and determined the formation of primary and secondary oxidation products. Gas chromatography (GC) coupled with mass spectrometry (MS) and liquid chromatography (LC) coupled with MS with high mass resolution were applied to quantify these compounds. The potential influence of antioxidants, partly emptied bottles and repeated opening to allow air contact was determined. Finally, an analytical study was performed on aged hydroalcoholic fine fragrance samples and deodorants/antiperspirants recalled from consumers in order to obtain information on linalool hydroperoxide content in typical aged products.

Materials and methods Hydroalcoholic formulations Linalool content relative to the total fragrance oil content was assessed in 861 hydroalcoholic market formulations. Among these, 30 % of the formulas contained linalool at 10,000– 30,000 μg/g, 20 % contained linalool at 30,000–50,000 μg/g and 21 % contained linalool at more than 50,000 μg/g (expressed relative to the total fragrance content), with a median content of 23,000 μg/g. The fragrance formulations tested in detail in the stability study were designed or selected to reflect these typical contents as shown in Table 1. Formulations A50, A20 and A5 were designed by a Givaudan perfumer with the same overall formulation and decreasing content of linalool. They were formulated at 10 % in an 80:20 ethanol/ demineralized water mix as is the case for most hydroalcoholic products on the market. Commercial fragrance C was selected on the basis of a linalool content typical for a market product. In addition, we tested synthetic linalool (a commercial manufacturing quality) and natural linalool (used for naturalgrade fragrances) at 10 % in the same hydroalcoholic formulation. These latter treatments reflect the extreme case of a fragrance consisting entirely of linalool, and they allow comparison of the difference between naturally derived and synthetic linalool. Given the high linalool content, they also allow monitoring of components formed at low concentration. Synthetic linalool is a racemic mixture of 50 % (-)-linalool and 50 % (+)-linalool, whereas natural linalool has a 98:2 isomeric excess of (-)-linalool. The different variants of fragrances A were stored in the presence of three stabilizing agents, which are present in typical commercial products, namely 0.05 % antioxidant tert-butyl hydroxytoluene, 0.05 % chelating agent ethylenediaminetetraacetate,

6167

and 0.2 % UV-filtering formulation Covabsorb (ethylhexyl methoxycinnamate/butyl methoxydibenzoylmethane/ethylhexyl salicylate). The two linalool products were stored both in the presence and in the absence of these stabilizing agents. All samples (12.5- or 25-ml initial volume) were stored in 30-ml glass bottles. Temperature stability of hydroalcoholic products The products described in the previous section were stored in an oven at 45 °C in order to accelerate the ageing process by around fourfold as compared with ambient temperatures, whereas control samples were stored at 5 °C. All products were stored in parallel (1) in full bottles, (2) in half-emptied bottles which were opened only for sampling and (3) in halfemptied bottles opened every 14 days (opened for approximately 1 min to allow gas exchange) over the 9-month study period to maximize air exposure. Commercial products are sold with pump nebulizers, and under normal product use, only the liquid volume used is replaced with ambient air. Thus, opening the bottles should represent a worst case, allowing greater air contact as compared with normal product use. The full experimental set-up was monitored for the 2month study, which reflects a typical industry procedure for stability tests. The treatments with the highest oxygen exposure (half-emptied bottles opened every 14 days) were monitored for a further 7 months (9 months in total). All storage samples at 45 °C were analysed in triplicate (three independent bottles sampled at each time point), with the exception of commercial fragrance C, which was analysed in duplicate. Samples stored at 5 °C were analysed in duplicate. Antiperspirant formulations and temperature stability study The antiperspirant formulations contained either 1 % synthetic linalool or 1 % fragrance A50, fragrance A20 or fragrance A5 and an antiperspirant base which is representative of a commercial product [antiperspirant agent (aluminium chlorohydrate, 10 %), emollient (cyclomethicone, 12 %), suspending agent (quaternium-18 hectorite, 0.8 %), propellant gas (isobutane/propane, 76.2 %), tert-butyl hydroxytoluene (0.005 %), ethylenediaminetetraacetate (0.005 %) and Covabsorb (0.2 %)]. Pressurized antiperspirant/deodorant aerosol cans were stored under the conditions described earlier, and were destructively sampled after 1 and 2 months. Studies on products recalled from consumers Consumers were asked to bring in partly used samples of hydroalcoholic products stored in their homes for at least 2 years. Thirty-nine samples were obtained as listed in the Electronic supplementary material. These samples were then directly used for a detailed analytical investigation by GC–MS

