| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Departments of Medicine (A.J.O., A.M., M.A.B.) and Renal Medicine (A.M., M.A.B.) University of New South Wales, St. George Hospital, Sydney 2217, Australia
Address all correspondence and requests for reprints to: Dr. Anthony J. O’Sullivan, Department of Medicine, St. George Hospital, Kogarah, NSW, 2217, Australia. E-mail: A.OSullivan{at}unsw.edu.au
Abstract
There is a sexual dimorphism in body fat in humans. Adipose tissue increases with puberty and early pregnancy in women, suggesting gonadal steroids can influence body fat. Previously, we have observed that oral estrogen, compared with transdermal estrogen, reduced postprandial lipid oxidation and increased body fat, possibly due to suppressed hepatic lipid oxidation. If estrogen effects lipid oxidation, we predicted that subjects with significantly different endogenous estrogen production would oxidize lipids at different rates. The aim of this study was to compare energy metabolism in 12 pregnant (19 wk gestation, 29 ± 1 yr, 1.66 ± 0.02 m, 73.5 ± 2.4 kg), 11 nonpregnant premenopausal (29 ± 2 yr, 1.68 ± 0.02 m, 63.1 ± 1.8 kg), and 28 postmenopausal (58 ± 1 yr, 1.62 ± 0.01 m, 69.9 ± 1.0 kg) women who were not receiving estrogen, and to relate these findings to endogenous estrogen concentrations. All women underwent indirect calorimetry under identical situations in the basal and postprandial state following a standard mixed meal. Basal (5998 ± 184 vs. 5712 ± 184 vs. 5800 ± 121 kJ·24 h, respectively) and postprandial energy expenditure (7172 ± 239 vs. 6964 ± 210 vs. 6955 ± 147 kJ·24 h) was similar among groups. However, basal lipid oxidation was reduced in pregnant (45.3 ± 6.1 mg/min, P < 0.05) and nonpregnant women (44.5 ± 6.3 mg/min, P < 0.05) compared with postmenopausal women (58.4 ± 2.9 mg/min). Postprandial lipid oxidation differed among groups, being least in pregnant women (8.8 ± 6.2 mg/min) compared with nonpregnant (28.9 ± 6.4 mg/min, P < 0.04) and postmenopausal (48.1 ± 4.0 mg/min, P = 0.0001) women. There was a significant reciprocal increase in postprandial carbohydrate oxidation. Mean postprandial glucose levels were slightly but nonsignificantly higher in pregnant women. Insulin levels were significantly higher in postmenopausal compared nonpregnant, but not pregnant, women. In a multiple regression analysis, serum estradiol (log transformed) correlated negatively with postprandial lipid oxidation (r = -0.66, P = 0.0001) and positively with postprandial nonesterified free fatty acid levels, whereas no correlation was found with postprandial insulin, glucose, fat free mass, and fat mass. In summary, postprandial lipid oxidation is reduced in pregnancy compared with that in healthy nonpregnant women, who in turn have lower postprandial lipid oxidation than postmenopausal women. This implies that the premenopausal years and early pregnancy are states of efficient fat storage, possibly mediated through reduced lipid oxidation due to estrogen, therefore increasing body fat for reproduction, thus supporting the notion that fat mass can be regulated.
THERE IS A sexual dimorphism in body composition in humans. It is well established that females have a smaller lean body mass compartment but a larger fat body mass compartment than males (1), a difference that appears at puberty and is then maintained throughout adult life (2, 3, 4). The difference in subcutaneous fat measurement between boys and girls, which correlates with percent body fat (1), is within 1–2 mm up to the age of 10 yr (3). From age 10 yr, subcutaneous fat thickness increases quicker in girls than boys, finishing 44–93% greater by 16.5 yr, depending on the site of measurement (3). Although it is thought that gonadal steroids account for the greater degree of body fatness in females the mechanism is not known (5). Early pregnancy is another state where fat mass is increased (6), a desirable change required for late pregnancy fetal development and future lactation (5).
