- Email: [email protected]
Serum leptin concentration in women: effect of age, obesity, and estrogen administration
FERTILITY AND STERILITYt VOL. 70, NO. 3, SEPTEMBER 1998 Copyright ©1998 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.
Serum leptin concentration in women: effect of age, obesity, and estrogen administration V. Daniel Castracane, Ph.D.,* Robert R. Kraemer, Ed.D.,† Mary Ann Franken, M.D.,* Ginger R. Kraemer,‡ and Terry Gimpel* Texas Tech University Health Sciences Center, Amarillo, Texas; Southeastern Louisiana University, Hammond, Louisiana; and North Oaks Obstetrics and Gynecology, Hammond, Louisiana
Objective: To compare serum leptin levels in normally cycling reproductive females (20 –35 years old) with those in age-matched males, in women who were receiving oral contraceptives, and in older (postmenopausal) women (50 – 65 years old) who were or who were not receiving hormone replacement therapy. Design: Case-control study. Setting: Obstetrics and Gynecology Clinic, Texas Tech University Health Sciences Center-Amarillo, or the Exercise Physiology Laboratory at Southeastern Louisiana University. Patient(s): Normally cycling women between the ages of 20 –35 years and age-matched controls who were receiving oral contraceptives. Postmenopausal women between the ages of 50 – 65 years who were or who were not receiving hormone replacement therapy. Received September 10, 1997; revised and accepted April 28, 1998. Supported in part by a Faculty Development grant from Southeastern Louisiana University, Hammond, Louisiana. Presented at the 44th Annual Society for Gynecologic Investigation meeting in San Diego, California, March 19 –22, 1997. Reprint requests: V. Daniel Castracane, Ph.D., Department of Obstetrics and Gynecology, Texas Tech University Health Sciences Center, 1400 Wallace Boulevard, Amarillo, Texas 79106 (FAX: 806-354-5516; Email: [email protected] .ttuhsc.edu). * Department of Obstetrics and Gynecology, Texas Tech University Health Sciences Center. † Department of Kinesiology and Health Studies, Southeastern Louisiana University. ‡ North Oaks Obstetrics and Gynecology. 0015-0282/98/$19.00 PII S0015-0282(98)00187-3
472
Main Outcome Measure(s): Serum leptin concentration. Result(s): In all groups, serum leptin concentrations were correlated significantly with body mass index. Leptin levels were significantly higher in young women than young men (P ,.001), but no other statistically significant differences were found for the other three comparisons. Conclusion(s): Serum leptin concentrations expressed as a measure of adiposity (body mass index) are greater in young normally cycling females (20 –35 years old) than in age-matched males. There is no difference in levels of serum leptin between young and postmenopausal (50 – 65 years old) women. Estrogen administration, either in young women who are receiving estrogen-progestin oral contraceptives or in postmenopausal women who are receiving hormone replacement therapy, does not effect serum leptin concentrations. (Fertil Sterilt 1998;70:472–7. ©1998 by American Society for Reproductive Medicine.) Key Words: Leptin, adiposity, estrogen, menopause, oral contraceptives, hormone replacement therapy, aging, males
Numerous studies have demonstrated the absence of leptin in the genetically (ob/ob) obese mouse and the reversal of this obese condition with leptin administration, which also corrects the sterility associated with this genetic defect (1–3). The role of leptin in the human remains unresolved. Many studies now have demonstrated that circulating levels of leptin in the human are related directly to measures of obesity such as body mass index (BMI) or percent body fat (4 – 6). These reports generally span a wide range with respect to age, although gender differences have been reported, with serum leptin levels always reported to be higher in females than males (5–7). In the present case-control study, we investigated whether any differences exist between
leptin levels in normally cycling young women (20 –35 years of age) and men in the same age group. However, we were more interested in the effect that estrogen might have in females because estrogen receptors have been reported in adipose tissue (8 –10), suggesting that estrogens may be involved in the regulation of leptin production. Therefore, we also compared this population of young women with an agematched group of women who were receiving standard estrogen-progesterone combination oral contraceptives (OCs) as well as older women who were or who were not receiving estrogen replacement therapy (ERT). These data demonstrate that the relationship of serum leptin in women to BMI does not change with age or estrogen administration.
