Occurrence of and Dermal Exposure to Benzene, Toluene and Styrene in Sunscreen Products Marketed in the United States

Sample Collection

Fifty sunscreen products were purchased from retail and grocery stores (n = 39) in New York State and online sources (n = 11) during April – July 2021. Further details of the samples are shown in Figure 1. Information regarding manufacturers, brand name, lot number, and country of manufacture were obtained from product labels (Table S1). The sunscreen products represented 44 brands and 94% of them (n = 47) were manufactured in the United States whereas two samples were manufactured in Japan and one in Korea. Twenty sunscreen products were labeled as SPF < 50, 20 were SPF −50 and 10 were SPF > 50 (Table S2). The sunscreen products contained ZnO (38% of the samples), TiO₂ (20%), octisalate (52%), octocrylene (52%), avobenzone (52%), homosalate (50%), oxybenzone (10%), octyl salicylate (4%), octinoxate (4%) as active ingredients. Twenty-seven sunscreens listed only organic UV filters (oxybenzone, octisalate, octocrylene, avobenzene) as active ingredients, whereas 16 of them contained only inorganic UV filters (ZnO and TiO₂). Four sunscreen products had a combination of organic and inorganic UV filters. Further details of the ingredients listed on the sunscreen products are shown in Table S2.

Chemicals and Reagents

Analytical standards of benzene, toluene and styrene of purities ≥99.9%, ≥99.9% and ≥98.0%, respectively, as well as three corresponding isotopically labeled internal standards (ISs: benzene-d₆ [purity: ≥99.0%], toluene-d₈ [≥99.0%] and styrene-d₈ [≥98.0%]) were purchased from Sigma-Aldrich (St. Louis, MO, USA). HPLC-grade acetonitrile (≥99.9%) and water were purchased from J.T. Baker (Center Valley, PA, USA).

Analysis of Benzene, Toluene and Styrene

Benzene, toluene and styrene concentrations were determined by following a method described by the United States FDA with slight modifications (US FDA, 2007). Briefly, 0.5 g of sunscreen was weighed into a 15-mL polypropylene (PP) tube (Falcon^™, Fisher Scientific, Waltham, MA, USA) and 4.95 mL of acetonitrile and 50 μL of 1000 ng/mL IS mixture (in acetonitrile) were added. The mixture was shaken in a reciprocal shaker (Eberbach Corp., Ann Arbor, MI, USA) at 280 oscillations/min for 10 min. The supernatant was transferred into a new PP tube, and a 1-mL aliquot of the supernatant was transferred into a glass vial for high-resolution gas chromatography coupled with high-resolution mass spectrometry (HRGC-HRMS) analysis.

Identification and quantification of target analytes were accomplished using an Agilent 7890A HRGC (Agilent Technologies, Santa Clara, CA, USA) coupled with a JEOL JMS-800 D Ultra FOCUS^™ HRMS (JEOL USA, Inc., Peabody, MA, USA). Chromatographic separation was accomplished using a DB-Select 624 Ultra Inert column (30 m length, 0.250 mm i.d., 1.40 μm film thickness; Agilent Technologies, Santa Clara, CA, USA). Helium was used as the carrier gas. Two microliters of the sample extracts were injected in the pulsed split mode (50:1 split ratio with pulsed pressure of 25 psi for 0.5 min and split flow rate of 50 mL/min). The carrier gas flow was maintained at 1 mL/min. The inlet temperature was set at 250°C. The column oven temperature was programmed as follows: initial hold at 40° C for 5 min, increased to 240°C at a rate of 30°C/min, and then held for 4 min. The equilibration time was 0.5 min and the total run time was 15.667 min. The MS ion source temperature was 230°C. An electron impact positive ionization selected ion monitoring was used for compound acquisition. The ions were monitored at m/z 78.0464 for benzene, m/z 84.0841 benzene-d₆, m/z 92.0620 for toluene, m/z 100.1123 for toluene-d₈, m/z 104.0620 for styrene and m/z 112.1123 for styrene-d₈. Polyfluorokerosene was used as the reference compound for the calibration of HRMS. Quantification of the target analytes was achieved using an isotope dilution method, by taking the ratio of absolute response of each native analyte to that of corresponding isotope-labeled IS. Peak integration, calibration and quantitation were performed using DioK software (JEOL USA, Inc., Peabody, MA, USA).

