点击显示 收起
1 From the Psychiatric Center GGZ Delfland, Delft, Netherlands (EJG); the Department of Endocrinology, Andrology Unit, VU University Medical Center, Amsterdam (LJGG and AWFTT); and the Wageningen Center for Food Sciences, Wageningen University, Wageningen, Netherlands (MBK and PLZ)
2 Supported by the Wageningen Centre for Food Sciences (to MBK and PLZ), which is an alliance of major Dutch food industries, Maastricht University, TNO Nutrition and Food Research, and Wageningen University and Research Centre that receives financial support from the Dutch government. 3 Address reprint requests to EJ Giltay, Psychiatric Center GGZ Delfland, PO Box 5016, 2600 GA Delft, Netherlands. E-mail: giltay{at}dds.nl.
ABSTRACT
Background: During pregnancy there is a high demand for docosahexaenoic acid (DHA), which is needed for formation of the fetal brain. Women who do not consume marine foods must synthesize DHA from fatty acid precursors in vegetable foods.
Objective: We studied sex differences in DHA status and the role of sex hormones.
Design: First, DHA status was compared between 72 male and 103 female healthy volunteers who ate the same rigidly controlled diets. Second, the effects of sex hormones were studied in 56 male-to-female transsexual subjects, who were treated with cyproterone acetate alone or randomly assigned to receive oral ethinyl estradiol or transdermal 17ß-estradiol combined with cyproterone acetate, and in 61 female-to-male transsexual subjects, who were treated with testosterone esters or randomly assigned for treatment with the aromatase inhibitor anastrozole or placebo in addition to the testosterone regimen.
Results: The proportion of DHA was 15 ± 4% (
Key Words: n–3 Fatty acids • docosahexaenoic acid • estrogen administration • testosterone administration
INTRODUCTION
Eicosapentaenoic acid (EPA, 20:5n–3) and docosahexaenoic acid (DHA, 22:6n–3) are n–3, long-chain, highly unsaturated fatty acids (HUFAs). During pregnancy maternal DHA concentrations increase, and the fetus needs substantial quantities of DHA, especially for development of the central nervous system (1-4). There is a minor pathway of biosynthesis from the precursor -linolenic acid (ALA, 18:3n–3) to EPA and DHA (5-9), and this pathway occurs predominantly in the liver (10). Vegetarians and persons who do not eat fish may depend on this pathway because the source of EPA and DHA is fatty fish, whereas ALA is found in vegetables, walnuts, and vegetable oils (eg, canola and soybean oils). Moreover, a high dietary intake of DHA and EPA, and maybe also of ALA, may lower the risk of fatal ischemic heart disease and sudden death (11-17).
With regard to the biosynthesis of DHA from precursors, the endogenous conversion of 13C-labeled ALA into DHA seems to be greater in women than in men (9, 18), and in biopsies of subcutaneous adipose tissue, DHA concentrations were lower in men than in women (19). Previous studies in women have shown shifts in fatty acid composition—which suggest effects on elongation and desaturation—during the menstrual cycle, pregnancy, and menopause (20-22) and changes in fatty acid composition induced by several estrogenic, progestational, and antiestrogenic compounds but have not examined n–3 HUFAs (23-29). Moreover, no study has directly compared plasma n–3 HUFAs in men and women eating the same controlled diets or investigated the effects of sex hormones in a controlled trial in humans.
We chose to investigate the effects of hormones in transsexual subjects, because all are in a similar age range and are eugonadal before their sex reassignment (ie, they have sex steroid concentrations that are appropriate for their genetic sex). Estradiol and testosterone concentrations are naturally low in men and women, respectively, and cross-sex hormone administration induces subsequently profound changes in hormonal status, which provides a unique model for studying sex hormone effects on n–3 HUFAs in human subjects.
