Effect of Oral Contraceptive Progestins on Serum Copper Concentration

Effect of oral contraceptive progestins on serum copper concentration

G Berg1, L Kohlmeier2 and H Brenner1

1Department of Epidemiology, University of Ulm, Albert-Einstein-Allee 43, 89069 Ulm, Germany; and 2Department of Nutrition, University of North Carolina, Chapel Hill, NC 27514, USA

Objectives: Recent epidemiologic studies have shown an increased mortality from cardiovascular diseases in people with higher serum copper levels. Even though higher serum copper concentration in women using oral contraceptives is well known, there is still uncertainty about the in¯uence of newer progestin compounds in oral contraceptives on serum copper concentration. This issue is of particular interest in the light of recent ®ndings of an increased risk of venous thromboembolism in users of oral contraceptives containing newer progestins like desogestrel compared to users of other oral contraceptives. Design: Cross-sectional epidemiologic study. Examinations included a detailed questionnaire on medical history and lifestyle factors, a seven day food record, and blood samples. Setting: National health and nutrition survey among healthy people living in private homes in West Germany in 1987 ± 1988. Subjects: Nonpregnant and nonlactating women aged 18 ± 44 y (n 610). Results: Overall, the use of oral contraceptives was positively associated with serum copper concentration in by bi- and multivariable linear regression models with log-transformed values of serum copper concentration as dependend variable and oral contraceptive preparations and potential confounding variables as independent variables. Serum copper concentration in women using oral contraceptives varied more strongly by different progestin compounds than by estrogen contents. The highest increase of serum copper was seen in women using oral contraceptives containing antiandrogen progestins (55%; 95% CI: 37 ± 76%), followed by desogestrel (46%; 95% CI: 36 ± 56%), norethisteron/lynestrenol (42%; 95% CI: 29 ± 57%), and levonorgestrel (34%; 95% CI: 24 ± 45%). Conclusion: While elevated serum copper concentration was found in users of all types of oral contraceptives, elevation was more pronounced among women taking oral contraceptives with antiandrogen effective progestins like antiandrogens or third generation oral contraceptives containing desogestrel. Further investigation is required to shed light on the possible role of high serum copper concentration in increasing cardiovascular or thrombotic risk of women using oral contraceptives. Sponsorship: The national health and nutrition survey was funded by the German Ministry of Research and Technology. Descriptors: oral contraceptives; copper; cardiovascular disease; progestin

Introduction generation oral contraceptives which were introduced in the early 1960s and which contained very high concentrations There is growing evidence that high serum copper concen- of estrogens and progestins (Margen & King, 1975). In the tration is an independent cardiovascular risk factor (Kok et 1970s, second generation drugs were developed which al, 1988; Salonen et al, 1991; Reunanen et al, 1992, 1996) contained reduced doses of estrogens and progestins. although results are not unequivocal (Klevley et al, 1984; Since 1980, third generation oral contraceptives were Klevley, 1989). introduced containing gonadal progestins like desogestrel, The use of oral contraceptives has consistently been gestoden and norgestimate. Little is known about the effect found to be associated with an increase of serum copper of these new progestins on serum copper concentration. concentration (Margen & King, 1975; Prasad et al, 1975; Taking into consideration that oral contraceptives have Horwitt et al, 1975; Sing et al, 1978; Prema et al, 1980; been linked to an increase in the incidence of vascular Crews et al, 1980; Vir & Love, 1981; Hinks et al, 1983; disease (Realini & Goldzieher 1985; La Vecchia, 1991), the Heese et al, 1987; Stauber & Florence, 1988; Lukaski, analysis of copper concentration of women using oral 1989; Fischer & L'Abbe, 1990; Johnson et al, 1992; New- contraceptives is important. Recent studies have suggested house et