Quality Assurance and Quality Control (QA/QC)

Analysts took care not to use any lotions or sunscreens before and during the analysis. Powder-free, Purple Nitrile^™ gloves (Kimtech^™; Kimberly-Clark Corporation, Neenah, WI, USA) were worn by analysts during sample preparation and analysis. Additional precautions such as rapid processing of samples and sealing of vials quickly after transferring the samples from containers, to minimize volatilization loss, were taken. Addition of labelled IS prior to extraction captures any lose during sample preparation and that was accounted for from recovery rate calculations. A procedural blank (n = 10) was prepared using acetonitrile, which was subjected to entire analytical procedure to check for the contamination arising from laboratory materials and solvents. The mean absolute concentrations of benzene and toluene found in procedural blanks were 6.20 ng/mL (range = 4.52–8.77 ng/mL) and 29.7 ng/mL (range = 25.5–33.9 ng/mL), respectively (Supplementary Figure, Figure S1). Styrene was not found in procedural blanks. The mean concentrations of benzene and toluene found in procedural blanks were subtracted from concentrations measured in samples. Concentrations of benzene and toluene were carefully monitored and only those samples with responses at least twice those found in blanks were reported as detectable values. A sunscreen sample was fortified with target analytes at concentrations of 10 and 100 ng/mL and analyzed through the entire procedure. Recoveries of benzene, toluene and styrene in the fortified sample were in the ranges of 70–130% with a mean recovery of 99%, 107% and 101%, respectively. A 10-point, linear, non-forced-through-zero, standard calibration curve was prepared at a concentration range of 1–1000 ng/mL (regression coefficient >0.99 for each analyte) for quantification. The limits of detection (LODs) were calculated using the International Council for Harmonization (ICH) guidelines for analytical procedures, based on the standard deviation of the response (Sy) of the calibration curve and the slope of the calibration curve (S) at levels approximating the LOD according to the formula; LOD = 3.3 * (Sy/S) (ICH, 2005). The LODs of benzene, toluene and styrene were 0.100, 0.009 and 0.008 ng/g, respectively. For the calculation of mean and median concentrations, analyte concentrations below the LOD were substituted with the LOD value divided by the square root of 2 (Pal et al., 2022a).

Exposure and Risk Assessments

The details of exposure and risk assessment calculations are provided, with an example, in the Supplementary Text (Text S1). We estimated the dermal exposure of benzene, toluene and styrene from sunscreen products using the following equation (1) (Pal et al., 2022a; US EPA, 2011):

Where DED is the potential dermal exposure dose (ng/kg-bw/d), C is the concentration (ng/g) of the target analyte in each sunscreen sample, A is the average amount of sunscreen applied per day (g/day) in the United States and BW is the body weight (kg). The sunscreen usage rate was 85 g/day, based on the values recommended (between 10 a.m. and 4 p.m.) for a typical summer day outing in a beach (Li and Kannan, 2022). The average body weights of children/teenagers and adults were 57 and 80 kg, respectively, according to the US EPA’s exposure factors handbook (Pal et al., 2022a; ATSDR, 2016). The dermal permeability of benzene can increase significantly with sunscreen application (Nakai et al., 1997), as many active and inactive ingredients (containing surfactants, hydrocarbons, alcohols, alkanes, esters, amines and amides) present in these products can act as vehicles or penetration enhancers to increase permeation of VOCs across the skin. Several previous studies have examined dermal permeation of benzene (Adami et al., 2006; Hanke et al., 2000; Blank and McAuliffe, 1985). However, data pertaining to dermal uptake of benzene in the presence of complex organic solvent mixtures such as sunscreen products is still unavailable (Williams et al., 2011). In absence of relevant permeation rates for VOCs from sunscreens, we assumed a conservative value of 100% dermal permeation in our calculations.

The non-cancer risks from dermal exposure to benzene, toluene and styrene were calculated as hazard quotients (HQs) using the following equation (2) (Pal et al., 2022a; Lin et al., 2020; US EPA, 2011; ATSDR, 2016):

Where DED is the dermal exposure dose (ng/kg-bw/d) and RfD is the reference dose (ng/kg-bw/d). The RfDs of benzene, toluene and styrene were 4 × 10³, 8 × 10⁴ and 2 × 10⁵ ng/kg-bw/d, established on the basis of decreased lymphocyte count, increased kidney weight and hematological effects, respectively (Lim et al., 2014; Chaiklieng et al., 2019; Qin et al., 2019). A HQ value >1 suggests potential non-cancer risks.