Body adipose tissue stores are determined by a balance between energy intake and energy expenditure (5, 7, 8), although individual variation may also be determined by genetics (9). If increased energy intake were to account for the increase in fat mass in females at puberty and in early pregnancy, then an increase in total caloric intake or fat intake would be expected. However, puberty is associated with a reduction in caloric intake per kg body weight (10) that persists into adult life (2), and early pregnancy is not associated with a major increase in caloric intake (6). Percent fat intake remains unchanged throughout puberty and adult life in females (2, 10). A reduction in energy expenditure may contribute to fat gain with aging (7), however, it is unlikely to contribute to fat gain in puberty or early pregnancy because these states are also associated with increases in fat free mass (2, 6), which is associated with greater basal energy expenditure (7).
A reduction in fat (lipid) oxidation is a possible means of adipose tissue gain (8), a mechanism that has been proposed to explain the increased adiposity associated with moderate ethanol intake (11). Females are reported to have reduced postprandial free fatty acid flux compared with males, suggesting that female sex steroids can influence fatty acid metabolism (12). Previously, we have observed that oral estrogen reduced postprandial lipid oxidation compared with transdermal estrogen in postmenopausal women (13). We hypothesized this was due to a suppression of hepatic lipid oxidation due to a first pass effect of oral estrogen. Oral estrogen was also associated with a small but significant increase in body fat. Against this background, we hypothesized that subjects with significantly different endogenous estrogen production would oxidize lipids at different rates as a means of finely adjusting body fat mass for biological advantages. Hence, we have assessed lipid oxidation in premenopausal women, postmenopausal women not on hormone replacement therapy, and pregnant women at approximately 19 wk gestation and related these findings to endogenous serum estradiol (E2) concentrations.
Materials and Methods
Subjects
Subjects were recruited from the general population (Table 1
). All were in good health and not
taking medications. Premenopausal women were studied in the follicular
stage of the menstrual cycle, and no subject was taking oral
contraceptive steroids. Postmenopausal women were not on hormone
replacement therapy. The pregnant women were studied at 19 ± 1 wk
gestation, and all were singleton pregnancies. All pregnant women
underwent a glucose tolerance test during pregnancy, and women with
gestational diabetes mellitus were excluded. Subjects with obesity were
not excluded. None of the 12 nonpregnant subjects, 1 of the 11 pregnant
subjects, and 5 of the 28 postmenopausal subjects had a body mass index
(BMI) greater than 30. The study was approved by the Research Ethics
Committee of South Eastern Sydney Area Health Service and the Committee
on Experimental Procedures Involving Human Subjects of the University
of New South Wales. All subjects gave written informed consent.
|
A cross-sectional study was performed. All subjects underwent measurement of energy expenditure and substrate oxidation in the basal and postprandial states, assessment of body composition, and blood sampling. Subjects attended our Clinical Research Room at 0800 h after a 12-h overnight fast. After height and weight were measured, a small plastic iv cannula was inserted under local anesthetic for blood withdrawal. Energy expenditure and substrate oxidation were measured by indirect calorimetry, during the fasting and postprandial states, using the Deltatrac Metabolic Monitor (Datex Instrumentarium Corp., Helsinki, Finland) as described previously (13). Subjects were then rested comfortably in the recumbent position for at least 40 min in preparation for measurement of basal energy expenditure and substrate oxidation, which was performed over the next 30 min. A standardized mixed meal (Ensure, Abbott Laboratories, Sydney, Australia; 14.0% protein, 31.5% fat, and 54.5% carbohydrate) was then administered over 5 min, and indirect calorimetry was performed during the 30- to 60- and 90- to 120-min time intervals after the nutrient load. The mixed meal consisted of the caloric equivalent of 40.0% of the subject’s measured basal energy expenditure. Separate urine samples were collected during the basal and meal periods, from which urine urea nitrogen excretion was measured. Energy expenditure and substrate oxidation were estimated from the following equations (14):
energy expenditure = 3.91 VO2 + 1.10 VCO2 - 3.34 Nu
carbohydrate oxidation = 4.55 VCO2 - 3.21 VO2 - 2.87 Nu
lipid oxidation = 1.67 VO2 - 1.67 VCO2 - 1.92 Nu
Carbohydrate and lipid oxidation is expressed as grams per minute. VO2 represents oxygen consumption, and VCO2 represents carbon dioxide production in liters per minute. Nu represents urinary urea nitrogen excretion in grams per minute. The monitor was calibrated against standardized gases before each study. The coefficient of variation (CV) from five subjects studied on three separate occasions for energy expenditure was 4.2%, and substrate oxidation was 4%.
