Concentrations and Potential Health Risks of Metals in Lip Products

Background: Metal content in lip products has been an issue of concern. Objectives: We measured lead and eight other metals in a convenience sample of 32 lip products used by young Asian women in Oakland, California, and assessed potential health risks related to estimated intakes of these metals. Methods: We analyzed lip products by inductively coupled plasma optical emission spectrometry and used previous estimates of lip product usage rates to determine daily oral intakes. We derived acceptable daily intakes (ADIs) based on information used to determine public health goals for exposure, and compared ADIs with estimated intakes to assess potential risks. Results: Most of the tested lip products contained high concentrations of titanium and aluminum. All examined products had detectable manganese. Lead was detected in 24 products (75%), with an average concentration of 0.36 ± 0.39 ppm, including one sample with 1.32 ppm. When used at the estimated average daily rate, estimated intakes were > 20% of ADIs derived for aluminum, cadmium, chromium, and manganese. In addition, average daily use of 10 products tested would result in chromium intake exceeding our estimated ADI for chromium. For high rates of product use (above the 95th percentile), the percentages of samples with estimated metal intakes exceeding ADIs were 3% for aluminum, 68% for chromium, and 22% for manganese. Estimated intakes of lead were < 20% of ADIs for average and high use. Conclusions: Cosmetics safety should be assessed not only by the presence of hazardous contents, but also by comparing estimated exposures with health-based standards. In addition to lead, metals such as aluminum, cadmium, chromium, and manganese require further investigation.

Cosmetic products contain thousands of chemicals, some of which have been associ ated with reproductive, developmental, or other health effects based on human or animal studies, including phthalates, formaldehyde, methylene chloride, acetone, acetonitrile, methacrylates, toluene, xylene, ethyl ether, and lead (Pb) (De Cock et al. 2012;Dodson et al. 2012;Ferret et al. 2012;Heisterberg 2012;Lefebvre et al. 2012;Moyer and Nixon 2012;Ulker et al. 2012). Lip products have been suggested as a particular concern because of the potential for exposure through inges tion (Loretz et al. 2005).
In October 2007, the Campaign for Safe Cosmetics tested 33 popular brands of lipsticks and reported that 61% contained Pb, with levels up to 0.65 parts per million, indicating a cause for concern (Campaign for Safe Cosmetics 2007). Although the Campaign for Safe Cosmetics report was not peer reviewed, it brought attention to the issue of chemicals in cosmetic and personal care products and their safety. Since then, two other studies evaluated Pb in eye shad ows and lipsticks: a U.S. Food and Drug Administration (FDA) study that detected Pb in all tested lipsticks (Hepp et al. 2009) and a study (AlSaleh et al. 2009) that identi fied several cosmetic products containing Pb > 20 ppm, the FDA limit of Pb as an impu rity in color additives for cosmetics (FDA 2011). Studies conducted in other countries have also detected Pb and cadmium (Cd) in some lipstick samples (AdepojuBello et al. 2012;Brandao et al. 2012;Gondal et al. 2010;Solidum and Peji 2011).
In the present study we extended the FDA study and the Campaign for Safe Cosmetics by testing for Pb, aluminum (Al), Cd, cobalt (Co), chromium (Cr), copper (Cu), man ganese (Mn), nickel (Ni), and titanium (Ti) in lipsticks and lip glosses used by young women, estimating potential daily intakes, and comparing the estimates to existing health guidelines.

Method
Sample collection. A convenience sample of lipsticks and lip glosses was selected based on information provided by 12 members of the Asian Communities for Reproductive Justice (ACRJ) Youth Program, a group of Asian girls 14-19 years of age who lived in low income neighborhoods in Oakland, California. Specifically, the girls were asked to record the brand and product names of the lipsticks and lip glosses they were carrying and had in their bathrooms at home, which represented prod ucts used by their sisters as well. The reported products were then purchased by researchers at a chain drug store (26 products), a major department store (4 products), and a chain specialty store (2 products).
