Environ. Sci. Technol. 2008, 42, 7709–7714
Validation of Dietary Intake of Dichlorodiphenyltrichloroethane and Metabolites in Two Populations from Beijing and Shenyang, China Based on the Residuals in Human Milk S H U T A O , * ,† Y A N X I N Y U , † W E N X I N L I U , † XUEJUN WANG,† JUN CAO,† BENGANG LI,† XIAOXIA LU,† AND MING H. WONG‡ Laboratory for Earth Surface Processes, College of Urban and Environmental Sciences, Peking University, Beijing 100871, China, Croucher Institute for Environmental Sciences, and Department of Biology, Hong Kong Baptist University, Kowloon Tong, Hong Kong, PR China
Received May 3, 2008. Revised manuscript received August 4, 2008. Accepted August 11, 2008.
This paper presents the results of an investigation on association between dietary intakes and human milk concentrations of dichlorodiphenyltrichloroethane and metabolites (DDTs) of two populations from Beijing and Shenyang, China. We analyzed human milk samples from 76 women and 287 composite food samples covering major food categories for concentrations of DDTs. We also collected information on food consumptions and calculated dietary intakes of DDTs of the participants on individual basis. The median values of the measured DDTs in human milk were 125 ng/g lipid and 132 ng/g lipid for the samples from Beijing and Shenyang, respectively. The mean ((standard deviation) daily dietary intakes of DDTs by the two groups were 32.0 ( 14.2 ng/kg · day and 27.9 ( 11.3 ng/kg · day, respectively. The temporal trends of decreasing in DDTs and increasing in DDE/DDT ratio suggested that the residuals were primarily from historical application. We found a significant correlation between human milk concentration and daily dietary intake of DDTs, while the dietary intake could explain 22% of the variation in the DDTs in human milk. In addition to dietary exposure, we also found that maternal body mass index (body weight divided by the squared height), body weight, body height, and mother’s age contributed significantly to the variation of DDTs in human milk after intake normalization. The result of a probabilistic risk assessment indicated that the exposure of infants to DDTs through breast feeding would be a public health concern for years to come, although breast feeding is still recommended.
Introduction Despite considerable declines in levels of dichlorodiphenyltrichloroethane and metabolites (DDT, DDD, and DDE, * Corresponding author phone: 0086-10-62751938; fax: 0086-1062751938; e-mail:
[email protected]. † Peking University. ‡ Hong Kong Baptist University. 10.1021/es801219v CCC: $40.75
Published on Web 09/20/2008
2008 American Chemical Society
DDTs as a total) in the global environment, their residuals in various media including biota remain to be a matter of concern in many places because of their high persistence and possible toxic effects (1-3). Over the past half-century, more than 400 000 tons of DDT were produced in China, accounting for approximately 20% of the total world production (2). Most of the products were used in agriculture leading to a widespread occurrence in the environment (3, 4). Since the general population is primarily exposed to DDTs through ingestion of contaminated foods, the occurrence of DDTs in food commodities is a key issue in public health (5). Yet, the information on food contamination and human dietary exposure to DDT in China is scarce. Human breast milk is one of the most widely used mediums for monitoring human burdens of DDTs and other organochlorine pesticides (OCPs) (6). Wong et al. have reviewed the use and human burden of OCPs in China and indicated that the level of DDTs in human breast milk of the Chinese population was higher than those of many other countries, and this is likely due to higher background levels in the environment (4). The presence of DDTs in human milk is also a matter of concern for infant’s health (7, 8). The association between dietary exposure and body loading of OCPs was extensively studied (9, 10). However, the quantitative evidence obtained were mostly the significant correlations between the residuals of OCPs in human tissues and the consumption of certain foods or food categories rather than the direct correlation between body burden and dietary intake (9, 10). To our knowledge, Vaz was the only one who reported a correlation between the OCPs in Swedish human milk and average dietary intake of OCPs from animal origin foods (11). The objective of this study was to gather the first-hand evidence on causal relationship between dietary intake and human milk residuals of DDTs. DDT and metabolites studied included p,p′-DDT, p,p′-DDD, p,p′-DDE, o,p′-DDT, o,p′-DDD, and o,p′-DDE. The target population consisted of two subsets of mothers from Beijing and Shenyang, China (locations are shown in Figure S1 of the Supporting Information (SI)). We determined the dietary intake by the populations based on the dietary frequencies of all the participating individuals and the measured DDTs in major foodstuffs. Human milk samples were collected and measured for DDTs. Based on these data, we tested a hypothes that human milk can serve as a quantitative predictor for validating dietary exposure to DDTs. In addition, we evaluated the influence of nondietary lifestyle factors on the levels of DDTs in human milk and addressed the potential risk of the accumulated DDTs on breastfed infants on a probabilistic basis.