SUBJECTS AND METHODS
Comparison between men and women
This study aimed to investigate whether women have higher circulating DHA concentrations than do men who consume the same diet. We studied 175 healthy volunteers who consumed strictly controlled diets for 3-wk periods of dietary intervention. The 3 experiments ran from 1987 to 1992. Most subjects were university students, and the women and the men were similar in age. The diets contained no n–3 fatty acids from fish (ie, no EPA or DHA) (30). All the subjects were normolipemic and apparently healthy and had a median age of 23 y (range: 18–67 y) and a mean (±SD) body mass index (in kg/m2) of 22.2 ± 2.6. We categorized the subjects into 3 groups: 72 men, 71 women not taking oral anticontraceptives (OACs), and 32 women taking OACs. The diets contained constant amounts of ALA. The compositions of other macro- and micronutrients were exactly the same for the men and the women. All diets supplied on average between 11.0 and 12.4 MJ (2630 and 2950 kcal)/d and 2.8% of total energy intake as stearic acid (18:0), 0.1% as trans fatty acids, 10% as palmitic acid (16:0), 10% as oleic acid (18:1n–9), and 10% as myristic acid for the different controlled diets. Energy requirements were estimated from 3-d food records, and each subject received a diet that maintained his or her body weight. Body weight was recorded twice weekly, and energy intake was adjusted when necessary. The subjects were asked to maintain their usual pattern of activity and not to change their smoking habits or use of oral contraceptives. They recorded in diaries any sign of illness, any medications or oral contraceptives taken, and any deviations from their diets. Inspection of the diaries showed that the participants followed the instructions carefully.
All experiments used special fats developed by the Unilever Research Laboratory (Vlaardingen, Netherlands). Fasting blood samples were collected 2 or 3 times after 3 wk of each experimental diet, and EDTA plasma was stored at –80 °C. We showed previously that cholesterol ester DHA stabilizes within 2 wk (31). Plasma cholesteryl ester fatty acids were measured in serum samples and then averaged for each subject. The studies were approved by the Ethical Review Committees of Wageningen University.
Intervention studies with sex steroids
To test the role of sex hormones, the prospective effects of cross-sex hormone administration were studied in 56 male-to-female (MF) and 61 female-to male (FM) transsexual subjects. These studies ran from 1994 to 2003. Psychological criteria for diagnosis and treatment followed the guidelines provided by the Harry Benjamin International Sex Dysphoria Association (32). All of these subjects were eugonadal, had never taken sex steroids, and were aged between 16 and 50 y (median: 29 y). They had a mean body mass index of 22.3 ± 3.3 (33, 34). We studied 8 groups (Tables 1 and 2). Thirty MF transsexual subjects were randomly assigned with open labels to receive either oral ethinyl estradiol [Lynoral, 100 µg/d; Organon, Oss, Netherlands; group 1 (n = 15)] or transdermal 17ß-estradiol [Estraderm TTS 100, 100 µg 2 times/wk; CIBA-Geigy, Basel, Switzerland; group 2 (n = 15)], both in combination with cyproterone acetate (Androcur, 100 mg/d; Schering, Berlin). Cyproterone acetate is an antiandrogen with progestational properties. A control group of 10 MF transsexual subjects received cyproterone acetate alone (group 3). Another group of 16 MF transsexual subjects were treated with oral ethinyl estradiol plus cyproterone acetate for 12 mo (group 4).
View this table:
TABLE 1. Baseline characteristics of the male-to-female transsexual subjects in the intervention studies
View this table:
TABLE 2. Baseline characteristics of the female-to-male transsexual subjects in the intervention studies
Thirty ovariectomized, FM transsexual subjects receiving treatment with exogenous testosterone were randomly assigned in double-blind fashion to receive the selective aromatase inhibitor anastrozole [Arimidex, 1 mg/d; Zeneca, Wilmington, DE; group 5 (n = 16)] or placebo [group 6 (n = 14)]. In ovariectomized FM transsexual subjects, virtually all circulating testosterone originates from exogenous administration, and anastrozole inhibits the conversion of administered testosterone to 17ß-estrogen (35). Thirty-one FM transsexual subjects were treated intramuscularly with 250 mg testosterone esters (Sustanon; Organon) every 2 wk according to the standard treatment at our clinic. In 17 FM transsexual subjects, blood was collected after 4 mo (group 7), and in 14 FM transsexual subjects, blood was also collected after 12 mo (group 8).
