Xenoestrogenic activity in blood of European and Inuit populations

1476-069X-5-12 1476-069X Research Xenoestrogenic activity in blood of European and Inuit populations Bonefeld-Jorgensen C Eva ebj@mil.au.dk Hjelmborg S Philip phj@mil.au.dk Reinert S Thayaline line_job1@hotmail.com Andersen S Birgitte bsa@mil.au.dk Lesovoy Vladimir dimusic@ic.kharkov.ua Lindh H Christian Christian.Lindh@med.lu.se Hagmar Lars lars.hagmar@ymed.lu.se Giwercman Aleksander aleksander.giwercman@kir.mas.lu.se Erlandsen Mogens mogens@biostat.au.dk Manicardi Gian-Carlo manicardi.giancarlo@unimo.it Spanò Marcello spanomtc@mail.casaccia.enea.it Toft Gunnar gutof@as.aaa.dk Bonde Peter Jens jpbon@as.aaa.dk

Unit of Cellular and Molecular Toxicology, Department of Environmental and Occupational Medicine, Institute of Public Health, Vennelyst Boulevard 6, Build. 1260, University of Aarhus, DK-8000 Aarhus, Denmark

Regional Clynical Center of Urology and Nephrology, Kharkiv, Ukraine

Department of Occupational and Environmental Medicine, University Hospital, SE-22185 Lund, Sweden

Scanian Fertility Centre, Malmö University Hospital, SE-20502 Malmö, Sweden

Department of Biostatistics, Institute of Public Health, University of Aarhus, Vennelyst Boulevard 6, Build. 1260, DK-8000 Aarhus, Denmark

Laboratory of Genetics, Department of Agricultural Sciences, University of Modena and Reggio Emilia, Viale Kennedy 17, I-42100 Reggio Emilia, Italy

Section of Toxicology and Biomedical Sciences, BIOTEC-MED, ENEA Casaccia, Via Anguillarese 301, I-00060 Rome, Italy

Department of Occupational Medicine, Aarhus University Hospital, Noerrebrogade 44, Build.2C, DK-8000 Aarhus, Denmark

Environmental Health: A Global Access Science Source 1476-069X 2006 5 1 12 http://www.ehjournal.net/content/5/1/12 16674832 10.1186/1476-069X-5-12 16 11 2005 05 5 2006 05 5 2006 2006 Bonefeld-Jorgensen et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

Human exposure to persistent organic pollutants (POPs) is ubiquitous and found in all individuals. Studies have documented endocrine disrupting effects and impact on reproduction. The aim of the present study was to compare the level of xenoestrogenic activity in serum of groups with varying POP exposure, and to evaluate correlations to the POP biomarkers, 2,2',4,4',5,5'-hexachlorobiphenyl (CB-153) and 1,1-dichloro-2,2-bis (p-chlorophenyl)-ethylene (p,p'-DDE).

Methods

The study included 358 men: Greenlandic Inuit's, Swedish fishermen, and Warsaw (Poland) and Kharkiv (Ukraine) inhabitants. Xenoestrogenicity of serum extracts alone (XER) and XER competitive (XERcomp) effect on 17β-estradiol induced estrogen receptor (ER) transactivity were assessed in the hormone free, lipophilic serum fraction containing the POPs using the MVLN human breast cancer cell line.

Results

No agonistic XER activity was exhibited for Inuit serum samples, while 12 – 24% of the European samples had detectable agonistic XER activity. On the contrary, 71% of Inuit serum samples antagonized XERcomp compared to 7 – 30 % in the other regions. XER and XERcomp were not or weakly correlated to the two POP markers. XER activity of Inuit samples was negatively associated to levels of CB-153 and p,p'-DDE. For the Warsaw group a positive and negative correlation between XER and p,p'-DDE and estradiol equivalence level and CB-153 levels was found.

Conclusion

No strong consistent association between xenoestrogenic net activity and the two POP markers was found. The results showed that the selected POP markers alone can not predict the integrated xenoestrogenic serum activity. Correlations to the POP markers were found at the extreme edge; the Inuit's and Warsaw study groups eliciting high frequency of samples with ER antagonistic and agonistic activity, respectively. We suggest that the variation in xenoestrogenic serum activity reflects differences in POP exposure mixture, genetic factors and/or life style factors.

1. Introduction

Human exposure to environmental contaminants is ubiquitous and affects also individuals living far away from the source of contaminants such as industries, combustion and waste disposal sites. Mainly caused by dietary exposure and due to their resistance to environmental and biotic degradation persistent organochlorine pollutants (POPs), e.g. chlorinated pesticides and polychlorinated biphenyls (PCBs), accumulate in body fat tissues of humans and animals. Many POPs can mimic hormone activities and are, therefore, potential endocrine disrupters (EDs) suspected to increase the risk of cancer, birth defects, reproductive and neuro-immune disorders 123. To date, no clear-cut evidence for adverse endocrine-related human health effects of POPs have been obtained at the individual or population level. However, data from studies on wild-life species, laboratory animals and biomarker effects in vitro have strengthened the need for further research to address the concern.

The sex steroid receptors such as the estrogen receptor α (ERα) and ERβ and the androgen receptor (AR) belongs to the nuclear receptor family and are generally ligand-dependent transcription factors 456. Their genomic mediated pathway include steps as binding of ligand to receptor, translocation into nucleus and binding of the receptor-ligand complex to a specific DNA response element causing a gene response. Androgens and estrogens are not male or female hormones only. Both hormones are important in both sexes and an absolute necessity for reproductive development including male fertility 7. The steroid androgen hormones, testosterone and dihydrotestosterone, are necessary for the normal male phenotype and the presence of ERα and ERβ in human foetal testis and epididymis cells indicates that estrogens play a physiological role in regulations of spermatogenesis in mammals 8910.

