Open Access

Association between dioxin concentrations in breast milk and food group intake in Vietnam

  • Kae Saito1Email author,
  • Dang Duc Nhu1,
  • Hiroyuki Suzuki1,
  • Teruhiko Kido2,
  • Rie Naganuma2,
  • Chiaki Sakakibara2,
  • Kenji Tawara3,
  • Muneko Nishijo4,
  • Hideaki Nakagawa4,
  • Kaoru Kusama5,
  • Phung Tri Dung6,
  • Le Hong Thom6 and
  • Nguyen Ngoc Hung6
Environmental Health and Preventive Medicine200915:48

https://doi.org/10.1007/s12199-009-0106-9

Received: 20 March 2009

Accepted: 3 August 2009

Published: 15 September 2009

Abstract

Objectives

To clarify the association between dioxin concentrations in breast milk and food group intake in herbicide-sprayed and nonsprayed areas in Vietnam.

Methods

This survey was conducted in August 2007 in sprayed and nonsprayed areas, respectively. The interviews were performed using a questionnaire to obtain information on personal characteristics and usual dietary intake. Eighty mothers of sprayed area and 42 mothers of nonsprayed area participated in the study. Breast milk was analyzed for concentration of polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs).

Results

Multiple regression analysis showed that location (sprayed or nonsprayed area) has the highest association with the toxic equivalents (TEQ)-PCDDs, TEQ-PCDFs, and TEQ-Total rather than other factors. In the sprayed area, the adjusted R 2 values of regression were approximately 0.1. On the other hand, the adjusted R 2 values in the nonsprayed areas were higher than those in the sprayed area, i.e., between 0.2 and 0.3, and showed that there were significant associations with body mass index (BMI) in all models.

Conclusions

Dioxin exposure was less affected by usual dietary intake in the sprayed area than in the nonsprayed area in Vietnam. It was clear that past exposure rather than present dietary intake may affect present dioxin concentrations in breast milk in the sprayed area in Vietnam. This study suggests that present dioxin concentrations in breast milk were maintained by continuous past exposure even after 30–40 years had passed.

Keywords

DioxinBreast milkFoodDietary intakeVietnam

Introduction

During the Vietnam War, herbicide was sprayed over forests and villages in Central and Southern Vietnam to defoliate the vegetation between 1961 and 1971. The primary mixture used was Agent Orange, which contained 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) [1]. In general, dioxins refer to polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and coplanar polychlorinated biphenyls (co-PCBs), which are lipophilic compounds that bind to sediment and organic matter in the environment and which have a tendency to accumulate in the fatty tissues of animals and human beings.

It has been reported that the largest source of dioxin contaminants is dietary intake, which accounts for more than 90% of total exposure [2]. Because dioxins are lipid soluble and tend to accumulate in adipose tissue, such as seafood, meat, dairy products, and eggs, the highest concentrations are accumulated in human tissue through the food chain. These studies also suggest that human intake of dioxin has been decreasing for several decades [26]. Moreover, several studies have reported positive correlations between the intake of fish and dairy products and dioxin concentrations in breast milk, which are the specific factors influencing the accumulation of dioxin in humans [79]. There have, however, been few studies in Vietnam, in spite of the fact that a great deal of herbicide was sprayed there.

The 2003 study of dioxin contamination in food eaten by the Vietnamese people carried out by Schecter et al. [10] was the first document focusing on Agent Orange in the 30–40 years that have passed since the herbicides were sprayed. That study showed a total TEQ for ducks from 286 to 343 ppt wet weight, for chickens from 0.35 to 48 ppt wet weight, and for fish from 0.19 to 66 ppt wet weight, against usual TCDD levels in food of less than 0.1 ppt. Moreover, previous studies have reported dioxin in samples of soil, food, human blood, and breast milk in southern Vietnam, even though the Agent Orange contamination occurred 30–40 years before sampling [1113]. Therefore, there is potential for continuous dioxin exposure through the consumption of fish and meat from contaminated areas.

Previously in Vietnam animal meat was examined; however, the relationship between the consumption of different foods in the diet and the concentration of dioxin in breast milk remains unclear. In addition, nutritional investigations in herbicide-sprayed areas in Vietnam have been inadequate. In Japan the main source of dietary intake of dioxins is fish, while the main sources in Europe and the USA are meat and dairy products, reflecting differences in dietary habits [14]. For these reasons, more detailed information is needed to elucidate the role of routine dietary intake in human exposure to dioxin 30–40 years after Agent Orange use was discontinued in Vietnam. Such information may also prove useful in discussing the prevention of exposure to dioxin through dietary intake. The purpose of this study is to clarify the association between dioxin concentrations in breast milk and food group intake in herbicide-sprayed and nonsprayed areas in Vietnam and also to investigate the specific factors influencing dioxin levels in breast milk.

