Document e5rxjGz05mzEpkLLb081JrQMM

DDT Metabolite and Androgens in African-American Farmers Author(s): Stephen A. Martin, Jr., Siobn D. Harlow, Mary Fran Sowers, Matthew P. Longnecker, David Garabrant, David L. Shore, Dale P. Sandler Source: Epidemiology, Vol. 13, No. 4 (Jul., 2002), pp. 454-458 Published by: Lippincott Williams & Wilkins Stable URL: http://www.jstor.org/stable/3703770 Accessed: 24/08/2010 15:19 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=lww. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. Lippincott Williams & Wilkins is collaborating with JSTOR to digitize, preserve and extend access to Epidemiology. http://www.jstor.org DDT Metabolite in African-American and Androgens Farmers Matthew Stephen A. Martin, Jr.,1'2 Siobdn D. Harlow,2 Mary Fran Sowers,2 P. Longnecker,1 David Garabrant,2 David L. Shore,3 and Dale P. Sandler1 Background. The ubiquitous dichlorodiphenyltrichloroethane (DDT) metabolite l,l-dichloro-2,2-bis(|)-chlorophenyi)ethylene (DDE) is an androgen receptor antagonist. Data on po? tential antiandrogenic activity of DDE in humans are limited. Methods. The relations between concentrations of plasma DDE and several serum androgens (total testosterone, bioavailable testosterone, 5a-dihydrotestosterone, and free androgen index) were examined in 137 North Carolina black male farmers, using multiple linear regression. Results. Participants ranged in age from 30 to 88 years (mean = 62 years). Most had farmed for about 30 years and 27% reported having used DDT. The median DDE level was 7.7 ug per liter (1213 fig per kg lipid), slightly higher than in other recent studies. Overall, concentrations of DDE and androgens were unrelated. Total testosterone decreased 2% (95% confi? dence limits [CL] = ?9%, 5%) per increase in interquartile distance of lipid-adjusted DDE. The percentage change in other hormones was similarly negligible. However, among those whose DDE level was in the top tenth percentile, com? pared with all others, total testosterone and free androgen index were lower by 23% (CL= -40%, 1%) and 22% (CL =?41%, 4%) respectively. Plasma androgen levels decreased with age, a relation that has previously been studied only in whites. Conclusions. Studies of more highly exposed populations may be needed to evaluate effects, if any, of DDE. (Epidemiology 2002;13:454-458) Key words: anti-androgen, androgens, hypogonadism, blacks, farmers, pesticide exposure, DDT, DDE. major metabolite of dichlorodiphenyltrichloroethane (DDT), l,l-dichloro-2,2-bis(f)-chloroThe phenyl)ethylene (DDE), is a persistent environ? mental contaminant that is ubiquitous among people worldwide. DDE was recently found to bind with the androgen receptor in male rats1 and to inhibit the bind? ing of androgen to the androgen receptor, androgeninduced transcriptional activity, and androgen action.1-4 Kelce and Wilson5 suggested that DDE levels in humans can exceed the levels that inhibit human androgen receptor transcriptional activation in vitro. If DDE acts as Fromthe 'NationalInstituteof EnvironmentHalealthSciences,Epidemiology BranchR, esearchTriangleParkN, C;2DepartmenoftEpidemiologUy,niversity of MichiganSchool of PublicHealth,Ann Arbor,Ml; and 3Westat,Inc, DurhamN, C. Addresscorrespondencteo: Dale P. SandlerE, pidemiologByranchN, ational Instituteof EnvironmentaHl ealthSciences,P.O. Box 12233,MD A3-05, ResearchTriangleParkN, C 27709;sandler@niehs.nih.gov Thisworkwasfundedthroughthe EpidemiologByranchat theNationalInsti? tuteof EnvironmentaHlealthScienceswithadditionaslupporftromthe Na? tionalInstituteosf HealthOfficeof Researchon MinorityHealth. Submitted26June2001;finalversionaccepted4 March2002. Copyrigh?t 2002byLippincotWt illiams& Wilkins,Inc. DOI: 10.1097/01.EDE.0000016382.31505.66 an androgen receptor antagonist in humans, it could affect normal sexual differentiation and fertility in males5 or induce a compensatory