Body composition was measured by dual-energy x-ray absorptiometry (DEXA; Lunar DPX, Lunar Corp., Madison, WI) in the postmenopausal women (15) and by bioimpedance analysis in the premenopausal women and the pregnant women (16). DEXA provides a threecompartment model allowing total body bone mineral content, lean and fat body mass to be quantified for the whole body. To calculate fat free mass, total body bone mineral content and lean body mass were added (15). In our laboratory, the CV from 10 subjects scanned four times was 1.5% for bone mineral density, 2.9% for fat mass, and 1.4% for fat free mass. Bioimpedance analysis was performed using the Bodystat 1500 (Bodystat Ltd., Douglas, Isle of Man, UK). Bioimpedance analysis provides a two-compartment model derived from the resistance to flow of a single electric current (50 kHz) through body fluids (16). In our laboratory, the CV derived from 10 subjects studied twice within 10 min was 0.5% for fat mass and 0.2% for fat free mass, and the CV derived from 10 subjects studied twice within 4 wk was 4.0% for fat mass and 1.6% for fat free mass. DEXA scans and bioimpedance analysis was performed on the same day in 18 women, 3–5 d postpartum. Both techniques correlated well for fat mass (r2 = 0.92, P = 0.0001), however, bioimpedance analysis overestimated fat mass by 2.1 kg compared with DEXA (32.7 ± 2.5 kg vs. 30.6 ± 2.4 kg, P < 0.05). Body composition was expressed as absolute fat free and fat mass in kg, and percent body fat.
Blood samples were taken for fasting and postprandial glucose levels every 30 min and insulin and nonesterified free fatty acids (NEFAs) levels every 60 min whereas E2, cholesterol, and triglyceride levels were only performed on the basal sample.
Laboratory analysis
Insulin was measured by double-antibody RIA as described previously (17). Glucose concentrations were analyzed on a glucose analyzer (model 23AM; Yellow Springs Instrument Co., Yellow Springs, OH). Serum triglyceride levels were measured with an automated enzymatic method. Urinary urea nitrogen was measured by an enzymatic UV method (Roche Molecular Biochemicals, Mannheim, Germany). NEFAs were determined by an acyl CoA oxidase-based colorimetric method (Wako, Osaka, Japan). E2 was measured by a double-antibody RIA (Clinical Assays Estradiol-2; Sorin Biomedica, Saluggia, Italy) with 125I-E2 as the tracer. The interassay and intra-assay CV were 12% and 8%, respectively.
Statistical analysis
Statistical analysis was performed with the aid of the SAS statistical package (SAS/Stat 1990; SAS Institute, Inc.Cary, NC) using an ANOVA model. Relationships between variables were determined by multiple regression analysis. E2 levels were measured on a fasting serum sample and were log-transformed for statistical analysis. Results are expressed as mean ± SE.
Results
Demographic data and body composition (Table 1
)
The postmenopausal women were significantly shorter than the premenopausal women whereas the premenopausal women weighed less than the pregnant and postmenopausal women. The postmenopausal women had significantly less fat free mass than premenopausal or pregnant women, whereas the premenopausal women had less body fat. Therefore, percent body fat was lowest in the premenopausal women compared with the pregnant women, who had in turn had lower percent body fat than the postmenopausal women.