Materials and Methods Population. The studied population was 40 women (average age 28.7) recruited from Tiantan Hospital in Beijing and 36 women (average age 27.5) from Liaoning Maternity and Child Hospital and Shenyang Maternity and Child Hospital in Shenyang. All participants had lived in the current addresses for at least six years. Tables S1 and S2 in the SI list socialdemographic characteristics of the target populations. We asked each participant to complete a written informed consent prior to the sample collection. As a Hong Kong Research Grants Council supported project, the research protocol was approved by the Committee on the Use of Human and Animal Subjects in Teaching and Research (HASC) of Hong Kong. We asked each participant to answer two questionnaires with the assitance of qualified nurses. A nondietary determinant questionnaire covers information on lifestyle and social-demographic characteristics, and a VOL. 42, NO. 20, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7709
dietary questionnaire addresses food consumption frequencies during a one-year period prior to childbirth and the average amount of food consumed on each occasion. Table S3 in the SI summarizes the food consumption information. Sample Collection. We collected samples from June 2005 to December 2006 in Beijing and from June 2006 to July 2007 in Shenyang, respectively. Each participant donated approximately 50 mL of human milk with a breast pump from 1 to 75 days after childbirth. The human milk samples were stored in chemical-free glass-bottles with Teflon seals with potassium dichromate (final concentration of 1.4 mg/mL) spiked, frozen immediately, and stored at -18 °C prior to analysis. According to the officially reported food consumption pattern of Chinese population, major foods sampled include fruits (apple, banana, pear, grape, and orange), vegetables (Chinese cabbage, cabbage, spinach, cucumber, carrot, green pepper, eggplant, lettuce, potato, and bean), cereals (rice and flour), fish (carp, grass carp, crucian, and bighead), meat (pork, chicken, beef, and mutton), eggs, milk (five brands), and oil (one brand). We sampled 144 and 143 composite samples of various foods from randomly selected markets and supermarkets in Beijing and Shenyang, respectively. Each composite sample was a mixture of at least four subsamples. For example, five composite carp samples were prepared from 35 individual fish and two composite apple samples were from 10 apples. All nonliquid samples were freeze-dried (EYELA-FDU-830, Tokyo Rikakikai, Japan). Reagents. Analytical grade solvents including acetonitrile, n-hexane, and dichloromethane were obtained from Beijing Reagent Company, China and purified by distillation. 2,4,5,6tetrachloro-m-xylene (TCMX) and 4,4′-dichlorobiphenyl (J&K Chemical Ltd., U.S.) were used as internal standard and surrogate, respectively. Stock standard and working standard solutions were prepared by diluting a commercial mixed standard (J&K Chemical Ltd., U.S.) with n-hexane. Granular anhydrous sodium sulfate was heated at 600 °C for 6 h and stored in a sealed desiccator prior to use. Florisil (60∼80 mesh, PR grade, Dikma Technologies, U.S.) was precleaned for 6 h at 650 °C and dried in a 130 °C oven for at least 16 h before use. Silica gel (60∼80 mesh, Beijing Chemical Reagent Co., China) was heated at 450 °C for 4 h, kept in a sealed desiccator and reactivated at 130 °C for 16 h immediately prior to use. All glassware were cleaned in an ultrasonic cleaner (KQ-500B, Kunshan Ultrasonic Instrument, China) and heated at 400 °C for 6 h. Sample Extraction. For animal-origin samples except eggs, we applied U.S. Environmental Protection Agency (USEPA) 3630 method with a slight modification. Briefly, we Soxhlet extracted 3 g samples homogenized with 5 g of anhydrous sodium sulfate in a mixture of n-hexane (20 mL) and dichloromethane (80 mL) at 55 °C for 24 h. We further extracted the extract twice with n-hexane saturated acetonitrile (30 mL) for 1 min each time. We then shook the extract with 300 mL of 5% sodium sulfate and 30 mL of n-hexane for 10 min and finally extracted with 30 mL of n-hexane. We used a slightly modified USEPA-600/8-80-038 method for analysis of human milk, cow milk, eggs, and oil samples. We extracted 5 mL samples three times with 12 mL of acetonitrile for 5 min each time. We then added 120 mL of 12% sodium sulfate solution to the extract and shook for 5 min. We finally extracted the solution twice with 30 mL of n-hexane for 15 min each time. We adopted USEPA-600/8-80-038 method for vegetable, fruit, and cereal analyses. We extracted pulverized samples (40 g of vegetable or fruit, 10 g of cereal) with acetonitrile (80 mL for vegetable or fruit, 25 mL for cereal) for 30 min. We then extracted the filtered solution (after shook for 10 min with 300 or 100 mL 12% sodium sulfate solution added for vegetable or fruit and cereal, respectively) twice with 40 mL of n-hexane for 15 min each time. 7710