increase of testoster? one.6 Blocking the androgen receptor causes a decrease in androgenic effects that would be detected by the hypothalamus, leading to a compensatory increase in levels of gonadotropins and thus androgens.7 Flutamide is an antiandrogen that, like DDE, acts by blocking androgen binding with its receptor. Administration of flutamide to human males causes androgen levels to increase.8'9 We hypothesized that DDE would have a similar effect on androgens, so that in a cross-sectional study DDE and androgen levels would be directly associated. In this study, we examined the relationship between concurrent concentrations of plasma DDE and several serum androgens in a group of black male farmers and farm workers in North Carolina. Farmers and farm work? ers are likely to be more highly exposed to pesticides than is the general population.10 Furthermore, blacks have been found to have higher DDE concentrations than whites.11,12 Thus, North Carolina black farmers and farm workers were potentially more highly exposed to DDT than other groups in the United States. 454 Epidemiology july 2002, Vol. 13 No. 4 dde and androgens in black farmers 455 Methods The Agricultural Health Study is an ongoing prospec? tive study of licensed pesticide applicators from Iowa and North Carolina.13 To include more African Americans, in 1995-1996 we recruited additional participants at 118 predominantly black churches in five rural North Caro? lina counties. Eligibility criteria included being a resi? dent of North Carolina, not having been previously recruited for the Agricultural Health Study, and having had at least 2 years of adult farming experience or a spouse with such. About 2,300 potentially eligible men and women completed a brief screening questionnaire, and 1,602 were found to be eligible. Of these, 1,186 (74%) completed a telephone interview about their farm experiences and health status. A total of 389 (33%) black men and 797 (67%) black women completed the interview, reflecting the predominance of women among the church attendees. We recontacted the 389 men in 1999 and asked them to complete a follow-up telephone questionnaire. The follow-up interview included more detailed questions on previous farming and years of pesticide and DDT expo? sure. Fifty-five men did not complete the follow-up interview as a result of poor health (N = 20), death (N = 30), or lack of telephone service at the place of residence (N = 5). Of the remaining 334 men, 275 (82%) completed the questionnaire, 33 (10%) refused, and 26 (8%) could not be traced. We attempted to collect a blood sample from inter? viewed men who lived in four contiguous northeastern counties. We excluded 27 of the 228 eligible men from a blood draw because they were currently taking anticoagulant medication or they had had a seizure in the past. Of the 201 remaining men, 141 (70%) attended one of 12 examination sessions. Blood samples could not be collected from three attendees because of very high blood pressure or other medical reasons. Thus, we ob? tained blood samples from 138 men. Fasting blood for DDE analysis was collected in a metal-free Vacutainer containing EDTA (lavender top) and was kept cool until later the same day when the plasma was frozen in glass at ? 20?C until analyzed. The serum from blood collected in a red-topped Vacutainer was refrigerated until analyzed for lipids within 48 hours. Samples for hormone analysis were refrigerated until later on the day of collection, when they were frozen at -80?C until analyzed. The study was reviewed and approved by the Institu? tional Review Boards at the National Institute of Envi? ronmental Health Sciences and the National Institutes of Health. Participants gave verbal informed consent for the telephone interview and provided written informed consent for the blood draw. Organochlorine Analysis DDE was extracted from 2 