The cancer risk (CR) from dermal exposure to benzene was calculated using the following equation (3) (Pal et al., 2022a; Lim et al., 2014; US EPA, 2011; ATSDR, 2016) :

Where CR is the cancer risk, DED is the daily dermal exposure dose (ng/kg-bw/d) and DSF is the dermal cancer slope factor (kg-bw/d/ng) for benzene, which is 5 × 10⁻⁸ kg-bw/d/ng, derived based on data from occupationally exposed leukemia patients (US EPA, 2012). The CR denotes the upper-bound excess lifetime cancer risk estimated to result from continuous exposure to an agent over a lifetime, and risk-specific doses are derived from the slope factor to estimate the dose associated with a specific risk level, for example, one-in-a-million (i.e., 1.0 × 10⁻⁶) (Qin et al., 2019). The CR calculation for benzene was based on the assumption that sunscreens were used daily over the lifetime. Since toluene and styrene have no known carcinogenicity in humans, cancer risks were not calculated for these two compounds.

Statistical Analyses

All statistical analyses were performed using Excel 2019 (Microsoft Office Professional Plus, Redmond, WA, USA), SPSS 19.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism v.9.1.1 (GraphPad Software, San Diego, CA, USA). The significance level was set at p < 0.05. Shapiro-Wilk test was applied to test the normality of data. Since majority of the variables were not normally distributed, non-parametric tests were applied. The analyte concentrations in different groups of products were compared using Mann-Whitney test and Kruskal-Wallis test. Spearman’s rank correlation was used to determine relationships among VOCs. ∑3 VOCs represents sum concentrations of benzene, toluene and styrene.

Concentrations of Benzene, Toluene and Styrene

Benzene, toluene and styrene were found in 80%, 92% and 58% of the sunscreens analyzed, respectively. The distribution of benzene, toluene and styrene concentrations found in 50 sunscreens is shown in Figure S2. The mean concentrations of benzene, toluene, styrene and ∑3 VOCs in sunscreens were 45.8, 89.0, 161 and 296 ng/g, respectively (Table 1). The range, geometric mean and median ∑3 VOCs concentrations in sunscreens were 1.60–2005, 131 and 134 ng/g, respectively (Table 1). The distribution of benzene in various categories of sunscreens stratified on the basis of active ingredients (e.g., organic or inorganic) did not vary significantly (Figure S3). Furthermore, benzene was also found in sunscreens with inorganic UV filters as active ingredients (Figure S3). These results imply that organic active ingredients (e.g., benzophenone) are not the only sources of benzene in sunscreens. The consumer products can have multiple sources of VOCs that can include petroleum derived inactive ingredients, gelling agents such as carbomers, and/or unlabeled fragrance ingredients (Hudspeth et al., 2022).

Prior to our study, Valisure, a Connecticut-based pharmacy product testing company, reported the occurrence of benzene in sunscreen and after-sun care products (Hudspeth et al., 2022; Valisure, 2021). That study reported that ~10% of the sunscreens contained benzene at concentrations > 2000 ng/g (Hudspeth et al., 2022). However, that study did not report actual benzene concentrations in products, but simply categorized them as percentage of samples that contained benzene above or below 2000 ng/g. Thus, our study is the first to quantitively determine benzene in sunscreens.

The mean concentrations of benzene, toluene, styrene and ∑3 VOCs in sunscreen products manufactured in the United States (n = 47) were 46.0, 84.7, 171 and 302 ng/g, respectively (Table S1); the corresponding mean concentrations in two samples from Japan were 61.8, 99.1, 0.006 and 161 ng/g and that in a sample from Korea were 5.76, 272, 0.006 and 278 ng/g, respectively (Table S1). Although the number of samples originated from Japan and Korea was small, they consistently showed low concentrations of styrene in comparison to those found in products from the United States.