Energy metabolism
Basal and postprandial energy expenditure (diet-induced
thermogenesis) was similar among the three groups (Fig. 1). Basal lipid oxidation was
significantly higher in the postmenopausal women compared with the
premenopausal and pregnant women (Fig. 2). Postprandial lipid oxidation was
significantly different between the three groups with the lowest level
in pregnant women, intermediate in premenopausal women, and highest in
postmenopausal women. These significant differences in postprandial
lipid oxidation persisted when expressed per kilogram body weight or
fat free mass.
|
|
). Postprandial carbohydrate oxidation
seemed higher in the pregnant women compared with the premenopausal
women, however, this did not reach statistical significance
(P = 0.1). Protein oxidation derived from urine urea
nitrogen excretion was not significantly different among pregnant
(33.1 ± 4.0 mg/min), premenopausal (36.6 ± 3.5 mg/min), and
postmenopausal (48.2 ± 5.4 mg/min) women.
|
)Glucose and basal insulin levels were similar among the three subject groups, however, postmenopausal women had significantly higher postprandial insulin levels than premenopausal women. Compared with both other groups, basal NEFA levels were highest in postmenopausal women, but postprandial NEFA levels were lowest. Serum E2 levels differed markedly, as expected, among the three groups. Basal insulin sensitivity derived from basal glucose and insulin levels using the homeostasis model assessment (HOMA) model (18) was similar among pregnant (1.4 ± 0.2, no units), premenopausal (1.2 ± 0.2), and postmenopausal (1.5 ± 0.2) women.
|
In a multiple regression analysis performed using subjects from
all three groups, fat free mass (r = 0.55, P =
0.0002) and fat mass (r = 0.37, P = 0.02) were
significantly and positively related to basal energy expenditure, as
expected, but not to basal insulin, glucose, NEFA, and E2 levels.
Postprandial energy expenditure was significantly related only to fat
free mass (r = 0.76, P = 0.0001). Basal lipid
oxidation was significantly related to basal NEFA levels (r =
0.62, P = 0.0001) whereas basal insulin, glucose,
triglyceride, and E2 levels where not related. Postprandial lipid
oxidation was significantly and negatively related to serum E2 levels
(r = -0.66, P = 0.0001) and positively related to
postprandial NEFA levels (P = 0.03), but not to
postprandial insulin, glucose, fat free mass, and fat mass. Because
pregnancy is an extreme state as regards E2 levels, we analyzed our
data with the pregnant subjects excluded, and postprandial lipid
oxidation remained significantly and negatively related to serum E2
levels (r = -0.39, P = 0.01) as shown in Fig. 4. Basal carbohydrate oxidation was
negatively related to basal NEFA levels (r = -0.46,
P = 0.002) only while postprandial carbohydrate
oxidation was related to fat free mass (r = 0.48,
P = 0.0004) and E2 levels (r = 0.65,
P = 0.0001).
|
This cross-sectional study has shown that energy metabolism is significantly different in women according to their endogenous estrogen status. Whereas basal and postprandial energy expenditure was similar among groups, substrate oxidation of an identical meal was significantly different. Postprandial lipid oxidation was lowest in the pregnant women (highest E2 levels), intermediate in nonpregnant premenopausal women, and highest in postmenopausal women (lowest E2 levels). No significant difference in glucose tolerance was detected between the groups, however, postprandial insulin levels were highest in the postmenopausal women, probably because of increased body fat. Mean postprandial NEFA levels were significantly lower in the postmenopausal women compared with the pregnant women, and the reason is not clear, but it is unlikely to relate solely to differing insulin sensitivity because all groups had similar glucose tolerance and HOMA values. We and others have shown that postprandial free fatty acid metabolism is affected by sex steroids independent of effects on insulin action following meal ingestion (12) and during hyperinsulinemic clamping (17). Glucose, insulin, and body fat were not significantly related to lipid oxidation in a regression analysis, and an indirect measure of insulin sensitivity (HOMA) was not different among the groups. Taken together, these findings suggest that the alteration in postprandial lipid oxidation in pregnancy is unlikely to be due to a perturbation in insulin action.