6168

S. Kern et al.

Table 1 Study set-up and products tested Product Hydroalcoholic products Fragrance A50 Fragrance A20 Fragrance A5 Commercial fragrance C Synthetic linalool Natural linalool Samples recalled from consumers Antiperspirant products Fragrance A50 Fragrance A20 Fragrance A5 Synthetic linalool Samples recalled from consumers

Content of linalool in the fragrance part

5 % synthetic linaloola 2 % synthetic linaloola 0.5 % synthetic linaloola 2.8 % linalool, origin of linalool unknownb 100 % synthetic linaloola 100 % naturally derived linalool from Cinnamomum camphoraa 39 partly used samples, aged 2 years to more than 10 years; for details and linalool content, see Table S5 5 % synthetic linaloolc 2 % synthetic linaloolc 0.5 % synthetic linaloolc 100 % synthetic linaloolc 29 used products and 5 fresh commercial samples

a

The fragrance mixture and the pure linalool samples were diluted tenfold in 80 % ethanol in water to make up the final eau de toilette product.

b

Commercial eau de toilette with a fragrance content of 11 % on the basis of analysis c

The fragrance mixture and the pure linalool samples were formulated at 1 % in an antiperspirant propellant base.

and LC–MS. Similarly, 29 partly used antiperspirants and deodorants and five fresh commercial samples were collected for analysis of linalool and linalool hydroperoxide. Reference compound A 45:55 mixture of (5E)-7-hydroperoxy-3,7-dimethylocta-1,5diene-3-ol 2 (Fig. 1) and 6-hydroperoxy-3,7-dimethylocta-1,7diene-3-ol 3 (consisting of cis and trans diastereomers which are not separated by GC–MS or LC–MS; Fig. 1) was prepared from synthetic linalool by oxidation in a microemulsion made from water, sodium dodecyl sulfate, n-butanol and CH2Cl2, using 35 % aqueous Na2MoO4 with H2O2 as a source of singlet oxygen [7]. This linalool hydroperoxide reference contained 80.2 % total linalool hydroperoxides (55:45 mixture of 6hydroperoxide and 7-hydroperoxide) as determined with a quantitative NMR experiment with anisic aldehyde as an internal calibration reference. It was used for linalool hydroperoxide identification and for spiking experiments. NMR data were in agreement with previously reported data [3, 5, 7]: –

Compound 2. Rf (1:1 hexane/methyl tert-butyl ether) 0.38. H NMR (400 MHz, CDCl3): δ (ppm) 8.00–7.97 (br s, OOH), 5.92 (dd, J=10.6, 17.4 Hz, 1H), 5.69 (ddd, J=6.3, 7.3, 15.9 Hz, 1H), 5.62 (d, J=15.9 Hz, 1H), 5.21 (dd, J=1.3,