Materials and methods

Study population

Study areas were designated in the north-central area of Vietnam, namely the Cam Chinh commune in the Cam Lo district of Quang Tri province, where herbicides were sprayed during the war, and the Cam Phuc commune in the Cam Xuyen district of Ha Tinh province, which was not sprayed with herbicides during the war. These two communes were once separated by the demilitarized zone (DMZ), the 17th parallel that divided the country during the Vietnam War. A large number of herbicides were sprayed on Cam Chinh commune on the southern side of the DMZ, but they were not sprayed on Cam Phuc commune on the northern side.

In this study, participants were selected from a 2002–2003 survey of the long-term effects of dioxin on human health. Briefly, this investigation sought to clarify the relationship between dioxins and ecological and human health in Vietnam. Ninety lactating mothers in a sprayed area and 72 lactating mothers in a nonsprayed area were recruited for the investigation in September 2002 and July 2003. Participants were between 20 and 30 years old and provided 10 to 20 ml breast milk. Samples were analyzed for dioxin concentrations. The methods and results of that study were reported previously [1517].

This survey was conducted in August of 2007 at local district health centers in sprayed and nonsprayed areas, respectively. The interviews were performed using a questionnaire to obtain information on personal characteristics and usual dietary intake by Vietnamese researchers specially trained for this study and who belonged to the 10-80 Division of the Vietnam Ministry of Health.

Measurements

Information on age, family size, number of children, years of residence, monthly income of husband, level of education, occupation, and smoking status were obtained by the original questionnaire. Body height and weight were measured and body mass index (BMI) was calculated. Dietary intake was assessed by a food frequency questionnaire (FFQ) for Vietnamese. This questionnaire was developed by Kusama et al. [18] as a tool to estimate the habitual nutrient intake of the Vietnamese population. The FFQ is a standard tool in nutritional epidemiology and calculates the intake of nutrients. The intake of calories and nutrients were computed by multiplying the frequency of intake for each food item by the nutrient content of the specified portion size. The reproducibility and validity of this FFQ were established by using 24-h dietary recalls (24HRs) and repeated FFQ. This FFQ consisted of a 116-item food list (foods and dishes) and questions about breakfast and ingredients for dishes eaten at lunch and dinner, and includes foods and dishes, consumption frequency, and portion size. Question responses were given on a scale from 1 to 10 for frequency of consumption: never, less than once per month, 2–3 times per month, 1–2 times per week, 3–4 times per week, 5–6 times per week, once per day, 2–3 times per day, 4–5 times per day, and more than 6 times per day. Portion size was categorized into three sizes: small (approximately half of the median size), median, and large (1.5 times the median size). Participants were asked about the average consumption frequency and portion size of each of the food items listed in the questionnaire using a book with full-sized photographs of all food items at the median size for the FFQ interview to improve the accuracy of the estimation of portion size.

Data analysis

Analysis focused on dioxin concentrations in breast milk in years 2002 and 2003 as a dependent variable. Independent variables were food group intake and subject characteristics. Concentrations of dioxins were presented as toxic equivalent (TEQ) levels. Calculation of TEQ was based on World Health Organization (WHO) 1998 toxic equivalency factors (TEFs) [19].

For each individual, dietary intake was calculated by the Vietnam EIYOKUN dietary assessment system. Each nutrient intake was used to calculate the energy-adjusted nutrient intake as the residual from the regression, with nutrient intake as the dependent variable and energy as the independent variable.

Median with 25 and 75 percentiles of dioxin concentrations in breast milk, intake of nutrient and food groups, and subject characteristics were calculated, and differences between herbicide-sprayed and nonsprayed areas were assessed by Student t test, Welch’s test, and Wilcoxon signed-rank test. Chi-square was used to compare differences in proportion of education levels, occupation, and smoking status between these areas. The correlation between dioxin concentrations and characteristic data and food group intake were examined using Spearman’s correlation coefficients. Before calculating the Pearson correlation coefficient, the distribution of all data was carefully checked, and if any data proved unsuitable for normal distribution, values were log-transformed to improve normality. To identify the major sources of different food group intake contributing to dioxin concentrations in breast milk, stepwise multiple regression analysis of food group intake for dietary habits was used to seek the most significant combination of variables. Data analyses were carried out using JMP®6 software (SAS Institute, Japan), and the statistical level for significant difference was set at 0.05.

Ethical approvals

The purpose of the present study was explained thoroughly, and written informed consent was obtained from each participant through their local people’s authorities committee. All data were transformed to codes in the analysis process for individuals and they were not identified. To conduct this survey, we obtained permission from the Medical Ethics Committee of Kanazawa University (permission no: Health-89).