In groups 4 and 8, a validated Dutch semiquantitative food-frequency questionnaire [of the European Prospective Investigation into Cancer and Nutrition (EPIC)] was used at baseline and after 1 y to estimate the daily intake of total energy and fats (36). Nutrient intakes were calculated with the use of the Dutch nutrient database. All studies were approved by the Ethical Review Committees of the VU University Medical Center.
Laboratory tests
Blood was drawn before testosterone treatment in genetic women between days 5 and 9 of the follicular phase and, during testosterone treatment, within 5–9 d after the most recent testosterone injection (in groups 6 and 7). Before and 2 and 4 mo after initiation of hormone administration, blood was collected in evacuated tubes containing the calcium-chelator EDTA, which prevents not only coagulation but possibly also fatty acid oxidation. All blood was collected after a 12-h fast, immediately placed on ice, centrifuged at 3500 x g for 30 min at 4 °C, and stored within 1 h at –80 °C until analysis. The fatty acid composition was measured in plasma cholesteryl esters as previously described (30, 31), because n–3 fatty acids in plasma cholesteryl esters are highly correlated with dietary intake (37) and with the proportion found in skeletal muscle phospholipids (38). Plasma concentrations of HDL cholesterol and triacylglycerols were determined by using enzymatic colorimetric methods. Serum 17ß-estradiol and testosterone concentrations were determined by using standardized radioimmunoassays, and luteinizing hormone and follicle-stimulating hormone concentrations were determined by using immunometric luminescence assays.
Statistical analysis
For the comparison between the men, the women not taking OACs, and the women taking OACs, an analysis of covariance (ANCOVA) for independent samples was used (with age and dietary experiment as covariates). Because the women taking OACs were significantly younger than the women not taking OACs, age was also a covariate. Post hoc paired comparisons were made by using a Sidak test to identify significant differences between groups. Data are presented as estimated marginal means (adjusted for age and dietary experiment) with 95% CIs (Table 3).
View this table:
TABLE 3. Percentages by weight of fatty acids in cholesteryl esters in fasting serum samples from healthy men and women who consumed a controlled experimental diet for 3 wk1
For groups 1, 2, 5, and 6, who received intervention with sex steroids, a two-factor repeated-measures analysis of variance with interaction was used to compare effects between the randomly assigned groups over time (group 1 compared with group 2, and group 5 compared with group 6). Post hoc comparisons were made by using a Sidak test to identify significant differences between time points. Data are presented as means ± SDs or SEMs or, for right-skewed data, as geometric means with 10th–90th percentiles in parentheses. For groups 3, 4, 7, and 8, results are reported by group, and repeated-measures analysis of variance or Student's t test for paired samples was used to analyze the effects of hormone administration over time, as appropriate. Groups 1–4 and groups 5–8 were compared at 4 mo in separate ANCOVA analyses (with baseline variables as covariates). A post hoc Sidak test was used if the ANCOVA was significant. Proportional changes were correlated by using the Pearson's correlation coefficient. A two-tailed P value < 0.05 was considered statistically significant. The software used was SPSS 10.0 (SPSS Inc, Chicago).