ml of plasma using solid- phase extraction (C18) by the Centre de Toxicologie du Quebec. After a washing step, DDE was eluted with isooctane. The extract was then analyzed by gas chro? matography with electron capture detection. Identifica? tion and quantification of DDE was confirmed by mass spectrometry. DDE 13C was used as an internal standard. The limits of quantification and detection for total DDE were 0.5 and 0.2 jxg per liter, respectively. All study samples were above the detection limit for DDE. A serum DDE 5 jixgper liter standard and a serum DDE 25 /utgper liter standard were analyzed with each batch. The corresponding between-batch coefficient of variation (CV) for the standards was 5.2% and 8.3%, respectively (N = 24 batches); recovery averaged 97%. Lipid and Hormone Analysis Standard enzymatic assays for total cholesterol and triglycerides were performed at the Duke University CARL Clinical Laboratory, where androgens were also measured. Total testosterone (TT) was determined using a chemiluminescent competitive immunoassay (Immu- lite 2000 Immunoassay System, Diagnostic Products Corp, Los Angeles, CA). Sex hormone binding globulin (SHBG) was quantified using a chemiluminescent im- munometric assay (Immulite Immunoassay System, Di? agnostic Products Corp). Bioavailable testosterone (BT) was measured using a competitive enzyme immunoassay (American Laboratory Products Company, Windham, NH). 5a-Dihydrotestosterone (DHT) was measured us? ing a direct enzyme-linked immunosorbent assay (Amer? ican Laboratory Products Company). The mean be? tween-batch CV% was 5.2% for TT, 2.6% for SHBG, 10.3% for BT, and 9.3% for DHT. Additional Covariates A trained technician measured height, weight, and waist and hip circumferences of participants at the time blood was drawn. This information was used to calculate body mass index (BMI, kg/m2) and waist-to-hip ratio (WHR). Statistical Analysis To express the DDE plasma concentrations on a lipid basis, total lipids (TL) were estimated as follows14: TL = 2.27 (total cholesterol) + triglycerides + 0.623. We calculated the free androgen index using the following formula15: free androgen index = (TT [nmol/L] X 100)/ (SHBG [nmol/L]). Because of the skewed distributions of plasma DDE, lipid-adjusted DDE, and serum hormone concentrations, the values were loge transformed before testing for trend across age categories. Also, the Spear? man correlation coefficient between lipid-adjusted DDE and BMI (and WHR) was calculated. 456 Martin et al. o < tT V jI EPIDEMIOLOGY July 2002, Vol. 13 No. 4 Models of loge-transformed values of each serum hor? mone concentration were fitted with coefficients for untransformed DDE concentration (lipid basis), age at blood draw, and BMI (or WHR, where noted). To increase interpretability, we also present the model re? sults on hormone-DDE associations as the percentage change in interquartile distance of DDE, with the change expressed using hormone values in their natural scale. To evaluate linearity, we examined models of hormone levels with DDE levels categorized by quartile or dichotomized at the 90th percentile (vs all other values) of the distribution. One subject had an unusually high plasma DDE con? centration of 232 jiLg per liter. The sample was reextracted and reassayed and the concentration on repeat analysis was 231 jutgper liter. As this concentration was an outlier, the observation was excluded from our re? ported statistical analysis. Including this participant did not alter the findings. Results The average age at blood draw among the 137 men was 62 years (standard deviation [SD] =13 years; range 30-88). The average BMI was 28.7 kg/m2 (SD = 4.7), and the average WHR was 0.95 (SD = 0.06). Fifty-five per cent of the men had farmed for 25 years or more, 74% reported ever having used any pesticides, and 27% reported ever-use of DDT. The average number of