There is extensive data of in vitro mechanism based models on endocrine-related POPs effects on sex hormonal systems 11121314. Although much less potent compared to 17β-estradiol (E2) the three PCB congeners CB-138, CB-153 and CB-180, comprising up to 50% of the bio-accumulated sum of PCBs, elicit a receptor mediated antiestrogen and/or antiandrogen in vitro activity 15, and similarly some hydroxylated PCBs exert low potency estrogenic and/or antiestrogenic effects 1617. Generally PCBs do not bind to the sex hormone binding globulin (SHBG) or at very low affinity 18, which may cause a higher bio-availability of the compounds. Animal studies have indicated that prenatal exposure to POPs as PCBs, dioxins and p,p'-DDE is associated with reduced male fertility 192021222324, whereas the effect of POPs on human fertility is still controversial 2526, and therefore an integrated risk assessment of EDs is needed 27. In vitro studies of xenoestrogen mixtures present at or below their no-observed-effect concentration (NOEC) or sub-NOEC was demonstrated to cause a dramatic additive enhancement of hormone actions 2829. Rather than monitoring the level of identified xenoestrogens, the integrated biological activity was used to estimate the xenobiotic burden of serum samples using 17303132.

The present study was a part of the EU supported research project Inuendo 33. The specific aim of the present study was to compare xenoestrogenic activities across Inuit's and European regions, and to evaluate whether the serum levels of the POP biomarkers, CB-153 and p,p'-DDE, were associated to the actual xenoestrogenic net activity in blood.

2. Subjects and methods

2.1 Study population and collection of blood samples

Based on earlier reports of POP body burdens four populations were included in the Inuendo study 33 for a cross-sectional fertility study representing a wide range of POP burdens. For the present study the aim was to recruit 300 male spouses to pregnant women in Greenland, Warsaw (Poland) and Kharkiv (Ukraine), and a subgroup (n = 100) of an already established cohort of fishermen from Sweden was also included in the study 25. A general criterion for eligibility was that the participants had to be born in the country of study and at least 18 years of age. Serum samples were randomly selected from the main Inuendo study groups, and the intension was to analyse serum samples including high and low POP exposure within each study group. However, because of the limit in project timing and in some cases too little blood was collected for the analysis, serum samples were randomly selected before POP determination, and fewer serum samples were analyzed. Males were recruited during 2002–2004 in Greenland (n = 72; Sisimiut, West Coast n = 50, and Tasiilaq, East Coast n = 22), in Warsaw, Poland (n = 99) and in Kharkiv, Ukraine (n = 88) 25. Earlier studies have reported high POP levels in Greenlandic Inuit's and differences in levels between habitants of the East and West Coast of Greenland 3435. However to keep similar power in the four study groups the statistical evaluation was carried out on the combined Greenlandic data only.

Information about demographic and lifestyle factors as age, body mass index (BMI), intake of seafood, coffee, smoking habits, and alcohol consumption was collected by interviews (Table 1) 2535. Venous blood samples were collected into 10 ml vacuum tubes and after centrifugation the serum was transferred to Nunc tubes and stored at -80°C for later analysis 35. The study was approved by the local ethical committees representing all participating populations and all subjects signed an informed consent.

Table 1

Demographic and life style characteristics of men in the study groups

Greenland n = 72

Warsaw n = 98

Sweden n = 100

Kharkiv n = 88

All n = 358

Age

median

30 (70)

30 (97)

47 (96)

25 (85)

31 (348)

years

min-max

18–46

24–46

24–68

16–45

16–68

BMI

median

26 (72)

25 (96)

26 (99)

24 (86)

25 (353)

Kg/m²

min-max

12–38

19–38

22–37

19–36

12–38

Alcohol

median

2.0 (36)

3.5 (86)

n.a

2.5 (66)

3.0 (188)

drink/week

min-max

0–21

0–30

0.2–15

0–30

Smoking ever

89 (71)

50 (97)

60 (98)

82 (87)

68 (353)

Seafood

median

1.5 (68)

1.0 (89)

n.a

4.0 (86)

2.0 (243)

days/week

min-max

0–9.0

1.0–9.0

0–9.0

Coffee

median

3.0 (62)

2.0 (83)

n.a

2.0 (37)

2.0 (182)

cups/day

min-max

0–20

0–6.0

0.5–7.0

0–20

Total testosterone

median

15 (36)

13 (76)

12 (99)

18 (88)

14 (299)

nmol/l

min-max

3.2–25

6.5–23

4.2–28

8.4–32

3.2–32

Estradiol

median

59 (35)

71 (76)

67 (99)

78 (88)

70 (298)

nmol/l

min-max

31–89

45–297

25–154

33–144

25–297

The values in the bracket are the numbers of individuals with data for the specific demographic and life style characteristics.

2.2 Determination of CB-153 and p,p'-DDE in serum

Serum concentrations of CB-153 and p,p'-DDE were determined using solid phase extraction (SPE) and on-column degradation of lipids followed by analysis with gas chromatography mass spectrometry as described 35. CB-153 and p,p'-DDE levels were adjusted for serum lipids. Levels of detection, coefficients of variation and participation in quality control programs have been described in detail elsewhere 3536.

2.3 SPE-HPLC fractionation of the serum samples

To obtain the serum fraction containing the actual mixture of bio-accumulated POPs a solid phase extraction (SPE) followed by high-performance liquid chromatography (HPLC) fractionation was performed on 3.6 ml serum 37. Similar to the described methods 173038 the POPs were extracted from the serum by SPE separation on a Spherisorb Si 60 analytical column (250 × 4.6 mm i.d., 5 μm particle size, Waters, Milford, MA, USA). This is a normal phase column and thus the separation is based on lipophilicity. This crude serum extract was then further processed using HPLC in order to separate the POPs from the endogenous hormones. The first fraction (F1: 0.00–5.30 min, protected from light in brown tubes) was set up to include most POPs while leaving out endogenous hormones by testing 18 lipophilic chemicals including PCBs, pesticides, akylphenols, bisphenol A, 6 endogenous hormones (estrogens and androgens), and 1 oral contraceptive on the HPLC system using retension times as discriminating factor. This F1 SPE-HPLC fraction was evaporated to near dryness and frozen for later ER mediated chemical activated luciferase gene expression (ER-CALUX) analysis. ER-CALUX analyses of the HPLC fractions F1, F2.1 (5.30 – 12.0 min) and F2.2 (12.00 – 14.5 min) verified that the endogenous hormones were separated from F1 but found in F2.1 and F2.2. Moreover, comparison of SPE-HPLC F1 extracts obtained by our system with F1 extracts obtained from another laboratory (Institute of Public Health, University of Southern Denmark) elicited same responses in the ER-CALUX analyses 37.