We hypothesize that estrogen reduces postprandial lipid oxidation at puberty and in early pregnancy to facilitate efficient fat storage in preparation for fertility, birth, and lactation, respectively (5). This modification in lipid oxidation would enable fat storage to occur without needing significant changes in dietary fat and caloric intake (2, 6, 10). Kopp-Hoolihan et al. (6) followed women from before conception throughout pregnancy. At 24 wk gestation women had gained 3.9 kg of body fat without any change in energy intake. These authors estimated that the energy deposited as fat by the mother during pregnancy equates to approximately 130 MJ. The pregnant women in our study oxidized approximately 20.1 mg/min less lipid in the postprandial period than did premenopausal women. If the postprandial state is considered to be 180 min long and the subjects are eating three meals per day, this change would equate to a saving of 10.9 g of fat per day or 392 KJ. If these changes in lipid oxidation were to persist throughout a 40-wk pregnancy this would lead to a storage of approximately 110 MJ of fat without requiring a significant change in dietary fat intake. In support of this notion, we have previously reported that oral estrogen replacement, when compared with the transdermal route, resulted in a greater suppression of postprandial lipid oxidation (13). Moreover, oral estrogen therapy lead to a significant increase in fat mass over a 6-month period, which we speculated arose from chronic suppression of lipid oxidation. A similar mechanism (reduced lipid oxidation) has been proposed for the increased adiposity associated with moderate ethanol intake (11). Hence, one mechanism by which females may regulate body fat is through altering lipid (fatty acid oxidation).
If estrogen does influence body fat stores throughout female adult life, altered body fat mass should be seen in the postmenopausal years. Consistent with our hypothesis are the findings that fat mass (19) and BMI (2) fall in the decades postmenopause. In contrast, the early postmenopausal years are related to fat mass gain, but this has been accounted for by reduced physical activity (20), findings reported in a prospective study following women across the menopause. If estrogen were to facilitate efficient fat storage through reduced lipid oxidation this may in part account for the increased fat mass observed in adult women compared with men when their diets (in respect to fat content) are proportionally similar (2). Supporting this notion are observations that male-to-female transsexuals gain body fat when commenced on estrogen therapy (21).
A relationship between E2 and reduced postprandial lipid oxidation, as found in our study, is supported by a number of studies that have shown that estrogen does influence postprandial lipid metabolism. Jensen (12) reports that postprandial free fatty acid flux is reduced in premenopausal women compared with men, possibly related to gender differences in subcutaneous adipose tissue whereby male distribution fat tissue is less responsive to postprandial suppression of free fatty acid release. Oral estrogen replacement to postmenopausal women has been reported to alter postprandial elimination of lipoprotein remnants (22). We have also reported that oral estrogen replacement has significant effects on postprandial lipid metabolism in postmenopausal women (13). When compared with the transdermal route, oral estrogen therapy resulted in a greater transient suppression of postprandial lipid oxidation. Together, these studies are consistent with the findings in the present study that estrogen can have significant effects on postprandial lipid metabolism.
The mechanism by which oral estrogen may reduce postprandial lipid oxidation is not known. The mechanism may be due to effects on estrogen-sensitive hepatic proteins or it may involve direct inhibitory effects of estrogen on the liver, a major site of fatty acid metabolism. The latter proposal is supported by in vitro studies showing that pharmacological concentrations of estrogen reduce ketogenesis (a product of fatty acid oxidation) and increase fatty acid incorporation into triglycerides (23, 24). Because intrahepatic fatty acids are partitioned between oxidative and nonoxidative (fatty acid incorporation into triglycerides) pathways, estrogen may regulate the metabolic fate of intrahepatic free fatty acids by directing fatty acids away from oxidative into lipogenic pathways. Supporting this mechanism are clinical observations that oral (where lipid oxidation is reduced), but not transdermal, estrogen therapy stimulates hepatic triglyceride synthesis and increases triglyceride level (25) and reports that triglyceride levels increase during pregnancy (26), where lipid oxidation is also reduced. This potential mechanism is supported by the finding of a strong inverse relationship between lipid oxidation and triglyceride levels during treatment with various estrogen doses (27). Thus, pregnancy and oral E therapy in postmenopausal women are situations both associated with reduced postprandial lipid oxidation and increased triglyceride levels.