Results

The study population consisted of 122 mothers, namely 80 mothers (participation rate 88.9%) in herbicide-sprayed area and 42 mothers (participation rate 58.3%) in nonsprayed area. Table 1 presents the characteristics of subjects in herbicide-sprayed and nonsprayed areas. Median age was statistically different between the sprayed and nonsprayed areas (P = 0.016) at 32.0 and 30.0 years, respectively. Median height (149.9 versus 152.3 cm; P = 0.005) and body weight (44.0 versus 45.0 kg; P = 0.005) were also statistically different between the two areas. However, median BMI was 19.0 and 19.3 kg/m2, respectively, showing no statistical difference between the two areas. A statistically significant difference in education between the two groups was shown, but there were no significant differences in occupation or smoking status.
Table 1

Characteristics of subjects in herbicide-sprayed and nonsprayed areas

Item

Sprayed area (n = 80)

Nonsprayed area (n = 42)

P value

Age (years)

32.0 (28.0–36.8)

30.0 (27.0–32.0)

0.016a

Height (cm)

149.9 (147.6–154.3)

152.3 (149.5–155.0)

0.005a

Weight (kg)

44.0 (40.0–46.4)

45.0 (42.8–50.3)

0.005b

BMI (kg/m2)

19.0 (17.8–20.9)

19.3 (18.4–20.8)

0.268a

Family size (people)

5.0 (4.0–6.0)

4.0 (4.0–5.0)

0.001a

Number of children (people)

2.0 (2.0–3.0)

2.0 (2.0–2.3)

0.073a

Residence period (years)

29.0 (18.8–33.0)

27.5 (15.8–32.0)

0.301a

Monthly income of husband (×104 VND)

100 (76–150)

100 (50–150)

0.067a

Education level

 <Elementary school

23 (28.8%)

23 (54.8%)

0.006c

 >Junior high school

57 (71.3%)

19 (45.2%)

Occupation

 Farmer

70 (87.5%)

36 (85.7%)

0.783c

 Other

10 (12.5%)

6 (14.3%)

Smoking status

 Current

3 (3.8%)

0 (0%)

 Never

77 (96.3%)

41 (100%)

Data are shown as median (25th–75th percentile) or number (%)

BMI body mass index

aWilcoxon signed rank test

bWelch’s test

cChi-square

Table 2 shows the TEQ levels of PCDDs, PCDFs, and total (PCDDs + PCDFs) in breast milk in sprayed and nonsprayed areas. All dioxin concentrations values were log-transformed to improve normality. There were statistical differences (P < 0.001) in TEQ-PCDDs, TEQ-PCDFs, and TEQ-Total between sprayed and nonsprayed areas.
Table 2

Comparison of dioxin concentrations (pgTEQ/gFat) in breast milk between herbicide-sprayed and nonsprayed areas

Dioxins

Sprayed area (n = 80)

Nonsprayed area (n = 42)

P value

TEQ-PCDDs

4.54 (2.95–6.39)

1.88 (1.55–2.34)

<0.001a

TEQ-PCDFs

5.06 (3.37–7.91)

1.99 (1.40–2.39)

<0.001b

TEQ-Total

10.13 (6.46–14.20)

3.80 (2.97–4.81)

<0.001b

Data are shown as median (25th–75th percentile). Data of TEQ-PCDFs are log-transformed to improve normality

a t test

bWilcoxon signed-rank test

Table 3 summarizes the median intake (with 25 and 75 percentiles) of energy and nutrient and food group per day in sprayed and nonsprayed areas and a comparison of the difference between the two areas. Mean energy intake was not statistically different between sprayed and nonsprayed areas, measuring 1,854 and 1,793 kcal/day, respectively. Intake of lipid (P = 0.036) in the nonsprayed area was higher than in the nonsprayed area. In the two areas, intake of fruit and fruit juice was the highest; next came cereals, then meat and meat products, dark-green vegetables, sugars, confectioneries, and soft drinks. Intake of pulses (P < 0.001), yellow vegetables (P < 0.011), fruit and fruit juice (P = 0.044), and alcoholic beverages (P < 0.001) in the sprayed area were higher than in the nonsprayed area. On the other hand, there were no cases of significantly higher intake amounts in the nonsprayed area over the sprayed area.
Table 3

Dietary intake from FFQ in herbicide-sprayed and nonsprayed areas

Dietary intake

Sprayed area (n = 80)

Nonsprayed area (n = 42)

P valuea

Energy

1854 (1440–2423)

1793 (1389–2225)

0.620

Protein

87.2 (78.5–92.9)

87.6 (83.3–95.7)