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 20, 2008
Sample Cleanup. For all the samples except vegetable, fruit, and cereal, we concentrated the extract to 1 mL in a rotary evaporator and transferred with 1 mL of n-hexane to a chromatography column (30 cm × 10 mm i.d.) filled with the silica gel and eluted with 20 mL of n-hexane (discarded) and 35 mL dichloromethane in sequence at a rate of 2 mL/ min. For vegetable, fruit, and cereal samples, we eluted the extract with n-hexane (50 mL) and a mixture of n-hexane and dichloromethane (50 mL, 3:7 v/v) in sequence on a chromatography column (30 cm × 10 mm i.d.) filled with both florisil (4 g) and silica gel (6 g). For all the samples, we concentrated the eluate to a final volume of 0.8 mL and spiked with 125 µL of the internal standard (2 mg/L). Sample Analysis. We determined DDT and metabolites using a gas chromatograph (GC, Agilent GC 6890, U.S.) equipped with a 30 m HP-5 column (0.25 mm i.d., 0.25 µm film) and a 63Ni electron capture detector (µ-ECD). We maintained the oven temperature at 50 °C, raised to 150 °C at a rate of 10 °C/min, to 240 at 3 °C/min, and held for 5 min. The injector and the detector temperatures were 220 and 280 °C, respectively. We identified GC peaks based on the retention time of individual authentic standards ((0.3%) and quantified them by the internal standard which was added before the extraction. We determined lipid contents by weighting the samples before and after the extraction. Quality Control. We followed routine quality control procedures. We ran at least two procedural blanks simultaneously with every set of the sample analysis by going through the same extraction and cleanup procedures, and the measured procedure blanks were mostly more than 1 order of magnitude lower than the sample measurements. We also ran a standard solution for each batch of sample analysis to check the possible degradation of p,p′-DDT. For the majority of samples, we measured two or three replicates depending on sample availability to check for reproducibility. We determined detection limits for four categories of samples individually: (1) fruit and vegetable; (2) cereal; (3) meat and fish; and (4) human milk, caw milk, and eggs and recoveries for three categories of samples individually (similar to detection limits with combined fruits, vegetable, and cereal). We used TCMX as the surrogate, and the recoveries were 94, 93, and 71% for the three sample categories, respectively. We also determined recoveries by spiking the samples with a working standard (100 ng/g sample media) and the average recoveries of the individual species in the spiked samples varied from 77.1 to 98.2%. The mean detection limits were 0.004∼0.211 ng/g (fresh weight) for various compounds and samples, and the detailed data are presented in Table S4 in the SI. Table S5 in the SI provides the detailed information on the recoveries for various categories of samples. Data Analysis. In addition to median values, we also reported arithmetic means and standard deviations for comparison with those reported in the literature, even though they would not be preferred as measures of central tendencies. We used log-transformed concentrations for statistical analysis. We used t test and Pearson correlation test for comparison between groups and for evaluation of association, respectively. In all cases, we considered the results significant at a probabilistic value of p < 0.05. We conducted conventional and forward stepwise multivariate regressions to evaluate the dependence of DDTs in human milk on the consumption of various foods and on lifestyle factors and correlation analysis on the relationship between residuals in the human milk and dietary intake of DDTs. We used Mont Carlo simulation (10 000 runs) for assessment of probabilistic risk on breastfed infants. For the Mont Carlo simulation, we used normal distributions for body weight and milk consumption and log-normal distributions for concentrations of DDTs, respectively. We derived the distribution parameters
TABLE 1. Lipid Normalized Concentrations of DDT and Metabolites in the Human Milk As Arithmetic Means, Standard Deviations (SD), Medians (M), Minimum (Min) and Maximum (Max), ng/g Lipida Beijing, n ) 40
city
p,p′-DDT p,p′-DDD p,p′-DDE o,p′-DDT o,p′-DDD o,p′-DDE DDTs lipid, g/ml a
Shenyang, n ) 36
mean ( sd
M
min
max
mean ( sd
M
min
max
4.95 ( 3.61 3.55 ( 3.33 169 ( 180 1.09 ( 0.91 N.D. 3.87 ( 2.70 183 ( 186 0.033 ( 0.019
3.89 2.74 112 0.80 N.D. 2.92 125 0.030
1.21 N.D. 30.2 N.D. N.D. 1.39 34.7 0.009
17.0 20.8 1010 3.90 N.D. 13.1 1050 0.091
4.48 ( 3.51 4.78 ( 5.57 136 ( 132 1.19 ( 1.43 2.44 ( 5.12 4.95 ( 5.23 154 ( 143 0.033 ( 0.014
3.51 2.64 117 0.78 N.D. 2.92 132 0.031
N.D. N.D. 15.65 N.D. N.D. N.D. 18.74 0.008
14.6 20.8 763 5.34 22.1 18.1 833 0.072
Note: N.D. below the detection limits.