Two pools of blood bank serum (Aarhus Sygehus, Denmark), one male (KHM) and one female (KHF), were distributed into tubes with 3.6 ml and stored at -80°C. One sample from each sex was weekly processed by the SPE-HPLC method in parallel with the project samples serving as serum controls for the cleanup procedure. The day to day inter assay coefficient of variation (CV) of the SPE-HPLC + ER-CALUX analyses of these control blood samples were ≤ 13%.

2.3.1 Dissolving the SPE-HPLC samples

The SPE-HPLC extracts samples (project samples and controls) were thawed and protected from light during handling. Sample solvent, 20 μl EtOH:H₂O:DMSO (50:40:10), was added to each sample tube and the samples were placed in an Eppendorf Thermomixer Comfort at 550 rpm and 37°C for 15 minutes. Then 200 μl growth media [0.5% Dextran-Charcoal treated foetal calf serum (DC-FCS, Hyclone, Bie & Berntsen, Aarhus, DK) in phenol red-free Dulbecco's Modified Eagle's Medium (DMEM) (Gibco, Invitrogen, Taastrup, DK) supplemented with 4 mM glutamine (Sigma-Aldrich, Vallensbaek Strand, DK) and 1.28 mg/ml garamycin (Schering-Plough, Dassel, Germany)] was added to each sample and the shaking at 550 rpm and 37°C was continued for another 15 minutes. Then each sample was transferred to two new test tubes (100 μl/tube) each containing 400 μl growth media with or without the normal ER ligand E2, respectively, and used for ER-CALUX determinations.

2.4 ER-CALUX assay

The stable transfected MVLN human breast cancer cell line (kindly provided by M. Pons, France) carrying the endogenous ERα and ERβ genes, and the introduced estrogen-response-element (ERE)-luciferase reporter vector 39 was used for determination of ER-CALUX activities as described 40. The measured luciferase data was corrected to cell protein (internal control), and expressed in relative light units per microgram protein (RLUs/μg protein). All controls and SPE-HPLC F1 serum extracts (100 μl/well) were analysed in triplicate. If one of the triplicate values deviated more than 30% the mean was calculated from two wells only. An E2 dose-response control in the concentration range from 0.05 pM to 500 pM (solvent 96% ethanol = 0.1%) was performed in parallel each analysis day as described 1440. The maximal effective concentration (EC₁₀₀), the half maximum effective concentration (EC₅₀), and EC₄₀of E2 was calculated to 150 pM, 33 pM and 25 pM, respectively, by fitting the dose-response data into Chapman, 4 parameter sigmoid equation curve using Sigma Plot 8.0 (SPSS, Chicago, IL, USA). The E2-EC₁₀₀and the E2-EC₄₀were used as parallel positive controls on each plate in each assay. Xenoestrogenic activity (XER) was determined by exposure of the cells to serum extracts alone and XERcomp activity upon co-exposure of the cells to serum extract + 25 pM E2 (EC₄₀). The solvent controls of XER and XERcomp activity were sample solvent only and the sample solvent plus 25 pM E2 (EC₄₀), respectively (see 2.3.1 for handling procedure), and for XER and XERcomp activity calculation the respective solvent controls (RLU/μg protein) were set to 1. Finally, the data was given as activity per ml serum and the values of the solvent controls were 3.13 RLU/ml serum.

The E2 equivalence (XER-EEQ) value of the agonistic XER activities was obtained by interpolation of data onto the Chapman 4 parameter sigmoid curve using the dose-response E2 Sigma Plot curve. Both the average ER-CALUX inter-CV of the solvent controls and intra-individual CV for serum sample extracts were below 5%.

No cell toxicity on MVLN determined by CellTiter 96 Cell Proliferation assay from Promega (Madison WI, US) 40 was observed after exposure to SPE-HPLC extracts.

2.5 Statistical analysis

The comparisons of means between XER, XERcomp and XER-EEQ were performed by the Oneway-ANOVA test. When ANOVA indicated significant group difference complementary multiple comparison ad hoc tests were used. The test for homogeneity of variance was performed with Levene's test. The least significant difference (LSD) pair wise multiple comparison test was used for the variables with equal variance (p ≥ 0.05) and Dunnett's T3 test was used for the variables with an unequal variance (p ≤ 0.05).

The association in each study group between POP marker and xenoestrogen activity was evaluated by means of Spearman's rank correlation. The overall association between the POP marker and xenoestrogen activity across the study groups (combined data) were evaluated by comparing the regression lines for each study group (multiple regression analysis).

Our hypothesis is that a potential determinant of POP bioaccumulation might also be a potential determinant for serum xenoestrogenic activity. As known from the literatures and also from an assessment of the total Inuendo study populations 2535 age and seafood are such determinants. Moreover, lifestyle characteristics (Table 1) were evaluated as potential determinants of XER and XERcomp levels of combined data and the separate study groups. Using multivariate linear regression analyses, assessing the impact of the POP biomarkers on XER and XERcomp, the impact of potential confounders were evaluated by entering blocks of variables together with either CB-153 or p,p'-DDE as follows: In the first step age and seafood intake (continuous variables) were included in the model, and in the second step additionally smoking status (smoked ever yes/no), BMI, and coffee intake was included as continuous variables. Alcohol consumption was only recorded for limited number (n = 188, see Table 1) of subject and not for the Swedish study group. Therefore, the second step was carried out with and without alcohol consumption in the potential confounder model, which did not change the overall pattern of non-adjusted to adjusted data. Due to many missing values on the potential confounders (Table 1) the number of available observations in the confounder analyses are much smaller than in the unadjusted analysis on the full dataset (full dataset (XER, XERcomp and POP markers) N = 348; first step of confounders: N = 231; second step: N = 172). A reduction of the number of observations with 50% might introduce serious selection problems, and hence the confounder analyses might lack greater validity.

The XER and XERcomp activities were determined in protein free serum extracts free of endogenous estrogens and testosterone. As a method verification Spearman's rank correlation analyses of XER and XERcomp against determined blood levels of estradiol and testosterone (total and free) 41 were performed on the combined study group data.

Normal distribution was assessed by Q-Q plots. To improve normality and homogeneity of variance, the XER and XERcomp activities as well as lipid adjusted CB-153 (lip-CB-153) and p,p'-DDE (lipDDE) were natural logarithmic transformed and the statistical analyses were performed on the ln-transformed data. The POP variables were treated as continuous variables. The statistical analysis was performed in SPSS 10.0 (SPSS Inc, Chicago, IL) with the significance level p ≤ 0.05.