MATERIALS AND METHODS Subjects Serum samples were obtained from 127 female subjects, usually as a fasted morning sample. At the time of collection, information recorded for all subjects included age, height, and weight as well as use of any medications that might interfere with reproductive or endocrine functions, with particular emphasis on estrogen medications, either OCs or hormone replacement therapy (HRT). Subjects were excluded from the study if they were receiving any medications with the exception of OCs and HRT; subjects who received these therapies formed distinct groups in this protocol. Subjects were then divided into a young (20 –35 years of age; mean [6 SD] age, 27.23 6 0.63 years) group of untreated normally cycling female subjects (n 5 58) with cycle length between 25 and 35 days. A similar age-matched group of women (mean [6 SD] age, 26.43 6 1.07 years) who were receiving standard low-dose estrogen-progestin OCs (n 5 23) for at least 3 months was used to compare the effect of estrogen administration in a young population. A third group of subjects aged 50 – 65 years was selected as postmenopausal women (n 5 46); this group was further subdivided into those who were receiving HRT (n 5 28) (mean [6 SD] age 56.5 6 0.92 years) and those who were not receiving HRT (n 5 18) (mean [6 SD] age 55.8 6 0.85 years). Those subjects who were receiving HRT had been taking that regimen for at least 3 months. A small group of age-matched males (n 5 13) (mean [6 SD] age 28.85 6
TABLE 1 Correlation coefficients, r 2, and P for comparing body mass index and leptin concentrations in the different groups. Group
Type of therapy received
Statistic
Young men
Young women
Older women
HRT
No HRT
OC
r r2 P
0.76 0.58 .0024
0.84 0.71 .0001
0.72 0.52 .0001
0.47 0.22 .0116
0.88 0.77 .0001
0.78 0.61 .0001
Note: HRT 5 hormone replacement therapy; OC 5 oral contraceptives.
1.13 years) was obtained to compare with the young group of normal females. All subjects gave their informed, written consent under different protocols approved by the Texas Tech University Health Sciences Center-Amarillo Institutional Review Board and Committee for the Use of Humans and Animals as Research Subjects, Southeastern Louisiana University Institutional Review Board. All serum samples were analyzed for serum leptin.
Assay methods Serum leptin was assayed with use of a second antibody RIA system available from Linco Diagnostics Incorporated (St. Louis, MO). The interassay coefficient of variation (CV)
FIGURE 1 Relationship of serum leptin concentration with BMI for male (n 5 13) and female (n 5 58) subjects between the ages of 20 and 35 years.
FERTILITY & STERILITYt
473
FIGURE 2 Relationship of serum leptin concentration with BMI for females 20 –35 years of age (n 5 58) with normal cycles and no form of contraception and postmenopausal women between the ages of 50 – 65 years of age (n 5 46). Postmenopausal subjects in the graph include both women on HRT and without HRT.
for low and high pools was 9.3% and 4.7%, respectively. The intra-assay CV averaged ,5% for all assays.
Statistics Pearson’s correlation coefficients were determined to assess the relationship between BMI and leptin for subgroups (young men, young women, older women, subjects receiving HRT, subjects not receiving HRT, and subjects receiving OC). Between-group analyses (independent t-tests) were conducted comparing group differences in gender, age, and HRT. Finally, to determine whether BMI affected the differences in leptin concentrations between groups (young males versus young females, young women versus older women, subjects receiving HRT versus those not receiving HRT, and subjects receiving OCs versus those not receiving OCs), four analyses of covariance (ANCOVA) were conducted with BMI as a covariate and leptin as the dependent variable.
RESULTS In all groups, leptin concentrations correlated significantly with BMI (Table 1; Figs. 1– 4). The t-tests indicated that leptin levels were significantly higher in young women versus young men (P 5 .004), older women versus younger women (P 5 .029), subjects who were receiving HRT versus those who were not (P 5 .0409), and subjects who were not receiving OCs versus those who were receiving them (P 5 474
Castracane et al.