al, 1993; Milne & Johnson, 1993). However, only that the incidence for venous thromboembolism may be few studies have analyzed the in¯uence of different pre- even higher in women who use third generation oral parations of oral contraceptives on serum copper concen- contraceptives (WHO, 1995; Jick et al, 1995; Spitzer et tration. The majority of those studies analyzed only ®rst al, 1996) compared to women using second generation oral contraceptives, although these results are not homogeneous Correspondence: Dr G Berg. (Farmer et al, 1997). The aim of this study was to analyze Received May 3 1998; accepted June 12 1998 the in¯uence of different types of oral contraceptives Oral contraceptives and copper G Berg et al 712 classi®ed by progestin compounds and estrogen contents on height in metres and was used as measure of relative serum copper concentration in a population based study body weight. from Germany. Statistical analysis In a ®rst step, the study population was described with respect Method to basic sociodemographic variables. Age, smoking habits, Study population alcohol consumption, nutrient intake, history of pregnancies, This analysis is based on a national health and nutrition and BMI were compared between current, past, and never survey among healthy people in Germany conducted in users of oral contraceptives. The relation of different types of 1987/88 (German Nutrition SurveyÐVERA-StudyÐ oral contraceptives and covariates (age, smoking, alcohol, Verbundstudie ErnaÈhrungserhebung und Risikofaktoren copper-intake, history of pregnancies, and BMI) with serum Analytik). A strati®ed random sample was drawn from copper concentration was assessed using simple and multiple the non-institutionalized population aged 18 ± 89 years in linear regression with the natural logarithm of serum copper West-Germany and Berlin. With a participation rate of concentration as the dependent variable. The percentage about 70%, 2006 men and women were recruited in the increase of serum copper concentration per unit change in study. There were only minor differences between the study the independent variables was estimated by exponentiating population and the target population with respect to sex and the corresponding regression coef®cients. The log-transfor- age distribution (Schneider et al, 1992). The present ana- mation of serum copper concentration was necessary due to lysis on the in¯uence of oral contraceptive use on serum its highly skewed distribution. Copper-intake was likewise copper concentration was restricted to nonpregnant and log-transformed. We used untransformed values for BMI in nonlactating women aged 18 ± 44 years, who did not take all analyses. Forty-®ve out of 610 women were excluded from other sex hormones (n 610). the multivariable analysis because of missing values in one or more variables. One additional person was excluded because her serum copper concentration exceeded the mean by more Data collection than ®ve standard deviations even after log transformation. Participants were recruited during all seasons and in all This woman was therefore regarded as extreme and poten- regions of West-Germany and Berlin. They completed a tially in¯uential outlier. After this exclusion, the log-trans- detailed questionnaire on medical history and various formed values of serum copper concentration were close to sociodemographic and lifestyle factors including diet, normally distributed. Collinearity diagnostics in the multi- smoking habits and use of oral contraceptives. In addition, variable models were performed using the Belsley ± Kuh ± a seven day food record and fasting blood samples were Welsch-criteria (Belsley et al, 1980). Collinearity was not taken. The latter were collected in heparinized tubes found to be of concern in these analyses. between 7 and 10 am. Serum copper was measured with an electrothermic atomic absorption spectrophotometry. A detailed description of the survey methodology has been Results published elsewhere (Speitling et al, 1992). Use of oral contraceptives was categorized into current, Table 1 summarizes the bivariate associations of oral past and never use. Information on the use of oral contra- contraceptive use with the covariates: 195 