- Email: [email protected]
Organochlorines and endometriosis: A mini-review
Available online at www.sciencedirect.com
Chemosphere 71 (2008) 203–210 www.elsevier.com/locate/chemosphere
Review
Organochlorines and endometriosis: A mini-review Jean-Franc¸ois Heilier a
a,*
, Jacques Donnez
b,c
, Dominique Lison
a
Universite´ catholique de Louvain, Faculty of Medicine, Industrial Toxicology and Occupational Medicine Unit, Brussels, Belgium b Universite´ catholique de Louvain, Cliniques Universitaires St-Luc, Department of Gynecology, Brussels, Belgium c Universite´ catholique de Louvain, Faculty of Medicine, Gynecology unit, Brussels, Belgium Received 2 April 2007; received in revised form 26 September 2007; accepted 27 September 2007 Available online 19 November 2007
Abstract Organochlorines (polychlorinated biphenyls and dioxin-like compounds) are suspected to play a role in the etiopathogenesis of endometriosis. This hypothesis, based on experimental data, has been circulating for years in the scientific community and several epidemiologic surveys have attempted to obtain confirmatory human data. The purpose of this mini-review is to provide an overview of the twelve epidemiological studies that have assessed the relationship between endometriosis and organochlorine exposure. Several studies did not observe a significant association between peritoneal endometriosis and organochlorines. The deep nodular form of endometriosis appears associated with a higher serum level of both dioxin-like compounds and polychlorobiphenyls. The type of control women, the nature of the chemicals measured, and the definition of the disease could modulate the ability to detect the possible relationship between endometriosis and organochlorine exposure. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Endometriosis; Deep endometriotic nodules; Dioxins; Tetrachlorodibenzodioxin; Polychlorinated biphenyls; Epidemiology
Contents 1. 2. 3.
4.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dioxins-like compounds and PCBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dioxins, PCBs and endometriosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Type of controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Peritoneal endometriosis or deep endometriotic nodules of the recto-vaginal septum. . . . . . . . . . . . . . . . . . . . 3.3. Nature of chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Issues for future epidemiologic studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Which epidemiological design? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. How to define and assess the presence of endometriosis? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. How to select controls? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. How to assess organochlorine (dioxin) exposure? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Which confounding factor should be controlled?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
204 204 204 204 207 207 208 208 208 208 208 209
Abbreviations: AhR, aryl hydrocarbon receptor; CALUX, chemically activated luciferase expression; OR, odds ratio; PCBs, polychlorinated biphenyls; PCDDs, polychlorinated dibenzo-p-dioxins; PCDFs, polychlorinated dibenzofurans; TCDD (2,3,7,8-TCDD), 2,3,7,8-tetrachlorodibenzo-p-dioxin; TEF, toxic equivalency factor; TEQ, toxic equivalent. * Corresponding author. Tel.: +32 2 764 53 30; fax: +32 2 764 53 38. E-mail address: [email protected] (J.-F. Heilier). 0045-6535/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2007.09.044
204
5.
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
1. Introduction
3. Dioxins, PCBs and endometriosis
Endometriosis is a gynaecological disease characterised by the extra-uterine growth of endometrial tissue which causes internal bleeding, inflammation and scarring, and often leads to infertility (Donnez et al., 2002). Although the exact aetiology of the disease remains unknown, a role of both environmental (Birnbaum and Cummings, 2002; Gerhard and Runnebaum, 1992) and genetic factors (Treloar et al., 2005) has been suspected. Recently, some organochlorines (polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs) and polychlorinated biphenyls (PCBs)) previously suspected as risk factors (Rier et al., 2001), were found to be significantly associated with the disease (Heilier et al., 2004, 2005; Porpora et al., 2006).
The role of PCBs and/or dioxin-like compounds in the aetiopathogenesis of endometriosis is controversial. This hypothesis has been initially based on experimental data reported by Rier et al. (1993) demonstrating, that rhesus monkeys chronically exposed to 2,3,7,8-TCDD (5 and 25 ppt per day during four years) exhibited, 10 years after termination of exposure, peritoneal endometriosis which incidence and severity directly correlated with exposure and dose. For a long time, this experiment appeared as the best evidence of a role for organochlorines in the onset or growth of endometriosis. Rier et al. (2001) reconsidered, some years later, blood samples from this experiment and detected significant amounts of dioxin-like PCBs that originated from a contamination of TCDD, indicating a possible role of dioxin-like compounds rather than TCDD in the endometriotic effect. Criticism on the statistical analysis of Rier’s data has also been raised, after a reappraisal by Guo (2004) who concluded to the absence of relationship with endometriosis. The hypothesis of a role of PCBs and/or dioxins in the onset or the growth of endometriosis has also been addressed in several epidemiologic studies. The purpose of this mini-review is to summarize the existing evidence of a relationship between endometriosis and PCDD/Fs and/or PCBs in humans, with a particular attention to factors that could modify this relation, if any. Factors that are particularly of concern are (i) the selection of control women, (ii) disease definition and (iii) the nature of the chemicals investigated. The main results of 12 publications including 10 casecontrol, 1 cross-sectional and 1 cohort studies are summarized in Table 1. All reported on environmental exposures probably reflecting the fact that occupational exposures have essentially involved men (Zober et al., 1990). Eight studies reported odds ratio (OR), adjusted or not, for linear increment or cut-off in organochlorine blood content. In all studies but one (Fierens et al., 2003) patients and controls underwent gynaecological examination and were classified according to the presence of endometriosis or not. Fierens et al. (2003) classified participants according to selfreported information.
2. Dioxins-like compounds and PCBs Dioxins are made up of 210 chlorinated hydrocarbons divided into 75 PCDDs and 135 PCDFs. Dioxins are mainly by-products of industrial processes (e.g. waste incineration or iron and steel industries), formed when a thermal process (300–800 °C) occurs in the presence of chlorinated organic substances. Seventeen dioxin congeners which preferentially bio-accumulate in the food chain, bind to a specific receptor (aryl hydrocarbon receptor or AhR) that mediates most of their toxic effects and activates several genes, including cytochromes P450 (Hahn, 2003; Mimura and Fujii-Kuriyama, 2003). Several PCBs were manufactured until the 1970s because of their physicochemical properties (stability, viscosity, . . .), and used mostly in closed applications such as heat transfer in capacitors and transformers. Among the 209 PCB congeners (for a full list see http://www.epa.gov/toxteam/pcbid/ table.htm), 12 share with dioxins an activity on the AhR, and are refereed to as WHO-PCBs, dioxin-like PCBs, coplanar PCBs or non (mono)-ortho-PCBs because of the absence of chlorine substitution in ortho positions (1, 4, 6, 9) that gives the molecule a planar configuration. All compounds able to bind the AhR (7 PCDD, 10 PCDF and 12 dioxin-like PCBs) have been classified relatively to 2,3,7,8-TCDD (the most toxic dioxin congener) and received a toxic equivalency factor (TEF) which reflects their respective biological potency. Concentrations of these compounds, often present as mixtures, are frequently expressed as pg TEQ/g lipids, which is the sum of the products of the concentration of each compound multiplied by its TEF. The biological concentration of organochlorines is expressed per g lipids because these compounds are essentially stored in the fatty compartments of the body (van den Berg et al., 1998).
3.1. Type of controls Most studies have included only women undergoing laparoscopic examination (for infertility, tubal ligation, pelvic pain, . . .) who were divided into cases or controls according to the presence of endometriosis or not. A large majority of studies are hospital-based and controls have been recruited in clinics for reproductive medicine or gynaecological
Table 1 Epidemiological studies investigating the possible role of organochlorines in endometriosis Population (n, age [mean ± SD or median (range)])
Diagnosis method (inclusion criteria for cases)
Chemicals (analytical method)
OR [CI95%]/concentration of chemicals (pvalue)
Mayani et al, 1997 (Israel) case-control
Cases: Infertile women (clinic for reproductive medicine) (n = 44, 33.6 ± 6.7 y.) Ctls: idem (n = 35, 34.4 ± 5.8 y.) Cases: women with pelvic pain, infertility or undergoing tubal fulguration (n = 86; 32.2 ± 5.9 y.) Ctls: idem (n = 70, 33.2 ± 7.0 y.) Cases: Infertile women (clinic for reproductive medicine) (n = 42, 31.0 (25–42)) Ctls: idem (n = 27, 32.0 (24–41)) Cases: Women resident of Seveso <30 years in 1976 (n = 19)
Laparoscopy (endometriosis)
2,3,7,8-TCDD (GC–MS)
OR: 7.6 [0.87–169.7] (cut-off 0.4 ng TCDD/ liter blood)
PCBs (# 28; 52; 99; 101; 105; 118; 153; 156; 170; 180; 183; 187), pesticides (GC-ECD)
OR (sum of PCBs): non significantly different (adjusted for age, BMI, number of children and indication of laparoscopy)
PCDD/Fs + WHO-PCBs (CALUXÒ) + PCBs (#118, 138, 153, 180) (GC-ECD)
OR : 4.6 [0.48–43.62] (adjusted for BMI and alcohol consumption)
2,3,7,8-TCDD (GC-HRMS)
OR: 1.52 [0.34–6.82] (cut-off 20 ppt (pg TCDD/liter blood)) OR: 1.78 [0.7–4.53] (cutoff 100 ppt (pg TCDD/liter blood)) (only cases and non diseased)
PCDD/PCDFs, WHO-PCBs, PCBs (GC-HRMS)
PCDD/Fs Cases: 26.2 (18.2–37.7) Ctls: 25.6 (24.3–28.9) pg TEQ/g fat; WHO-PCBs Cases: 7.97 (5.05–12.6) Ctls: 7.45 (6.69–8.30) pg TEQ/g fat; marker PCBs Cases: (294 (215– 401) Ctls: 372 (351–403) ng/g fat
PCDD/Fs, WHO-PCBs (GC-HRMS)
PCDD/Fs Ctls It 8.9 ± 1.3 Ctls Be 24.7 ± 3.7 Cases It 10.4 ± 1.1 Cases Be 18.2 ± 2.7 pg TEQ/g fat; non-ortho-PCBs Ctls It 3.92 ± 0.58 Ctls Be 9.4 ± 1.4 Cases It 3.9 ± 0.4 Cases Be 8.6 ± 1.3; pg TEQ/g fat; mono-ortho-PCBs Ctls It 4.83 ± 0.71 Ctls Be 10.4 ± 1.5 Cases It 4.0 ± 0.4 Cases Be 7.7 ± 1.1 pg TEQ/g fat;
Marker PCBs (GC-HRMS)
OR: DEN vs. Ctls 3.12 [1.09–8.90]; OR: PE vs. Ctls 1.10 [0.43–2.82]; OR: Both diseases vs. Ctls 1.95 [0.87–4.34] adjusted for an age of 30 y; for an increased of 100pg/g fat
PCDD/Fs and WHO-PCBs (GC-HRMS)
OR: PE vs. Ctls 1.89 [0.93–3.81]; OR: DEN vs. Ctls 3.31 [1.44–7.57]; OR: Both diseases 2.61 [1.30–5.27] for an increased of 10 pg TEQ/g lipids
Lebel et al., 1998 (Que´bec, Canada) case-control
Pauwels et al., 2001(Belgium) casecontrol Eskenazi et al., 2002 (Seveso, Italy) cohort
Fierens et al., 2003 (Belgium) cross-sectional
De Felip et al., 2004 (Italy) case-control
Heilier et al., 2004 (Belgium) case-control
Heilier et al., 2005 (Belgium) case-control
Ctls: idem (n = 277) Cases: Belgian women participating to a study assessing organochlorine body burden in general population (n = 10; 49.0 y. (40.3–57.7))
Idem Laparoscopy (endometriosis)
Idem Laparoscopy (endometriosis)
Idem Laparoscopy/laparotomy (endometriosis) or ultrasound (cyst or mass characteristic of endometriosis) Idem No diagnosis (patient knowledge)
Ctls: idem (n = 132; 51.2 y. (49.3–52.8)) Cases: Belgian (Be) and Italian (It) women of reproductive age, nulliparous suspected of endometriosis (n = 23 (4 pools analyzed))
Idem Laparoscopy (endometriosis)
Ctls: Belgian and Italian women of reproductive age suspected of benign adnexal mass (n = 17 (2 pools analyzed)) Cases: Women attending for endometriosis/deep endometriotic nodules surgical intervention (DEN n = 10; 29.0 y. (24–44); PE n = 7; 27.0 y. (22–45)) Ctls: Women attending routine examinations (n = 10; 33.5 y. (25–39) Cases: Women attending for endometriosis/deep endometriotic nodules surgical intervention (DEN n = 25; 30.8 ± 7.0 y.; PE n = 25; 31.8 ± 6.5 y.) Ctls: Women attending routine examinations (n = 21; 33.1 ± 6.0 y.)