All methods were approved by the University of California, Berkeley, Institutional Review Board (IRB). The young women who provided information on the lip products they used, and their parents or guardians, signed an informed consent form approved by the University of California, Berkeley, IRB before the study. We have complied with all appli cable IRB regulations and requirements.
Analytical method. Sample analysis fol lowed National Institute for Occupational Safety and Health (NIOSH) standard method for metals (Method 7303: Hot Block/HCl/HNO3 Digestion) with slight modifications (NIOSH 2003). Approximately 0.5 g of each sample was transferred into a clean, 50mL hot block digestion tube and digested with 2.0 mL concentrated nitric acid (HNO 3 ) on a block digester (LACHAT Instruments, Loveland, CO) at 130°C for 15 hr, with the tubes covered with glass fun nels to allow for nitric acid reflux during the digestion. Samples were diluted to 12.5 mL with distilled, deionized water and then filtered to remove material that did not completely dissolve, including waxy material floating on the top of the digest and white or light yellow precipitates that were most likely sili cates. Solutions were analyzed by inductively coupled plasma optical emission spectrometry (ICP/OES) (Optima 5300DV; PerkinElmer, Waltham, MA). Metals examined included Al, Cd, Co, Cr, Cu, Mn, Pb, Ni, and Ti. Reagent blanks and media blanks were also analyzed.
Data analysis. Measured metal concentra tions (parts per million, weight/weight) were Background: Metal content in lip products has been an issue of concern. oBjectives: We measured lead and eight other metals in a convenience sample of 32 lip products used by young Asian women in Oakland, California, and assessed potential health risks related to estimated intakes of these metals. Methods: We analyzed lip products by inductively coupled plasma optical emission spectrometry and used previous estimates of lip product usage rates to determine daily oral intakes. We derived acceptable daily intakes (ADIs) based on information used to determine public health goals for exposure, and compared ADIs with estimated intakes to assess potential risks. results: Most of the tested lip products contained high concentrations of titanium and aluminum. All examined products had detectable manganese. Lead was detected in 24 products (75%), with an average concentration of 0.36 ± 0.39 ppm, including one sample with 1.32 ppm. When used at the estimated average daily rate, estimated intakes were > 20% of ADIs derived for aluminum, cadmium, chromium, and manganese. In addition, average daily use of 10 products tested would result in chromium intake exceeding our estimated ADI for chromium. For high rates of product use (above the 95th percentile), the percentages of samples with estimated metal intakes exceeding ADIs were 3% for aluminum, 68% for chromium, and 22% for manganese. Estimated intakes of lead were < 20% of ADIs for average and high use. conclusions: Cosmetics safety should be assessed not only by the presence of hazardous contents, but also by comparing estimated exposures with health-based standards. In addition to lead, metals such as aluminum, cadmium, chromium, and manganese require further investigation.  to estimated daily metal intakes (micrograms per day) based on lip product use data from a study of cosmetic product use among 360 women (ages 19-65 years) from ten different U.S. geographical regions (Loretz et al. 2005). The investigators reported that on average the women used lipsticks 2.35 times per day (range, 0-20 times) and applied 10 mg of product at each use (range, 0-214 mg), resulting in average daily use of 24 mg of lip products (range, 0-214 mg; 95th percen tile = 87 mg/day). We assumed that all applied lip products were ingested, and thus estimated metal intakes for average use (24 mg/day) and high use (87 mg/day) of lip products.