women (32.1% ceptives was almost complete; only two persons refused to of the study population) were current users of oral contra- answer the questions about their contraception methods. ceptives, 322 women (53.0%) had used oral contraceptives The name of the currently used oral contraceptive prepara- in the past, and 91 women (15.0%) were never-users. tion was given by 152 out of 195 women using oral Current users of oral contraceptives on the average were contraceptives (78%). Preparations were classi®ed with younger than past and never-users. Forty percent of current regard to progestin compounds according to their chemical users were below age 25 years, compared to 12.1% and structure into: desogestrel, levonorgestrel, lynestrenol/norethisteron, and antiandrogen containing oral contraceptives. The estrogen contents in oral contraceptives was dichot- Table 1 Associations of oral contraceptive use with covariates omized in up to 45 mg ethylestradiol (30 ± 45 mg) and equal Use of oral contraceptives to 50 mg ethylestradiol. None of the preparations listed by the participants had a higher ethylestradiol content. Three Total Current Past Never types of preparations contained mestranol rather than (n 608) (n 195) (n 322) (n 91) ethylestradiol which was converted into the comparable Age (%) amount of ethylestradiol as described by Murad & Kuret, 18 ± 24 y 23.4 40.0 12.1 28.6 1991). 25 ± 34 y 43.2 47.2 45.0 26.4 35 ± 44 y 33.4 12.8 42.9 45.0 The following potential confounders were considered in Current smoker (%) 52.8 53.9 53.4 48.4 the analysis: Age, smoking status (current smoker/non- Alcohol consumption (%) smoker), alcohol consumption (any alcohol consumption, no alcohol 12.0 9.2 13.7 9.9 > 0 and 15 g alcohol/d, < 15 g alcohol/d) copper intake, > 0 and 15 g alcohol/d 62.5 68.2 59.9 60.4 history of pregnancies (ever pregnant/never pregnant), and > 15 g alcohol/d 25.5 22.6 26.4 29.7 Copper intake (g/day) 1.8 1.8 1.9 1.8 body mass index (BMI). Alcohol consumption and copper- Median (5th ± 95th percentile) (1.0 ± 2.8) (1.0 ± 2.9) (1.0 ± 3.0) (1.0 ± 2.8) intake were calculated from a seven day food record using History of pregnancies (%) 63.3 57.7 74.5 57.2 the data base of the computerized German food-code (BLS: Body mass index (BMI) (%) BundeslebensmittelschluÈssel). Records were assessed for < 19 kg/m2 8.9 9.2 9.3 6.6 19 and < 24 kg/m2 61.5 67.7 59.3 56.0 completeness before coding. The BMI was calculated by 24 kg/m2 29.6 23.7 31.4 37.4 dividing the body weight in kilograms by the squared Oral contraceptives and copper G Berg et al 713 28.6% of past and never-users, respectively. Two thirds of Table 3 Crude and adjusted relation between use of different current users of oral contraceptives were under 30 years of contraceptives de®ned by progestin compounds and serum copper age. Smoking and alcohol consumption were very common concentration and unrelated to oral contraceptive use in this study % increaseb (95% CI) population with an overall prevalence of about 50% and Meana 90%, respectively. Examination of copper intake showed n (mmol/l) Crude Adjusted c no major differences between current, past and never users No use of oral contraceptive 379 16.9 reference reference of oral contraceptives. A history of pregnancies was more Progestin compound frequent in past users of oral contraceptives (74.5%) than in antiandrogen 16 26.3 56 (37 ± 77) 56 (37 ± 77) current or never-users (57.7% and 57.2% respectively). The desogestrel 61 24.7 46 (36 ± 56) 46 (36 ± 57) percentage lynestrenol/norethisteron 28 23.7 41 (28 ± 56) 42 (28 ± 56) 2 levonorgestrel 44 22.4 33 (23 ± 44) 34 (24 ± 45) of overweight women (BMI 24 kg/m ) was highest unknown 39 20.8 23 (13 ± 34) 24 (14 ± 35) in women who had never used oral contraceptives, whereas r2 0.28 r2 0.31 current users more frequently had a BMI within a the recommended range of 19 ± 24 kg/m2 than other Mean values: log-transformed copper concentration was back- transformed to the natural scale. women. Only a minority of women (8.9%) had a BMI b% increase: the increase of log-transformed copper concentration below 19 kg/m2. compared to non-users of oral contraceptives (reference) was back- One hundred and ®fty-two (77.9%) of the 195 women