RESULTS
Comparison between men and women
The 72 men had a mean (±SD) age of 29.6 ± 12.9 y (range: 18–67 y), and the 103 women had a mean age of 27.4 ± 10.2 y (range: 18–67 y) (P = 0.20 for sex difference). DHA concentrations were 14.6 ± 3.9% (
Fifty-six MF transsexual subjects were treated with oral ethinyl estradiol plus cyproterone acetate for 4 mo [group 1 (n = 15)], transdermal 17ß-estradiol plus cyproterone acetate for 4 mo [group 2 (n = 15)], cyproterone acetate alone for 4 mo [group 3 (n = 10)], and oral ethinyl estradiol plus cyproterone acetate for 12 mo [group 4 (n = 16)] (Table 4). After estrogen administration, circulating sex hormones were profoundly altered (groups 1–4, Table 4). HDL-cholesterol and triacylglycerol concentrations increased after oral administration of estrogens, and HDL-cholesterol concentrations decreased after transdermal estrogen administration (P < 0.05 for interaction between groups 1 and 2; Table 4), which indicated different effects on hepatic lipid metabolism (39, 40). In M
View this table:
TABLE 4. Hormone and lipid concentrations, percentages by weight of n–3 fatty acids in EDTA plasma cholesteryl esters, and energy and fat intakes in 30 male-to-female (MF) transsexual subjects randomly assigned to receive ethinyl estradiol plus cyproterone acetate (CA) or 17ß-estradiol plus CA (n = 15 in each group), in 10 MF transsexual subjects treated with CA, and in 16 MF transsexual subjects treated with ethinyl estradiol plus CA for 12 mo1
No further changes occurred between 4 and 12 mo (group 4) during oral administration of ethinyl estradiol [–6.3 ± 4.4% (
Sixty-one FM transsexual subjects were randomly assigned to receive intramuscular testosterone esters plus anastrozole [group 5 (n = 16)] or intramuscular testosterone esters plus placebo [group 6 (n = 14)] or were treated with intramuscular testosterone esters for 4 [group 7 (n = 17)] or 12 [group 8 (n = 14)] mo (Table 5). In the FM transsexual subjects who were already being treated with testosterone, anastrozole (group 5) decreased 17ß-estradiol concentrations and increased luteinizing hormone and follicle-stimulating hormone concentrations in comparison with placebo (group 6) (all P <0.005 for interaction; Table 5). Anastrozole had no significant effect on testosterone concentrations (group 5 compared with group 6; Table 5). DHA concentrations decreased 9.4%, though not significantly (P = 0.089 for interaction), when anastrozole (group 5) in comparison with placebo (group 6) was added to the testosterone regimen [–5.7 ± 6.6% (
View this table:
TABLE 5. Hormone concentrations, percentages by weight of n–3 fatty acids in EDTA plasma cholesteryl esters, and energy and fat intakes in 30 female-to-male (FM) transsexual subjects who were already receiving testosterone treatment and who were randomly assigned to receive anastrozole (n = 16) or placebo (n = 14) and in 17 and 14 FM transsexual subjects starting treatment with intramuscular testosterone esters for 4 and 12 mo, respectively1
After 4 mo of testosterone administration, circulating testosterone concentrations increased, whereas mean 17ß-estradiol, luteinizing hormone, and follicle-stimulating hormone concentrations decreased significantly (groups 7 and 8 combined; all P < 0.05). In the 31 F
DISCUSSION
We found that DHA concentrations in plasma cholesteryl esters were higher in the women than in the men and that this difference was independent of dietary differences. The difference in DHA is equal to what can be achieved by an extra intake of 62 mg DHA/d (31), which is provided by 1 regular fish-oil capsule once every other day or 1–2 fatty fish meals/mo. Such differences may be relevant with respect to pregnancy, which is associated with high circulating concentrations of both estrogens and DHA, and the fetus depends on the DHA supply from the pregnant mother (1-4). Furthermore, the inclusion of fatty fish meals in the diet is recommended for the prevention of cardiovascular disease (11-17).
Previous studies suggesting sex differences in n–3 fatty acid status are difficult to interpret because the diets were not kept the same for men and women; thus, any observed differences in n–3 fatty acid status may have reflected differences in dietary intake (19, 41). However, the sex difference in DHA observed in our study was not caused by sex differences in dietary intake, because subjects received the same controlled diets and maintained their body weight. The men and the women, therefore, consumed comparable amounts of n–3 HUFAs within the same meal context. Notably, we also found differences between the sexes and between the women who were or were not taking OACs in the proportions of several saturated and monounsaturated fatty acids. These findings are in line with those from previous intervention studies in (oophorectomized) postmenopausal women that showed that several estrogenic compounds decreased concentrations of stearic acid and oleic acid and increased palmitic acid in serum lecithin or cholesteryl esters (23-29).