FIGURE 1. Time trends of p,p′-DDT, p,p′-DDE and p,p′-DDE/ p,p′-DDT ratio in human milk in Beijing from 1982 to 2006. The data from 2005 to 2006 (Beijing) and 2006-2007 (Shenyang, for comparison) were collected in the current study (symbols marked gray), whereas the other data were from Yu et al. (open symbols) (14). The dynamic changes of p,p′-DDT and p,p′-DDE were described using a logistic model, whereas the change in p,p′-DDE/p,p′-DDT ratio was fitted with a simple exponential model.
FIGURE 2. Comparison between the multivariate regression model predicted and the measured DDTs in human milk for the two groups from Beijing and Shenyang in log-scale. The prediction was based on the consumption of major foodstuffs. The data from Beijing and Shenyang were pooled together. from the data collected in this study and assumed that the average diet of breastfed infants was 800 ( 160 g milk per day (12).
Results and Discussion Concentrations of DDT and Metabolites in Human Milk. Table 1 presents means, standard deviations, and medians of DDT and metabolites in human milk, normalized to lipid contents. Although there were significant differences in dietary intakes of DDTs between the two groups from Beijing and Shenyang, the differences in human milk concentrations were not significant (p > 0.05). Temporal Trends of DDTs in Human Milk in Beijing. Over the past 30-40 years, notable decreases in DDTs in human milk in many places around the world were reported
(4, 6, 13). Yu et al. had monitored p,p′-DDT and p,p′-DDE in human milk in Beijing for 20 years and reported that p,p′DDT+p,p′-DDE dropped from 6210 ng/g lipid in 1982 to 737 ng/g lipid in 2002 (14). We plot p,p′-DDT, p,p′-DDE (logscaled), and the ratio of the two in human milk against time in Figure 1 using Yu’s data and ours together. Steady decline trends of both p,p′-DDT and p,p′-DDE are clearly demonstrated, reflecting the decrease in DDTs in the environment. Nakata et al. also reported a downward trend of DDTs in human adipose tissue in Beijing (15). Historically, DDTs in human milk from Beijing and other Chinese cities were significantly higher than those from developed countries where DDT was banned in early 1970s (4). The results of this study indicate that DDTs in human milk in Beijing and Shenyang are currently gradually approaching the same order of magnitude as those in some developed countries (8). It was well established that DDE/DDT ratio is an indication of exposure history and a high DDE/DDT ratio implies past exposure, whereas a low DDE/DDT ratio suggests recent input (16). In this study, we found that the dominant compound in human milk was p,p′-DDE for both Beijing (92.6%) and Shenyang (88.4%), whereas p,p′-DDT contributed to very small fractions of the total (2.7% and 2.9% for Beijing and Shenyang, respectively). While both p,p′-DDT and p,p′DDE decreased, p,p′-DDE/p,p′-DDT increased over time (Figure 1), implying that the detected DDTs were mainly originated from the historical application. We quantified the temporal trends of p,p′-DDT, p,p′-DDE and their ratio with curvilinear regression (the curves in Figure 1) and present the detailed results of the regression in the SI. We also calculated the population-level half-lives of p,p′DDT and p,p′-DDE in human milk in Beijing following the procedure of Caudill et al. (17). Since DDT was not restricted untill 1983 and was not totally banned for agricultural use untill 1992 in China, a sharp decline did not occur until the early 1990s. Therefore, data for the 1980s (from 1982 through 1989) were not used for the half-life calculation. The estimated population level half-lives of p,p′-DDT and p,p′-DDE in human milk in Beijing were 5.3 and 7.6 years, respectively. The values are in the same range of those estimated by the others (6, 15). We predicted conservatively that the median values of p,p′-DDT and p,p′-DDE in human milk in Beijing would drop to 0.39 ng/g lipid and 0.34 ng/g lipid (0.10∼1.6 ng/g lipid and 11∼110 ng/g lipid as 95% confidence intervals) in 2020, respectively. Dependence of DDTs in Human Milk on Food Consumption. We believed that the variations in the measured DDTs in the human milk among the individuals (102% and 93% as the coefficient of variations for Beijing and Shenyang, respectively) were primarily caused by the difference in their dietary habit. To test this hypothesis, we performed a multivariate linear regression using the measured DDTs in human milk and the consumptions of various foodstuffs, VOL. 42, NO. 20, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7711