transformed to the natural scale. using oral contraceptives supplied information about their cAdjusted for age, smoking, alcohol consumption, copper intake, history of contraceptive medication. Use of different types of oral pregnancies, BMI. contraceptives is shown in Table 2. None of the women reported to take the `minipill' which only contains proges- Table 4 Crude and adjusted relation between use of different tin. About 70% of the women with known oral contra- contraceptives de®ned by estrogen contents and serum copper concentration ceptive preparations used oral contraceptives containing % increaseb (95% CI) less than 45 mg ethylestradiol (median dosage during the 21 d of using: 32.4 mg). The most common progestin n Meana (mmol/l) Crude Adjusted c compound was desogestrel (41%) followed by levonorges- No use of oral 379 16.9 reference reference trel (30%) and lynestrenol/norethisteron (19%). Only 16 contraceptive women (11%) reported to use antiandrogen containing oral estrogen content contraceptives. There was a variety of combinations of 45 mg estrogen content 105 24.5 45 (37 ± 53) 45 (37 ± 54) 50 mg estrogen content 44 22.9 36 (25 ± 47) 36 (26 ± 48) progestin compounds and estrogen contents in this study unknown 39 20.8 23 (13 ± 34) 24 (14 ± 35) population. Each combination between categories of pro- r2 0.28 r2 0.30 gestin compounds and estrogen contents was used. Deso- gestrel, levonorgestrel and antiandrogen, in contrast to aMean values: log-transformed copper concentration was back-transformed to the natural scale. lynestrenol/norethisteron compounds, were more often b% increase: the increase of log-transformed copper concentration compared used with low ethylestradiol concentration. to non-users of oral contraceptives (reference) was back-transformed to the Tables 3 and 4 show the results of bivariate and multiple natural scale. regression models in which oral contraceptives were class- cAdjusted for age, smoking, alcohol consumption, copper intake, history of i®ed by progestin compounds and estrogen contents, pregnancies, BMI. respectively (the limited sample size did not allow joint classi®cation by both components in a single model). The of oral contraceptives containing desogestrel, lynestrenol/ mean serum copper concentration was higher among users norethisteron, and levonorgestrel, was associated with an of all types of oral contraceptive preparations than among increase of serum copper concentration by 46%, 42%, and non-users, but it varied by progestin compounds. The 34%, respectively, after adjustment for multiple co-vari- highest copper concentrations were seen in women taking ates. Even compared to women using levonorgestrel con- oral contraceptives containing antiandrogens. This was taining oral contraceptives, serum copper concentration equivalent to an increase of 56% of serum copper concen- was signi®cantly higher among women taking antiandrogen tration compared to non-users of oral contraceptives. Use containing oral contraceptives (P-value 0.048). Regard- ing the estrogen content, serum copper concentrations Table 2 Used oral contraceptives by estrogen content and progestin differed only marginally. However, use of oral contracep- compound tives containing lower doses of ethylestradiol was asso- Estrogen contenta ciated with a higher increase of serum copper. In both multiple regression models, a signi®cant positive relation- 30 ± 45 mg 50 mg ship between BMI and serum copper concentration was ethylestradiol ethylestradiol Total observed. The multivariable analyses showed no signi®cant Progestin compound n n n (%) association of age, smoking habits, alcohol consumption, copper intake, and history of pregnancies with serum Desogestrel 43 19 62 (41%) copper concentration. Overall, the impact of the covariates Levonorgestrel 37 8 45 (30%) was small, and adjustment for the covariates had little Lynestrenol/ 14 15 29 (19%) norethisteron effect on the estimates of oral contraceptive associated Antiandrogens 13 3 16 (11%) increase of serum copper concentration. In the bi- and Total 107 (70%) 45 (30%) 152 (100%) multivariable model in which oral contraceptives were classi®ed by progestin compound 28% and 31% of the aPreparation containing mestranol whose