0.227

Lipid

52.8 (45.3–60.0)

56.7 (52.3–61.7)

0.036

Carbohydrate

290.4 (270.4–312.4)

276.9 (269.1–294.7)

0.059

Fiber

12.8 (10.9–16.0)

15.0 (12.4–19.4)

0.018

Cereals

255.2 (205.5–330.6)

239.9 (185.6–324.6)

0.435

Potatoes and starches

31.0 (10.5–66.0)

40.8 (25.6–70.5)

0.070

Nuts and seeds

25.0 (15.1–26.2)

25.0 (12.8–25.0)

0.984

Pulses

0 (0–27.7)

0 (0–0)

<0.001

Soybean products

15.1 (7.6–30.6)

15.1 (7.6–30.2)

0.617

Dark-green vegetables

175.4 (126.8–254.0)

197.2 (120.8–299.7)

0.309

Yellow vegetables

45.3 (26.8–69.3)

32.0 (13.8–43.9)

0.011

Other vegetables

126.8 (86.1–207.1)

94.0 (57.9–179.0)

0.063

Fruit and fruit juice

504.7 (402.0–694.1)

459.7 (290.1–604.3)

0.044

Fats

6.3 (2.7–12.0)

7.2 (1.7–10.8)

0.436

Vegetable oils

16.1 (10.5–21.6)

15.0 (8.9–23.0)

0.663

Meat and meat products

206.0 (130.1–316.7)

262.1 (172.0–372.3)

0.125

Fish and fish products

52.6 (33.4–84.6)

59.6 (30.7–109.0)

0.777

Shellfish and shellfish products

14.0 (5.1–31.0)

13.9 (5.0–28.3)

0.944

Eggs

30.2 (16.3–49.4)

33.4 (12.5–76.0)

0.668

Milk and dairy products

0.6 (0–7.1)

0 (0–18.8)

0.984

Sugars, confectioneries, soft drink

159.5 (122.3–222.8)

162.1 (112.9–183.9)

0.236

Alcohol beverages

0.4 (0–7.1)

0 (0–0)

<0.001

Seasonings

41.3 (28.2–54.1)

44.2 (30.1–57.0)

0.508

Data are shown as median (25th–75th percentile). Units of energy data are kcal/day; others are in g/day. Nutrient intake is adjusted for energy intake by the residual method

aWilcoxon signed-rank test

The food group intake that correlated statistically with dioxin concentrations in breast milk in the sprayed area is shown in Table 4. In the sprayed area, no statistical correlations were found between intake of any food group intake and dioxin concentrations in breast milk. In contrast, in the nonsprayed area positive statistical correlations were shown between intake amounts of fats (r = 0.35, P = 0.023), shellfish and shellfish products (r = 0.31, P = 0.048), milk and dairy products (r = 0.31, P = 0.043), and seasonings (r = 0.39, P = 0.011) and breast milk TEQ-PCDFs. Moreover, soybean product (r = 0.32, P = 0.039) and seasoning (r = 0.32, P = 0.042) intake were positively correlated with breast milk TEQ-Total. However, no food group intake was significantly correlated with breast milk TEQ-PCDDs.
Table 4

Spearman correlation coefficients between dioxin concentrations in breast milk and food group intake

Food group

Sprayed area (n = 80)

Nonsprayed area (n = 42)