17.2 Hz, 1H), 5.07 (dd, J=1.3, 10.6 Hz, 1H), 2.33 (dd, J= 5.9, 13.5 Hz, 1H), 2.27 (dd, J=6.8, 13.6 Hz, 1H), 2.00–1.94 (br s, OH), 1.33 (s, 2Me), 1.31 (s, Me), 13C NMR (100 MHz, CDCl3): δ (ppm) 144.60, 137.96, 126.37, 112.14, 81.94, 72.73, 45.10, 27.58, 24.32, 24.19. MS (electron impact, tR 6.68 min.): m/z (relative intensity) 152 (0.1, [M]+· −H2O2), 137 (1), 125 (1), 109 (2), 85 (8), 83 (20), 82 (100), 71 (93), 67 (42), 59 (5), 55 (23), 43 (70), 41 (20), 39 (13), 27 (12). Compound 3 (1:1 diastereomeric mixture). Rf (1:1 hexane/ methyl tert-butyl ether) 0.38. 1H NMR (400 MHz, CDCl3): δ (ppm) 8.47–8.44 (br s, OOH), 8.43–8.41 (br s, OOH), 5.88 (dd, J=10.9, 17.4 Hz, 1H), 5.87 (dd, J=10.9, 17.8 Hz, 1H), 5.22 (dd, J=1.0, 17.3 Hz, 1H), 5.21 (dd, J=1.4, 17.4 Hz, 1H), 5.07 (dd, J=1.3, 10.6 Hz, 2H), 5.02–4.98 (m, 4H), 4.32–4.25 (m, 2H), 1.85–1.83 (br s, OH), 1.76–1.74 (br s, OH), 1.74–1.72 (br s, 2Me), 1.71–1.49 (m, 8H), 1.29 (s, Me), 1.29 (s, Me). 13C NMR (100 MHz, CDCl3): δ (ppm) 144.53, 144.51, 143.52 (2C) , 114.14, 114.04, 112.09 (2C), 89.47, 89.29, 73.13, 73.09, 37.75, 37.53, 27.99, 27.93, 25.08 (2C), 17.46, 17.40. MS (electron impact, tR 6.92 min.): m/z (relative intensity) 171 (1, [M]+· –·CH3), 153 (2, [M]+· −H2O −·CH3), 137 (5), 125 (3), 123 (4), 109 (9), 107 (9), 98 (10), 97 (10), 93 (13), 83 (18), 82 (26), 81 (19), 71 (100), 69 (28), 67 (57), 55 (41), 43 (75), 41(52), 39 (31), 27 (16).

Analytical methods For all the hydroalcoholic formulations, 1.25 g of the samples was weighed in analytical vials, spiked with 50 μl of 10 % methyl octanoate in hexane as a reference compound and directly injected into the GC–MS or LC–MS system without further dilution. The antiperspirants tested in the stability study were cooled to -20 °C, the cans were pierced and the propellant gas was allowed to evaporate for 1 h. One gram of the residual formulation was mixed with 10 ml of pentane and 50 μl of 10 % methyl octanoate in hexane. The sample was stirred for 30 min, pentane was evaporated and the residual sample was filtrated and used for analysis. In the antiperspirants/deodorants recalled from consumers not containing solvent in addition to the propellant gas, 0.1 g of the residual sample after evaporation of the propellant gas was mixed with 900 μl of 80 % ethanol/20 % water and centrifuged (1.5-ml tubes; 20,800g; 10 min at ambient temperature) and filtrated prior to LC–MS analysis. Samples containing solvent were not diluted. In each case, a parallel sample was spiked with synthetic linalool hydroperoxide at a final level of 16 μg/ml. GC–flame ionization detection (FID) and GC–MS to determine the stability of linalool in the 2-month stability studies (GC–MS/FID method 1)

1

GC analysis was done in parallel using an HP 6890 series GC system (Agilent) linked to either a 5973 mass-selective