Metals in cosmetic products are not currently regulated in the United States. Therefore, as a point of comparison for potential health risks, we estimated accept able daily intakes (ADIs) for Al, Cd, Cu, Ni, and Pb based on information used by the California Environmental Protection Agency (Cal/EPA) to determine Public Health Goals (PHGs) for drinking water (Cal/EPA 2001a, 2001b (Table 1). Specifically, we derived ADIs based on the following no observed adverse effect lev els (NOAELs) or lowest observed adverse effect levels (LOAELs) and uncertainty fac tors (UF) used to determine PHGs: Al NOAEL/LOAEL = 125 mg/day and UF = 100 (10 for duration of study, 10 for interindi vidual variation and sensitive subgroups) (Cal/EPA 2001a); Cd NOAEL = 19 µg/day and UF = 50 (5 for protecting sensi tive individuals, 10 for cancer risk due to oral exposure to Cd) (Cal/EPA 2006); Cu NOAEL = 426 µg/kgday and UF = 3 for uncertainties in study data (Cal/EPA 2008b); Ni NOAEL = 1.12 mg/kgday and UF = 1,000 (10 for interspecies extrapolation, 10 for intra species variability, and 10 for potential carcino genicity of oral exposure) (Cal/EPA 2001b); Pb NOAEL/LOAEL = 2.86 µg/day and UF = 3 for uncertainty in protectiveness of this level and small sample size (Cal/EPA 2009). For NOAEL/LOAEL reported according to body weight per day (micrograms per kilogram per day) we assumed a body weight of 50 kg for young Asian women to determine the ADI.
Our ADI for Cr was based on the PHG derived by the Cal/EPA for carcinogenic risks associated with hexavalent Cr accord ing to the standard risk calculation [concen tration = risk/(potency × dose)] (Cal/EPA 2011b), such that ADI = risk/P o , where risk = a default risk level of 1 in 1 million, or 10 -6 , and P o = 0.5/mg/kgday, the oral cancer potency for hexavalent Cr, result ing in an estimated ADI of 0.1 µg/day for a 50kg woman. Mn does not have a PHG, so we used the California Reference Exposure Level (REL) for systemic effects of Mn via inhalation of 20 m 3 of air per day (Cal/EPA 2008a), assuming that toxicokinetic differ ences between oral and inhalational routes of exposure were not significant. Co and Ti have no PHGs or RELs because they are not regu lated by California or federal standards; there fore we did not derive ADIs for these metals.
Finally, we compared estimated metal intakes via lip products to the derived accept able daily intakes. We derived relative intake indices (RIIs) for metals via lip products as a percentage of the ADI: Hence, for each metal, intake at the ADI would yield an RII of 100%. RIIs were cal culated assuming average use of lip products (intake of 24 mg of product/day) and high use (87 mg/day).

Results
Lip product information. We tested 32 indi vidual products in this preliminary study, including 8 lipsticks and 24 lip glosses sold by a total of 7 distinct companies. Prices ranged from $5.59 to $24. The tested products were representative of those used by young women in the ACRJ Youth Program.
Metal concentrations in lip products. Mn, Ti, and Al were detected in all examined prod ucts, with Ti and Al present in the highest concentrations of the metals tested (Table 2 and Figure 1). Pb was detected in 75% of products, with an average concentration of 0.36 ± 0.39 ppm (median, 0.151 ppm; maxi mum, 1.32 ppm). Approximately half (47%) of the samples contained Pb at concentrations higher than the FDArecommended maxi mum level of 0.1 ppm for Pb in candy likely to be consumed frequently by small children (FDA 2005). Co had the lowest average con centration among the examined metals (0.28 ± 0.31 ppm, mean ± SD). Metal concentra tions varied substantially across the products (Table 2). For example, product L1014 had the highest Cr concentration (9.72 ppm) and the second highest concentrations of Cd, Mn, and Pb (2.16, 35.3, and 1.25 ppm, respec tively). Products L1021 and L1029 had the highest concentrations of Pb (1.32 ppm) and Al (27,032 ppm), respectively, and both had high Cr and Mn levels. However, we did not observe clear patterns indicating that metal concentra tions were related to specific brands, product type (lipstick vs. lip gloss), color, or cost.