TEQ-PCDDs

TEQ-PCDFs

TEQ-Total

TEQ-PCDDs

TEQ-PCDFs

TEQ-Total

r

P value

r

P value

r

P value

r

P value

r

P value

r

P value

Cereals

−0.02

0.877

−0.12

0.281

−0.10

0.362

−0.08

0.606

0.16

0.304

0.05

0.745

Potatoes and starches

0.09

0.438

0.06

0.626

0.07

0.522

−0.01

0.949

0.16

0.314

0.07

0.673

Nuts and seeds

−0.14

0.211

−0.14

0.217

−0.15

0.185

0.19

0.226

0.18

0.250

0.20

0.207

Pulses

0.02

0.832

0.06

0.573

0.05

0.641

0.20

0.208

0.19

0.238

0.20

0.203

Soybean products

0.01

0.901

0.03

0.803

0.02

0.848

0.27

0.079

0.28

0.076

0.32

0.039

Dark-green vegetables

0.02

0.866

0.07

0.534

0.07

0.552

0.10

0.545

0.06

0.722

0.08

0.618

Yellow vegetables

−0.11

0.311

−0.16

0.148

−0.15

0.176

0.17

0.275

0.26

0.097

0.23

0.149

Other vegetables

0.05

0.690

0.10

0.359

0.11

0.354

0.02

0.923

0.09

0.583

0.07

0.653

Fruit and fruit juice

−0.05

0.648

−0.07

0.520

−0.07

0.538

0.00

0.996

0.20

0.213

0.11

0.484

Fats

−0.10

0.401

−0.12

0.275

−0.12

0.270

0.15

0.346

0.35

0.023

0.27

0.079

Vegetable oils

−0.07

0.537

−0.05

0.652

−0.06

0.608

0.06

0.723

0.30

0.051

0.21

0.188

Meat and meat products

−0.02

0.854

−0.01

0.949

−0.01

0.908

−0.04

0.786

0.13

0.418

0.06

0.693

Fish and fish products

0.03

0.807

0.14

0.222

0.10

0.397

0.19

0.237

0.10

0.549

0.15

0.344

Shellfish and shellfish products

0.00

0.989

−0.05

0.674

−0.03

0.801

0.07

0.681

0.31

0.048

0.23

0.148

Eggs

0.14

0.202

−0.02

0.882

0.06

0.586

−0.16

0.320

0.12

0.441

−0.03

0.838

Milk and dairy products

0.14

0.207

0.07

0.532

0.10

0.365

0.12

0.446

0.31

0.043

0.24

0.122

Sugars, confectioneries, soft drinks

−0.02

0.837

−0.06

0.572

−0.06

0.569

0.06

0.727

0.15

0.358

0.13

0.419

Alcohol beverages

0.11

0.324

0.05

0.651

0.09

0.446

0.05

0.754

0.15

0.335

0.10

0.531

Seasonings

−0.05

0.658

−0.07

0.520

−0.07

0.522

0.18

0.246

0.39

0.011

0.32

0.042

All data are log-transformed to improve normality

r correlation coefficients

Multiple linear regression analysis was used to determine which food groups were significantly correlated with breast milk TEQ-PCDDs, TEQ-PCDFs, and TEQ-Total as a dependent variable. The analysis focused on breast milk TEQ-PCDDs, TEQ-PCDFs, and TEQ-Total. Independent variables were subject characteristics (area, age, BMI, and monthly income of husband) and food group intake. To reduce the number of variables in the multivariate model, prognostic factors were included in the model by mixed-direction stepwise analysis. Stepwise criteria were that the variable could enter the equation when its F statistic probability was greater than 2.0.

Table 5 presents the multiple linear regression in all subjects (sprayed and nonsprayed areas). Multiple linear regression analysis found that area was the only variable associated with TEQ-PCDDs (β = 0.610, adjusted R 2 = 0.367), TEQ-PCDFs (β = 0.643, adjusted R 2 = 0.408), and TEQ-Total (β = 0.647, adjusted R 2 = 0.413).
Table 5

Stepwise multiple linear regression of TEQ-PCDDs, TEQ-PCDFs, and TEQ-Total levels in breast milk and food group intake in all subjects (n = 122)

Variables

Standardized coefficients

P value

Adjusted R 2

TEQ-PCDDs

 Constant

 

<0.0001

0.367

 Area

0.610

<0.0001

TEQ-PCDFs

 Constant

 

<0.0001

0.408

 Area

0.643

<0.0001

TEQ-Total

 Constant

 

<0.0001

0.413

 Area

0.647

<0.0001

All data are log-transformed for analysis. Variables: area code (1 for sprayed area, 0 for nonsprayed area), age, BMI, monthly income of husband, cereals, potatoes and starches, nuts and seeds, pulses, soybean products, dark-green vegetables, yellow vegetables, other vegetables, fruit and fruit juice, fats, vegetable oils, meat and products, fish and products, shellfish and products, eggs, milk and dairy products, sugars, confectioneries, soft drink, alcohol beverages, seasonings

In the sprayed area the associations between food group intake and TEQ-PCDDs (adjusted R 2 = 0.072), TEQ-PCDFs (adjusted R 2 = 0.052), and TEQ-Total (adjusted R 2 = 0.026) concentrations showed that adjusted R 2 values of regression were small (Table 6).
Table 6

Stepwise multiple linear regression of TEQ-PCDDs, TEQ-PCDFs, and TEQ-Total levels in breast milk and food group intake in herbicide-sprayed and nonsprayed areas

Variables

Standardized coefficients

P value

Adjusted R 2

Sprayed area (n = 80)

 TEQ-PCDDs

  Constant

 

<0.0001

0.072

  Fats

−0.308

0.038

  Eggs

0.262

0.064

  Alcohol beverages

0.231

0.055

  Nuts and seeds

−0.206

0.073

 TEQ-PCDFs

  Constant

 

0.000

0.052

  Fats

−0.204

0.097

  Fish and products

0.167

0.135

  Alcohol beverages

0.159

0.188

  Nuts and seeds

−0.145

0.199

 TEQ-Total

  Constant

 