3. Results

3.1 Basic characteristics, CB-153, p,p'-DDE and xenoestrogenic serum levels

The distribution of demographic and lifestyle factors (Table 1) and serum CB-153 and p,p'-DDE median levels (Table 2) that may potentially influence the ER-mediated activities of the 358 adult males in this study were similar with that obtained for the total Inuendo study population 2535. Correlation between serum concentration of CB-153 and p,p'-DDE was found with higher correlations in Greenland (r_s= 0.94) and Sweden (r_s= 0.75), while relatively lower correlation was observed for the study group of Kharkiv (r_s= 0.45) and Warsaw (r_s= 0.28).

Table 2

Xenoestrogenic serum activities, estradiol equivalents and lipid adjusted CB-153 and p,p'-DDE in serum of the study groups

Greenland

Warsaw

Sweden

Kharkiv

ALL* ⁴

All study group data* ⁵

XER* ¹ RLU/ml serum

100

358

0.003

Median

2.89

3.09

3.04

3.15

3.05

Mean

2.92

3.30

3.22

3.17

Min

1.0

2.4

1.0

1.1

Max

6.0

6.5

8.0

% agonist

% antagonist

XER-EEQ pg/g lipid* ²

1^♣

0.63

Median

103

76♥

139

114

Mean

166

161

179

171

Min

Max

516

364

580

XER comp* ³ RLU/ml serum

348

< 0.001

Median

2.65

2.96

2.90

2.88

2.86

Mean

2.69

3.29

2.89

2.87

2.95

Min

2.0

1.8

1.0

1.1

1.0

Max

3.8

7.0

6.8

4.5

7.0

% add/syn

% antagonist

CB-153 ng/g lipid

100

354

< 0.001

Median

220

210

Min

5.1

4.1

5.5

3.3

Max

5500

130

1500

200

5500

DDE ng/g lipid

100

354

< 0.001

Median

630

570

240

880

560

Min

240

324

Max

13000

2100

2300

12000

13000

In the independent assays significant differences between the triple serum extract determinations and their respective solvent controls (% agonists/additive/synergistic) and % antagonists, were tested by the Student's t-test (Microsoft Office Excel, significance level p = 0.05). *¹: XER: Xenoestrogenic agonistic or antagonistic activity of serum extract alone: the solvent control = 3.13 RLU/ml serum. % agonistic and % antagonistic indicates the % of samples eliciting a significantly increase or decrease in XER activity compared to solvent control, which was set to 1. *²: XER-estradiol equivalence (XER-EEQ) of the samples eliciting significantly agonistic activity; data calculated by interpolation to the E2 dose response curve using the sigma plot program. ♥ : One Swedish sample was too high to calculate the XER-EEQ. ♣ : One serum sample from Greenland only had an agonistic action, thus no XER-EEQ value is given. *³: XERcomp; XER competitive activity of serum extract + 25 pM 17β-estradiol (E2-EC₄₀); % add/syn: additive/synergistic and % antagonistic indicates the % of samples responding with a further increase or a decrease of the E2-EC₄₀induced activity, respectively, compared to the E2-EC₄₀solvent control = 3.13 RLU/ml serum, which was set to 1;. *⁴: XER activity includes 358 samples, however, only 348 of theses samples were applied to the XERcomp analyses. All of the samples had CB-153 and p,p'-DDE determined. *⁵: One-way ANOVA (p value) evaluation of the combined study group data except for XER-EEQ including only the European groups. In transformed data was used.

The Greenlandic serum extracts elicited predominantly antagonistic effect on XER (35%) and XERcomp (71%), whereas the European samples elicited relatively higher frequency of XER agonistic activity (12–21%) of which the Warsaw samples had the highest mean of XER/XERcomp (Table 2 and Figure 1).

Figure 1

The xenoestrogenic CALUX activity of study groups

The xenoestrogenic CALUX activity of study groups. (A) Agonistic activity of serum extracts alone (XER) and (B) competitive XER activity upon coexposure with 25 pM E2 (EC₄₀) (median, quartiles (25% and 75%) and extreme variables). For the Swedish study group an extreme agonistic (A) RLU value of 12.02 was determined (not shown). The reference line of the respective solvent controls ± SD (3.13 ± 0.16) are given as dotted lines.

Oneway ANOVA analyses showed significant differences between the study groups (Table 2). Multiple comparisons of means showed that Inuit XER activities significantly differed from each of the three European study groups (p ≤ 0.03); whereas the XER activity and XER-EEQ of the European study groups did not mutually differ. Only one sample from Greenland elicited XER agonistic activity and thus no XER-EEQ data was given for the Inuit's. The XERcomp activity of the Warsaw group differed from all other study groups (p ≤ 0.004).

3.2 Associations between xenoestrogenic activity and CB-153 and p,p'-DDE

XER activity for Inuit's showed an inverse correlation to the p,p'-DDE levels (r_s= -0.29; p = 0.02) and borderline inverse to CB-153 (r_s= -0.22; p = 0.07). For the Warsaw study group the XER activity showed a significant positive correlation to the p,p'-DDE data (r_s= 0.21; p = 0.04), and for XER-EEQ a significant negative correlation to CB-153 (r_s= -0.45; p = 0.04) was found (See Additional file 1: Spearman's correlation analyses between xenoestrogenic serum activities and the level of CB-153 and p,p'-DDE). No further correlations between XER, XERcomp or XER-EEQ and any of the two POP markers were found. Adjustment for the potential confounders in the multivariate regression model did not give any different results for the impact of the POP biomarkers on XER and XERcomp as compared with the unadjusted models, neither for the combined data or each study group (data not shown).

Additional File 1

Title: Spearman's correlation analyses between xenoestrogenic serum activities and the level of CB-153 and p,p'-DDE. ln-transformed and POP lipid adjusted data was used. For definition of XER, XERcomp and XER-EEQ see legend to Table 2. Statistical significant data is given in bold. ¹: Spearman's inter-correlation between CB-153 and p,p'-DDE, respectively.