Detection of potentially skin sensitizing hydroperoxides

detector (for identification) or a flame ionization detector (FID) (for quantification). An Rtx-1 GC column (100 % polydimethylsiloxane; Restek, Lisses, France) was used, with an inner diameter of 0.25 mm, a length of 60 m and a film thickness of 0.25 μm. The injector temperature was 250 °C. A 2-μl sample was injected in a 1:85 split ratio under constant pressure (18.0 psi, 1 ml/min initial flow rate) conditions (carrier gas, helium). The temperature of the column oven was ramped from 70 °C at 2 °C/min to 240 °C, and this temperature was then held for 45 min. GC–MS to analyse samples for linalool and secondary oxidation products on prolonged ageing (GC–MS method 2) GC–MS analysis was done using an HP 6890 series GC system (Agilent) with a 5973 mass-selective detector. The GC column used was a BPX5 column (SGE) containing 5 % phenyl and 95 % polydimethylsiloxane, with an inner diameter of 0.22 mm, a length of 12 m and a film thickness of 0.25 μm. The injector temperature was 230 °C. A 1-μl sample was injected in a 1:10 split ratio under constant flow (1.0 ml/min) conditions (carrier gas, helium). The temperature of the column oven was ramped from 50 °C (hold for 2 min) at 10 °C/ min to 150 °C, and then at 35 °C/min to 270 °C. Quantification was done using the single ion monitoring method with three ions for each component. Detailed conditions for the different target analytes are summarized in Table S1. LC–MS to analyse samples for linalool hydroperoxide on prolonged ageing All LC–MS/MS measurements were performed using a Dionex UltiMate XRS 3000 high-performance LC system coupled to a Q Exactive Orbitrap mass spectrometer (Thermo Scientific) with electrospray ionization in positive mode. For LC separation, an XBridge C18 column with dimensions of 2.1 mm×50 mm and a particle size of 2.5 μm with a 2.1 mm×10 mm precolumn of the same material (Waters) was used. The flow rate was 200 μl/min. Eluent A consisted of water containing 0.1 % formic acid, and eluent B consisted of methanol containing 0.1 % formic acid. A linear gradient was run from 90 % eluent A (hold for 1 min) to 100 % eluent B within 13 min (held for 2 min), back to 90 % eluent A within 2 min, followed by 2 min equilibration. The injection volume of the sample was 1 μl. The resolution of the high-resolution MS (HR-MS) spectra was set to 70,000. The mass accuracy was below 5 ppm. Data-dependent high-resolution product ion spectra (HR-MS/MS) were recorded at a resolution of 17,500. The ion source parameters adjusted were as follows: sheath gas flow (35 arbitrary units), auxiliary gas flow (15 arbitrary units), capillary temperature (300 °C) and source voltage (4 kV). Fragmentation was obtained from dissociation in an octopole collision cell using a higher-energy collision dissociation setting of 35 (arbitrary units). Detailed conditions for the

6169

different target analytes are summarized in Table S1. The LC– MS method was applied only to the samples stored for 9 months and the fragrances and antiperspirants recalled from consumers, as it was not available at the beginning of the study. To validate the method and calculate the analytical recovery, a number of spiking experiments with linalool hydroperoxide added to the fragrances described in Table 1 were done (see Table S2). Identification and quantification With use of the LC–MS method, 2 is chromatographically not separated from 3, and we thus report the values for the sum of the isomers, which we refer to as ‘linalool hydroperoxide’. Identification was based on (1) the retention time, (2) full-scan mass analysis of the sodium adduct ion of linalool hydroperoxide using high resolution and high mass accuracy, (3) MS/ MS analysis of this ion (209.1148) and its main fragment (177.0886) using high resolution and high mass accuracy, and (4) the intensity ratio of this fragment to its precursor ion. Quantification was done on signals obtained from the extracted mass of 209.1148 within a 5-ppm mass accuracy window (209.1138–209.1158) at the given retention time. The instrumental limit of detection (LOD) for linalool hydroperoxide using the LC–MS method is 0.5 μg/g for the reference standard and 1 μg/g for samples containing matrix. The limit of quantification (LOQ) is in the region of 3 μg/g (see Fig. 2). For 7hydroxylinalool 7, 6-hydroxylinalool 8, and trans/cis-linalool oxide 5 (Fig. 1) analysed by GC–MS, the LOD is 1 μg/g and the LOQ is 3 μg/g on the basis of the calibration range of the reference sample without sample matrix. For the GC–MS method, the instrumental LOD for linalool hydroperoxide is in the region of 16 μg/g for the reference standard (i.e. without sample matrix) on the basis of the lowest observable signal from the calibration function. The LOQ is around 50 μg/g (see Fig. S1). Data evaluation All data were analysed using Xcalibur 2.2 (Thermo Scientific) including QuanBrowser.