Idem
Laparoscopy/laparotomy (endometriosis/deep endometriotic nodules) Pelvic examination, vaginal echography, CA-125 (<35 U/ml) Laparoscopy/laparotomy (endometriosis/deep endometriotic nodules)
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
Reference, country and study design
Pelvic examination, vaginal echography, CA-125 (<35 U/ml) 205
(continued on next page)
206
Table 1 (continued) Population (n, age [mean ± SD or median (range)])
Diagnosis method (inclusion criteria for cases)
Chemicals (analytical method)
OR [CI95%]/concentration of chemicals (p-value)
Louis et al., 2005 (USA) case-control
Cases: Women undergoing laparoscopy for tubal sterilization or diagnostic (regardless of preoperative diagnosis) (n = 32; 32.7 ± 4.44 y.)
Laparoscopy (endometriosis)
PCBsa (GC-ECD)
OR: tertile I (<0.85 ng/g serum) 1.00 reference; OR: tertile II (0.85–1.36 ng/ g serum) 2.26 [0.61– 8.34]; OR: tertile III (>1.36 ng/g serum) 1.44 [0.40–5.15]; adjusted for gravidity (0/1), current cigarette smoker (yes/no) and serum lipids
Tsukino et al., 2005 (Japan) case-control
Ctls: idem (n = 52; 31.6 ± 4.96 y.) Cases: Japanese infertile women (n = 58 (52 blood samples analyzed); 32.4 ± 3.4)
Idem Laparoscopy (endometriosis stage II–IV)
PCDD/Fs, WHO-PCBs, PCBs, Pesticides (GC-HRMS)
quartile I (<20.27pg TEQ/g lipids) OR: 1.00 reference (Sum of TEQ values); quartile II (20.27–25.07 pg TEQ/g lipids) OR: 0.97 [0.36– 2.63]; quartile III (25.07–31.84 pg TEQ/g lipids) OR: 0.38 [0.12–1.17]; quartile IV (>31.84 pg TEQ/g lipids) OR: 0.41 [0.14–1.27] adjusted for menstrual regularity (regular or irregular) and average cycle length (days)
Reddy et al., 2006 (India) case-control
Ctls: idem (n = 81 (70 blood samples analyzed); 32.9 ± 3.9) Cases: infertile women of South Indian origin (n = 85; 30.6 ± 5.4 y.)
Laparoscopy (endometriosis stage 0–I) Laparoscopy (endometriosis)
PCBs #1, 5, 29, 98 phthalate esters (GC)
PCB#1 Ctls 0.04 ± 0.13 Stage I 0.23 ± 0.26 Stage II 0.42 ± 0.29 Stage III 0.60 ± 0.27 Stage IV 0.84 ± 0.56 lg/ml; PCB#5 Ctls 0.01 ± 0.05 Stage I 0.10 ± 0.12 Stage II 0.24 ± 0.22 Stage III 0.62 ± 0.39 Stage IV 0.75 ± 0.43 lg/ml; PCB#29 Ctls 0.02 ± 0.09 Stage I 0.13 ± 0.15 Stage II 0.02 ± 0.31 Stage III 0.50 ± 0.34 Stage IV 0.99 ± 0.54 lg/ml; PCB#98 Ctls 0.00 ± 0.02 Stage I 0.03 ± 0.10 Stage II 0.11 ± 0.19 Stage III 0.37 ± 0.32 Stage IV 0.26 ± 0.31 lg/ml
Idem
Porpora et al., 2006 (Italy) case-control
Ctls: Women of south India origin undergoing laparoscopy for tubal sterilization with proven fertility (n = 88; 30.9 ± 6.1 y.) Cases: Italian nulliparous women (n = 40; 29 y. (22–40))
PCBs# 28; 52; 101; 105; 118; 138; 153; 156; 167; 170; 180 (GC–MS/ MS)
OR = 1 (<250 ng/g) Referent; OR 6.5 (1.5–28) 250–360 ng/g; OR 5.3 (1.3–23) >360 ng/g (adjusted by age and smoking habits)
Ctls: idem (n = 40; 34 y. (20–40))
Idem
Laparoscopy (endometriosis)
CTL: control; DEN: deep endometriotic nodules; PE: peritoneal endometriosis; GC–MS: gas chromatography–mass spectrometry; GC-ECD: gas chromatography-electron capture detector; CALUXÒ: chemically activated luciferase expression; GC-HRMS: gas chromatography–high resolution mass spectrometry. a PCBs# 6, 18, 19, 22, 25, 28, 31, 33, 40, 42, 44, 45, 47, 48, 50, 52, 55, 59, 60, 64, 70, 74, 82, 94, 97, 99, 101, 105, 114, 118, 126, 128, 129, 132, 134, 135, 136, 138, 141, 147, 151, 153, 167, 169, 170, 172, 174, 176, 177, 179, 180, 181, 183, 185, 187, 188, 189, 190, 194, 195, 205, 206.