Click here for file

3.3 Multiple regression of xenoestrogenic activity on POP data across the study groups

Scatter plots of XER or XERcomp against CB-153 or p,p'-DDE for the four study groups are shown in Figure 2. Multiple regression analysis of both response variables (XER and XERcomp) showed homogeneity of slopes between study groups and CB-153/p,p'-DDE (Table 3), i.e. parallel regression lines among study groups. Furthermore, a model with parallel regression lines showed significant differences between the intercepts of the study groups (Table 3). Thus the differences in XER/XERcomp between study groups found by Oneway ANOVA (Table 2) and multiple comparisons still exist after adjustment for CB-153 or p,p'-DDE. Finally, we note that although correlations was observed between XERs and the POP markers for the Greenlandic and Warsaw study groups (Supplement), for the combined data the associations between XER/XERcomp and CB-153/DDE, as measured by the common slopes, are very weak and not statistically significant (Table 3).

Figure 2

Schematic illustration of xenoestrogenic activities related to the POP markers

Schematic illustration of xenoestrogenic activities related to the POP markers. Xenoestrogenic (XER) and competitive XER (XERcomp) serum activities related to CB-153 and p,p'-DDE levels for the four country based study groups as a relation between (A) XER and p,p'-DDE, (B) XER and CB-153, (C) XERcomp and p,p'-DDE, and (d) XERcomp and CB-153. The values are given as ln transformed data.

Table 3

Multiple regressions of the combined study group data

Response variable

Exposure variable

Homogeneity of slopes (p-value)

Common slope Estimate (SE) p-value

Common intercept (p-value)

Adjusted R- square

XER (N = 348)

CB-153

0.86

-0.01 (0.01), 0.34

0.05

0.029

p,p'- DDE

0.24

-0.03 (0.02), 0.12

0.005

0.034

XERcomp (N = 338)

CB-153

0.20

0.01 (0.02), 0.45

<0.001

0.087

p,p'- DDE

0.15

-0.006 (0.02), 0.71

<0.001

0.086

Response and exposure variables are ln-transformed in the analysis. Homogeneity of slopes: p-value for the test for homogeneity of slopes between study group and exposure variable. Common slope: the estimated common slope across study groups assuming no interaction. Common intercept: p-value for the test of a common intercept across study groups assuming a common slope. Adjusted R-square assuming a common slope.

3.4 Correlations between xenohormone activities and endogenous hormone levels

To verify the exclusion of endogenous hormones from the SPE-HPLC- F1 serum extracts used for CALUX activity measurements, the XER, XER-EEQ and XERcomp results were evaluated for possible correlation to blood estradiol (pmol/L) and testosterone (nmol/L) (free and total) levels. No correlations were found between the serum XER/XERcomp activities and blood sex hormone levels neither for the combined data nor for the separate study groups.

4. Discussion

In the present study we determined the actual xenoestrogenic net activity in serum of adult men from Greenland, Sweden, Warsaw (Poland) and Kharkiv (Ukraine), representing populations with different POP exposure patterns (Table 2) 34354243. For the combined data no strong consistent correlations between serum xenoestrogenic activity and the two POP proxy markers, CB-153 and p,p'-DDE, was observed. Significant correlations between marker POPs and xenoestrogenic serum activity was found for Inuit's and the Warsaw group, having the lowest and highest ratio of agonist:antagonist samples, respectively; an inverse association of XER activity to the p,p'-DDE and CB-153 levels for Inuit's, and for the Warsaw group a positive and a negative association of XER to p,p'-DDE and XER-EEQ to CB-153, respectively. Because of the high inter-correlation between CB-153 and p,p'-DDE for the Inuit's it can not be assessed which of the two POPs exert the main impact. However, for the Warsaw group the very low concentration of CB-153 and low inter-correlation of marker POPs indicates that p,p'-DDE might have a positive effect on XER activity, and CB-153 an inverse impact on agonistic XER as elicited by XER-EEQ. Overall, these data might be explained by the high level of PCBs in Greenland with predominant anti-estrogenic actions 1517444546 and relatively high level of p,p'-DDE in the Warsaw group having weak estrogenic action 17284748. However, that does not explain why no correlations were observed for the Swedish and Kharkiv groups since they had a CB-153 and p,p'-DDE marker profile similar to Greenland and Warsaw, respectively. Multiple regression analyses showed, that the determined differences of XER and XERcomp activity between the study groups could not be explained by the two selected POP proxy markers alone neither in combination with the measured potential confounders in the present study. However, the combined XER vs. CB-153 data showed homogeneity across the study groups, had common slopes and similar intercept (p = 0.05, Table 3, Figure 2), and assuming a common intercept a significant (β = -0.132, p = 0.01) negative correlation between CB-153 and XER activity was found. Future studies might elucidate whether our observations are serendipitous.

Between the groups the XER and XERcomp activities differed significantly. The Inuit's clearly differed from the European study groups eliciting no agonistic but high frequency of samples (71%) with antagonistic action. In contrast, the European serum samples exerted both agonistic (12 – 21% of the samples) and antagonistic effects (7 – 30% of the samples).

The heterogeneous pattern of xenoestrogenic correlations to CB-153/p,p'-DDE levels and also the differences in xenoestrogenic activity among the study groups might be explained by different exposure profiles e.g. : i) the estrogenic effects elicited by ER transactivity is a consequence of the combined net effects of many POPs having either ER agonistic or ER antagonistic actions; ii) the two selected POP markers do not reflect and/or represent the chemicals primarily responsible for the determined xenoestrogenic effects; iii) the combined response of the serum mixture is significant affected although the concentration of the single compounds found in human serum is lower than that needed to elicit a response of the compound on its own in an in vitro system; and finally iv) the net XER/XERcomp activities is the net response of exposure patterns, genetic background and life style factors. In summary, based on the two selected POP markers only, it is hard to predict the net xenoestrogenic impact in vivo as well as in the present ex vivo xenoestrogenic measurements.

The use of CB-153 and p,p'-DDE in serum as index biomarkers of POP exposure has been supported by several previous studies 49505152. However, it must be taken into consideration that the relative concentrations of non-coplanar and co-planar PCBs can differ between regions depending on varying exposure sources. The ratio between serum levels of CB-153 (and other non-coplanar congeners) and co-planar PCBs was higher in Canadian Inuit's than in Canadian Caucasians from the Arctic area 53, which calls for caution using CB-153 as a global exposure marker for POP as well as for xenoestrogenic serum activities. In the present study it might be expected that the ratio between non-dioxin-like, non-coplanar PCBs (e.g. CB-153) and co-planar PCB (e.g. CB-126) also is higher for the Greenlandic Inuit's compared to the three European study groups.

