j. Soc. Cosmet. them., 39, 329-346 (November/December 1988) Quantitative assay of volatile and non-volatile N-nitrosarnines by gas chromatography with an electrolytic conductivity detector. I. Method development and assay of N-nitrosodiethanolarnine (NDELA) in creams and lotions STEVEN W. COLLIER, STANLEY R. MILSTEIN, and DONALD S. ORTH, The Andrew Jergens Company, 2535 Spring Grove Avenue, Cincinnati, OH 45214, and KOKA JAYASIMHULU, University of Cincinnati Medical Center, Institute of Environmental Health, Cincinnati, OH 45267. Received June 6, 1988. Presented in preliminary form at the 18th Ohio Valley Chromatography Symposium (Hueston Woods State Park), June 25, 1987. Synopsis A gas chromatographic method for the measurement of N-nitrosodiethanolamine (NDELA) has been devel- oped. Selective detection of the N-nitroso moiety was accomplished through the use of an electrolytic conductivity detector, employing a ceramic reaction tube operated in the non-catalytic mode and optimized for NDELA detection. Adequate volatility of NDELA was obtained by derivatization to the bis-trimethyl- silyl or bis-acetate ester. These structures were verified by mass spectrometry. The detection limit of NDELA was found to be 0.25 ng eluted as the bis-acetate ester and 0.38 ng when eluted as the bis-tri- methylsilyl ester. A protocol for the cleanup of cosmetic samples was applied successfully in tandem with this method to the analysis of a cosmetic cream and lotion, spiked in part with 20 and 100 ppb exogenous NDELA, respectively. The recovery of the NDELA from the cream matrix was calculated to be 81%. The sensitivity and relative low cost of the gas chromatography/electrolytic conductivity detector system makes it an attractive means for examining raw materials and finished cosmetic products for dedicated trace level nitrosamine screening. INTRODUCTION The contamination of cosmetics by N-nitrosamines is a problem of international con- cern. Industrial and regulatory agencies are interested in developing analytical methods for the detection of one or more nitrosamines in cosmetic products. In the United States, the Cosmetic, Toiletry & Fragrance Association (CTFA) is presently conducting round-robin tests in participation with industrial and government laboratories to vali- date a total nitrosamines methodology with a sensitivity of ca. 20 parts-per-billion (ppb). The Division of Colors and Cosmetics of the Food & Drug Administration (FDA) Steven W. Collier's present address is Food and Drug Administration, 200 C Street SW, Washington, DC 20204. 329

330 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS has developed a chemiluminescent method for total nitrosamines assessment with a limit of determination of 50 ppb for both polar nitrosamines such as N-nitrosodiethan- olamine (NDELA) and non-polar nitrosamines such as N-nitrosomethyl-tetradecyla- mine (1). A proposal before the West German government would ban secondary amines from cosmetic products in an effort to prevent the in situ formation of N-nitrosamines this same proposal would mandate the use of 99 + % grades of triethanolamine in cosmetic products and place a ceiling of 10 ppb NDELA on marketed cosmetic products (2). Most of the ca. 130 N-nitrosamines (NNA) tested to date have proven to be potent carcinogens in a number of animal species, including primates (3), and they have been ranked as second in carcinogenic potential only to the aflatoxins (4). The N-nitrosamine of chief concern to the cosmetic and toiletries industry and regulatory agencies is NDELA, whose carcinogenicity was first reported by Druckrey et al. in 1976 (5) and later identified as a potent animal carcinogen (6). NDELA's occurrence as a trace con- taminant in a number of diverse cosmetics and toiletries was first brought to the in- dustry's attention in 1977 by Fan et al. (7). NDELA, in common with other NNA, appears to function as a "procarcinogen," requiring metabolic activation prior to be- coming an active carcinogen. The proximate carcinogen, probably an electrophilic car- bonium-ion or its labile diazonium-ion precursor, is believed (8) to initiate carcino- genesis by a mutagenic event involving the alkylation of a nucleic acid guanine or cytosine residue, thus altering the transcription or translation of the normal chromo- somal genetic code of the affected cell(s). Recent Hansch-Taft quantitative structure-ac- tivity relationship (QSAR) studies (9) have lent support to a predominant mechanism of metabolic activation involving microsomal [3-oxidation, although N-nitroso-2-hydrox- ymorpholine, an or-oxidation product of NDELA, has also been detected after incuba- tion of NDELA with rat liver 9000 X g supernatant (10). Reports of widespread NDELA occurrence in a variety of cosmetics and toiletries by Fan et al. (11), the FDA (12), and Spiegelhalder and Preussman (13) have taken on added significance with the demonstration that NDELA is absorbed percutaneously (14-16) and has been found to be excreted in the urine of a human subject, following topical application of an NDELA-contaminated cosmetic (17). Qualitatively, topical administration of NDELA has been shown to induce the same types of tumors in hamsters as either oral swabbing or subcutaneous injection (18). In response to the problem of NNA in cosmetic products, the cosmetics and toiletries industry (19) and FDA (20) have recognized the importance of twin goals: 1) assessment of the conditions conducive to the formation of NNA in raw materials and finished products and 2) inhibition of NNA in cosmetics and toiletries. The kinetics of forma- tion and methods of inhibition of NDELA in oil-in-water emulsions (21), nonaqueous triethanolamine systems (22), and NNA formation in the presence of micelies have been discussed (23). Several analytical methods have been used to quantitate the concentra- tion of NDELA in cosmetics. The thermal energy analyzer (TEA) has been employed widely as a sensitive and selective detector for NDELA when coupled with high perfor- mance liquid chromatography (HPLC) (11,24,25) and gas chromatography (GC) (26). HPLC in combination with ultraviolet spectrophotometric detection (27-30) or differ- ential pulse polarography (DPP) (31,32), and GC in combination with electron capture detection (ECD) (33,34) and flame ionization detection (FID) (35) have also been used