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
Reference, country and study design
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
wards where they were referred to for infertility (Mayani et al., 1997; Lebel et al., 1998; Pauwels et al., 2001; Tsukino et al., 2005), gynaecological problems (e.g. presence of benign adnexal mass) (De Felip et al., 2004; Porpora et al., 2006), tubal sterilization (Louis et al., 2005), or routine examination (Heilier et al., 2004, 2005). In the study of Tsukino et al. (Tsukino et al., 2005), patients suffering from endometriosis at stage I as defined by American Society for Reproductive Medicine (1997) were used as controls. In these hospital-based studies, controls did therefore not constitute a random subset of the underlying population (Holt and Weiss, 2000) which could contribute to obscure the association between endometriosis and organochlorines (Zondervan et al., 2002). With population-based controls, for which a laparoscopic examination is not possible, the rate of asymptomatic endometriosis must be defined a priori and Zondervan et al. (2002) noted that a prevalence <2% might be expected, if controls are selected as free of moderate to severe pelvic syndrome. The probability of misclassification in the absence of laparoscopic examination of controls would, therefore, be limited. Heilier et al. (2004, 2005) addressed the problem by recruiting patients without any pelvic pain, presenting normal pelvic examination, vaginal echography and laboratory test (CA-125 < 30 U/ml) but who did not undergo laparoscopic examination. Porpora et al. (2006) recruited controls in patients suffering from gynaecological problems different from endometriosis. 3.2. Peritoneal endometriosis or deep endometriotic nodules of the recto-vaginal septum Beside peritoneal endometriosis, deep endometriotic (adenomyotic) nodules of the recto-vaginal septum have been described, as a distinct clinical entity with a specific histopathogenesis (Donnez et al., 1996). Heilier et al. (2004, 2005) observed a significantly increased risk for this specific form of the disease associated with PCBs and dioxin-like compounds concentration in serum. This association was not statistically significant for peritoneal endometriosis, suggesting that, in humans, organochlorines might preferably induce the deep endometriotic form of the disease. Most investigators have classified patients according to the revised American Society for Reproductive Medicine classification (ASRM, 1997) which does not take deep endometriotic nodules as a distinct entity into consideration. It might therefore be important to separate, in future studies, peritoneal endometriosis and deep endometriotic nodules as distinct entities to assess the possible etiological contribution of organochlorines. 3.3. Nature of chemicals All 17 PCDD/Fs and 12 dioxin-like PCBs able to bind AhR were measured in four studies (Fierens et al., 2003; De Felip et al., 2004; Heilier et al., 2005; Tsukino et al.,
207
2005) by gas chromatography–high resolution mass spectrometry and in another study (Pauwels et al., 2001) by CALUXÒ (Chemically Activated Luciferase Expression – a receptor binding assay which reports the TEQ concentration of the mixture). Two other studies (Mayani et al., 1997; Eskenazi et al., 2002) have considered 2,3,7,8-TCDD only. Selected dioxin-like PCBs, but not all 12, were also measured in others studies. Among the studies which took AhR ligands into consideration (Mayani et al., 1997; Pauwels et al., 2001; Eskenazi et al., 2002; Fierens et al., 2003; De Felip et al., 2004; Heilier et al., 2005; Tsukino et al., 2005), 2 reported non significantly different concentrations in serum, 4 non significant OR (3 with OR > 1 and 1 with OR < 1) and one study (Heilier et al., 2005) reported a significantly increased risk to develop deep endometriotic nodules associated with higher serum concentrations of dioxin-like compounds. The study of Eskenazi et al. (2002) is unique because these patients were accidentally exposed to TCDD. They were living in Seveso (Italy) around a plant that produced organochlorine herbicides (2,4,5-trichlorophenoxyacetic acid (2,4,5-T)), when in July 1976, an uncontrolled exothermic reaction lead to the release in the air of a chemical cloud that deposited its content (TCDD among others compounds) over several square kilometers of a populated countryside (Bertazzi et al., 1989). Serum concentrations observed in this population were elevated (>1000 ppt – (Mocarelli et al., 1991)), Eskenazi et al. (2002) observed only a trend to endometriosis with TCDD concentrations in serum but no significant association. ortho-PCBs, which are not able to activate AhR, were measured in 8 studies (Lebel et al., 1998; Pauwels et al., 2001; Fierens et al., 2003; Heilier et al., 2004; Louis et al., 2005; Tsukino et al., 2005; Porpora et al., 2006; Reddy et al., 2006). Congeners measured in 7 of them belong to the ‘‘marker PCB group’’ (PCBs# 28, 52, 101, 138, 153, 180, 170, 194, 206, 209). Reddy et al. (2006) measured others congeners (PCBs# 1, 5, 29, 98). In addition, in studies by Louis et al. (2005) and Tsukino et al. (2005) marker PCBs were measured among other ortho-PCBs (32 and 62 congeners, respectively) and individual measurements for marker PCBs were not available. Concentrations in serum are reported for the study of Fierens et al. (2003) who did not observe any difference in the sum of marker PCBs between women reporting endometriosis and controls. Odds ratio have been calculated (but not systematically reported) in five studies (Lebel et al., 1998; Heilier et al., 2005; Louis et al., 2005; Tsukino et al., 2005; Porpora et al., 2006) but only Porpora et al. (2006) reported significant OR for PCBs (Table 1). Similarly for dioxin-like compounds, Heilier et al. (2004) reported a significant OR, for an increase of 10 pg/g lipids, for deep endometriotic nodules but not for peritoneal endometriosis. Although structurally similar, dioxin-like compounds and marker PCBs do not exert a similar pattern of toxicological effects essentially because of differences in their mechanism of action. Dioxin-like compounds act
208
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
(probably exclusively) through the interaction with AhR (Hahn, 2003) whereas marker PCBs do not. Whether these compounds combine their effects in endometriosis remains elusive. In the examined studies, none has demonstrated an association between dioxin-like compounds and peritoneal endometriosis but only trends or non significant OR. Only one study (Porpora et al., 2006) reported a significant relationship with marker PCBs. The explanation of these observations might lie in the high correlation often reported among the concentrations of the different compounds in serum (Gladen et al., 2003). Human exposure to both types of compounds occurs essentially (>95%) through the consumption of fatty foodstuffs contaminated by mixtures of organochlorines. In addition, as both groups of organochlorines share similar toxicokinetics (long half-life time), the concentration of marker PCBs in serum has been proposed as a proxy for that of dioxin-like compound in serum (Longnecker et al., 2000). 4. Issues for future epidemiologic studies The studies that we have reviewed did not clearly support or refute a role for organochlorines in the etiopathogenesis of endometriosis. The absence of consistent conclusions results from epidemiologic choices that did not perfectly meet all requirements for analyzing a complex trait such as endometriosis. Some guidelines could be derived from the reviewed studies and from the recommendations of Holt and Weiss (2000) and Zondervan et al. (2002), in order to better delineate this relationship, future investigators should consider five issues.
nance imaging) and/or biochemical (CA-125 and/or new markers measured in peritoneal fluid such as cytokines (TNF-alpha, IL-6, IL-1) (Gupta et al., 2006). If for gynecologists, a laparoscopic examination remains the gold standard to ascertain the presence of the disease, it must be defined, as the third question, how to assess the absence of disease in controls, but this is closely related to their selection. 4.3. How to select controls? Holt and Weiss (2000) have extensively discussed control selection both in hospital and population-based casecontrol studies. They recommended that controls, to minimize bias, should arise from the same population as the cases, and identified several kinds of controls i.e. (i) women referred to for infertility who do not suffer from endometriosis; (ii) women attending a gynecological ward for routine examination and (iii) friends or family of cases, or women living in the same area. Women belonging to the two first groups, undergo clinical examination that ascertains the presence of (peritoneal) endometriosis, but they constitute a non random subset of the underlying population for the analysis of several risk factors. In the search of an association between endometriosis and organochlorines (as well as for any environmental factors), the recruitment of women living in the same area should be preferred because this option allows cases and controls to be similarly exposed to organochlorine (also by consumption of contaminated food). However this type of controls do not undergo laparoscopic examination. 4.4. How to assess organochlorine (dioxin) exposure?
4.1. Which epidemiological design? Etiological studies (i.e. case-control and cohort studies) are probably more appropriate to identify an association between endometriosis and organochlorines which could not be uncovered with a cross-sectional design. However, although cohort studies are more accurate (given an unbiased risk estimator and real prevalence of the disease), and although this design has already been used for the identification of some risk factors (e.g. with the Nurse Health Study II, (Missmer et al., 2004)), a case-control design appears more appropriate for endometriosis because of the relatively low prevalence of the disease. 4.2. How to define and assess the presence of endometriosis? Three issues should be address. Firstly the definition of the disease which includes the three forms of endometriosis (i.e. peritoneal endometriosis, ovarian endometriosis and deep endometriotic (adenomyotic) nodules of the rectovaginal septum) that is not taken into consideration with ASRM classification and is of primordial importance (Nisolle and Donnez, 1997). Secondly the diagnostic criteria: clinical (coelioscopy; ultrasonography; magnetic reso-
Exposure to organochlorines could be assessed either by the measurement of dioxin concentrations in blood (both by classical analytical methods (GC–MS) or by CALUX) or by the assessment of fatty foodstuffs consumption (e.g. by food frequency questionnaire) that are most likely to be contaminated by organochlorines. It is important to keep in mind that (i) organochlorines (especially dioxins) are persistent pollutants that accumulate both in the environment and in the human body and (ii) that in the frame of a case-control study, the assessment of exposure is always retrospective whereas it could be theoretically prospective in a cohort design. In addition, as endometriosis is often diagnosed in 25–35 years old women, the relevant period of exposure to organochlorines could stretch from in utero live to the day of diagnosis. Therefore assessment of present organochlorines body burden or fatty foodstuffs consumption habits could be insufficient to characterize organochlorine exposure. Recent advances in exposure models (e.g. probabilistic models of exposure (Vrijens et al., 2002) or elimination models (Aylward et al., 2005)) could allow to reconstruct exposure, but these models do not have been already used to explore the relationship between endometriosis and organochlorines.
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
4.5. Which confounding factor should be controlled? This point is the hardest to address for a multifactorial disease such as endometriosis. Confounding factors are variables which bias the estimate of the effect of exposure on disease. They must meet two criteria: (i) to be related to the disease (but not a consequence of the disease) and (ii) to be related to the risk factor (but not a consequence of the risk factor) (Woodward, 1999). Some authors have added an additional criterion, stressing that confounders must not be an intermediate step in the causal pathway of the disease (McNamee, 2003). In order to identify and adjust exposure for these variables, they must have been recorded and therefore suspected during the design of the study. All agents that could be involved in the onset or the growth of endometriosis and that accumulate in the environment, such as organochlorines, meet this definition and are therefore potential confounding variables. This highlights the benefit of tools that assess the activity of toxicants rather than their concentrations.