The concentrations of benzene and toluene in sunscreens were positively correlated (ρ = 0.247, p < 0.05), whereas the concentrations of benzene and styrene (ρ = 0.224, p > 0.05) or toluene and styrene (ρ = 0.096, p > 0.05) showed no significant correlations (Figure 2). These results suggest that the sources of the three VOCs in sunscreens vary. The sunscreens contain a number of active and inactive ingredients (Table S2) and these chemicals are the sources of VOCs. However, the concentrations of three VOCs did not vary with the type of active ingredient or SPF of sunscreens (p > 0.05) (Figure 3; Figure S4). Lack of significant correlations (p > 0.05) among the three VOCs and the type of active ingredients suggest that VOCs arise from multiple ingredients present in sunscreen formulations (Figure S4). It is worth noting that styrene/acrylates copolymers have been used in cosmetics and personal care products for their water-resistant properties (Cosmetic Ingredient Review 2002). This may explain higher concentrations styrene than those of benzene and toluene in sunscreens from the United States. Information on fate of these copolymers in sunscreens is not known and presence of styrene in these products can be due to degradation of these copolymers.

Exposure and Risk Assessments

For exposure assessment, the AAD recommended sunscreen application dose of 85 g on a typical 6–7 h outing on a summer day in beach was used. The calculated mean DEDs of benzene, toluene, styrene and ∑3 VOCs for children/teenagers were 68.3, 133, 240 and 441 ng/kg-bw/d, respectively (Table 1); the corresponding mean values for adults were 48.7, 94.6, 171 and 314 ng/kg-bw/d (Table S3). Children/teenagers had slightly higher exposure doses than adults which can be explained by the smaller bodyweights of the former. A previous study estimated average exposure dose to benzene of 320 μg/day (for a person weighing 70 kg, it would be 4570 ng/kg-bw/d) in the United States with indoor air inhalation accounting for >90% of the exposure (Wallace, 1989; WHO, 2000). Cigarette smoking can add as much as 1800 μg/day and passive smoking at 50 μg/day of benzene exposure (Wallace, 1989; WHO, 2000). The mean dermal exposure doses of benzene calculated from sunscreen products in the current study were two orders of magnitude lower.

The non-cancer risk quotients (i.e., HQ) from benzene in all sunscreens were below unity (Figure 4). The HQ values of benzene ranged from 2.61 × 10⁻⁶ to 3.21 × 10⁻¹ (mean = 1.71 × 10⁻²) for children/teenagers and from 1.86 × 10⁻⁶ to 2.29 × 10⁻¹ (mean = 1.22 × 10⁻²) for adults (Table S4). However, the cancer risk from dermal exposure to benzene exceeded the benchmark of one-in-a-million (i.e., 1.0 × 10⁻⁶) for 44% of the sunscreens in children/teenagers and 38% of the sunscreens in adults (Figure 4). The overall cancer risk estimates of benzene in sunscreens ranged from 5.22 × 10⁻¹⁰ to 6.43 × 10⁻⁵ (mean = 3.42 ×10⁻⁶) for children/teenagers and from 3.72 × 10⁻¹⁰ to 4.58 × 10⁻⁵ (mean = 2.43 × 10⁻⁶) for adults (Table S4). Chronic exposure to benzene is linked to decreased red blood cell counts and increased blood cancer risks (Li et al., 2021; Pal and Kannan, 2023). Thus studies describing chronic exposure to VOCs present in sunscreens and associated health effects need further investigation.

The HQ for non-carcinogenic risks of toluene and styrene in all sunscreen samples were also below unity. The HQ values of toluene ranged from 1.12 × 10⁻⁷ to 8.76 × 10⁻³ (mean = 1.66 × 10⁻³) for children/teenagers and from 7.97 × 10⁻⁸ to 6.24 × 10⁻³ (mean = 1.18 × 10⁻³) for adults (Table S4). Whereas the HQ for styrene ranged from 4.47 × 10⁻⁸ to 1.23 × 10⁻² (mean = 1.20 × 10⁻³) for children/teenagers and from 3.19 × 10⁻⁸ to 8.77 × 10⁻³ (mean = 8.56 × 10⁻⁴) for adults (Table S4). Chronic exposure to toluene is linked to reproductive and respiratory complications, whereas that of styrene poses risks to central nervous system and kidney (Li et al., 2021).