years of any pesticide use was 12.3, and the average for DDT use for agricultural purposes was 2.2 years. The median DDE level (7.7 jxg per liter) was slightly higher than the level observed among other recently studied groups of North Carolinians; Millikan et al.16 reported a median DDE level of 5.7 /xg per liter, and Vine et al.11 reported a median of 2 u,g per liter. Plasma DDE and lipid-adjusted DDE concentrations increased linearly with age (Table 1). The Spearman correlation coefficient between lipid-adjusted plasma DDE and BMI was 0.11 (95% confidence limits [CL] = -0.06, 0.28), and that between DDE and WHR was 0.19 (CL = 0.03, 0.36). Serum TT, BT, 5a-dihydrotestosterone (DHT), and free androgen index decreased linearly with age, whereas sex hormone binding globulin (SHBG) increased lin? early with age (Table 1). All of these associations were consistent with expectations. 12,18~22Four men had total testosterone levels less than 200 ng per deciliter, a value used clinically to define hypogonadism.23 All were 65 years of age or older. Thus, among men age 65 years or older, the prevalence of hypogonadism was 6%. Multivariate regression models of loge TT, BT, and DHT concentrations, and free androgen index showed that after adjusting for age and BMI, lipid-adjusted plasma DDE was not associated with any androgen (Table 2). Adjust- Epidemiology july 2002, Vol. 13 No. 4 DDE and Androgens in Black Farmers 457 TABLE 2. Multiple Linear Regression Analysis Between Serum Androgens and Variables of Interest in the Agricultural Health Study?Special Recruitment Study (N = 137) fD*oPDrmaEr=utilaal,rl1l-Fv5daA:ilcIuh=elsoa(rrToeT-s2h[,on2wm-bnois.l(/lp]X-ch10lo0r)o/(pShHeBnyGl[)nemthoySll/Eel]n=)we, ;shtaernedTaTred=rrtoort/;aul-tge/skt=ogsptearrotsnpee.rbillion.Thefreeandrogenindex(FAI)wascalculatedusingthefollowing t Percentagcehangein hormoneperincreasein interquartidleistanceforlipid-adjusteDdDE(/xg/kg)andits95%confidencelimits. ment for WHR instead of BMI gave similar results (not shown). In the model of free androgen index, the coeffi? cient for the interaction between age and DDE level was -3.5 X IO"6 (CL = -7.4 X IO"6,0.5 X 10~6), suggesting that at older ages, DDE may be associated with a lower free androgen index. For example, the model predicted that the free androgen index for an 85-year-old man with a plasma DDE level of 5,000 /xg/kg would be 18% lower than if his DDE level were 1,000 /mg/kg. When the multivariate regression analysis was re? peated using quartiles of lipid-adjusted plasma DDE, the results were consistent with a linear relation between DDE and each of the loge hormone concentrations. On the other hand, total testosterone was 23% lower (CL = ?40%, 1%) among those whose DDE level was in the top tenth percentile, compared with all others. Those in the top tenth percentile also differed in level of free testosterone ( ? 18%; CL = ?43, 17), 5a-dihydrotesterone (?0.5%; CL = ?25, 33), and free androgen index (-22%; CL = -41, 4) (not shown in table). Discussion Among the North Carolina black male farmers and farm workers in our study, overall androgen levels were unrelated to DDE levels overall. The absence of an association between lower levels of DDE and androgen level among adult males has recently been reported.24 Twenty-four young men from a malaria area of Mexico had a mean serum DDE level of -600 jxg per liter.25 DDE was inversely related to the ratio of bioavailable to total testosterone and positively related to SHBG level in these young men. We found neither relation in our data. Although neither association observed in the Mexico study was in the direction we hypothesized (that DDE, acting as an androgen receptor antagonist, might lead to a compensatory increase in androgens), each was con? sistent with some disruption of androgen metabolism by DDE. DDE was also associated with decreased semen volume and