.0026; Table 2). Using BMI as a covariate, ANCOVAs revealed that independent of BMI, leptin concentrations were significantly higher in younger women compared with young men (P 5 .0001), but no other statistically significant differences were found for the other three comparisons (Table 2). A post hoc decision was made to conduct power analyses to determine the extent to which we could detect statistically significant differences. We estimated power to detect moderate (0.50) and large (0.80) effect sizes for each analysis using the procedures recommended by Kraemer and Thiemann (1987) (11) for unequal group designs. Power to detect moderate effect sizes ranged from 0.40 to 0.70. To detect large effect sizes, the power of all analyses exceeded 0.70 (2 analyses were greater than 0.95) with the exception of the comparison between those receiving and not receiving HRT, which was 0.70. Acceptable range for power is considered to be 0.70 – 0.90 (12), and all analyses reached this level to detect significant differences for large effect sizes.
DISCUSSION The relationship of serum leptin concentration with some measure of adiposity (e.g., BMI, percent body fat) has been demonstrated in numerous reports (3–5). We have used the same approach to study serum leptin concentration with different physiologic states in women as well as comparison
Serum leptin in women: age and estrogen
Vol. 70, No. 3, September 1998
FIGURE 3 Relationship of serum leptin concentration with BMI for normal cycling females (20 –35 years) (n 5 58) and an age-matched group of women on low-dose oral contraceptives (n 5 23).
with age-matched normal males. We attempted to determine whether estrogen would have an effect on serum leptin concentrations. It has been reported that estrogen receptors are present in adipose tissue (8 –10) and therefore might serve as a mechanism to regulate leptin production. In a recent study of rats and humans it was suggested that serum leptin levels may increase in response to estrogen treatment (13). Comparison with males represents an obvious distinction in which females would have elevated levels of estrogen over males. The other comparisons in this study involve normally cycling reproductive aged women compared with postmenopausal women who were or were not receiving estrogen replacement as well as a younger group who were receiving estrogen-progestin OCs. These groups were studied because of the dynamics of estrogen in these subjects. We have examined the relationship of females to males and with a small group of males have demonstrated, as have other investigators (5–7), that males consistently have lower serum leptin concentrations than do females (Fig. 1). In this study we have confined our subjects to a tighter age range of 20 –35 years than was used in these other studies. Because a large number of reports already have compared males and females, this aspect of our study was not of paramount importance but serves to demonstrate the validity of assay procedures and patient sampling. The factors that regulate leptin levels in males are not clear, although the suggestion that visceral versus subcutaFERTILITY & STERILITYt
neous adipose tissue, the distribution of which is different in males than females, may be a contributing factor (14). Similarly, in hypogonadal males, leptin levels are elevated and return to lower levels after testosterone treatment (15), suggesting that an endocrine involvement may be at the basis of the distinction between males and females. We were interested in the possibility that estrogens might represent a physiologic regulator of leptin production and/or secretion. Therefore, we obtained serum samples from a group of normally cycling younger reproductive aged women, 20 –35 years of age, which served as controls and compared serum leptin concentrations in this group with a group of older postmenopausal women, 50 – 65 years of age. In the original comparison (Fig. 2) we included women in the postmenopausal group, both with and without postmenopausal HRT. This comparison allows us to determine whether aging changes the relationship of serum leptin to BMI. This relationship was not different between the young and older group of women. Ostlund et al. (6) presented data on leptin concentration over a wide age range and suggested that circulating leptin levels are related inversely to age and were reduced by as much as 53% in subjects .60 years. Several factors make these data difficult to interpret. Data are presented not only as means 6 SE, which we think is inappropriate for such data, but in some cases (e.g., age-related data), leptin is plotted as the means 6 SE with apparently mixed male and female values. Because male leptin values are lower than 475
FIGURE 4 Relationship of serum leptin concentration to BMI for postmenopausal subjects who were receiving HRT (n 5 28) and who were not receiving HRT (n 5 18).