TABLE 2. Daily Dietary Intakes (Means, Standard Deviations and Medians) of DDTs by the Two Target Groups from Beijing and Shenyang, ng/kg · day Ppopulation Beijing n ) 40 Shenyang n ) 36
mean ( SD M mean ( SD M
p,p′-DDT
p,p′-DDD
p,p′-DDE
o,p′-DDT
o,p′-DDD
o,p′-DDE
DDTs
15.3 ( 8.9 13.9 11.6 ( 5.2 10.28
3.00 ( 1.2 2.58 5.7 ( 3.7 4.61
7.73 ( 3.2 6.78 3.5 ( 1.2 3.38
3.43 ( 1.4 3.49 2.3 ( 0.79 2.28
1.15 ( 0.37 1.07 4.7 ( 3.8 3.92
1.40 ( 0.67 1.25 0.2 ( 0.1 0.176
32.0 ( 14.2 28.4 27.9 ( 11.3 25.6
both log-transformed, as dependent and independent variables, respectively. Figure 2 presents the predicted human milk DDTs using the model against the measured ones. With the coefficients of determination of 0.56 and 0.62 for Beijing and Shenyang, respectively (0.59 for the pooled data), we confirmed that food consumption was a good predictor for body burden of DDTs. Dependence of DDTs in the Human Milk on Dietary Intake. We calculated the daily dietary intake of DDT and metabolites by the target populations using the food consumption data and the measured concentrations in various foodstuffs ( Tables S6 and S7 in the SI) and present the results as means, standard deviations (SD), and medians (M) in Table 2. The detailed information on the intake is provided in the SI. According to the result of a t test, the dietary intake of DDTs by the Beijing group was significantly higher than that by the Shenyang group, which can be explained by the differences in both food consumption rates and residuals of DDTs in the foods. For example, the average consumption of cow milk by the Beijing group (198 ( 191 g/day) was more than twice of that by the Shenyang group (81 ( 74 g/day), while p,p′-DDE concentrations in pork, chicken and crucian fish as fresh weight from Beijing (1.19 ( 0.20 ng/g, 0.54 ( 0.25 ng/g and 2.98 ( 3.32 ng/g, respectively) were significantly higher than those from Shenyang (0.21 ( 0.10 ng/g, 0.14 ( 0.04 ng/g, and 0.59 ( 0.20 ng/g, respectively). With the measured residuals of DDTs in various foodstuffs available, we could further relate the human milk DDTs to dietary intake so as to test the key hypothesis of this study: human milk samples can serve as quantitative predictors for validating the dietary exposure to DDTs. We found significant correlations (p < 0.001) between the log-transformed dietary intakes and the human milk concentrations of DDT and metabolites for the both populations from Beijing and Shenyang. We noticed that the correlations were more significant without lipid normalization (ng/g), likely suggesting that the accumulation of DDTs was not associated with the accumulation of body fat, and the latter may dilute DDTs in the body to a certain extent. As an example of such association, we plot DDTs in human milk against the dietary intake of DDTs in Figure 3. If all the samples from the two populations were pooled together, the correlation was significant at a level of 10-5, and 22% of the variation in human milk DDTs among the studied individuals could be explained by dietary intake alone. The difference in the slops of the regression lines between the two groups from Beijing and Shenyang agrees with the observed fact that there were significant differences in the dietary intake of DDTs but not in human milk DDTs between them. Other Factors Affecting DDTs in Human Milk. In addition to dietary exposure, many other factors may affect body burdens of OCPs. For example, associations were often reported between OCPs in human milk and a number of lifestyle or social-demographic factors including maternal age, number of births, lifetime lactation, birth weight, body mass index (BMI, body weight divided by the squared height) of mother (10, 18-21). We performed a multiple stepwise regression to assess the possible influences of maternal age, height, weight and BMI of the mothers, number of births, 7712
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 20, 2008