5. Discussion It is generally admitted that the detection rate of endometriosis increases. However, this has not been formally evidenced in the general population, essentially because a strict diagnosis of endometriosis involves a laparoscopic examination. In contrast, levels of dioxins are decreasing in the environment (European Commission DG Environment, 1999) and it could, therefore, be argued that if dioxins were involved in the etiology of endometriosis, they should rather be associated with a decreasing prevalence of the disease. Two arguments should, however, be taken into account when considering this hypothesis: (i) It cannot be excluded that dioxins, like diethylstilbestrol, could act during in utero life, implying a lag time in the temporal relationship. (ii) To observe a parallel decrease of dioxin levels and endometriosis, it would need that dioxins were a necessary and sufficient cause of the disease. Firstly, dioxins cannot be a necessary cause of endometriosis because both forms of the disease (peritoneal endometriosis and deep endometriotic nodules) have been described well before dioxins accumulated in the environment (Benagiano and Brosens, 1991). Secondly, they could also not be a sufficient cause because all women that have been heavily exposed have not developed endometriosis. Serum concentrations observed in women exposed in Seveso were higher than those in the serum of monkeys exposed to TCDD by Rier (Rier et al., 1993; Bois and Eskenazi, 1994). The absence of a significant association in this population is, however, compatible with the toxicological and epidemiological principle of dose–response relationship when considering dioxin-like compounds and/or PCBs only as component causes of endometriosis and/or deep endometriotic nodules (Rothman and Greenland, 1998; Klaassen, 2001). The role of other environmental components should
209
also be investigated carefully. In this specific context, concern about dietary phytoestrogens is rising because they seem to be associated with endometrial hyperplasia (Unfer et al., 2004) and demonstrate in vitro activity on aromatase (Edmunds et al., 2005) which plays a key role in the endometriotic process (Attar and Bulun, 2006). The relationship between exposure and the disease observed in some studies could also be interpreted as evidence of reverse causality (Rothman and Greenland, 1998). Reverse causality means that the outcome (endometriosis) may have caused the exposure i.e. an excessive organochlorine body burden. According to this hypothesis, breast-feeding, that is known to reduce he organochlorine body burden, could explain why controls women (that are not infertile) have lower organochlorine body burdens. This hypothesis, while applying mainly to populationbased studies (women recruited in infertility clinics are often nullipare), stresses the need to take breastfeeding into account in future studies. In conclusion, the majority of studies dealing with peritoneal endometriosis did not observe a significant association with organochlorines. In contrast, it appears that deep endometriotic nodules are associated with a higher serum level of dioxin-like compounds and marker PCBs, which is compatible with a nosological distinction between peritoneal endometriosis and deep endometriotic nodules. It remains, however, to clarify the mechanism(s) potentially underlying this association. References American Society for Reproductive Medicine (ASRM) 1997. Revised American society for reproductive medicine classification of endometriosis: 1996, Fertil. Steril. 67, pp. 817–821. Attar, E., Bulun, S.E., 2006. Aromatase and other steroidogenic genes in endometriosis: translational aspects. Hum. Reprod. Update 12, 49–56. Aylward, L.L., Brunet, R.C., Starr, T.B., Carrier, G., Delzell, E., Cheng, H., Beall, C., 2005. Exposure reconstruction for the TCDD-exposed NIOSH cohort using a concentration- and age-dependent model of elimination. Risk Anal. 25, 945–956. Benagiano, G., Brosens, I., 1991. The history of endometriosis: identifying the disease. Hum. Reprod. 6, 963–968. Bertazzi, P.A., Zocchetti, C., Pesatori, A.C., Guercilena, S., Sanarico, M., Radice, L., 1989. Ten-year mortality study of the population involved in the Seveso incident in 1976. Am. J. Epidemiol. 129, 1187–1200. Birnbaum, L.S., Cummings, A.M., 2002. Dioxins and endometriosis: a plausible hypothesis. Environ. Health Perspect. 110, 15–21. Bois, F.Y., Eskenazi, B., 1994. Possible risk of endometriosis for Seveso, Italy, residents: an assessment of exposure to dioxin. Environ. Health Perspect. 102, 476–477. De Felip, E., Porpora, M.G., di Domenico, A., Ingelido, A.M., Cardelli, M., Cosmi, E.V., Donnez, J., 2004. Dioxin-like compounds and endometriosis: a study on Italian and Belgian women of reproductive age. Toxicol. Lett. 150, 203–209. Donnez, J., Nisolle, M., Smoes, P., Gillet, N., Beguin, S., Casanas-Roux, F., 1996. Peritoneal endometriosis and ‘‘endometriotic’’ nodules of the rectovaginal septum are two different entities. Fertil. Steril. 66, 362– 368. Donnez, J., Van Langendonckt, A., Casanas-Roux, F., Van Gossum, J.P., Pirard, C., Jadoul, P., Squifflet, J., Smets, M., 2002. Current thinking on the pathogenesis of endometriosis. Gynecol. Obstet. Invest. 54 (Suppl. 1), 52–58.
210
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
Edmunds, K.M., Holloway, A.C., Crankshaw, D.J., Agarwal, S.K., Foster, W.G., 2005. The effects of dietary phytoestrogens on aromatase activity in human endometrial stromal cells. Reprod. Nutr. Dev. 45, 709–720. Eskenazi, B., Mocarelli, P., Warner, M., Samuels, S., Vercellini, P., Olive, D., Needham, L.L., Patterson Jr., D.G., Brambilla, P., Gavoni, N., Casalini, S., Panazza, S., Turner, W., Gerthoux, P.M., 2002. Serum dioxin concentrations and endometriosis: a cohort study in Seveso, Italy. Environ. Health Perspect. 110, 629–634. European Commission DG Environment. 1999. Compilation of EU Dioxin Exposure and Health Data – Summary Report available at. Fierens, S., Mairesse, H., Heilier, J.F., De Burbure, C., Focant, J.F., Eppe, G., De Pauw, E., Bernard, A., 2003. Dioxin/polychlorinated biphenyl body burden, diabetes and endometriosis: findings in a population-based study in Belgium. Biomarkers 8, 529–534. Gerhard, I., Runnebaum, B., 1992. The limits of hormone substitution in pollutant exposure and fertility disorders. Zentralbl. Gynakol. 114, 593–602. Gladen, B.C., Doucet, J., Hansen, L.G., 2003. Assessing human polychlorinated biphenyl contamination for epidemiologic studies: lessons from patterns of congener concentrations in Canadians in 1992. Environ. Health Perspect. 111, 437–443. Guo, S.W., 2004. The link between exposure to dioxin and endometriosis: a critical reappraisal of primate data. Gynecol. Obstet. Invest. 57, 157– 173. Gupta, S., Agarwal, A., Sekhon, L., Krajcir, N., Cocuzza, M., Falcone, T., 2006. Serum and peritoneal abnormalities in endometriosis: potential use as diagnostic markers. Minerva Ginecol. 58, 527–551. Hahn, M.E., 2003. Evolutionary and physiological perspectives on Ah receptor function and dioxin toxicity. In: Schecter, A., Gasiewicz, T.A. (Eds.), Dioxins and health. Wiley-Interscience, Hoboken. Heilier, J.F., Nackers, F., Verougstraete, V., Tonglet, R., Lison, D., Donnez, J., 2005. Increased dioxin-like compounds in the serum of women with peritoneal endometriosis and deep endometriotic (adenomyotic) nodules. Fertil. Steril. 84, 305–312. Heilier, J., Ha, A.T., Lison, D., Donnez, J., Tonglet, R., Nackers, F., 2004. Increased serum polychlorobiphenyl levels in Belgian women with adenomyotic nodules of the rectovaginal septum. Fertil. Steril. 81, 456–458. Holt, V.L., Weiss, N.S., 2000. Recommendations for the design of epidemiologic studies of endometriosis. Epidemiology 11, 654–659. Klaassen, C.D., 2001. Casarett and Douls Toxicology. The basic science of poison. Mc Graw and Hill, New York. Lebel, G., Dodin, S., Ayotte, P., Marcoux, S., Ferron, L.A., Dewailly, E., 1998. Organochlorine exposure and the risk of endometriosis. Fertil. Steril. 69, 221–228. Longnecker, M.P., Ryan, J.J., Gladen, B.C., Schecter, A.J., 2000. Correlations among human plasma levels of dioxin-like compounds and polychlorinated biphenyls (PCBs) and implications for epidemiologic studies. Arch. Environ. Health 55, 195–200. Louis, G.M., Weiner, J.M., Whitcomb, B.W., Sperrazza, R., Schisterman, E.F., Lobdell, D.T., Crickard, K., Greizerstein, H., Kostyniak, P.J., 2005. Environmental PCB exposure and risk of endometriosis. Hum. Reprod. 20, 279–285. Mayani, A., Barel, S., Soback, S., Almagor, M., 1997. Dioxin concentrations in women with endometriosis. Hum. Reprod. 12, 373–375. McNamee, R., 2003. Confounding and confounders. Occup. Environ. Med. 60, 227–234. Mimura, J., Fujii-Kuriyama, Y., 2003. Functional role of AhR in the expression of toxic effects by TCDD. Biochim. Biophys. Acta 1619, 263–268. Missmer, S.A., Hankinson, S.E., Spiegelman, D., Barbieri, R.L., Marshall, L.M., Hunter, D.J., 2004. Incidence of laparoscopically confirmed endometriosis by demographic, anthropometric, and lifestyle factors. Am. J. Epidemiol. 160, 784–796.
Mocarelli, P., Needham, L.L., Marocchi, A., Patterson Jr., D.G., Brambilla, P., Gerthoux, P.M., Meazza, L., Carreri, V., 1991. Serum concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin and test results from selected residents of Seveso, Italy. J. Toxicol. Environ. Health 32, 357–366. Nisolle, M., Donnez, J., 1997. Peritoneal endometriosis, ovarian endometriosis, and adenomyotic nodules of the rectovaginal septum are three different entities. Fertil. Steril. 68, 585–596. Pauwels, A., Schepens, P.J., D’Hooghe, T., Delbeke, L., Dhont, M., Brouwer, A., Weyler, J., 2001. The risk of endometriosis and exposure to dioxins and polychlorinated biphenyls: a case-control study of infertile women. Hum. Reprod. 16, 2050–2055. Porpora, M.G., Ingelido, A.M., di Domenico, A., Ferro, A., Crobu, M., Pallante, D., Cardelli, M., Cosmi, E.V., De Felip, E., 2006. Increased levels of polychlorobiphenyls in Italian women with endometriosis. Chemosphere 63, 1361–1367. Reddy, B.S., Rozati, R., Reddy, S., Kodampur, S., Reddy, P., Reddy, R., 2006. High plasma concentrations of polychlorinated biphenyls and phthalate esters in women with endometriosis: a prospective case control study. Fertil. Steril. 85, 775–779. Rier, S.E., Martin, D.C., Bowman, R.E., Dmowski, W.P., Becker, J.L., 1993. Endometriosis in rhesus monkeys (Macaca mulatta) following chronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Fund. Appl. Toxicol. 21, 433–441. Rier, S.E., Turner, W.E., Martin, D.C., Morris, R., Lucier, G.W., Clark, G.C., 2001. Serum levels of TCDD and dioxin-like chemicals in Rhesus monkeys chronically exposed to dioxin: correlation of increased serum PCB levels with endometriosis. Toxicol. Sci. 59, 147–159. Rothman, K.J., Greenland, S., 1998. Modern Epidemiology. Lippincott Williams & Wilkins, Philadelphia, PA. Treloar, S.A., Wicks, J., Nyholt, D.R., Montgomery, G.W., Bahlo, M., Smith, V., Dawson, G., Mackay, I.J., Weeks, D.E., Bennett, S.T., Carey, A., Ewen-White, K.R., Duffy, D.L., O 0 connor, D.T., Barlow, D.H., Martin, N.G., Kennedy, S.H., 2005. Genomewide linkage study in 1176 affected sister pair families identifies a significant susceptibility locus for endometriosis on chromosome 10q26. Am. J. Hum. Genet. 77, 365–376. Tsukino, H., Hanaoka, T., Sasaki, H., Motoyama, H., Hiroshima, M., Tanaka, T., Kabuto, M., Niskar, A.S., Rubin, C., Patterson Jr., D.G., Turner, W., Needham, L., Tsugane, S., 2005. Associations between serum levels of selected organochlorine compounds and endometriosis in infertile Japanese women. Environ. Res. 99, 118–125. Unfer, V., Casini, M.L., Costabile, L., Mignosa, M., Gerli, S., Di Renzo, G.C., 2004. Endometrial effects of long-term treatment with phytoestrogens: a randomized, double-blind, placebo-controlled study. Fertil. Steril. 82, 145–148. van den Berg, M., Birnbaum, L., Bosveld, A.T., Brunstrom, B., Cook, P., Feeley, M., Giesy, J.P., Hanberg, A., Hasegawa, R., Kennedy, S.W., Kubiak, T., Larsen, J.C., van Leeuwen, F.X., Liem, A.K., Nolt, C., Peterson, R.E., Poellinger, L., Safe, S., Schrenk, D., Tillitt, D., Tysklind, M., Younes, M., Waern, F., Zacharewski, T., 1998. Toxic equivalency factors (TEFs) for PCBs, PCDDs, PCDFs for humans and wildlife. Environ. Health Perspect. 106, 775–792. Vrijens, B., De Henauw, S., Dewettinck, K., Talloen, W., Goeyens, L., De Backer, G., Willems, J.L., 2002. Probabilistic intake assessment and body burden estimation of dioxin-like substances in background conditions and during a short food contamination episode. Food Addit. Contam. 19, 687–700. Woodward, M., 1999. Epidemiology study design and data analysis. Chapman & Hall/CRC, Boca Raton. Zober, A., Messerer, P., Huber, P., 1990. Thirty-four-year mortality follow-up of BASF employees exposed to 2,3,7,8-TCDD after the 1953 accident. Int. Arch. Occup. Environ. Health 62, 139–157. Zondervan, K.T., Cardon, L.R., Kennedy, S.H., 2002. What makes a good case-control study? Design issues for complex traits such as endometriosis. Hum. Reprod. 17, 1415–1423.
