Chlorinated Paraffins in Human Milk from Urban Sites in China, Sweden, and Norway

Short-, medium-, and long-chain chlorinated paraffins (SCCPs, MCCPs, and LCCPs) were analyzed in human milk from the Yangtze River Delta (YRD) and Scandinavia. Individual samples were collected from Shanghai, Jiaxing, and Shaoxing (China), Stockholm (Sweden), and Bodø (Norway) between 2010 and 2016. Mean concentrations (range) of SCCPs, MCCPs, and LCCPs in samples from the YRD were 124 [<limit of detection (LOD)-676], 146 (<LOD-1260), and 19.1 (<LOD-184) ng g–1 fat, respectively, all of which were significantly (p < 0.05) higher than 15.9 (<LOD-120), 45.0 (<LOD-311), and 5.50 (<LOD-29.0) ng g–1 fat, respectively, in samples from Scandinavia. MCCPs predominate in most samples, and LCCP concentrations exceed reported for polybrominated diphenyl ethers in human milk from the same regions. This study is the first to confirm LCCP exposure via breastfeeding. Principal component analysis showed that the YRD samples were more influenced by SCCPs than the Scandinavian samples, which mirror different exposures to CPs between the regions. Because of a large variation in concentrations among individuals, SCCP intake via breastfeeding indicated a potential health concern in the 90th percentile among Chinese infants. Further, CP concentrations in the YRD samples from first-time mothers were on average three times higher than from second-time mothers. In order to limit the worldwide CP contamination, the inclusion of SCCPs as persistent organic pollutants in the Stockholm Convention needs to be followed up, with the inclusion of MCCPs and LCCPs as well.


Sample storage
All samples were collected 2-12 weeks after giving birth. The target volume was at least 50 mL from each mother. Samples were subdivided into several portions and stored in the freezer below -18°C in glass bottles until chemical analysis. The Swedish samples were initially stored at -18ºC in plastic bags and bottles. Upon being shipped to the Environmental Specimen Bank at the Swedish Museum of Natural History in Stockholm, samples were then thawed, transferred to pre-washed glass bottles with lids covered with aluminum foil, and stored at -20ºC prior to analysis.

Sample extraction and clean up
The denaturation and extraction method followed Fängström, et al. 1 with some minor modifications. 13 C 10 -1,5,5,6,6,10-hexachlorodecane (10 ng) was spiked as surrogate standard to the human milk samples (about 5 g) in a test tube (18 mL). The samples were homogenized with a solvent mixture of formic acid (6 M, 1 mL), isopropanol (6 mL) and diethyl ether mixed with n-hexane (6 mL, 1:1 v/v). After centrifugation for 5 minutes at 3000 rpm, the organic phase was transferred to a new tube (18 mL). The aquatic phase was extracted once more with isopropanol (1.5 mL) and diethyl ether/n-hexane (3 mL, 1:1 v/v).
The combined organic phase was further washed twice with a sodium chloride solution (4 mL, 0.9%), whereas the aquatic phase was removed to a 15 mL tube. The combined aquatic phase was further extracted with diethyl ether/n-hexane (3 mL, 1:1 v/v) to get the analytes recovered to the organic phase. Finally, the combined organic phase was transferred to a pre-weighted beaker and lipid content was measured gravimetrically.
The lipids were dissolved with n-hexane (4 mL) in a test tube. Sulfuric acid (4 mL) was added to remove the lipids. The upper organic phase was taken out after inverting and centrifugation. The lower acid phase was partitioning once more with 4 mL n-hexane. The combined organic phase was concentrated to about 0.5 mL by a gentle nitrogen flow. A multi-layer column prepared in a Pasteur pipette packaged with activated silica gel (0.1) and silica gel impregnated with sulfuric acid (0.9 g, 2:1 w/w) from bottom to top was applied to further clean up. The column was preconditioned with n-hexane (3 mL), afterwards, the sample was applied to the column and eluted with n-hexane/dichloromethane (15 mL, 1:1 v/v). The collected solvent was evaporated to 1 mL by nitrogen flow. Silica gel column (deactivated with 3% water w/w, 4g, 0.063-0.2 mm Kiselgel 60) was applied to separate CPs from other nonpolar POPs described in detail elsewhere. 2 The column was conditioned by n-hexane (30 mL). The first fraction eluted with n-hexane (28 mL) was discarded to remove S4 nonpolar POPs such as PCBs and HCB. The second fraction containing CPs was collected with n-hexane/diethyl ether (30 mL, 3:1 v/v). Prior to instrumental analysis, Dechlorane 603 (10 ng) was added to samples as volumetric standard. Three empty tubes were prepared as field blanks in China and tested for possible contamination. One procedural blank was analyzed in parallel with each batch of five samples to assess any potential contamination through the laboratory work.

Instrumental analysis of CPs
The recovery of 13 C 10 -1,5,5,6,6,10-hexachlorodecane was analyzed by a Varian 450 gas chromatograph equipped with an electron capture detector (GC-ECD) and a Varian CP-8400 auto-sampler. Injections (1 μL) were performed on a programmable temperature vaporizing (PTV) injector operating in splitless mode at a temperature of 250 ℃. CP-Sil 5CB column (15 m x 0.25 mm i.d. x 0.25µm film thickness; Varian capillary column) was used. Helium and nitrogen were used as carrier gas and make-up gas, respectively. The column oven temperature program was set as: 80 ℃ hold for 2 min, 15 ºC/min to 300ºC and hold for 8 min.
Native CPs in human milk samples were analyzed using direct injection dichloromethane-enhanced APCI-QTOF-MS (QTOF Premier, Waters, UK) with same instrumental settings described in a previous study. 3 We screened chloride adduct ions of 280 CP congener groups from C 9 H 17 Cl 3 to C 31 H 52 Cl 14 (C 9-31 , Cl 2-14 ) in individual samples. C 9 and up to C 27 CPs were confirmed. Instrumental responses of the CP congener groups represented their profile of each sample.
We compared CP chain length ranges in the samples with those in a set of CP technical products (n = 65) and selected a sub-set of 16 products (Table S3) for quantification in this study, consisting of 5 SCCPs, 6 MCCPs, and 5 LCCPs. The CP congener group profile of each human milk samples was reconstructed by a deconvolution algorithm 4 from the profiles of the selected technical products. Then percentage contributions of the technical products were used to calculate instrumental response factors of SCCPs, MCCPs, and LCCPs in the sample. Here concentrations of C 9 CPs were included in those of SCCPs, as they were considered as impurities of SCCPs. 5 The performance of CP quantification was evaluated by the goodness of fit (R 2 ) between the reconstructed profile and the native one. R 2 ≥0.50 indicates a valid quantification. 4,6 Quantification of all samples fulfilled the criterion (Table   S4). Table S1. Theoretical number of positional isomers for polychlorinated n-alkanes by assuming no more than one chlorine atom bound to any carbon atom. The calculation is based on Tomy et al. (1997). 7        high consumption 540000 2110000 1480000

S5
The EDI was estimated based on the calculation method given in EFSA (2019), and the BM DL 10 value for SCCPs and M CCPs proposed by EFSA is 2.3 and 36 mg (kg BW) -1 day -1 , respectively. 7 The LOEL value for M CCPs and LCCPs, respectively, proposed by Environment and Climate Change Canada, is 6 and 71 mg (kg BW) -1 day -1 . 8 S16 Figure S1.