DETERMINATION OF LEAD IN LIPSTICKS 165 lipsticks resulted in relative percent differences (RPD) of more than 10%, but only two lipsticks exhibited an absolute difference of more than 0.30 mg Pb/kg, suggesting few infl uences from the matrix. Repeatability was demonstrated by preparing lipsticks in duplicate. The average value of differences between duplicate preparations was less than the estimated precision of the method (0.05 mg Pb/kg). As another check of method performance and sample homoge- neity, the contractor was asked to collect a second lipstick of the same shade and lot number from each manufacturer. Of these pairs, 18 of the 19 results agreed within 0.10 mg Pb/kg and therefore showed good method repeatability and homogeneity of the lip- sticks. One lipstick initially showed a difference of more than 3.0 mg Pb/kg between tubes. However, a repeat analysis showed agreement within 0.3 mg Pb/kg, and the initial difference was attributed to a preparation batch with mislabeled analytical solutions. The discrepancy prompted the FDA to follow up with several additional repeat analyses and preparations. CHECK ANALYSES To make sure that correct values were reported, the contractor reanalyzed 10% of its ana- lytical solutions and the FDA repeated the preparations and analyses of several lipsticks, including the lipsticks with the 15 highest lead values. This process not only identifi ed the mislabeled preparation batch described above, but also revealed a second preparation batch whose lead values were calculated with an incorrect dilution factor. The data were corrected accordingly. STATISTICAL COMPARISON OF MEAN LEAD VALUES AMONG COMPANIES We found that Company G lipsticks had lead values signifi cantly lower than those from each of the other six companies. Lipsticks from Companies A, E, and H had signifi cantly lower mean lead values than lipsticks from both Companies C and D, but their mean lead values were not signifi cantly different from each other. The mean lead value for Company F lipsticks was only signifi cantly different (higher) than Company G’s. DISCUSSION The FDA is aware that two other groups have reported results for lead in lipsticks using the FDA’s method. One group from the U.S. analyzed 56 lip products including lip- sticks, lip glosses, moisturizing sticks, and lip stains, and reported an average of 0.9932 mg Pb/kg (9). An E.U. group surveyed 81 lipsticks and 32 lip glosses purchased in the E.U. in 2010 and 2011 and found an average lead content of 0.86 mg Pb/kg for the lip- sticks, with a range of 0.05 to 3.75 mg/kg (10). The lip glosses contained less lead (averaging 0.44 mg Pb/kg), which the authors considered to be a result of their lower pigment content. In our previous work (1), we similarly attributed higher lead content to higher mineral or pigment content. The authors illustrated the lower lead content of the

JOURNAL OF COSMETIC SCIENCE 166 lip glosses by plotting lead content versus product type. The curve for lip glosses fell about 1 mg Pb/kg below the curve for lipsticks. The E.U. group looked for trends relating lead content to lipstick shade and found that the lead content in the pink shades was highest, followed by purple, and then red. The authors did not provide an explanation for this trend, but it is possible to understand the trend for higher lead in shades with higher levels of mineral pigments acting as lighten- ing toners for the darker red shades. We found it diffi cult to categorize the lipstick shades we surveyed because the hues blend from one color to another, and we did not attempt to fi nd associations between lead content and shade. Lipsticks with lead content greater than 3.36 mg Pb/kg had a variety of shades. The E.U. group also compared lead content to lipstick price but did not fi nd any trend. We made no attempt to associate lead content with price, in part because the same shades are available at multiple prices. The higher mean lead values for Companies C and D may be a refl ection of the source of raw materials, but this would need further investigation and proprietary information from the lipstick manufacturers. It is interesting to note that one manufacturer (Company G) was able to attain signifi cantly lower lead values than the other companies (See Table II and the sta- tistical analyses). Both the U.S. and E.U. groups found an average lead content in lipsticks that is consistent with the average of 1.11 mg Pb/kg determined in this study. The E.U. group also concluded that lip products in the E.U. normally contain lead in the range of 1–2 mg Pb/kg. Our results are consistent with this fi nding, as is evident from Figure 1, which shows that nearly 90% of the lipsticks we surveyed contain less than 2 mg Pb/kg. CONCLUSIONS The mean lead content found in lipsticks from the initial (1.07 mg Pb/kg) and expanded (1.11 mg Pb/kg) surveys was approximately the same. The median lead content from the expanded survey was 0.89 mg Pb/kg. Only ten lipsticks (2.5%) in the expanded survey, and none in the initial survey, were found to contain lead above the 99% upper confi dence limit of 3.36 mg Pb/kg. Some statistically signifi cant associations between lead level and parent company were found. ACKNOWLEDGMENTS I thank the FDA’s William Mindak for his guidance and advice on analytical aspects, Dennis Ruggles for performing statistical calculations, Beth Meyers for editing the large table of lipstick and lead results, Julie Barrows for signifi cant editorial advice, Patricia Hansen for contract advice, and John Gasper for coordinating efforts between the FDA and the contractor. REFERENCES (1) N. M. Hepp, W. R. Mindak, and J. Cheng, Determination of total lead in lipstick: Development and validation of a microwave-assisted digestion, inductively coupled plasma-mass spectrometric method, J. Cosmet. Sci., 60, 405–414 (2009).