Chemosphere 71 (2008) 203–210 www.elsevier.com/locate/chemosphere
Review
Organochlorines and endometriosis: A mini-review Jean-Franc¸ois Heilier a
a,*
, Jacques Donnez
b,c
, Dominique Lison
a
Universite´ catholique de Louvain, Faculty of Medicine, Industrial Toxicology and Occupational Medicine Unit, Brussels, Belgium b Universite´ catholique de Louvain, Cliniques Universitaires St-Luc, Department of Gynecology, Brussels, Belgium c Universite´ catholique de Louvain, Faculty of Medicine, Gynecology unit, Brussels, Belgium Received 2 April 2007; received in revised form 26 September 2007; accepted 27 September 2007 Available online 19 November 2007
Abstract Organochlorines (polychlorinated biphenyls and dioxin-like compounds) are suspected to play a role in the etiopathogenesis of endometriosis. This hypothesis, based on experimental data, has been circulating for years in the scientific community and several epidemiologic surveys have attempted to obtain confirmatory human data. The purpose of this mini-review is to provide an overview of the twelve epidemiological studies that have assessed the relationship between endometriosis and organochlorine exposure. Several studies did not observe a significant association between peritoneal endometriosis and organochlorines. The deep nodular form of endometriosis appears associated with a higher serum level of both dioxin-like compounds and polychlorobiphenyls. The type of control women, the nature of the chemicals measured, and the definition of the disease could modulate the ability to detect the possible relationship between endometriosis and organochlorine exposure. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Endometriosis; Deep endometriotic nodules; Dioxins; Tetrachlorodibenzodioxin; Polychlorinated biphenyls; Epidemiology
Contents 1. 2. 3.
4.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dioxins-like compounds and PCBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dioxins, PCBs and endometriosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Type of controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Peritoneal endometriosis or deep endometriotic nodules of the recto-vaginal septum. . . . . . . . . . . . . . . . . . . . 3.3. Nature of chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Issues for future epidemiologic studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Which epidemiological design? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. How to define and assess the presence of endometriosis? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. How to select controls? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. How to assess organochlorine (dioxin) exposure? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Which confounding factor should be controlled?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
204 204 204 204 207 207 208 208 208 208 208 209
Abbreviations: AhR, aryl hydrocarbon receptor; CALUX, chemically activated luciferase expression; OR, odds ratio; PCBs, polychlorinated biphenyls; PCDDs, polychlorinated dibenzo-p-dioxins; PCDFs, polychlorinated dibenzofurans; TCDD (2,3,7,8-TCDD), 2,3,7,8-tetrachlorodibenzo-p-dioxin; TEF, toxic equivalency factor; TEQ, toxic equivalent. * Corresponding author. Tel.: +32 2 764 53 30; fax: +32 2 764 53 38. E-mail address: [email protected] (J.-F. Heilier). 0045-6535/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2007.09.044
204
5.
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
1. Introduction
3. Dioxins, PCBs and endometriosis
Endometriosis is a gynaecological disease characterised by the extra-uterine growth of endometrial tissue which causes internal bleeding, inflammation and scarring, and often leads to infertility (Donnez et al., 2002). Although the exact aetiology of the disease remains unknown, a role of both environmental (Birnbaum and Cummings, 2002; Gerhard and Runnebaum, 1992) and genetic factors (Treloar et al., 2005) has been suspected. Recently, some organochlorines (polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs) and polychlorinated biphenyls (PCBs)) previously suspected as risk factors (Rier et al., 2001), were found to be significantly associated with the disease (Heilier et al., 2004, 2005; Porpora et al., 2006).
The role of PCBs and/or dioxin-like compounds in the aetiopathogenesis of endometriosis is controversial. This hypothesis has been initially based on experimental data reported by Rier et al. (1993) demonstrating, that rhesus monkeys chronically exposed to 2,3,7,8-TCDD (5 and 25 ppt per day during four years) exhibited, 10 years after termination of exposure, peritoneal endometriosis which incidence and severity directly correlated with exposure and dose. For a long time, this experiment appeared as the best evidence of a role for organochlorines in the onset or growth of endometriosis. Rier et al. (2001) reconsidered, some years later, blood samples from this experiment and detected significant amounts of dioxin-like PCBs that originated from a contamination of TCDD, indicating a possible role of dioxin-like compounds rather than TCDD in the endometriotic effect. Criticism on the statistical analysis of Rier’s data has also been raised, after a reappraisal by Guo (2004) who concluded to the absence of relationship with endometriosis. The hypothesis of a role of PCBs and/or dioxins in the onset or the growth of endometriosis has also been addressed in several epidemiologic studies. The purpose of this mini-review is to summarize the existing evidence of a relationship between endometriosis and PCDD/Fs and/or PCBs in humans, with a particular attention to factors that could modify this relation, if any. Factors that are particularly of concern are (i) the selection of control women, (ii) disease definition and (iii) the nature of the chemicals investigated. The main results of 12 publications including 10 casecontrol, 1 cross-sectional and 1 cohort studies are summarized in Table 1. All reported on environmental exposures probably reflecting the fact that occupational exposures have essentially involved men (Zober et al., 1990). Eight studies reported odds ratio (OR), adjusted or not, for linear increment or cut-off in organochlorine blood content. In all studies but one (Fierens et al., 2003) patients and controls underwent gynaecological examination and were classified according to the presence of endometriosis or not. Fierens et al. (2003) classified participants according to selfreported information.
2. Dioxins-like compounds and PCBs Dioxins are made up of 210 chlorinated hydrocarbons divided into 75 PCDDs and 135 PCDFs. Dioxins are mainly by-products of industrial processes (e.g. waste incineration or iron and steel industries), formed when a thermal process (300–800 °C) occurs in the presence of chlorinated organic substances. Seventeen dioxin congeners which preferentially bio-accumulate in the food chain, bind to a specific receptor (aryl hydrocarbon receptor or AhR) that mediates most of their toxic effects and activates several genes, including cytochromes P450 (Hahn, 2003; Mimura and Fujii-Kuriyama, 2003). Several PCBs were manufactured until the 1970s because of their physicochemical properties (stability, viscosity, . . .), and used mostly in closed applications such as heat transfer in capacitors and transformers. Among the 209 PCB congeners (for a full list see http://www.epa.gov/toxteam/pcbid/ table.htm), 12 share with dioxins an activity on the AhR, and are refereed to as WHO-PCBs, dioxin-like PCBs, coplanar PCBs or non (mono)-ortho-PCBs because of the absence of chlorine substitution in ortho positions (1, 4, 6, 9) that gives the molecule a planar configuration. All compounds able to bind the AhR (7 PCDD, 10 PCDF and 12 dioxin-like PCBs) have been classified relatively to 2,3,7,8-TCDD (the most toxic dioxin congener) and received a toxic equivalency factor (TEF) which reflects their respective biological potency. Concentrations of these compounds, often present as mixtures, are frequently expressed as pg TEQ/g lipids, which is the sum of the products of the concentration of each compound multiplied by its TEF. The biological concentration of organochlorines is expressed per g lipids because these compounds are essentially stored in the fatty compartments of the body (van den Berg et al., 1998).