Treatment with oral ethinyl estradiol, but not with transdermal 17ß-estradiol, increased DHA concentrations. This supports the finding that women of reproductive age seem to have a greater capacity to convert ALA to DHA than do men (9, 18). Conversely, testosterone administration decreased DHA. This may be an effect of testosterone itself or of the decrease in plasma 17ß-estradiol associated with the administration of testosterone. Our findings support the idea that estrogen is the responsible hormone. This is in line with the positive correlation between the proportional changes in plasma DHA and serum 17ß-estradiol and with our finding that the aromatase inhibitor anastrozole—which blocks the conversion of androgens to estrogens and decreases 17ß-estradiol concentrations—further decreased DHA, although the decrease was of borderline significance. Moreover, in the group treated with transdermal 17ß-estradiol, plasma testosterone decreased to almost nil, whereas no effect of this reduction in testosterone was observed on DHA.
Endogenous synthesis of DHA from ALA via EPA requires elongases and desaturases, and isotope studies in adults (6, 9, 18, 42, 43) and infants (44) show that humans can indeed convert ALA to DHA in vivo, predominantly in the liver (but also in the lung, heart, and skeletal muscle) (10). These data also suggest that the rate of conversion of ALA to longer chain n–3 HUFAs is too low to affect health, yet most of these studies were done in men. We found that oral ethinyl estradiol administration increased DHA, whereas administration of transdermal 17ß-estradiol had no effect on DHA. The hepatic effect of synthetic ethinyl estradiol—because of its first pass through the liver (45) and compound-specific effects (46)—may be greater than that of transdermally absorbed 17ß-estradiol (47). Therefore, we speculate that hepatic synthesis of DHA represents the major source of the increase in DHA in women compared with men (10). The initial decrease in EPA, an intermediate in DHA synthesis, also suggests an increased conversion of EPA to DHA after initiation of ethinyl estradiol treatment. The alternative explanation, ie, that the increase in DHA is the result of a specific decrease in DHA clearance, which leads to high DHA concentrations in estrogenic milieus, seems less likely. From an evolutionary perspective, it seems plausible that an increase in endogenous maternal DHA biosynthesis during pregnancy and lactation enabled optimal fetal and neonatal growth and brain development (1-3). Vegetarian and non-fish-eating mothers may depend totally on this biosynthetic pathway to acquire DHA from ALA (1, 2).
The interpretation of the effects of cross-sex hormone administration is limited by the inclusion of a relatively small number of subjects, the open-label design, and the lack of a true placebo group, which was due to the nature of the study population and the treatment indication. Therefore, we do not know if the effects of sex steroids are similar to those found in eugonadal subjects who take sex hormones appropriate to their biological sex. Yet, the effects of cross-sex hormone administration were strong and consistent with the sex difference found in the healthy male and female volunteers.
In summary, our data suggest that biosynthesis of DHA is greater in women than in men because of the effects of sex hormones, presumably estrogens. A strong stimulus with estrogens induced an increase in DHA status, whereas a testosterone stimulus induced a decrease in DHA. Epidemiologic studies pointed toward a protective effect of 1–2 weekly servings of fish, especially fatty fish (11-15), against ischemic heart disease and sudden death. These protective effects were confirmed in the Diet and Reinfarction Trial (DART; 16) and the GISSI (Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto miocardico)-Prevenzione trial (17) in coronary patients who received either dietary advice to consume fish or fish-oil capsules, respectively. Because DHA status is lower in men, who also happen to have a higher risk of heart disease, dietary guidelines may need to recommend higher fish consumption for men than for women (48, 49). Moreover, these findings also suggest that maternal synthesis of DHA is under estrogenic control, which may contribute to the physiologic increase in maternal DHA concentrations during pregnancy (1, 2, 4). Placental transfer of DHA may subsequently provide the fetus with adequate amounts of DHA (1-3).
ACKNOWLEDGMENTS
EJG and LJGG participated in the planning, design, and execution of the intervention studies in transsexual subjects, and AWFTT participated in the planning, design, and execution of the study on anastrozole. MBK and PLZ participated in the design and execution of the cross-sectional part of the study in which men and women were compared. All authors participated in the overall analysis and the writing of the paper. None of the authors had any conflicts of interest.
REFERENCES