In vitro studies demonstrated that the combined effect of several xenoestrogens including POPs at sub-NOEC exerted an additive effect, which led to dramatic change in estrogenic in vitro activity 2829. To elucidate potential human and wildlife responses to the combined impact of accumulated xenoestrogens it remains to be demonstrated mechanistically how chemicals with antiestrogen effects can modify the response of chemicals acting estrogenic in a concerted action. It can be argued that xenoestrogenic equivalents contribute only a few percent to the endogenous hormone activity – even when possible higher bioavailability of the xenobiotics is taken into consideration. However, considering the further increase of XERcomp above the reference control (25 pM E2) we calculated (using Sigma Plot and the data of the E2 dose-response curve) for this subgroup of men (n = 86) a further mean activity increase of 47% and 21% in relation to the minimum (34 pM) and mean (70 pM) endogenous estradiol level, respectively. Since the level of endogenous estrogens is much lower in male than in female the xenoestrogenic activity might have higher impact on health risk in males.

In studies using similar approach as the present no correlation between neither CB-153 nor p,p'-DDE and xenoestrogenicity were found. In contrast to our determination of the more specific impact of serum xenoestrogens on ER transactivation, the other studies used the E-screen MCF-7 cell proliferation as end point 17303132. Moreover, we analyzed the xenoestrogenic action in male blood samples whereas the other four studies were focused on female adipose tissues 303132 and female serum from pregnant and non-pregnant women 17. These studies determined a higher frequency (~ 60–70%) of subjects with significantly increased estrogenic activity in the non-polar serum fraction of Spanish and Faroese women. For non-pregnant Danish women the frequency of subjects exceeding the background level was 22.7%, similarly to the frequency determined for the European males in the present study. In support to our data it was recently reported that high levels of PCBs in Slovakia male serum samples were associated with a decreased ER mediated activity and increased AhR mediated activity 46. Recently, estrogenicity of adipose tissue extracts including bio-accumulated xenoestrogens was shown associated with higher risk of breast cancer in leaner women 32.

4.1 Conclusions and perspectives

Across the four study groups no strong consistent correlation between xenoestrogenicity and the two POP proxy markers was found. Thus, the CB-153 and p,p'-DDE alone are not optimal global POP markers of the integrated xenoestrogenic serum activity. However, between the study groups a clear difference in serum xenoestrogenicity as well as differences in correlation to the CB-153 and p,p'-DDE proxy markers was found. Correlations to the POP markers were found for the two study groups at the extreme edge: Inuit's with high sample frequency of antiestrogenicity, and the Warsaw group with higher frequency of samples with estrogenicity. In support to POP biomarker determinations, the present ex vivo biomarker xeno-activity assay may be useful in assessment of geographical surveys for the impact of environmental xenoestrogens, reflecting differences in POP exposure patterns.

On the other hand, exposure assessment based upon measurements of the actual xenoestrogenic serum activity of the isolated SPE-HPLC, F1- POP- fraction may fail to detect effects accomplished 1) by other non-lipophilic compounds in serum, 2) through other receptors or 3) by other mechanisms. More work still needs to be done to fully understand to which extent ex vivo measurements of receptor-transactivation reflect the in vivo situation.

Competing interests

The author(s) declare that they have no competing interests.

Authors' contributions

ECB-J drafted the work and was the main responsible for design, performance and data evaluation of the specific project; PHJ, TSR and BSA performed the mechanistic work; CHL performed POP determinations in blood; ME was the main responsible for the statistical work; JPB, AG and LH designed the overall Inuendo project. JPB and GT coordinated the execution of the project and GT had main responsibility for creating the joint database. All authors participated in the design of the study and commented on the final manuscript.

Acknowledgements

Thanks to all CMT group members; Inger Sørensen for technical assistance, Tanja Krüger, Manhai Long and Mandana Ghisari for scientific support. A special thanks to Jan K. Ludwicki for coordination of the blood sampling in Warsaw. Also a special thanks to Anna Rignell-Hydbom, Katarzyna Goralczyk, Henning S. Pedersen for taking the blood samples in Sweden, Warswa and Greenland, respectively. Thanks to Héléne Åkesson, Berit Holmskov and Christina Held for performing chemical analyses in a skillful way. This study is part of the Project "INUENDO 33- Biopersistent organochlorines in diet and human fertility. Epidemiological studies of time to pregnancy and semen quality in Inuit and European populations", supported by The European Commission to the 5^thFramework Programme Quality of Life and Management of Living Resources, Key Action 4 on Environment and Health (Contract no. QLK4-CT-2001-00202) and INTAS (contract no 2001–2005). The work has also been funded by the Danish Environmental Protection Agency, the Swedish Research Council and the Swedish Council for Environment, Agricultural Sciences and Spatial Planning.