dosage was converted to comparable amounts of ethylestradiol (30): 80 mg mestranol correspond variance of the natural logarithm of serum copper concen- to 50 mg ethylestradiol. tration, respectively, was explained by the predictors. Oral contraceptives and copper G Berg et al 714 Discussion 31%. In particular, in the multiple regression models, we observed no signi®cant relationship of age, smoking habits, This study con®rms the well-known increase in serum alcohol consumption, copper intake and history of pregnan- copper concentration among oral contraceptive users and cies with serum copper concentration. There was a signi®- suggests that this increase is particularly high in women cant positive association between BMI and serum copper using oral contraceptives containing antiandrogen effective concentration. Adipose tissue is strongly related to sex progestins. The increase also appears to be relatively high hormones (Siiteri, 1987) and BMI is closely correlated to among women taking oral contraceptives containing the adipose tissue (Willett, 1990). Therefore, the signi®cant third generation progestin desogestrel which has a partially association between BMI and serum copper concentration antiandrogen capacity (Erkolla et al, 1990). The differential could also be explained by sex hormones. Our ®ndings are increase in serum copper concentration by progestin com- in agreement with other studies which also concluded that pound is unlikely to be explained by different estrogen factors affecting hormonal status (including pregnancies, contents of the preparations. menopausal status and use of oral contraceptives) are the The impact of oral contraceptives on serum copper key determinants of serum copper concentration (Grandejan concentration has been ascribed to an estrogen-induced et al, 1992). release of ceruloplasmin in the liver (O'Leary & Spellacy, This study provides a more detailed analysis of the 1969; Prema et al, 1980). Ceruloplasmin is the major relation of use of different types of oral contraceptives transporting protein of serum copper. We found only with serum copper concentration than previous studies. marginal differences in serum copper concentration, how- Nevertheless, the number of women using the different ever, comparing users of oral contraceptives with different types of oral contraceptives was limited which led to estrogen contents. The increase of serum copper seemed to relative wide con®dence intervals around the estimates of be even more pronounced in women using low estrogen the impact on serum copper. Sample size constraints also dose oral contraceptives ( 45 mg ethylestradiol) than in hindered analyses of the impact of additional factors, such women using oral contraceptives containing 50 mg ethyl- as duration of oral contraceptive use or the joint assesment estradiol. This might be explained by an opposite effect of of the impact of progestin compounds and estrogen con- progestins in oral contraceptives on serum copper or cer- tents. Larger studies are needed to address these aspects in uloplasmin: Levonorgestrel reduces ceruloplasmin in serum more detail. and reduces the estrogen effectÐprobably due to its partial androgen activity (Si et al, 1989) whereas desogestrel is more likely to be less androgen or partially antiandrogen Conclusion (Liukko et al, 1988; Erkolla et al, 1990). Liukko and colleagues (1988) found that women using desogestrel Our study suggests that oral contraceptive use, particularly containing oral contraceptives had a higher increase in use of antiandrogen effect-ive oral contraceptives, is serum ceruloplasmin than women using levonorgestrel strongly associated with increased serum copper concentra- containing oral contraceptives. tion. Given the increased evidence on the cardiovascular The question whether elevated serum copper concentra- risk associated with high serum copper concentration, the tion in young women using oral contraceptives leads to serum copper concentration in women who use oral contra- health risks such as an increased cardiovascular risk cannot ceptives, especially antiandrogen effective oral contracep- be answered with the data at hand. Nevertheless, in cohort tives, should not be neglected. studies