<0.0001

0.026

  Nuts and seeds

−0.197

0.080

Nonsprayed area (n = 42)

 TEQ-PCDDs

  Constant

 

0.636

0.216

  Yellow vegetables

0.442

0.011

  BMI

0.333

0.025

  Sugars, confectioneries, soft drinks

−0.285

0.083

  Soybean products

0.238

0.116

 TEQ-PCDFs

  Constant

 

0.009

0.296

  BMI

0.413

0.005

  Shellfish and products

0.350

0.020

  Seasonings

0.269

0.074

  Pulses

0.231

0.096

 TEQ-Total

  Constant

 

0.627

0.182

  BMI

0.349

0.022

  Yellow vegetables

0.271

0.073

  Soybean products

0.267

0.085

All data are log-transformed for analysis. Variables: age, BMI, cereals, potatoes and starches, nuts and seeds, pulses, soybean products, dark-green vegetables, yellow vegetables, other vegetables, fruit and fruit juice, fats, vegetable oils, meat and products, fish and products, shellfish and products, eggs, milk and dairy products, sugars/confectioneries/soft drinks, alcohol beverages, seasonings

In the nonsprayed area the adjusted R 2 values of regression were higher than in the sprayed area for all models using TEQ-PCDDs (adjusted R 2 = 0.216), TEQ-PCDFs (adjusted R 2 = 0.296), and TEQ-Total (adjusted R 2 = 0.182) as a dependent variable (Table 6). In the nonsprayed area all models showed an association between BMI and TEQ-PCDDs, TEQ-PCDFs, and TEQ-Total (β = 0.333, 0.413, 0.349). TEQ-PCDDs was most highly associated with intake of yellow vegetables (β = 0.442), and TEQ-PCDFs was associated with intake of shellfish and shellfish products (β = 0.350), while no association was found between TEQ-Total and any factor except BMI.

Discussion

This study adopted the concentration data for PCDDs and PCDFs in breast milk measured in years 2002 and 2003. Consequently, the lapse of time between data collection and this study is 4–5 years. Normally, coincident measure of dioxin concentrations and dietary survey are required. Nevertheless, the half-life of dioxin in human body has been estimated to be as long as 7.5 years [20]; therefore, the decrease in dioxin concentrations in subject breast milk at the time of data collection and the present study can be expected to exhibit no great difference. In addition, the economic-social and food distribution systems in the study areas have not undergone great change during the 5 years between data collection and the present study. Furthermore, the FFQ estimates long-term intake for usual dietary habit. Hence, we consider it possible to clarify the association between dioxin concentrations in 2002–2003 and present dietary intake in spite of the limitations of this method.

Correlation coefficients between dioxin concentrations in breast milk and food group intake were shown by FFQ nutrition survey in sprayed and nonsprayed areas in Vietnam. Although the mean value and range of dioxin concentrations in foods were different among countries, previous studies have reported that the main source of dietary intake of dioxin is adipose tissue and fish [24]. This process of biological condensation explains the high accumulation of dioxin in animals.

However, low dioxin exposure was seen in these studies under normal conditions, making it difficult to compare this study with previous studies due to the fact that our research areas were sprayed by herbicides, including dioxin, during the Vietnam War. There were no statistical correlations between dioxin concentrations and the intake of each food group in the sprayed area. In contrast, in the nonsprayed area, there were statistical relationships between the intake of fat, shellfish and shellfish products, milk and dairy products, and seasonings with TEQ-PCDFs in breast milk, although the correlation coefficient was small, and the intake of soybean products and seasoning with TEQ-Total in breast milk. However, all correlation coefficients were approximately 0.3; therefore, these results do not show a clear relation between dioxin concentrations in breast milk and food group intake.

The difference in correlations between dioxin concentrations in breast milk and food group intake in the two areas is attributed to the presence or absence of exposure to herbicide. In the sprayed area, although there is dietary exposure to dioxins in the sprayed and nonsprayed areas similar to studies by Huisuman et al. [7] and Takekuma et al. [8], past exposure may act as a stronger factor than present dietary intake in the sprayed area. Hence, it is implied that there was no statistical association between dioxin concentrations in breast milk and present dietary intake.

Multiple regression analysis showed that location (sprayed or nonsprayed area) made the highest contribution to dioxin concentrations in breast milk. It is noted that difference of location is a stronger factor than present dietary intake; that is, it is the strongest contributor to dioxin concentrations in breast milk, regardless of type. This indication stems from the fact that there was no statistical correlation between breast milk dioxin concentrations in simple correlation coefficients, and the fact that the economic-social systems of both areas are similar.