and lactation time on human milk DDTs (log-transformed). To eliminate the influence of the dietary intake in the analysis, we used an accumulation index (AI, day) defined as the ratio of DDTs in human milk (ng/g, without lipid normalization) to daily dietary intake (ng/kg · day) as a dependent variable. AI is actually the intake normalized concentration. Among the parameters tested, BMI (p ) 0.016), mother body weight (p ) 0.020), mother height (p ) 0.027) and age (p ) 0.031) entered the model sequentially. We plot the model predicted AI (AIpred) against the observed AI (AIobs) in Figure 4 and found that approximately 25.4% of the variation in AI was accountable by these factors. Furberg et al. also found that BMI was a positive predictor for OCPs in female plasma in Norway (18). Jackson et al. suggested that BMI was a direct measure of body fatness and high level of body fat was favorable for accumulation of OCPs in human body (21). Although BMI was the first variable entered the stepwise regression model, inclusion of both body height and weight in the model suggests that the definition of BMI was not optimal in describing the influence on DDT accumulation in human body or the dependence was not linear. The positive effect of mother’s age on DDTs in human milk was also revealed and the effect can be explained by at least two reasons: (1) DDTs were persistent in the body and could bioaccumulate with age, and (2) considerable decline in DDTs in the environment, subsequently in foodstuffs and exposure, had occurred since early 1990s, which reduced the exposure level for younger mothers. Although the milk samples were collected at different times after delivery, no significant contribution of the lactation time on residuals of DDTs in human milk was revealed, likely due to effects of other factors and the relatively small sample size. Breast Feeding Risk. Breast feeding is generally recommended despite human milk is often contaminated with various pollutants including DDTs (19, 20). Based on the average weight of the infants delivered by the two target populations (3420 ( 390 g), median value of DDTs in human milk measured in this study (3.97 ng/g fresh weight) and an assumed milk consumption of 800 g/day, we estimated that the average daily dietary intake of DDTs by a breastfed infant was 0.84 µg/kg · day, which was much lower than the WHOrecommended acceptable daily intake (ADI) of 20 µg/kg · day (7). However, if the U.S. Agency for Toxic Substances and Disease Registry defined minimal risk levels (MRLs) for hazardous substances (22) or the U.S. EPA’s reference dose (RfD) for noncancer toxicity of individual chemicals (23), both of which are 0.50 µg/kg · day for p,p′-DDT, was applied, the estimated daily intakes of 74% of the individuals investigated in this study would exceed the limits. We further assessed the risk for the individuals in the upper tail of the distribution by a Monte Carlo simulation using the data collected in this study. Figure 5 displays the cumulative frequency of daily intakes of DDTs by breastfed infants. It appears that 0.03% and 71.0% of the breastfed infants exceed the ADI of WHO or the RfD of USEPA, respectively. Therefore, to evaluate the breast-feeding risk conservatively, the possible health implication can not be totally ignored. Still it does not serve as a reason to discourage breast-feeding. Without further input to the environment, the downward trend of DDTs would continue. Given an approximate half-life of
milk in Beijing are provided. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited
FIGURE 3. Relationship between the measured DDTs in human milk (without lipid normalization) and the dietary intake of DDTs by the two groups from Beijing and Shenyang. Both parameters were log-transformed.
FIGURE 4. Relationship between the measured and the stepwise regression model predicted accumulation indices (AI, ratio of DDTs in human milk and daily dietary intake of DDTs).
FIGURE 5. Cumulative frequencies of the estimated daily intake of DDTs by infant through breast feeding over a range from 10-2 µg/kg · day to 102 µg/kg · day (solid curve). The results are derived from a Monte Carlo simulation using the data collected from Beijing and Shenyang in 2005-2007. We also predicted the cumulative frequency in 2020 for a comparison (dash curve). The results are assessed based on the WHO recommended acceptable daily intake (ADI) and the USEPA defined reference dose for noncancer toxicity (RfD). seven years in human body, DDTs in human milk in the studied cities will drop to a quarter of the current levels in 2020. By then, the rates exceeding the ADI and the RfD in Beijing would drop to