Estimated daily intakes via use of lip products. We converted measured metal con centrations (parts per million) in the indi vidual lip product samples to metal intakes (micrograms per day) based on usage pat terns reported by Loretz et al. (2005), assum ing average and high use (resulting in oral intake of 24 and 87 mg of product per day, respectively) ( Table 3). RIIs comparing esti mated metal intakes from lip products to the derived acceptable daily intake are presented in Figure 2. When used at the average daily rate (24 mg/day), estimated Cr intake from 10 products (31%) exceeded the ADI for Cr (RII > 100%). Estimates based on high use (87 mg/day) suggested exposures exceeding the ADI for Al in 1 sample (3% of the prod ucts tested), Cr in 22 samples (68%), and Mn in 7 samples (22%). Estimated intakes for Ni, Cu, and Pb were well below their ADIs even for high use. Estimated Pb intake for the product with the highest Pb concentration (product L1021) was 3% and 12% of the ADI assuming average and high use, respectively.

Discussion
This preliminary study of metal content in lip products suggests potential public health concerns. However, metals in cosmetic products are not currently regulated by the FDA. Although metal concentrations in lip products have been reported by studies both in the United States and in other countries (AdepojuBello et al. 2012;AlSaleh and AlEnazi 2011;AlSaleh et al. 2009;Brandao et al. 2012;Gondal et al. 2010;Gunduz and Akman 2013;Hepp et al. 2009;Solidum and Peji 2011), interpreting how reported concentrations may be related to potential health risk is challenging. We used California Public Health Goals for drinking water con taminants to derive healthbased standards for ingestion exposure from drinking water (Cal/EPA 2011a). In the calculation of PHG, a relative source contribution is applied to adjust intake of the contaminant from sources other than drinking water. Vulnerable popu lations such as infants and children are also considered by adjusting water consump tion rate (dose) for the different age groups. PHGs provide information on concentra tions of drinkingwater contaminants that pose no significant health risks if the water is consumed for a lifetime. Although they are not regulatory standards, PHGs are con sidered more health based than Maximum Contaminant Levels, which are manda tory drinkingwater standards that take into account not only health risks but also the fea sibility and cost of monitoring and maintain ing standards in drinkingwater supplies. In deriving the ADIs we did not account for metal intakes from other sources, nor did we consider potential age and sexrelated vul nerabilities, although they may have been partially accounted for by the uncertainty fac tors used in deriving PHGs. We used 20% of RII as an additional comparison point because 20% is a typical relative source con tribution value used in developing PHGs for  Corresponding numeric data are provided in Table 2 for all samples (n = 32). Boxes extend from the 25th to the 75th percentile, horizontal bars inside the boxes represent the median, diamonds represent the mean, whiskers extend to maximum and minimum observations within 1.5 times the length of the interquartile range above and below the 75th and 25th percentiles, respectively, and outliers are represented as circles.