females, these data may be represented disproportionately in the lower BMI group. Moreover, it is not clear what the age distribution of the different genders might be. In our study, we have a balanced distribution of adiposity in each group to effectively scrutinize the difference between groups. The study by Kohrt et al. (16) conforms with our conclusion, that there was no effect of HRT on leptin concentration, although data presentation was not designed to evaluate the effects of HRT on leptin.
who were receiving OCs. We have used the same control group and have used an age-matched group of women on estrogen-progestin OCs. The comparison for these two groups shows that the regression lines for both sets of data (Fig. 3) are virtually superimposable. These results demonstrate no effect of OCs on serum leptin concentration compared with younger women and expressed on the basis of adiposity. This is the first study comparing the effect of OCs in normal subjects on leptin concentration.
Another approach to examine the effect of estrogen replacement as a possible physiologic regulator of leptin production and secretion was to study leptin levels in women
As a last comparison, we have subdivided subjects in the postmenopausal group to those receiving HRT and those with no postmenopausal treatment. In this comparison, there
TABLE 2 Body mass index and concentration of leptin in young men, young women, older women, subjects receiving HRT or OCs and subjects not receiving HRT or OCs. Group
Variable Mean (6 SD) BMI Mean (6 SD) leptin concentration (ng/mL) Adjusted mean leptin concentration (ng/mL)
Type of therapy received
Young men
Young women
Older women
HRT
No HRT
OC
28.09 6 3.28
29.19 6 7.94
29.33 6 5.47
30.16 6 4.11
28.01 6 7.00
23.66 6 4.016
9.67 6 5.14
25.12 6 15.20
27.76 6 14.75
30.80 6 12.86
23.04 6 16.57
15.43 6 8.63
11.09
25.24
27.61
29.21
25.51
21.83
Note: HRT 5 hormone replacement therapy; OC 5 oral contraceptives.
476
Castracane et al.
Serum leptin in women: age and estrogen
Vol. 70, No. 3, September 1998
was no statistically significant difference between groups, indicating that standard doses of postmenopausal estrogen replacement have no effect on production and secretory dynamics of leptin from adipose tissue in this age group. The power of this analysis to detect a large effect size was within the acceptable range (0.7– 0.9) (11). Some studies, e.g., Rosenbaum et al (5), have used means 6 SE of serum leptin without correction for BMI or fat mass to compare leptin levels in different groups. However, it is important to remember that the major determinant of leptin production is some measure of fat mass or adiposity and any group that is not balanced for distribution of body fat in its subjects or not corrected for body composition may be biased when expressed as means 6 SE. Our adjusted means clearly demonstrate that without using BMI as a covariate, all groups compared were significantly different; however, with BMI as a covariate, only the male-female comparison was significantly different. In summary, this study has demonstrated that serum leptin levels in young normally cycling females, when expressed on the basis of BMI, is greater than age-matched normal males as has been demonstrated by other investigators. More importantly, with an adequate number of subjects and a well-distributed BMI range, there is no effect of estrogen administration on this serum leptin distribution, either in subjects receiving OCs or HRT. This appears to be the first investigation to have studied OCs and leptin. We believe the advantage in this study is an adequate number of subjects as well as a better BMI distribution in our different populations.
Acknowledgments: We thank Edward Hebert, Ph.D. (Department of Kinesiology and Health Studies, Southeastern Louisiana University, Hammond, LA), for invaluable help in statistical interpretation. We also appreciate the
FERTILITY & STERILITYt
help of many individuals in recruiting subjects, in performing phlebotomy, and in manipulating data. We also thank Debbie Pallmeyer for help in preparing the manuscript.