Incidentally, because the average age of participants in this study was 31.8 ± 5.5 years, it is suggested that present dioxin accumulation in human tissue was influenced by breast-feeding by mothers who were exposed directly to a herbicide, and that dioxin concentrations in breast milk were influenced by intake of foods with high concentrations of dioxin, even if almost all subjects were born after 1977, well after the cessation of herbicide spraying. Schecter et al. [21] reported that serum dioxin concentrations in the sprayed area were higher than those in the nonsprayed area in a comparison of children who were born in the sprayed area and who moved to the sprayed area from the nonsprayed area after the spraying of herbicide. Furthermore, Tada [22] suggested that dioxin concentrations in adults might affect those in infants. For these reasons, it is implied that dioxin concentrations in breast milk are not influenced by present dietary intake.

Multiple regression analysis distinguishes area, given the fact that the adjusted R 2 values of regression were small, namely approximately 0.1, though some associations were statistically significant in sprayed area. We could not explain the influencing factor in dioxin concentrations in breast milk. In contrast, the adjusted R 2 values in the nonsprayed area were higher than those of the sprayed area, that is, between 0.2 and 0.3, and there were differences between the two areas. Although standardized coefficients were different for each model, there were associations with BMI in all models in the nonsprayed area. Body burdens for lipophilic chemicals are dependent on the weight and body fat of an animal. Dioxins are all highly lipophilic, resulting in their partitioning into fatty tissues [23]. One physiologically based pharmacokinetic model accounts for changes in BMI over time, with higher BMI being related to longer half-life for TCDD [24]. The results of the present study are consistent with these previous studies.

In consequence it was made clear that past exposure rather than present dietary intake affects present dioxin concentrations in breast milk in the sprayed area in Vietnam. In contrast, it is suggested that present dietary intake and BMI might affect those in the nonsprayed area. Therefore, dioxin exposure was less affected by usual dietary intake in the sprayed area than in the nonsprayed area in Vietnam. However, in order to clarify the relationship, more detailed information on background dioxin exposure, namely food distribution systems, waste disposal methods, and pesticide use, are needed. Furthermore, past accumulation of dioxin and reduction of dioxin concentrations in the body are issues to be considered. In addition, continued study is necessary due to the great potential for continuous exposure to dioxins through contamination of foods in high-exposure areas, called “hot spots,” in spite of the fact that this study did not indicate any relationship between dioxin concentrations in breast milk and dietary intake.

Declarations

Acknowledgments

This study was supported by grants from the Japan Society for the Promotion of Science [Grant-in-Aid for Scientific Research, (B) no. 17406016 and Grant-in-Aid for Scientific Research (A) no. 19209021]. The authors are deeply indebted to members of the 10-80 division and staff members of the local district health centers in Vietnam for their cooperation and encouragement during the study.

Authors’ Affiliations

(1)
Graduate School of Medical Science, Kanazawa University
(2)
Faculty of Health Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University
(3)
Department of Public Health, Hyogo College of Medicine
(4)
Department of Epidemiology and Public Health, Kanazawa Medical University
(5)
National Institute of Public Health
(6)
10-80 Division, Hanoi Medical University