Human health and chemical mixtures: an overview Carpenter DO Arcaro KF Bush B Niemi WD Pang S Vakharia DD Environ Health Perspect 1998 106 suppl 6 1263 1270 9860880 The Human Health Effect Programme in Greenland, a review Bonefeld-Jorgensen EC Sci Total Environ 2004 331 215 231 10.1016/j.scitotenv.2004.03.030 15325151 Toxicological Properties of POPs and Related Health Effects of Concern for the Arctic Populations (Chapter 6) Bonefeld-Jorgensen EC Ayotte P AMAP assessment 2002: Human Health in the Arctic Oslo, Norway 2003 57 74 Volume on-line http://www.amap.no. Arctic Monitoring and Assessment Programme (AMAP), Oslo, Norway. xiv + 137pp Selective DNA recognition by the androgen receptor as a mechanism for hormone-specific regulation of gene expression Verrijdt G Haelens A Claessens F Mol Genet Metab 2003 78 175 185 10.1016/S1096-7192(03)00003-9 12649062 Nuclear receptors: the evolution of diversity Schwabe JW Teichmann SA Sci STKE 2004 2004 pe4 14747695 Mechanisms of estrogen receptor signaling: convergence of genomic and nongenomic actions on target genes Bjornstrom L Sjoberg M Mol Endocrinol 2005 19 833 842 10.1210/me.2004-0486 15695368 Estrogen in the adult male reproductive tract: a review Hess RA Reprod Biol Endocrinol 2003 1 52 179885 12904263 10.1186/1477-7827-1-52 Immunolocalization of androgen receptor and estrogen receptors alpha and beta in human fetal testis and epididymis Shapiro E Huang H Masch RJ McFadden DE Wu XR Ostrer H J Urol 2005 174 1695 1698 10.1097/01.ju.0000176472.90432.5b 16148684 Estrogens–male hormones? Carreau S Folia Histochem Cytobiol 2003 41 107 111 13678329 Molecular mechanism of estrogen action in the male: insights from the estrogen receptor null mice Couse JE Mahato D Eddy EM Korach KS Reprod Fertil Dev 2001 13 211 219 10.1071/RD00128 11800160 Effects of currently used pesticides in assays for estrogenicity, androgenicity, and aromatase activity in vitro Andersen HR Vinggaard AM Rasmussen TH Gjermandsen IM Bonefeld-Jorgensen EC Toxicol Appl Pharmacol 2002 179 1 12 10.1006/taap.2001.9347 11884232 Effects of currently used pesticides in the AhR-CALUX assay: comparison between the human TV101L and the rat H4IIE cell line Long M Laier P Vinggaard AM Andersen HR Lynggaard J Bonefeld-Jorgensen EC Toxicology 2003 194 77 93 10.1016/j.tox.2003.08.001 14636698 Effects of the pesticides prochloraz and methiocarb on human estrogen receptor alpha and beta mRNA levels analyzed by on-line RT-PCR Hofmeister MV Bonefeld-Jorgensen EC Toxicol In Vitro 2004 18 427 433 10.1016/j.tiv.2003.12.008 15130599 Effect of in vitro estrogenic pesticides on human oestrogen receptor alpha and beta mRNA levels Grunfeld HT Bonefeld-Jorgensen EC Toxicol Lett 2004 151 467 480 10.1016/j.toxlet.2004.03.021 15261991 Effect of highly bioaccumulated polychlorinated biphenyl congeners on estrogen and androgen receptor activity Bonefeld-Jorgensen EC Andersen HR Rasmussen TH Vinggaard AM Toxicology 2001 158 141 153 10.1016/S0300-483X(00)00368-1 11275356 Characterization of potential endocrine-related health effects at low- dose levels of exposure to PCBs Brouwer A Longnecker MP Birnbaum LS Cogliano J Kostyniak P Moore J Schantz S Winneke G Environ Health Perspect 1999 107 Suppl 4 639 649 10421775 Assessment of xenoestrogenic exposure by a biomarker approach: application of the E-Screen bioassay to determine estrogenic response of serum extracts Rasmussen TH Nielsen F Andersen HR Nielsen JB Weihe P Grandjean P Environ Health 2003 2 12 270076 14613489 10.1186/1476-069X-2-12 Interaction between human plasma sex hormone-binding globulin and xenobiotic ligands Jury HH Zacharewski TR Hammond GL Journal of Steroid Biochemistry & Molecular Biology 2000 75 167 176 10.1016/S0960-0760(00)00168-0 Effect of polychlorinated biphenyl, Aroclor 1254 on rat epididymis Anbalagan J Kanagaraj P Srinivasan N Aruldhas MM Arunakaran J Indian J Med Res 2003 118 236 242 14870796 Effects of gestational and lactational exposure to Aroclor 1242 on sperm quality and in vitro fertility in early adult and middle-aged mice Fielden MR Halgren RG Tashiro CH Yeo BR Chittim B Chou K Zacharewski TR Reprod Toxicol 2001 15 281 292 10.1016/S0890-6238(01)00129-0 11390173 Persistent DDT metabolite p,p'-DDE is a potent androgen receptor antagonist Kelce WR Stone CR Laws SC Gray LE Kemppainen JA Wilson EM Nature 1995 375 581 585 10.1038/375581a0 7791873 Developmental and reproductive toxicity of dioxins and related compounds: cross-species comparisons Peterson RE Theobald HM Kimmel GL Crit Rev Toxicol 1993 23 283 335 8260069 Effect of early postnatal exposure to polychlorinated biphenyls (PCBs) on fertility in male rats Sager DB Shih-Schroeder W Girard D Bull Environ Contam Toxicol 1987 38 946 953 10.1007/BF01609078 3107640 Endocrine disruption and hypospadias Steinhardt GF Adv Exp Med Biol 2004 545 203 215 15086029 Fertility in four regions spanning large contrasts in serum levels of widespread persistent organochlorines: a cross-sectional study Toft G Axmon A Giwercman A Thulstrup A Rignell-Hydbom A Pedersen HS Ludwicki J Zvyesday V Zhinchuk A Spano M Environ Health 2005 4 26 1308837 16280075 10.1186/1476-069X-4-26 Epidemiological evidence on reproductive effects of persistent organochlorines in humans Toft G Hagmar L Giwercman A Bonde JP Reprod Toxicol 2004 19 5 26 10.1016/j.reprotox.2004.05.006 15336708 Integrated risk assessment and endocrine disrupters Bridges JW Bridges O Toxicology 2004 205 11 15 10.1016/j.tox.2004.06.033 15458785 Mixtures of four organochlorines enhance human breast cancer cell proliferation Payne J Scholze M Kortenkamp A Environ Health Perspect 2001 109 391 397 1240280 11335188 Combining xenoestrogens at levels below individual no-observed-effect concentrations dramatically enhances steroid hormone action Rajapakse N Silva E Kortenkamp A Environ Health Perspect 2002 110 917 921 1240992 12204827 Human exposure to endocrine disrupters: standardisation of a marker of estrogenic exposure in adipose tissue Rivas A Fernandez MF Cerrillo I Ibarluzea J Olea-Serrano MF Pedraza V Olea N Apmis 2001 109 185 197 10.1034/j.1600-0463.2001.090302.x 11430496 Assessment of total effective