Another possible mechanism for estrogen to influence body fat is
through direct effects on adipose tissue as estrogen receptors are
present in adipocytes. The estrogen receptor-
is associated with BMI
variation in healthy women (28). Estrogen also has a major
influence on body fat distribution and adipocytes from different
regions respond differently to lipolytic agents (29).
Estrogen may also have indirect metabolic effects because estrogen has
recently been reported to attenuate GH signaling (30) and
GH is reported to increase lipid oxidation (31) and reduce
body fat mass (32).
Further support that body fat mass is regulated is provided by the discovery of leptin (5, 8, 33). Leptin, the ob gene product secreted by adipocytes, at levels that parallel the amount of body fat may be involved in a negative feedback loop that helps control fat accumulation (34). If puberty and early pregnancy were states of efficient fat storage, according to our hypothesis, leptin levels should rise in these states. Leptin levels/kilogram fat mass are similar in prepubertal males and females (35), but by late puberty females have significantly higher leptin levels/kilogram fat mass than males (35), a difference that exists throughout adult life (5). Leptin levels rise again during pregnancy and return to normal postpartum (36). Although the changes in leptin levels are consistent with our hypothesis, the interaction between sex steroids, leptin, body composition, and energy homeostasis is complex and requires further study.
In summary, the present study shows, for the first time, that postprandial substrate oxidation is significantly different in females according to their endogenous estrogen status. Because pregnancy is associated with reduced postprandial lipid oxidation, this should facilitate efficient fat storage without the need for significant dietary changes. The storage of fat in preparation for fertility, fetal development, and postpartum lactation has obvious biological advantages and supports the notion that fat mass can be regulated. Our findings need to be confirmed with longitudinal studies, although following women throughout these life stages is obviously difficult. If postprandial lipid oxidation in humans can be regulated, pharmacological agents may also be able to influence postprandial lipid oxidation and could, therefore, be considered for obesity management.
Acknowledgments
We are indebted to Prof. Ken Ho from the Garvan Institute of Medical Research for his input into these studies, and to Jane Lawson and Leonie Crampton for help with the clinical studies.
Footnotes
This work was supported in part by the National Health and Medical Research Council of Australia and in part by the Ramaciotti Foundation and the Sylvia and Charles Viertel Charitable Foundation in Australia.
Abbreviations: BMI, body mass index; CV, coefficient(s) of variation; DEXA, dual-energy x-ray absorptiometry; E2, estradiol; HOMA, homeostasis model assessment; NEFA, nonesterified free fatty acid.
Received March 21, 2001.
Accepted July 9, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  J. D. Veldhuis, J. N. Roemmich, E. J. Richmond, A. D. Rogol, J. C. Lovejoy, M. Sheffield-Moore, N. Mauras, and C. Y. Bowers Endocrine Control of Body Composition in Infancy, Childhood, and Puberty Endocr. Rev., February 1, 2005; 26(1): 114 - 146. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.-C. Leung, G. Johannsson, G. M. Leong, and K. K. Y. Ho Estrogen Regulation of Growth Hormone Action Endocr. Rev., October 1, 2004; 25(5): 693 - 721. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. F. Gevers, N. Loveridge, and I. C. A. F. Robinson Bone Marrow Adipocytes: A Neglected Target Tissue for Growth Hormone Endocrinology, October 1, 2002; 143(10): 4065 - 4073. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
,
P < 0.04 premenopausal vs. pregnant
and postmenopausal.