high serum copper concentration unequivocally was associated with an increased cardiovascular mortality (Kok et al, 1988; Salonen et al, 1991; Reunanen et al, 1992, Acknowledgements ÐThe VERA-study was supported by Federal Ministry 1996). Furthermore, in one of the multinational case control of Research and Technology Grants No. 704752, 704754, and 704766. The studies about the effect of different progestins in low authors are responsible for the content of this publication. estrogen oral contraceptives on venous thromboembolic diseases, women who took the most common antiandrogen cyproteronacetat showed a higher thromboembolic riskÐ References similar to the third generation progestins, although this Belsley DA, Kuh E & Welsch RE (1980): Regression Diagnostics: ®nding was not signi®cant due to the small numbers Identifying In¯uential Data and Sources of Collinearity. New York: (WHO, 1995). The most used antiandrogen effective oral John Wiley & Sons. contraceptive in our analysis was `Diane' which also con- Crews MG, Taper LJ & Ritchey SJ (1980): Effects of oral contraceptive tains cyproteronacetat. It should be noted, however, that the agents on copper and zinc balance in young women. Am. J. Clin. Nutr. 33, 1940 ± 1945. different impact of various types of oral contraceptives on Erkkola R, Hirvonen E, Luikiu J, Lumme R, Mannikko H & Aydinlik, S the risk of cardiovascular diseases may also be mediated by (1990): Ovulation inhibitors containing cyproterone acetate or desoges- other mechanisms, such as the different impact on lipid trel in the treatment of hyperandrogenic symptoms. Acta. Obstet. metabolisms. The relative importance of different mechan- Gynecol. Scand. 69, 61 ± 65. Farmer RDT, Lawrenson RA, Thopson CR, Kennedy JG & Hambleton IR isms by which oral contraceptives may affect cardiovascu- (1997): Population based study of risk of oral thromboembolism lar diseases requires further research. associated with various oral contraceptives. Lancet 349, 83 ± 88. Our analysis suggests that sexual hormones are the main Fischer PWF & L'Abbe MR (1990): Effects of age smoking, drinking, predictors for serum copper concentration. In the regression exercise and estrogen use on indices of copper status in healthy adults. models the different progestin compounds alone explained Nutr. Res. 10, 1081 ± 90. Grandejan P, Nielsen GD, Jorgensen PJ & Horder M (1992): Reference the largest part (28%) of the total variance of serum copper intervals for trace elements in blood: signi®cance for risk factors. Scand. concentration. By including the different independent co- J. Clin. Lab. Invest. 52, 321 ± 337. variates in the multivariable analysis the total explained Hinks LJ, Clayton BE & Lloyd RS (1983): Zinc and copper concentration variance of serum copper concentration increased only to in leucocytes and erythrocytes in healthy adults and the effect of oral contraceptives. J. Clin. Pathol. 36, 1016 ± 1021. Oral contraceptives and copper G Berg et al 715 Heese H, Lawrence MA, Dempster WS & Pocock F (1987): Reference Prema K, Ramalakshimi BA & Babu S (1980): Serum copper and zinc in values for serum copper, ceruloplasmin and zinc and haematological hormonal contraceptive users. Fertil. Steril. 33, 267 ± 271. indices for healthy nulliparous women. S. Afr. Med. J. 72, 490 ± 493. Realini JP & Goldzieher JW (1985): Oral contraceptives and cardiovas- Horwitt MK, Harvey CC & Dahm CH (1975): Relationship between levels cular disease: A critique of the epidemiologic studies. Am. J. Obstet. of blood lipids, vitamins C, A, and E, serum copper compounds, and Gynecol 152, 729 ± 798. urinary excretions of tryptophan metabolites in women taking oral Reunanen A, Knekt P & Aaran RK (1992): Serum ceruloplasmin level and contraceptive therapy. Am. J. Clin. Nutr. 28, 403. the risk of myocardial infarction and stroke. Am. J. Epidemiol. 