sperm count in the men from Mexico, providing further support for some disruption of andro? gen metabolism by DDE. Our data suggested that higher levels of DDE were associated with lower levels of total testosterone and free androgen index. Overall, the avail? able data suggest that if DDE alters androgen levels in adult males, this may occur only at higher exposure levels. Among U.S. agricultural workers exposed to DDT before it was banned, serum levels of DDT and DDE combined were directly related to levels of several enzymes in serum that reflect liver function.26 That observation lends credibility to the possibility that ef? fects on liver function, such as enzyme induction, could be responsible for lower androgen levels. In our data, DDE level was inversely associated with free androgen index among older men. As with the associations in the Mexican men, this finding was not in the hypoth? esized direction. The relation of DDE with the ratio of bioavailable to total testosterone was not similarly modified by age. Without confirmation in other data, the impor? tance of the effect modification by age is questionable. DDE concentrations tend to remain fairly constant over time and a single measure for estimating exposure is highly reliable.27 Among 99 men with stable weight, the correlation among serum total testosterone levels mea? sured 1 year apart was 0.61, suggesting that a single androgen level is fairly reliable.28 Nonetheless, our abil? ity to measure an association of DDE with androgen levels was hampered by the limited sample size, and 458 Martin et al. EPIDEMIOLOGY July 2002, Vol. 13 No. 4 possibly by selection bias. Plasma DDE and serum hor? mone concentrations were available for only 138 of the 275 men. Those who did and did not provide a blood sample differed in level of education and a few farmingrelated characteristics. Whether the relation of DDE to androgen levels differs between the study group and the larger group of eligible men is unknown, although such a difference seems unlikely. The age-related decline in testosterone levels among males is well documented for whites but not for other racial/ethnic groups.23 Our data provide further informa? tion on the population distribution of androgen levels in African-American men and suggest that the decline with age observed in whites also occurs in African Americans. Whether the prevalence of hypogonadism in our population was unusual could not be determined because of the absence of data on the prevalence of low total testosterone levels among other populations with defined age distributions and because of uncertainties about the comparability of low testosterone levels mea? sured in different laboratories. Acknowledgments Wearegratefutlo GaryGrantandConcernedCitizensofTilleryt;oJoyPierce, Glenn HeartwellA, lesiaSanyika,SusieCovington,GwenMcKoy,Teresita GabrielA, nitaGalyeana, ndFikriYucel,whoconductedthefieldworkforthe AgriculturaHl ealthStudy?SpecialRecruitmenSt tudy;to BruceLobaugh, Jean-PhilippWe ebera, ndAlainLeBlanfcorassistancieninterpretinlgaboratory results;andto the staffsat the Centrede Toxicologiedu QuebecandDuke UniversityCARLClinicalLaboratory. References 1. Kelce WR, Stone CR, LawsSC, Gray LE,KemppainenJA, Wil? son EM.PersistentDDT metabolitep,p*-DDEis a potent androgen receptorantagonist.Nature 1995;375:581-585. 2. Kelce WR, LambrightCR, GrayLE,RobertsKP. Vinclozolin and p.p'-DDEalter androgen-dependentgene expression:in vivocon? firmation of an androgen receptor-mediatedmechanism. Toxicol Appl Pharmacol1997;142:192-200. 3. Dutour A. Dichlorodiphenyltrichloroethaneand androgenrecep? tor. Eur] Endocrinol1996;134:422-423. 4. Gaido KW, LeonardLS, Lovell S, et al. Evaluationof chemicals with endocrine modulatingactivity in a yeast-basedsteroid hor? mone receptor gene transcriptionassay. ToxicolAppl Pharmacol 1997;143:205-212. 