3.1. Type of controls Most studies have included only women undergoing laparoscopic examination (for infertility, tubal ligation, pelvic pain, . . .) who were divided into cases or controls according to the presence of endometriosis or not. A large majority of studies are hospital-based and controls have been recruited in clinics for reproductive medicine or gynaecological
Table 1 Epidemiological studies investigating the possible role of organochlorines in endometriosis Population (n, age [mean ± SD or median (range)])
Diagnosis method (inclusion criteria for cases)
Chemicals (analytical method)
OR [CI95%]/concentration of chemicals (pvalue)
Mayani et al, 1997 (Israel) case-control
Cases: Infertile women (clinic for reproductive medicine) (n = 44, 33.6 ± 6.7 y.) Ctls: idem (n = 35, 34.4 ± 5.8 y.) Cases: women with pelvic pain, infertility or undergoing tubal fulguration (n = 86; 32.2 ± 5.9 y.) Ctls: idem (n = 70, 33.2 ± 7.0 y.) Cases: Infertile women (clinic for reproductive medicine) (n = 42, 31.0 (25–42)) Ctls: idem (n = 27, 32.0 (24–41)) Cases: Women resident of Seveso <30 years in 1976 (n = 19)
Laparoscopy (endometriosis)
2,3,7,8-TCDD (GC–MS)
OR: 7.6 [0.87–169.7] (cut-off 0.4 ng TCDD/ liter blood)
PCBs (# 28; 52; 99; 101; 105; 118; 153; 156; 170; 180; 183; 187), pesticides (GC-ECD)
OR (sum of PCBs): non significantly different (adjusted for age, BMI, number of children and indication of laparoscopy)
PCDD/Fs + WHO-PCBs (CALUXÒ) + PCBs (#118, 138, 153, 180) (GC-ECD)
OR : 4.6 [0.48–43.62] (adjusted for BMI and alcohol consumption)
2,3,7,8-TCDD (GC-HRMS)
OR: 1.52 [0.34–6.82] (cut-off 20 ppt (pg TCDD/liter blood)) OR: 1.78 [0.7–4.53] (cutoff 100 ppt (pg TCDD/liter blood)) (only cases and non diseased)
PCDD/PCDFs, WHO-PCBs, PCBs (GC-HRMS)
PCDD/Fs Cases: 26.2 (18.2–37.7) Ctls: 25.6 (24.3–28.9) pg TEQ/g fat; WHO-PCBs Cases: 7.97 (5.05–12.6) Ctls: 7.45 (6.69–8.30) pg TEQ/g fat; marker PCBs Cases: (294 (215– 401) Ctls: 372 (351–403) ng/g fat
PCDD/Fs, WHO-PCBs (GC-HRMS)
PCDD/Fs Ctls It 8.9 ± 1.3 Ctls Be 24.7 ± 3.7 Cases It 10.4 ± 1.1 Cases Be 18.2 ± 2.7 pg TEQ/g fat; non-ortho-PCBs Ctls It 3.92 ± 0.58 Ctls Be 9.4 ± 1.4 Cases It 3.9 ± 0.4 Cases Be 8.6 ± 1.3; pg TEQ/g fat; mono-ortho-PCBs Ctls It 4.83 ± 0.71 Ctls Be 10.4 ± 1.5 Cases It 4.0 ± 0.4 Cases Be 7.7 ± 1.1 pg TEQ/g fat;
Marker PCBs (GC-HRMS)
OR: DEN vs. Ctls 3.12 [1.09–8.90]; OR: PE vs. Ctls 1.10 [0.43–2.82]; OR: Both diseases vs. Ctls 1.95 [0.87–4.34] adjusted for an age of 30 y; for an increased of 100pg/g fat
PCDD/Fs and WHO-PCBs (GC-HRMS)
OR: PE vs. Ctls 1.89 [0.93–3.81]; OR: DEN vs. Ctls 3.31 [1.44–7.57]; OR: Both diseases 2.61 [1.30–5.27] for an increased of 10 pg TEQ/g lipids
Lebel et al., 1998 (Que´bec, Canada) case-control
Pauwels et al., 2001(Belgium) casecontrol Eskenazi et al., 2002 (Seveso, Italy) cohort
Fierens et al., 2003 (Belgium) cross-sectional
De Felip et al., 2004 (Italy) case-control
Heilier et al., 2004 (Belgium) case-control
Heilier et al., 2005 (Belgium) case-control
Ctls: idem (n = 277) Cases: Belgian women participating to a study assessing organochlorine body burden in general population (n = 10; 49.0 y. (40.3–57.7))
Idem Laparoscopy (endometriosis)
Idem Laparoscopy (endometriosis)
Idem Laparoscopy/laparotomy (endometriosis) or ultrasound (cyst or mass characteristic of endometriosis) Idem No diagnosis (patient knowledge)
Ctls: idem (n = 132; 51.2 y. (49.3–52.8)) Cases: Belgian (Be) and Italian (It) women of reproductive age, nulliparous suspected of endometriosis (n = 23 (4 pools analyzed))
Idem Laparoscopy (endometriosis)
Ctls: Belgian and Italian women of reproductive age suspected of benign adnexal mass (n = 17 (2 pools analyzed)) Cases: Women attending for endometriosis/deep endometriotic nodules surgical intervention (DEN n = 10; 29.0 y. (24–44); PE n = 7; 27.0 y. (22–45)) Ctls: Women attending routine examinations (n = 10; 33.5 y. (25–39) Cases: Women attending for endometriosis/deep endometriotic nodules surgical intervention (DEN n = 25; 30.8 ± 7.0 y.; PE n = 25; 31.8 ± 6.5 y.) Ctls: Women attending routine examinations (n = 21; 33.1 ± 6.0 y.)
Idem
Laparoscopy/laparotomy (endometriosis/deep endometriotic nodules) Pelvic examination, vaginal echography, CA-125 (<35 U/ml) Laparoscopy/laparotomy (endometriosis/deep endometriotic nodules)
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
Reference, country and study design
Pelvic examination, vaginal echography, CA-125 (<35 U/ml) 205
(continued on next page)
206
Table 1 (continued) Population (n, age [mean ± SD or median (range)])
Diagnosis method (inclusion criteria for cases)
Chemicals (analytical method)
OR [CI95%]/concentration of chemicals (p-value)
Louis et al., 2005 (USA) case-control
Cases: Women undergoing laparoscopy for tubal sterilization or diagnostic (regardless of preoperative diagnosis) (n = 32; 32.7 ± 4.44 y.)
Laparoscopy (endometriosis)
PCBsa (GC-ECD)
OR: tertile I (<0.85 ng/g serum) 1.00 reference; OR: tertile II (0.85–1.36 ng/ g serum) 2.26 [0.61– 8.34]; OR: tertile III (>1.36 ng/g serum) 1.44 [0.40–5.15]; adjusted for gravidity (0/1), current cigarette smoker (yes/no) and serum lipids
Tsukino et al., 2005 (Japan) case-control
Ctls: idem (n = 52; 31.6 ± 4.96 y.) Cases: Japanese infertile women (n = 58 (52 blood samples analyzed); 32.4 ± 3.4)
Idem Laparoscopy (endometriosis stage II–IV)
PCDD/Fs, WHO-PCBs, PCBs, Pesticides (GC-HRMS)
quartile I (<20.27pg TEQ/g lipids) OR: 1.00 reference (Sum of TEQ values); quartile II (20.27–25.07 pg TEQ/g lipids) OR: 0.97 [0.36– 2.63]; quartile III (25.07–31.84 pg TEQ/g lipids) OR: 0.38 [0.12–1.17]; quartile IV (>31.84 pg TEQ/g lipids) OR: 0.41 [0.14–1.27] adjusted for menstrual regularity (regular or irregular) and average cycle length (days)
Reddy et al., 2006 (India) case-control
Ctls: idem (n = 81 (70 blood samples analyzed); 32.9 ± 3.9) Cases: infertile women of South Indian origin (n = 85; 30.6 ± 5.4 y.)
Laparoscopy (endometriosis stage 0–I) Laparoscopy (endometriosis)
PCBs #1, 5, 29, 98 phthalate esters (GC)
PCB#1 Ctls 0.04 ± 0.13 Stage I 0.23 ± 0.26 Stage II 0.42 ± 0.29 Stage III 0.60 ± 0.27 Stage IV 0.84 ± 0.56 lg/ml; PCB#5 Ctls 0.01 ± 0.05 Stage I 0.10 ± 0.12 Stage II 0.24 ± 0.22 Stage III 0.62 ± 0.39 Stage IV 0.75 ± 0.43 lg/ml; PCB#29 Ctls 0.02 ± 0.09 Stage I 0.13 ± 0.15 Stage II 0.02 ± 0.31 Stage III 0.50 ± 0.34 Stage IV 0.99 ± 0.54 lg/ml; PCB#98 Ctls 0.00 ± 0.02 Stage I 0.03 ± 0.10 Stage II 0.11 ± 0.19 Stage III 0.37 ± 0.32 Stage IV 0.26 ± 0.31 lg/ml
Idem
Porpora et al., 2006 (Italy) case-control
Ctls: Women of south India origin undergoing laparoscopy for tubal sterilization with proven fertility (n = 88; 30.9 ± 6.1 y.) Cases: Italian nulliparous women (n = 40; 29 y. (22–40))
PCBs# 28; 52; 101; 105; 118; 138; 153; 156; 167; 170; 180 (GC–MS/ MS)
OR = 1 (<250 ng/g) Referent; OR 6.5 (1.5–28) 250–360 ng/g; OR 5.3 (1.3–23) >360 ng/g (adjusted by age and smoking habits)
Ctls: idem (n = 40; 34 y. (20–40))
Idem
Laparoscopy (endometriosis)
CTL: control; DEN: deep endometriotic nodules; PE: peritoneal endometriosis; GC–MS: gas chromatography–mass spectrometry; GC-ECD: gas chromatography-electron capture detector; CALUXÒ: chemically activated luciferase expression; GC-HRMS: gas chromatography–high resolution mass spectrometry. a PCBs# 6, 18, 19, 22, 25, 28, 31, 33, 40, 42, 44, 45, 47, 48, 50, 52, 55, 59, 60, 64, 70, 74, 82, 94, 97, 99, 101, 105, 114, 118, 126, 128, 129, 132, 134, 135, 136, 138, 141, 147, 151, 153, 167, 169, 170, 172, 174, 176, 177, 179, 180, 181, 183, 185, 187, 188, 189, 190, 194, 195, 205, 206.