Results Detection of linalool hydroperoxide by different analytical methods At the beginning of the current study, only GC–MS was available to detect linalool hydroperoxide, with a LOQ of 50 μg/g (see Fig. S1). This limited sensitivity of GC–MS for linalool hydroperoxide was found to be too low to detect the typical levels in fragrance samples tested in the course of these investigations. Therefore, an additional method based on LC–

6170

HR-MS was developed. With use of this method, linalool hydroperoxide can be selectively detected and quantified by its exact mass ion [M+Na]+ =209.1148 (mass accuracy below 5 ppm). Specific confirmation of peak identification is possible on the basis of the retention time, the precursor ion [M+ Na] + = 209.1148 and its main fragment at 177.0886 ([C9H14O2 +Na]+). Figure 2 shows a chromatogram and the LC–MS/MS mass spectrum and indicates the analytical sensitivity and accuracy of the method in ethanol/water and when the hydroperoxide is added to experimental fragrance A50 used in the stability study. The LC–MS method is not linear at high concentrations, but shows good linearity in the range from 1 to 10 μg/g. In addition, the standard deviations from triplicate injections indicate that the reproducibility of this analytical method is high. In fragrance A50, the response is only slightly reduced. Stability of linalool and formation of primary and secondary oxidation products in a hydroalcoholic formulation The stability of synthetic linalool and a commercial natural linalool product obtained from Cinnamomum camphora (known by the trivial name ‘Ho wood’) in a typical hydroalcoholic formulation was first tested. Table 2 lists the analytical results after 2 months, and Table S3 gives the data for the 1-month time point for comparison (GC–MS/FID method 1). No significant degradation of linalool was observed during this period, independently of the presence of stabilizing agents or the repeated opening of the bottles. The secondary oxidation products linalool oxides 5 and 6 as well as 7-hydroxylinalool 7 (Fig. 1) were detected below 6 μg/g in samples containing synthetic linalool. Compounds 5 and 6 were detected at around 300 μg/g in the formulation containing natural linalool. These samples also contained detectable amounts of linalool hydroperoxide as determined by GC–MS. We did not see an increase over time or due to temperature, exposure to air or the absence of stabilizing agents for the oxidation products in both linalool products. A 2-month stability study at 45 °C reflecting the storage of a product for 8 months at room temperature is a standard procedure in industry. However, to investigate a scenario with higher oxidation risk, the study was prolonged for a further 7 months for the samples in half-empty bottles opened every second week (i.e. the samples with the highest oxidation risk). These samples were then analysed in more detail both by GC– MS method 2 and by the LC–HR-MS method, allowing sensitive detection of hydroperoxides (Table 3). The concentration of linalool had still not significantly decreased below the starting level of 100,000 μg/g after 9 months. Compound 2 was detected in all the samples with the more sensitive LC– MS method. It was found at 13–18 μg/g in the samples made from synthetic linalool and at 82–97 μg/g in the samples made from naturally derived linalool. The secondary oxidation

S. Kern et al.

product 5 was not detectable or was detected only in traces in the samples made from synthetic linalool, and it was found at around 80–150 μg/g in the samples made from naturally derived linalool. Similarly to the results obtained from evaluation at 1 and 2 months, neither the presence of stabilizers nor the storage temperature affected the stability of the parent linalool, and we did not observe an effect of these experimental parameters on the levels of the hydroperoxide or secondary oxidation products detected. Stability of linalool and formation of primary and secondary oxidation products in complex hydroalcoholic fragrances Linalool is never used in isolation, but always in the context of other fragrance materials. The fragrances described in Table 1 were thus subjected to a stability study as described earlier. Table S4 lists the results for the recovery of linalool from these samples after 1 and 2 months (GC–MS/FID method 1). We did not detect a significant decrease of the initial theoretical linalool content, neither owing to elevated temperature nor owing to exposure to oxygen in partly filled and repeatedly opened bottles. In commercial fragrance C, we found secondary oxidation product 5 at 6–8 μg/g in samples stored at 45 °C, but did not find it in the 5 °C control. The half-empty, repeatedly opened bottles were further incubated for another 7 months (still opened every 14 days), and then their contents were analysed in detail with the LC– MS method and GC–MS method 2 (Table 4). Overall, the samples still contained the theoretically expected levels of linalool in this independent analysis, indicating high stability also over the 9-month test. Linalool hydroperoxide was detected in the region of 2 μg/g in fragrances A5, A20 and A50, although these values were below the LOQ, whereas the secondary oxidation product 5 was not found. Commercial fragrance C contained compound 2 at around 2 μg/g when stored at 45 °C, but its level was 59 μg/g in the samples stored at 5 °C. In contrast, the sample at 5 °C contained only traces of 5, and its level was clearly elevated at 45 °C as observed before in the analysis at 2 months. This may indicate that this commercial fragrance contained low amounts of 2 on purchase, and this degrades to 5 on product storage at elevated temperature. Stability of linalool hydroperoxide in the different product matrices and at different temperatures Stability testing at 45 °C is routine in industry, and this approach was used here as it is known to enhance rates of lipid peroxidation (it is also widely used in oxidative stability studies on unsaturated oils). Yet the possibility exists that at this temperature the hydroperoxides may not be stable. We thus performed a 3-month stability test on the synthetic hydroperoxides. As shown in Fig. S2, linalool hydroperoxide is