xenoestrogen burden in adipose tissue and identification of chemicals responsible for the combined estrogenic effect Fernandez MF Rivas A Olea-Serrano F Cerrillo I Molina-Molina JM Araque P Martinez-Vidal JL Olea N Anal Bioanal Chem 2004 379 163 170 10.1007/s00216-004-2558-5 15024544 Breast cancer risk and the combined effect of environmental estrogens Ibarluzea Jm J Fernandez MF Santa-Marina L Olea-Serrano MF Rivas AM Aurrekoetxea JJ Exposito J Lorenzo M Torne P Villalobos M Cancer Causes Control 2004 15 591 600 10.1023/B:CACO.0000036167.51236.86 15280638 <url>http://www.inuendo.dk</url> The European Commission to the 5 th Framework Programme Quality of Life and Management of Living Resources, Key Action 4 on Environment and Health (Contract no. QLK4-CT-2001-00202) and INTAS (contract no 2001–2005). Biopersistent organochlorines in diet and human fertility. Epidemiological studies of time to pregnancy and semen quality in Inuit and European populations Biological Monitoring: Human Tissue Levels of Environmental Contaminants, Chapter 5 AMAP Assessment Human Health in the Arctic 2003 31 51 Volume Publications online http://www.amap.no (DEFINITIVE_POPS6.pdf). Oslo, Norway: Arctic Monitoring and Assessment Programme (AMAP), Oslo, Norway. xiv+137pp Inter-population variations in concentrations, determinants of and correlations between 2,2',4,4',5,5'-hexachlorobiphenyl (CB-153) and 1,1-dichloro-2,2-bis (p-chlorophenyl)-ethylene (p,p'-DDE): a cross-sectional study of 3161 men and women from Inuit and European populations Jonsson BA Rylander L Lindh C Rignell-Hydbom A Giwercman A Toft G Pedersen HS Ludwicki JK Goralczyk K Zvyezday V Environ Health 2005 4 27 1308838 16283941 10.1186/1476-069X-4-27 Exposure to CB-153 and p,p'-DDE and male reproductive function Rignell-Hydbom A Rylander L Giwercman A Jonsson BA Nilsson-Ehle P Hagmar L Hum Reprod 2004 19 2066 2075 10.1093/humrep/deh362 15284211 SPE-HPLC purification of endocrine disrupting compounds from human serum for assessment of xenoestrogenic activity Hjelmborg PS Ghisari M Bonefeld-Jorgensen EC Analytical and Bioanalytical Chemistry 2006 Development of a marker of estrogenic exposure in human serum Sonnenschein C Soto AM Fernandez MF Olea N Olea-Serrano MF Ruiz-Lopez MD Clin Chem 1995 41 1888 1895 7497650 A new cellular model of response to estrogens: a bioluminescent test to characterize (anti) estrogen molecules Pons M Gagne D Nicolas JC Mehtali M Biotechniques 1990 9 450 459 2257144 Effect of pesticides on estrogen receptor transactivation in vitro: A comparison of stable transfected MVLN and transient transfected MCF-7 cells Bonefeld-Jorgensen EC Grünfeld HT Gjermandsen IM Mol Cell Endocrinol 2005 244 20 30 10.1016/j.mce.2005.01.017 16219411 Reproductive Hormone Levels in Men Exposed to Persistent Organohalogen Polutants: A Study of Inuit and Three European Cohorts Giwercman Aleksander AR-H Gunnar Toft Lars Rylander Lars Hagmar Lindh C Henning S Pedersen Jan K Ludwicki Maryna Shevets Valentyna Zvyezday Marcello Spano Gian Carlo Manicardi Davide Bizarro Eva Bonefeld-Jorgensen Jens Peter Bonde INUENDO Environmental Health Perspectives 2006 Submitted the 19 th of December 2005; in revision for publication 1459925 16675426 Fish consumption and exposure to persistent organochlorine compounds, mercury, selenium and methylamines among Swedish fishermen Svensson BG Nilsson A Jonsson E Schutz A Akesson B Hagmar L Scand J Work Environ Health 1995 21 96 105 7618064 Mortality and cancer incidence among women with a high consumption of fatty fish contaminated with persistent organochlorine compounds Rylander L Hagmar L Scand J Work Environ Health 1995 21 419 426 8824747 Congener-specific accumulation of polychlorinated biphenyls in ovarian follicular wall follows repeated exposure to PCB 126 and PCB 153. Comparison of tissue levels of PCB and biological changes Gregoraszczuk EL Grochowalski A Chrzaszcz R Wegiel M Chemosphere 2003 50 481 488 10.1016/S0045-6535(02)00637-9 12685747 Hydroxylated polychlorinated biphenyls (PCBs) as estrogens and antiestrogens: structure-activity relationships Connor K Ramamoorthy K Moore M Mustain M Chen I Safe S Zacharewski T Gillesby B Joyeux A Balaguer P Toxicol Appl Pharmacol 1997 145 111 123 10.1006/taap.1997.8169 9221830 Impact of polychlorinated biphenyls contamination on estrogenic activity in human male serum Pliskova M Vondracek J Canton RF Nera J Kocan A Petrik J Trnovec T Sanderson T van den Berg M Machala M Environ Health Perspect 2005 113 1277 1284 1281266 16203234 Comparison of short-term estrogenicity tests for identification of hormone-disrupting chemicals Andersen HR Andersson AM Arnold SF Autrup H Barfoed M Beresford NA Bjerregaard P Christiansen LB Gissel B Hummel R Environ Health Perspect 1999 107 89 108 10229711 Transcriptional activation of the human estrogen receptor by DDT isomers and metabolites in yeast and MCF-7 cells Chen CW Hurd C Vorojeikina DP Arnold SF Notides AC Biochem Pharmacol 1997 53 1161 1172 10.1016/S0006-2952(96)00879-9 9175721 Monitoring of polychlorinated biphenyls in human blood plasma: methodological developments and influence of age, lactation, and fish consumption Grimvall E Rylander L Nilsson-Ehle P Nilsson U Stromberg U Hagmar L Ostman C Arch Environ Contam Toxicol 1997 32 329 336 10.1007/s002449900193 9096084 Serum concentrations of organochlorines in men: a search for markers of exposure Glynn AW Wolk A Aune M Atuma S Zettermark S Maehle-Schmid M Darnerud PO Becker W Vessby B Adami HO Sci Total Environ 2000 263 197 208 10.1016/S0048-9697(00)00703-8 11194153 Smoking as a determinant of high organochlorine levels in Greenland Deutch B Pedersen HS Jorgensen EC Hansen JC Arch Environ Health 2003 58 30 36 12747516 Polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans, and coplanar polychlorinated biphenyls in breast milk from two cities in Ukraine Gladen BC Schecter AJ Papke O Shkyryak-Nyzhnyk ZA Hryhorczuk DO Little RE J Toxicol Environ Health A 1999 58 119 127 10.1080/009841099157331 10522644 Breast cancer and organochlorines [letter] Dewailly E Ayotte P Brisson J Dodin S Lancet 1994 344 1707 1708 10.1016/S0140-6736(94)90491-X 7996987