Concentration (ppm)
volume 121 | number 6 | June 2013 • Environmental Health Perspectives the tested metals. We found that estimated intakes of Al, Cd, Cr, and Mn from some of the tested products were > 20% of their esti mated ADIs, assuming average daily use of lip products. The proportion of samples with RII > 20% substantially increased assuming high use (corresponding to the 95th percentile of lip product use based on a previous study) (Loretz et al. 2005), including 63% of the products tested for Al, 31% for Cd, 91% for Cr, and 66% for Mn. Cd and its compounds are known human carcinogens [International Agency for Research on Cancer (IARC) 1993]. Inhalation exposure of Cd has been associated with lung cancer and respiratory system damage (Chan et al. 1988;Davison et al. 1988;Nawrot et al. 2006;Smith et al. 1976;Stayner et al. 1992;Thun et al. 1985), and chronic oral expo sure may lead to kidney and bone impair ments (Åkesson et al. 2005;Nogawa et al. 1990). Animal studies indicate that young animals might absorb more Cd than adults and be more susceptible to bone impairments (Ogoshi et al. 1989). Animal studies also found that feeding rats and mice high level   Table 3 for all samples (n = 32). Boxes extend from the 25th to the 75th percentile, horizontal bars inside the boxes represent the median, diamonds represent the mean, whiskers extend to maximum and minimum observations within 1.5 times the length of the interquartile range above and below the 75th and 25th percentiles, respectively, and outliers are represented as circles. RII values represent the estimated daily intake for each metal as a percentage of the ADI values derived for this study. Although less than half (47%) of the tested lip products had detectable levels of Cd, our results sug gest that Cd intake could exceed 20% of our estimated ADI for Cd exposure via drinking water for one product assuming average use, and for 10 products assuming high use. Cr(VI) is a known human carcinogen; inhalation causes lung cancer and oral expo sure through drinking water has been linked with increased stomach tumors (ATSDR 2008b). Our measurements did not distin guish hexavalent Cr from less toxic forms, and the percentage of hexavalent Cr in the lip products is therefore unknown. However, high total estimated intakes of Cr from use of several lip products and the potential for additional exposure from other sources sug gests that Cr intake from lip products should be a priority for additional research.
Our acceptable daily intake value for Mn was derived from the Cal/EPA refer ence exposure limit for inhalational exposure because a public health goal is not avail able for Mn in drinking water. Inhalational exposure to high levels of Mn in occupa tional settings causes neurological effects in humans (ATSDR 2008c;Cook et al. 1974;Crossgrove and Zheng 2004). Although the evidence is inconclusive, Mn in drink ing water has been associated with neuro logical and neurobehavioral outcomes in children, which suggests that effects of oral exposure may be similar to effects associated with inhalational exposure (Bouchard et al. 2006;Kilburn 1987;Kondakis et al. 1989). Estimated Mn intake assuming high use of lip products exceeded our ADI value based on inhalational exposure for seven products (22% of tested products).
Although Pb was detected in 75% of the lip product samples, including 15 sam ples with concentrations higher than the FDA standard of 0.1 ppm for Pb in candy frequently consumed by children, RIIs for estimated Pb intakes appeared to be low compared with RIIs for Al, Cd, Cr, and Mn. Thus, although Pb in lip products has been intensively discussed (AlSaleh and AlEnazi 2011;AlSaleh et al. 2009;Bach and Newman 2010), other metals in the lip prod ucts should also be investigated. Nevertheless, it is generally accepted that there is no safe level of Pb intake (Centers for Disease Control and Prevention 2012), and the fed eral maximum contaminant level goal for Pb in drinking water is zero (U.S. Environmental Protection Agency 2012). The European Union Cosmetics Directive lists Cd, Cr, and Pb and their compounds as unacceptable constituents of cosmetic products (Salvador and Chisvert 2007).
The digestion method used in this study did not completely dissolve the lip product samples. The recent FDA study, which used a more complete digestion method to deter mine the total Pb concentration in lipsticks, reported an average Pb content of 1.07 ppm (range, 0.09-3.06 ppm) in 22 tested lipsticks (Hepp et al. 2009), in contrast with an aver age Pb concentration of 0.36 ppm (range, < 0.025-1.32 ppm) for our sample of 32 lip products. Differences between the studies may reflect variation in Pb content among the specific products tested, though incom plete digestion of our samples also may have reduced apparent concentrations relative to actual levels. Therefore, future studies should endeavor to measure total metal content to the degree possible.

Conclusion
Our data indicate the need for further studies to evaluate metal concentrations in lip products, as well as other cosmetics, and related potential health risks. In addition to Pb, metals such as Al, Cd, Cr, and Mn require further investigation. Cosmetics safety should be assessed not only by the presence of hazardous contents, but also by comparing estimated exposures with health based standards. This preliminary study of the metal content of 32 lip products suggests that toxic metals in cosmetics should be regulated to protect women's health in the United States, as has already been undertaken by the European Union through their Cosmetics Directive.