References 1. Pelleymounter MA, Cullen MJ, Baker MB, Hecht R, Winters D, Boone T, et al. Effects of the obese gene product on body weight regulation in ob/ob mice. Science 1995;269:540 –3. 2. Chehab FF, Lim ME, Ronghua L. Correction of the sterility defect in homozygous obese female mice by treatment with the human recombinant leptin. Nat Genet 1996;12:318 –20. 3. Weigle DS, Bukowsk TR, Foster DC, Holderman S, Kramer JM, Lasser G, et al. Recombinant ob protein reduces feeding and body weight in the ob/ob mouse. J Clin Invest 1995;96:2065–70. 4. Considine RV, Sinha MK, Heiman ML, Kriauciunas A, Stephens TW, Nyce MR, et al. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med 1996;334:292–5. 5. Rosenbaum M, Nicolson M, Hirsch J, Heymsfield SB, Gallagher D, Chu F, et al. Effects of gender, body composition and menopause on plasma concentrations of leptin. J Clin Endocrinol Metab 1996;81: 3424 –7. 6. Ostlund RE Jr, Yang JW, Klein S, Gingerich R. Relation between plasma leptin concentration and body fat, gender, diet, age and metabolic covariates. J Clin Endocrinol Metab 1996;81:3909 –13. 7. Kennedy A, Gettys TW, Watson P, Wallace P, Ganaway E, Pan Q, et al. The metabolic significance of leptin in humans: gender-based differences in relationship to adiposity, insulin sensitivity, and energy expenditure. J Clin Endocrinol Metab 1997;82:1293–300. 8. Wade GN, Gray JM. Cytoplasmic 17b-[3H]estradiol binding in rat adipose tissues. Endocrinology 1978;103:1695–701. 9. Pedersen SB, Hansen PS, Lund S, Andersen PH, Odgaard A, Richelsen B. Identification of oestrogen receptors and oestrogen receptor mRNA in human adipose tissue. Eur J Clin Invest 1996;26:262–9. 10. Prins JB, O’Rahilly S, Chatterjee VKK. Steroid hormones and adipose tissue. Eur J Clin Invest 1996;26:259 – 61. 11. Kraemer HC, Thiemann S. How many subjects? Statistical Power Analysis in Research. Newbury Park (CA): Sage Publications, 1987. 12. Grimm LG. Statistical applications for the behavior sciences. New York: John Wiley and Sons, 1993. 13. Shimizu H, Shimomura Y, Nakanishi Y, Futawatari T, Ohtani K, Sato N, et al. Estrogen increases in vivo leptin production in rats and human subjects. J Endocrinol 1997;154:285–92. 14. Hube F, Lietz U, Igel M, Jensen PB, Tornqvist H, Joost H-G, et al. Differences in leptin on RNA levels between omental and subcutaneous abdominal adipose tissue from obese humans. Horm Metab Res 1996; 28:690 –3. 15. Jockenhovel F, Blum WF, Vogel E, Englaro P, Muller-Wieland D, Reinwein D, et al. Testosterone substitution normalizes elevated serum leptin levels in hypogonadal men. J Clin Endocrinol Metab 1997;82: 2510 –3. 16. Kohrt WM, Landt M, Birge SJ Jr. Serum leptin levels are reduced in response to exercise training, but not hormone replacement therapy, in older women. J Clin Endocrinol Metab 1996;81:3980 –5.