References

  1. Stellman JM, Stellman SD, Christian R, Weber T, Tomasallo C. The extent and patterns of usage of Agent Orange and other herbicides in Vietnam. Nature. 2003;422(6933):681–7.View ArticlePubMedGoogle Scholar
  2. Liem AK, Fürst P, Rappe C. Exposure of populations to dioxins and related compounds. Food Addit Contam. 2000;17:241–59.View ArticlePubMedGoogle Scholar
  3. SOOP EU/EC Task 2.3.5 report. Assessment of dietary intake of dioxins and related PCBs by the population of EU member states. European Commission. Directorate-General Health and Consumer Protection, Brussels, Belgium; 2000.Google Scholar
  4. James RS, Martin DR. Dioxin and Dioxinlike PCBs in food. In: Schecter A, Gasiewicz TA, editors. Dioxin and health. Canada: Wiley; 2003. p. 89–136.Google Scholar
  5. Mato Y, Suzuki N, Katatani N, Kadokami K, Nakano T, Nakayama S, et al. Human intake of PCDDs, PCDFs, and dioxin like PCBs in Japan, 2001 and 2002. Chemosphere. 2007;67:S247–55.View ArticlePubMedGoogle Scholar
  6. Bilau M, Matthys C, Baeyens W, Bruckers L, Backer GD, Hond ED, et al. Dietary exposure to dioxin-like compounds in three age groups: results from the Flemish environment and health study. Chemosphere. 2008;70:584–92.View ArticlePubMedGoogle Scholar
  7. Huisman M, Eerenstein SE, Koopman-Esseboom C, Brouwer M, Fidler V, Muskiet FA, et al. Perinatal exposure to polychlorinated biphenyls and dioxins through dietary intake. Chemosphere. 1995;31:4273–87.View ArticlePubMedGoogle Scholar
  8. Takekuma M, Saito K, Ogawa M, Matumoto R, Kobayashi S. Levels of PCDDs, PCDFs and Co-PCBs in human milk in Saitama, Japan, and epidemiological research. Chemosphere. 2004;54:127–35.View ArticlePubMedGoogle Scholar
  9. Guan P, Tajimi M, Uehara R, Watanabe M, Oki I, Ojima T, et al. Associations between dietary intake and breast milk dioxin levels in Tokyo. Jpn Pediatr Int. 2005;47:560–6.View ArticleGoogle Scholar
  10. Schecter A, Quynh HT, Pavuk M, Päpke O, Malisch R, Constable JD. Food as a source of dioxin exposure in the residents of Bien Hoa City, Vietnam. J Occup Environ Med. 2003;45:781–8.View ArticlePubMedGoogle Scholar
  11. Dwernychuk LW, Cau HD, Hatfield CT, Boivin TG, Hung TM, Dung PT, et al. Dioxin reservoirs in southern Viet Nam-A legacy of Agent Orange. Chemosphere. 2002;47:117–37.View ArticlePubMedGoogle Scholar
  12. Schecter A, Quynh T, Päpke O, Tung KC, Constable JD. Agent Orang, Dioxins, and other chemicals of concern in Vietnam: update 2006. J Occup Environ Med. 2006;48:408–13.View ArticlePubMedGoogle Scholar
  13. Mai TA, Doan TV, Tarradellas J, de Alencastro LF, Grandjean D. Dioxin contamination in soils of Southern Vietnam. Chemosphere. 2007;67:1802–7.View ArticlePubMedGoogle Scholar
  14. Thoyama C, Masho R. The establishment of tolerable intake levels and the current situation of the risk assessment of dioxin and related compounds. Jpn J Hyg. 2005;60:60–9. (in Japanese).Google Scholar
  15. Kido T, Maruzeni S, Suzuki H, Odamae Y, Muranaka M, Naganuma R. Five years studies on the long term effects of war Agent Orange/dioxin on human health in Vietnam. Persistent Organic Pollutants (POPs) Research in Asia; 2008. p. 445–50.Google Scholar
  16. Tawara K, Honda R, Nishijo M, Nakagawa H. Pretreatment procedure of dioxin analysis for a small volume of human breast milk. J Kanazawa Med Univ. 2003;28:17–25.Google Scholar
  17. Tawara K, Nishijo M, Honda R, Maruzeni S, Seto T, Kido T, et al. Effects of maternal dioxin exposure on newborn size at birth among Japanese mother-infant pairs. Environ Health Prev Med. 1009;14:88–95.View ArticleGoogle Scholar
  18. Kusama K, Le DS, Hanh TT, Takahashi K, Hung NT, Yoshiike N, et al. Reproducibility and validity of a food frequency questionnaire among Vietnamese in Ho Chi Minh City. J Am Coll Nutr. 2005;24:466–73.PubMedGoogle Scholar
  19. Van den Berg M, Birnbaum L, Bosveld AT, Brunstrom B, Cook P, Feeley M. Toxic Equivalency Factors (TEFs) for PCBs, PCDDs, PCDFs for humans and wildlife. Environ Health Perspect. 1998;106:775–92.View ArticlePubMedGoogle Scholar
  20. Michalek JE, Tripathi RC. Pharmacokinetics of TCDD in veterans of Operation Ranch Hand: 15-years follow-up. J Toxicol Environ Health A. 1999;57:369–78.View ArticlePubMedGoogle Scholar
  21. Schecter A, Dai LC, Päpke O, Prange J, Constable JD, Matsuda M, et al. Recent dioxin contamination from Agent Orange in residents of a southern Vietnam city. J Occup Environ Med. 2001;43:435–43.View ArticlePubMedGoogle Scholar
  22. Tada H. The study on effect of dioxin concentration in breast milk for infants. Health and Labour Sciences Research Grants report of 2003; 2003 (in Japanese).Google Scholar
  23. Schecter A, Gasiewicz TA. Dioxin and health. Hoboken: Wiley; 2003.View ArticleGoogle Scholar
  24. Thomaseth K, Salvan A. Estimation of occupational exposure to 2, 3, 7, 8-tetrachlorodibenzo-p-dioxin using a minimal physiologic toxicokinetic model. Environ Health Perspect. 1998;106(Suppl.2):743–53.View ArticlePubMedGoogle Scholar

Copyright

© The Japanese Society for Hygiene 2009