5. Kelce WR, Wilson EM. Environmentalantiandrogens:develop? mental effects, molecularmechanisms,and clinical implications. ] Mol Med 1997;75:198-207. 6. You L, CasanovaM, Archibeque-EngleS, Sar M, Fan L, Heck H. Impairedmale sexual development in perinatal Sprague-Dawley and Long-Evanshooded ratsexposed in uteroand lactationallyto p,p'-DDE.ToxicolSci 1998;45:162-173. 7. HagmarL, BjorkJ, Sjodin A, BergmanA, ErfurthEM. Plasma levels of persistent organohalogensand hormone levels in adult male humans.ArchEnvironHealth2001;56:138-143. 8. MetzgerDL, KerriganJR. Androgen receptorblockadewith flut? amide enhances growthhormone secretion in late pubertalmales: evidence for independentactions of estrogenand androgen.J Clin EndocrinoMl etab1993;76:1147-1152. 9. Stone NN, Clejan SJ. Response of prostate volume, prostatespecificantigen, and testosteroneto flutamidein men with benign prostatichyperplasia.J Androl1991;12:376-380. 10. BrockJW, MelnykLJ,CaudillSP, Needham LL,BondAE. Serum levels of several organochlorinepesticides in farmerscorrespond with dietary exposure and local use history. Toxicol Ind Health 1998;14:275-289. 11. DaviesJE,EdmundsonWF, RaffonelliA, CassadyJC, MorgadeC. The role of social class in human pesticide pollution. Am ] Epi? demiol1972;96:334-341. 12. Stehr-GreenPA. Demographicand seasonalinfluenceson human serumpesticide residue levels. ] ToxicolEnvironHealth 1989;27: 405-421. 13. AlavanjaMCR, SandlerDP, McMasterSB, et al. The Agricultural Health Study. EnvironHealthPerspect1996;104:362-369. 14. Phillips DL, PirkleJL, BurseVW, BernertJT Jr, HendersonLO, Needham LL. Chlorinated hydrocarbonlevels in human serum: effects of fasting and feeding. ArchEnvironContamToxicol1989; 18:495-500. 15. KapoorP, LuttrellBM, WilliamsD. The freeandrogenindex is not validfor adultmales.] SteroidBiochemMolBiol1993;45:325-326. 16. Millikan R, DeVoto E, Duell EJ, et al. Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-Americanand white women in North Carolina. Cancer EpidemioBl iomarkerPs rev2000;9:1233-1240. 17. Vine MF, Stein L, Weigle K, et al. Plasmal,l-dichloro-2,2-bis(f>- chlorophenyl)ethylene (DDE) levels and immune response.AmJ Epidemio2l 001;153:53-63. 18. FreedmanDS, O'BrienTR, FlandersWD, DeStefano F, Barboriak JJ.Relation of serumtestosterone levels to high-density lipopro? tein cholesterol and other characteristics in men. Arterioscler Thromb1991;11:307-315. 19. Inmaculada Sanz-GallardoM, Guallar E, Van Tveer P, et al. Determinantsof p,p'-dichlorodiphenyldichloroethan(eDDE) con? centration in adipose tissue in women fromfive Europeancities. Arch EnvironHealth 1999;54:277-283. 20. KorenmanSG. Androgen function afterage 50 and treatmentof hypogonadism.CurrTherEndocrinoMl etab 1997;6:621-624. 21. Laden F, Neas LM, Spiegelman D, et al. Predictorsof plasma concentrations of DDE and PCBs in a group of U.S. women. EnvironHealthPerspect1999;107:75-81. 22. TsitourasPD, Bulat T. The aging male reproductivesystem. En? docrinolMetabClin NorthAm 1995;24:297-315. 23. Tenover JL. Male hormone replacement therapy including "andropause."EndocrinoMl etabClin NorthAm 1998;27:969-987. 24. Persky V, Turyk M, Anderson HA, et al. The effects of PCB exposureand fish consumptionon endogenoushormones.Environ HealthPerspect2001;109:1275-1283. 25. Ayotte P, Giroux S, Dewailly E, et al. DDT sprayingfor malaria control and reproductivefunction in Mexican men. Epidemiobgy 2001;12:366-367. 26. MorganDP, Lin LI. Bloodorganochlorinepesticideconcentrations, clinicalhematologyand biochemistryin workersoccupationallyex? posedto pesticides.ArchEnvironContamToxicol1978;7:423-447. 27. Gammon MD, Wolff MS, Neugut Al, et al. Temporalvariationin chlorinated hydrocarbonsin healthy women. Cancer Epidemiol BiomarkerPs rev 1997;6:327-332. 28. Dai WS, KullerLH, LaPorteRE, Gutai JP, Falvo-GerardL, CaggiulaA. The epidemiologyof plasmatestosteronelevels in middleaged men. Am] Epidemiol1981;114:804-816.