Detection of potentially skin sensitizing hydroperoxides

6171

Fig. 2 Analysis of linalool hydroperoxide by liquid chromatography (LC)–high-resolution mass spectrometry (MS). Calibrations curves in 80 % ethanol/20 % water (a) and in fragrance A50 (b). The low concentration range is zoomed in the insert. c The LC–MS chromatogram of the

extracted mass ion m/z 209.1148 and (d) and the MS/MS spectrum of m/z 209.11 of the reference sample containing 2 at 18 μg/g and 3 at 22 μg/g. RT retention time

intrinsically stable at 45 °C (no major temperaturedependent decline in fragrance A50 and in 10 % synthetic linalool). This control experiment validates the results obtained for the stability of linalool and experimental fragrances A5, A20 and A50 at elevated temperatures, as potential accumulation of the hydroperoxide would not be masked by its limited temperature stability. The data also indicate that the hydroperoxide is surprisingly stable in fine fragrances stored at room temperature.

presence of oxidation products under real-life conditions and at ambient temperatures, we conducted an additional study on samples recalled from consumers. Again, to study samples with significant oxidation risk, participants were asked to bring in at least 2-year-old and partly emptied fragrance bottles, and some samples were considerably older (Table S5). All the samples were analysed for linalool content (GC–MS method 2), the hydroperoxide (LC–MS) and secondary oxidation product 5 (GC–MS method 2) (Table S5, Fig. 3a). The samples contained a geometric mean linalool concentration of 1,888 μg/g (median 2,430 μg/g). Considering that most hydroalcoholic fragrances contain a perfume level of around 10 %, this is close to the median of 23,000 μg/g for linalool found in the total fragrance content of the 861 market products analysed (see “Materials and methods”), indicating that the 39 samples are representative of the market. We identified linalool and linalool hydroperoxide in 38 and 33 of 39 fragrances, respectively. The hydroperoxide levels determined on the basis of an external calibration curve ranged from below the LOQ to 49 μg/g, with a geometric mean

Detection of linalool and oxidation products in aged fragrances recalled from consumers The studies reported so far allowed monitoring of the fate of linalool under fully controlled conditions, and made it possible to study individual factors, such as temperature, time, stabilizing agents, bottle opening, admixture of other fragrance ingredients and partial filling. However, to glimpse the

6172

S. Kern et al.

Table 2 Stability of pure linalool formulated as a hydroalcoholic fragrance in a 2-month standardized stability test Linalool type

Storage Stabilizers Half Half Linalool temperature (°C) full full/opened (μg/g)a

Linalool7cis/trans-Linalool 7-Hydroxylinalool hydroperoxide (μg/g)b oxide (μg/g) (μg/g)

Synthetic Synthetic Synthetic Synthetic

45 45 45 5

+ + + +

105,091±33 105,978±7,708 97,330±1,666 100,003±1,405

Detection of potentially skin sensitizing hydroperoxides of linalool in fragranced products.

On prolonged exposure to air, linalool can form sensitizing hydroperoxides. Positive hydroperoxide patch tests in dermatitis patients have frequently ...
609KB Sizes 0 Downloads 5 Views