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
Reference, country and study design
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
wards where they were referred to for infertility (Mayani et al., 1997; Lebel et al., 1998; Pauwels et al., 2001; Tsukino et al., 2005), gynaecological problems (e.g. presence of benign adnexal mass) (De Felip et al., 2004; Porpora et al., 2006), tubal sterilization (Louis et al., 2005), or routine examination (Heilier et al., 2004, 2005). In the study of Tsukino et al. (Tsukino et al., 2005), patients suffering from endometriosis at stage I as defined by American Society for Reproductive Medicine (1997) were used as controls. In these hospital-based studies, controls did therefore not constitute a random subset of the underlying population (Holt and Weiss, 2000) which could contribute to obscure the association between endometriosis and organochlorines (Zondervan et al., 2002). With population-based controls, for which a laparoscopic examination is not possible, the rate of asymptomatic endometriosis must be defined a priori and Zondervan et al. (2002) noted that a prevalence <2% might be expected, if controls are selected as free of moderate to severe pelvic syndrome. The probability of misclassification in the absence of laparoscopic examination of controls would, therefore, be limited. Heilier et al. (2004, 2005) addressed the problem by recruiting patients without any pelvic pain, presenting normal pelvic examination, vaginal echography and laboratory test (CA-125 < 30 U/ml) but who did not undergo laparoscopic examination. Porpora et al. (2006) recruited controls in patients suffering from gynaecological problems different from endometriosis. 3.2. Peritoneal endometriosis or deep endometriotic nodules of the recto-vaginal septum Beside peritoneal endometriosis, deep endometriotic (adenomyotic) nodules of the recto-vaginal septum have been described, as a distinct clinical entity with a specific histopathogenesis (Donnez et al., 1996). Heilier et al. (2004, 2005) observed a significantly increased risk for this specific form of the disease associated with PCBs and dioxin-like compounds concentration in serum. This association was not statistically significant for peritoneal endometriosis, suggesting that, in humans, organochlorines might preferably induce the deep endometriotic form of the disease. Most investigators have classified patients according to the revised American Society for Reproductive Medicine classification (ASRM, 1997) which does not take deep endometriotic nodules as a distinct entity into consideration. It might therefore be important to separate, in future studies, peritoneal endometriosis and deep endometriotic nodules as distinct entities to assess the possible etiological contribution of organochlorines. 3.3. Nature of chemicals All 17 PCDD/Fs and 12 dioxin-like PCBs able to bind AhR were measured in four studies (Fierens et al., 2003; De Felip et al., 2004; Heilier et al., 2005; Tsukino et al.,
207
2005) by gas chromatography–high resolution mass spectrometry and in another study (Pauwels et al., 2001) by CALUXÒ (Chemically Activated Luciferase Expression – a receptor binding assay which reports the TEQ concentration of the mixture). Two other studies (Mayani et al., 1997; Eskenazi et al., 2002) have considered 2,3,7,8-TCDD only. Selected dioxin-like PCBs, but not all 12, were also measured in others studies. Among the studies which took AhR ligands into consideration (Mayani et al., 1997; Pauwels et al., 2001; Eskenazi et al., 2002; Fierens et al., 2003; De Felip et al., 2004; Heilier et al., 2005; Tsukino et al., 2005), 2 reported non significantly different concentrations in serum, 4 non significant OR (3 with OR > 1 and 1 with OR < 1) and one study (Heilier et al., 2005) reported a significantly increased risk to develop deep endometriotic nodules associated with higher serum concentrations of dioxin-like compounds. The study of Eskenazi et al. (2002) is unique because these patients were accidentally exposed to TCDD. They were living in Seveso (Italy) around a plant that produced organochlorine herbicides (2,4,5-trichlorophenoxyacetic acid (2,4,5-T)), when in July 1976, an uncontrolled exothermic reaction lead to the release in the air of a chemical cloud that deposited its content (TCDD among others compounds) over several square kilometers of a populated countryside (Bertazzi et al., 1989). Serum concentrations observed in this population were elevated (>1000 ppt – (Mocarelli et al., 1991)), Eskenazi et al. (2002) observed only a trend to endometriosis with TCDD concentrations in serum but no significant association. ortho-PCBs, which are not able to activate AhR, were measured in 8 studies (Lebel et al., 1998; Pauwels et al., 2001; Fierens et al., 2003; Heilier et al., 2004; Louis et al., 2005; Tsukino et al., 2005; Porpora et al., 2006; Reddy et al., 2006). Congeners measured in 7 of them belong to the ‘‘marker PCB group’’ (PCBs# 28, 52, 101, 138, 153, 180, 170, 194, 206, 209). Reddy et al. (2006) measured others congeners (PCBs# 1, 5, 29, 98). In addition, in studies by Louis et al. (2005) and Tsukino et al. (2005) marker PCBs were measured among other ortho-PCBs (32 and 62 congeners, respectively) and individual measurements for marker PCBs were not available. Concentrations in serum are reported for the study of Fierens et al. (2003) who did not observe any difference in the sum of marker PCBs between women reporting endometriosis and controls. Odds ratio have been calculated (but not systematically reported) in five studies (Lebel et al., 1998; Heilier et al., 2005; Louis et al., 2005; Tsukino et al., 2005; Porpora et al., 2006) but only Porpora et al. (2006) reported significant OR for PCBs (Table 1). Similarly for dioxin-like compounds, Heilier et al. (2004) reported a significant OR, for an increase of 10 pg/g lipids, for deep endometriotic nodules but not for peritoneal endometriosis. Although structurally similar, dioxin-like compounds and marker PCBs do not exert a similar pattern of toxicological effects essentially because of differences in their mechanism of action. Dioxin-like compounds act
208
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
(probably exclusively) through the interaction with AhR (Hahn, 2003) whereas marker PCBs do not. Whether these compounds combine their effects in endometriosis remains elusive. In the examined studies, none has demonstrated an association between dioxin-like compounds and peritoneal endometriosis but only trends or non significant OR. Only one study (Porpora et al., 2006) reported a significant relationship with marker PCBs. The explanation of these observations might lie in the high correlation often reported among the concentrations of the different compounds in serum (Gladen et al., 2003). Human exposure to both types of compounds occurs essentially (>95%) through the consumption of fatty foodstuffs contaminated by mixtures of organochlorines. In addition, as both groups of organochlorines share similar toxicokinetics (long half-life time), the concentration of marker PCBs in serum has been proposed as a proxy for that of dioxin-like compound in serum (Longnecker et al., 2000). 4. Issues for future epidemiologic studies The studies that we have reviewed did not clearly support or refute a role for organochlorines in the etiopathogenesis of endometriosis. The absence of consistent conclusions results from epidemiologic choices that did not perfectly meet all requirements for analyzing a complex trait such as endometriosis. Some guidelines could be derived from the reviewed studies and from the recommendations of Holt and Weiss (2000) and Zondervan et al. (2002), in order to better delineate this relationship, future investigators should consider five issues.
nance imaging) and/or biochemical (CA-125 and/or new markers measured in peritoneal fluid such as cytokines (TNF-alpha, IL-6, IL-1) (Gupta et al., 2006). If for gynecologists, a laparoscopic examination remains the gold standard to ascertain the presence of the disease, it must be defined, as the third question, how to assess the absence of disease in controls, but this is closely related to their selection. 4.3. How to select controls? Holt and Weiss (2000) have extensively discussed control selection both in hospital and population-based casecontrol studies. They recommended that controls, to minimize bias, should arise from the same population as the cases, and identified several kinds of controls i.e. (i) women referred to for infertility who do not suffer from endometriosis; (ii) women attending a gynecological ward for routine examination and (iii) friends or family of cases, or women living in the same area. Women belonging to the two first groups, undergo clinical examination that ascertains the presence of (peritoneal) endometriosis, but they constitute a non random subset of the underlying population for the analysis of several risk factors. In the search of an association between endometriosis and organochlorines (as well as for any environmental factors), the recruitment of women living in the same area should be preferred because this option allows cases and controls to be similarly exposed to organochlorine (also by consumption of contaminated food). However this type of controls do not undergo laparoscopic examination. 4.4. How to assess organochlorine (dioxin) exposure?
4.1. Which epidemiological design? Etiological studies (i.e. case-control and cohort studies) are probably more appropriate to identify an association between endometriosis and organochlorines which could not be uncovered with a cross-sectional design. However, although cohort studies are more accurate (given an unbiased risk estimator and real prevalence of the disease), and although this design has already been used for the identification of some risk factors (e.g. with the Nurse Health Study II, (Missmer et al., 2004)), a case-control design appears more appropriate for endometriosis because of the relatively low prevalence of the disease. 4.2. How to define and assess the presence of endometriosis? Three issues should be address. Firstly the definition of the disease which includes the three forms of endometriosis (i.e. peritoneal endometriosis, ovarian endometriosis and deep endometriotic (adenomyotic) nodules of the rectovaginal septum) that is not taken into consideration with ASRM classification and is of primordial importance (Nisolle and Donnez, 1997). Secondly the diagnostic criteria: clinical (coelioscopy; ultrasonography; magnetic reso-
Exposure to organochlorines could be assessed either by the measurement of dioxin concentrations in blood (both by classical analytical methods (GC–MS) or by CALUX) or by the assessment of fatty foodstuffs consumption (e.g. by food frequency questionnaire) that are most likely to be contaminated by organochlorines. It is important to keep in mind that (i) organochlorines (especially dioxins) are persistent pollutants that accumulate both in the environment and in the human body and (ii) that in the frame of a case-control study, the assessment of exposure is always retrospective whereas it could be theoretically prospective in a cohort design. In addition, as endometriosis is often diagnosed in 25–35 years old women, the relevant period of exposure to organochlorines could stretch from in utero live to the day of diagnosis. Therefore assessment of present organochlorines body burden or fatty foodstuffs consumption habits could be insufficient to characterize organochlorine exposure. Recent advances in exposure models (e.g. probabilistic models of exposure (Vrijens et al., 2002) or elimination models (Aylward et al., 2005)) could allow to reconstruct exposure, but these models do not have been already used to explore the relationship between endometriosis and organochlorines.
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
4.5. Which confounding factor should be controlled? This point is the hardest to address for a multifactorial disease such as endometriosis. Confounding factors are variables which bias the estimate of the effect of exposure on disease. They must meet two criteria: (i) to be related to the disease (but not a consequence of the disease) and (ii) to be related to the risk factor (but not a consequence of the risk factor) (Woodward, 1999). Some authors have added an additional criterion, stressing that confounders must not be an intermediate step in the causal pathway of the disease (McNamee, 2003). In order to identify and adjust exposure for these variables, they must have been recorded and therefore suspected during the design of the study. All agents that could be involved in the onset or the growth of endometriosis and that accumulate in the environment, such as organochlorines, meet this definition and are therefore potential confounding variables. This highlights the benefit of tools that assess the activity of toxicants rather than their concentrations.