136, Jick H, Jick SS, Gurewich V, Myers MW & Vasilakis C (1995): Risk of 1082 ± 1090. idiopathic cardiovascular death and nonfatal venous thromboembolism Reunanen A, Knekt P, Marniemi J, MaÈki J, Maatela J & Aromaa A (1996): in women using oral contraceptives with differing progestagen com- Serum calcium, magnesium, copper and zinc and the risk of cardiovas- ponents. Lancet 346, 1589 ± 1593. cular death. Eur. J. Clin. Nutr. 50, 431 ± 437. Johnson PE, Milne DB & Lykken GI (1992): Effects of age and sex on Salonen JT, Salonen R, Korpela H, Suntioinen S & Tuomilehto J (1991): copper absorption, biological half-life, and status in humans. Am. J. Serum copper and the risk of acute myocardial infarction: a prospective Clin. Nutr. 56, 917 ± 925. population study in men in Eastern Finland. Am. J. Epidemiol. 134, Klevley LM, Inman L, Johnson LAK, Lawler M, Mahalko JR, Milne DB, 268 ± 276. Lukaski HC, Bolonchuk W & Sandstead HH (1984): Increased choles- Schneider A, Eberhardt W, Heseker H & Moch DKJ: (1992): Die VERA- terol in plasma in a young man during experimental copper depletion. Stichprobe im Vergleich mit Mikrozensus, VolkszaÈhlung und anderen Metabolism 33, 1112 ± 1118. Nationalen Untersuchungen. VERA-Schriftenreihe, Bd II, KuÈbler W, Klevley LM (1989): Ischemic heart disease as copper de®ciency. Adv. Exp. Anders HJ, Heeschen W (eds). Niederkleen (Germany): Wissenschaf- Med. Biol. 258, 197 ± 208. tlicher Fachverlag Dr. Fleck. Kok FJ, van Duijn CM, Hofman A, Voet GB, Wolff FA, Paays CH & Si S, Jun-kang C, Mei-li H & Fortherby K (1989): Effect of some oral Valkenburg HA (1988): Serum copper and zinc and the risk of death from contraceptives on serum concentrations of sex hormone binding globu- cancer and cardiovascular disease. Am. J. Epidemiol. 128, 352 ± 359. line and ceruloplasmin. Contraception 39, 385 ± 399. La Vecchia C (1991): Review, Sex hormones and cardiovascular risk. Siiteri PK (1987): Adipose tissue as source of hormones. Am. J. Clin. Nutr. Human Reprod. 7, 162 ± 166. 45, 277 ± 282. Liukko P, Erkkola R & Berginkk EW (1988): Progestagen-dependent Sing EJ, Baccarini IM, O'Neill HJ & Olwin HJ (1978): Effects of effect on some plasma proteins during oral contraception. Gynecol. oral contraceptives on zinc and copper levels in human plasma Obstet. Invest. 25, 118 ± 122. and endometrium during the menstrual cycle. Arch. Gynecol. 226, Lukaski HC (1989): Effects of exercise training on human copper and zinc 303 ± 306. nutriture. Adv. Exp. Med. Biol. 258, 163 ± 170. Speitling A, HuÈppe R, Kohlmeier M, Matiaske B, Stelte W, Thefeld W Margen S & King JC (1975): Effect of oral contraceptive agents on & Wetzel S (1992): Methodical Handbook, Nutritional Survey and the metabolism of some trace minerals. Am. J. Clin. Nutr. 28, Risk Factors Analysis. VERA-Schriftenreihe, Bd Ib, KuÈbler W, Anders 392 ± 400. HJ, Heeschen W (eds). Niederkleen (Germany): Wissenschaftlicher Milne DB & Johnson PE (1993): Assessment of copper status, effect of Fachverlag Dr. Fleck. age gender on reference range in health adults. Clin. Chem. 39, Spitzer WO, Lewis MA, Heinemann LA, Thorogood M & MacRae K 4883 ± 4887. (1996): Third generation oral contraceptives and risk of venous throm- Murad F & Kuret JA (1991): Estrogens and Progestins 1384 ± 412. In: boembolic disorders: An international case-control study. Br. Med. J. The Pharmacological Basis of Therapeutics eds. Eighth Edn, A Good- 312, 83 ± 88. man Gilman, TW Rall AS Nies & P Taylor. New York: Pergamon Stauber JL & Florence TM (1988): A comperativ study of copper, lead, Press. cadmium, and zinc in human sweat and blood. Sci. Tot. Environmental Newhouse IJ, Clement DB & Lai C (1993): Effects of iron supplementa- 74, 235 ± 247. tion and discontinuation on serum copper, zinc, calcium, and magne- Vir SC & Love AHG (1981): Zinc and copper nutriture of women taking sium levels in women. Med. Sci. Sports Exerc. 25, 562 ± 571. oral contraceptive agents. Am. J. Clin. Nutr. 34, 1479 ± 1483. O'Leary JA & Spellacy W (1969): Zinc and copper levels in pregnant Willett W (1990): Nutritional Epidemiology. New York, Oxford: Oxford women and those taking oral contraceptives. J. Obstet. Gynecol 103, University Press. 131 ± 132. World Health Organization Collaborative Study of Cardiovascular Disease Prasad AS, Oberleas D, Lei KY, Moghissi KS & Stryker JC (1975): Effect and Steroid Hormone Concentration (1995): Effect of different proges- of oral contraceptive agents on nutrients: I. Minerals. Am. J. Clin. Nutr. tagens in low oestrogen oral contraceptives on venous thromboembolic 28, 377 ± 384. disease. Lancet 346, 1582 ± 1588.