5. Discussion It is generally admitted that the detection rate of endometriosis increases. However, this has not been formally evidenced in the general population, essentially because a strict diagnosis of endometriosis involves a laparoscopic examination. In contrast, levels of dioxins are decreasing in the environment (European Commission DG Environment, 1999) and it could, therefore, be argued that if dioxins were involved in the etiology of endometriosis, they should rather be associated with a decreasing prevalence of the disease. Two arguments should, however, be taken into account when considering this hypothesis: (i) It cannot be excluded that dioxins, like diethylstilbestrol, could act during in utero life, implying a lag time in the temporal relationship. (ii) To observe a parallel decrease of dioxin levels and endometriosis, it would need that dioxins were a necessary and sufficient cause of the disease. Firstly, dioxins cannot be a necessary cause of endometriosis because both forms of the disease (peritoneal endometriosis and deep endometriotic nodules) have been described well before dioxins accumulated in the environment (Benagiano and Brosens, 1991). Secondly, they could also not be a sufficient cause because all women that have been heavily exposed have not developed endometriosis. Serum concentrations observed in women exposed in Seveso were higher than those in the serum of monkeys exposed to TCDD by Rier (Rier et al., 1993; Bois and Eskenazi, 1994). The absence of a significant association in this population is, however, compatible with the toxicological and epidemiological principle of dose–response relationship when considering dioxin-like compounds and/or PCBs only as component causes of endometriosis and/or deep endometriotic nodules (Rothman and Greenland, 1998; Klaassen, 2001). The role of other environmental components should
209
also be investigated carefully. In this specific context, concern about dietary phytoestrogens is rising because they seem to be associated with endometrial hyperplasia (Unfer et al., 2004) and demonstrate in vitro activity on aromatase (Edmunds et al., 2005) which plays a key role in the endometriotic process (Attar and Bulun, 2006). The relationship between exposure and the disease observed in some studies could also be interpreted as evidence of reverse causality (Rothman and Greenland, 1998). Reverse causality means that the outcome (endometriosis) may have caused the exposure i.e. an excessive organochlorine body burden. According to this hypothesis, breast-feeding, that is known to reduce he organochlorine body burden, could explain why controls women (that are not infertile) have lower organochlorine body burdens. This hypothesis, while applying mainly to populationbased studies (women recruited in infertility clinics are often nullipare), stresses the need to take breastfeeding into account in future studies. In conclusion, the majority of studies dealing with peritoneal endometriosis did not observe a significant association with organochlorines. In contrast, it appears that deep endometriotic nodules are associated with a higher serum level of dioxin-like compounds and marker PCBs, which is compatible with a nosological distinction between peritoneal endometriosis and deep endometriotic nodules. It remains, however, to clarify the mechanism(s) potentially underlying this association. References American Society for Reproductive Medicine (ASRM) 1997. Revised American society for reproductive medicine classification of endometriosis: 1996, Fertil. Steril. 67, pp. 817–821. Attar, E., Bulun, S.E., 2006. Aromatase and other steroidogenic genes in endometriosis: translational aspects. Hum. Reprod. Update 12, 49–56. Aylward, L.L., Brunet, R.C., Starr, T.B., Carrier, G., Delzell, E., Cheng, H., Beall, C., 2005. Exposure reconstruction for the TCDD-exposed NIOSH cohort using a concentration- and age-dependent model of elimination. Risk Anal. 25, 945–956. Benagiano, G., Brosens, I., 1991. The history of endometriosis: identifying the disease. Hum. Reprod. 6, 963–968. Bertazzi, P.A., Zocchetti, C., Pesatori, A.C., Guercilena, S., Sanarico, M., Radice, L., 1989. Ten-year mortality study of the population involved in the Seveso incident in 1976. Am. J. Epidemiol. 129, 1187–1200. Birnbaum, L.S., Cummings, A.M., 2002. Dioxins and endometriosis: a plausible hypothesis. Environ. Health Perspect. 110, 15–21. Bois, F.Y., Eskenazi, B., 1994. Possible risk of endometriosis for Seveso, Italy, residents: an assessment of exposure to dioxin. Environ. Health Perspect. 102, 476–477. De Felip, E., Porpora, M.G., di Domenico, A., Ingelido, A.M., Cardelli, M., Cosmi, E.V., Donnez, J., 2004. Dioxin-like compounds and endometriosis: a study on Italian and Belgian women of reproductive age. Toxicol. Lett. 150, 203–209. Donnez, J., Nisolle, M., Smoes, P., Gillet, N., Beguin, S., Casanas-Roux, F., 1996. Peritoneal endometriosis and ‘‘endometriotic’’ nodules of the rectovaginal septum are two different entities. Fertil. Steril. 66, 362– 368. Donnez, J., Van Langendonckt, A., Casanas-Roux, F., Van Gossum, J.P., Pirard, C., Jadoul, P., Squifflet, J., Smets, M., 2002. Current thinking on the pathogenesis of endometriosis. Gynecol. Obstet. Invest. 54 (Suppl. 1), 52–58.
210
J.-F. Heilier et al. / Chemosphere 71 (2008) 203–210
Edmunds, K.M., Holloway, A.C., Crankshaw, D.J., Agarwal, S.K., Foster, W.G., 2005. The effects of dietary phytoestrogens on aromatase activity in human endometrial stromal cells. Reprod. Nutr. Dev. 45, 709–720. Eskenazi, B., Mocarelli, P., Warner, M., Samuels, S., Vercellini, P., Olive, D., Needham, L.L., Patterson Jr., D.G., Brambilla, P., Gavoni, N., Casalini, S., Panazza, S., Turner, W., Gerthoux, P.M., 2002. Serum dioxin concentrations and endometriosis: a cohort study in Seveso, Italy. Environ. Health Perspect. 110, 629–634. European Commission DG Environment. 1999. Compilation of EU Dioxin Exposure and Health Data – Summary Report available at
Mocarelli, P., Needham, L.L., Marocchi, A., Patterson Jr., D.G., Brambilla, P., Gerthoux, P.M., Meazza, L., Carreri, V., 1991. Serum concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin and test results from selected residents of Seveso, Italy. J. Toxicol. Environ. Health 32, 357–366. Nisolle, M., Donnez, J., 1997. Peritoneal endometriosis, ovarian endometriosis, and adenomyotic nodules of the rectovaginal septum are three different entities. Fertil. Steril. 68, 585–596. Pauwels, A., Schepens, P.J., D’Hooghe, T., Delbeke, L., Dhont, M., Brouwer, A., Weyler, J., 2001. The risk of endometriosis and exposure to dioxins and polychlorinated biphenyls: a case-control study of infertile women. Hum. Reprod. 16, 2050–2055. Porpora, M.G., Ingelido, A.M., di Domenico, A., Ferro, A., Crobu, M., Pallante, D., Cardelli, M., Cosmi, E.V., De Felip, E., 2006. Increased levels of polychlorobiphenyls in Italian women with endometriosis. Chemosphere 63, 1361–1367. Reddy, B.S., Rozati, R., Reddy, S., Kodampur, S., Reddy, P., Reddy, R., 2006. High plasma concentrations of polychlorinated biphenyls and phthalate esters in women with endometriosis: a prospective case control study. Fertil. Steril. 85, 775–779. Rier, S.E., Martin, D.C., Bowman, R.E., Dmowski, W.P., Becker, J.L., 1993. Endometriosis in rhesus monkeys (Macaca mulatta) following chronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Fund. Appl. Toxicol. 21, 433–441. Rier, S.E., Turner, W.E., Martin, D.C., Morris, R., Lucier, G.W., Clark, G.C., 2001. Serum levels of TCDD and dioxin-like chemicals in Rhesus monkeys chronically exposed to dioxin: correlation of increased serum PCB levels with endometriosis. Toxicol. Sci. 59, 147–159. Rothman, K.J., Greenland, S., 1998. Modern Epidemiology. Lippincott Williams & Wilkins, Philadelphia, PA. Treloar, S.A., Wicks, J., Nyholt, D.R., Montgomery, G.W., Bahlo, M., Smith, V., Dawson, G., Mackay, I.J., Weeks, D.E., Bennett, S.T., Carey, A., Ewen-White, K.R., Duffy, D.L., O 0 connor, D.T., Barlow, D.H., Martin, N.G., Kennedy, S.H., 2005. Genomewide linkage study in 1176 affected sister pair families identifies a significant susceptibility locus for endometriosis on chromosome 10q26. Am. J. Hum. Genet. 77, 365–376. Tsukino, H., Hanaoka, T., Sasaki, H., Motoyama, H., Hiroshima, M., Tanaka, T., Kabuto, M., Niskar, A.S., Rubin, C., Patterson Jr., D.G., Turner, W., Needham, L., Tsugane, S., 2005. Associations between serum levels of selected organochlorine compounds and endometriosis in infertile Japanese women. Environ. Res. 99, 118–125. Unfer, V., Casini, M.L., Costabile, L., Mignosa, M., Gerli, S., Di Renzo, G.C., 2004. Endometrial effects of long-term treatment with phytoestrogens: a randomized, double-blind, placebo-controlled study. Fertil. Steril. 82, 145–148. van den Berg, M., Birnbaum, L., Bosveld, A.T., Brunstrom, B., Cook, P., Feeley, M., Giesy, J.P., Hanberg, A., Hasegawa, R., Kennedy, S.W., Kubiak, T., Larsen, J.C., van Leeuwen, F.X., Liem, A.K., Nolt, C., Peterson, R.E., Poellinger, L., Safe, S., Schrenk, D., Tillitt, D., Tysklind, M., Younes, M., Waern, F., Zacharewski, T., 1998. Toxic equivalency factors (TEFs) for PCBs, PCDDs, PCDFs for humans and wildlife. Environ. Health Perspect. 106, 775–792. Vrijens, B., De Henauw, S., Dewettinck, K., Talloen, W., Goeyens, L., De Backer, G., Willems, J.L., 2002. Probabilistic intake assessment and body burden estimation of dioxin-like substances in background conditions and during a short food contamination episode. Food Addit. Contam. 19, 687–700. Woodward, M., 1999. Epidemiology study design and data analysis. Chapman & Hall/CRC, Boca Raton. Zober, A., Messerer, P., Huber, P., 1990. Thirty-four-year mortality follow-up of BASF employees exposed to 2,3,7,8-TCDD after the 1953 accident. Int. Arch. Occup. Environ. Health 62, 139–157. Zondervan, K.T., Cardon, L.R., Kennedy, S.H., 2002. What makes a good case-control study? Design issues for complex traits such as endometriosis. Hum. Reprod. 17, 1415–1423.