477
Serum leptin concentration in women: effect of age, obesity, and estrogen administration V. Daniel Castracane, Ph.D.,* Robert R. Kraemer, Ed.D.,† Mary Ann Franken, M.D.,* Ginger R. Kraemer,‡ and Terry Gimpel* Texas Tech University Health Sciences Center, Amarillo, Texas; Southeastern Louisiana University, Hammond, Louisiana; and North Oaks Obstetrics and Gynecology, Hammond, Louisiana
Objective: To compare serum leptin levels in normally cycling reproductive females (20 –35 years old) with those in age-matched males, in women who were receiving oral contraceptives, and in older (postmenopausal) women (50 – 65 years old) who were or who were not receiving hormone replacement therapy. Design: Case-control study. Setting: Obstetrics and Gynecology Clinic, Texas Tech University Health Sciences Center-Amarillo, or the Exercise Physiology Laboratory at Southeastern Louisiana University. Patient(s): Normally cycling women between the ages of 20 –35 years and age-matched controls who were receiving oral contraceptives. Postmenopausal women between the ages of 50 – 65 years who were or who were not receiving hormone replacement therapy. Received September 10, 1997; revised and accepted April 28, 1998. Supported in part by a Faculty Development grant from Southeastern Louisiana University, Hammond, Louisiana. Presented at the 44th Annual Society for Gynecologic Investigation meeting in San Diego, California, March 19 –22, 1997. Reprint requests: V. Daniel Castracane, Ph.D., Department of Obstetrics and Gynecology, Texas Tech University Health Sciences Center, 1400 Wallace Boulevard, Amarillo, Texas 79106 (FAX: 806-354-5516; Email: [email protected] .ttuhsc.edu). * Department of Obstetrics and Gynecology, Texas Tech University Health Sciences Center. † Department of Kinesiology and Health Studies, Southeastern Louisiana University. ‡ North Oaks Obstetrics and Gynecology. 0015-0282/98/$19.00 PII S0015-0282(98)00187-3
472
Main Outcome Measure(s): Serum leptin concentration. Result(s): In all groups, serum leptin concentrations were correlated significantly with body mass index. Leptin levels were significantly higher in young women than young men (P ,.001), but no other statistically significant differences were found for the other three comparisons. Conclusion(s): Serum leptin concentrations expressed as a measure of adiposity (body mass index) are greater in young normally cycling females (20 –35 years old) than in age-matched males. There is no difference in levels of serum leptin between young and postmenopausal (50 – 65 years old) women. Estrogen administration, either in young women who are receiving estrogen-progestin oral contraceptives or in postmenopausal women who are receiving hormone replacement therapy, does not effect serum leptin concentrations. (Fertil Sterilt 1998;70:472–7. ©1998 by American Society for Reproductive Medicine.) Key Words: Leptin, adiposity, estrogen, menopause, oral contraceptives, hormone replacement therapy, aging, males
Numerous studies have demonstrated the absence of leptin in the genetically (ob/ob) obese mouse and the reversal of this obese condition with leptin administration, which also corrects the sterility associated with this genetic defect (1–3). The role of leptin in the human remains unresolved. Many studies now have demonstrated that circulating levels of leptin in the human are related directly to measures of obesity such as body mass index (BMI) or percent body fat (4 – 6). These reports generally span a wide range with respect to age, although gender differences have been reported, with serum leptin levels always reported to be higher in females than males (5–7). In the present case-control study, we investigated whether any differences exist between
leptin levels in normally cycling young women (20 –35 years of age) and men in the same age group. However, we were more interested in the effect that estrogen might have in females because estrogen receptors have been reported in adipose tissue (8 –10), suggesting that estrogens may be involved in the regulation of leptin production. Therefore, we also compared this population of young women with an agematched group of women who were receiving standard estrogen-progesterone combination oral contraceptives (OCs) as well as older women who were or who were not receiving estrogen replacement therapy (ERT). These data demonstrate that the relationship of serum leptin in women to BMI does not change with age or estrogen administration.
MATERIALS AND METHODS Subjects Serum samples were obtained from 127 female subjects, usually as a fasted morning sample. At the time of collection, information recorded for all subjects included age, height, and weight as well as use of any medications that might interfere with reproductive or endocrine functions, with particular emphasis on estrogen medications, either OCs or hormone replacement therapy (HRT). Subjects were excluded from the study if they were receiving any medications with the exception of OCs and HRT; subjects who received these therapies formed distinct groups in this protocol. Subjects were then divided into a young (20 –35 years of age; mean [6 SD] age, 27.23 6 0.63 years) group of untreated normally cycling female subjects (n 5 58) with cycle length between 25 and 35 days. A similar age-matched group of women (mean [6 SD] age, 26.43 6 1.07 years) who were receiving standard low-dose estrogen-progestin OCs (n 5 23) for at least 3 months was used to compare the effect of estrogen administration in a young population. A third group of subjects aged 50 – 65 years was selected as postmenopausal women (n 5 46); this group was further subdivided into those who were receiving HRT (n 5 28) (mean [6 SD] age 56.5 6 0.92 years) and those who were not receiving HRT (n 5 18) (mean [6 SD] age 55.8 6 0.85 years). Those subjects who were receiving HRT had been taking that regimen for at least 3 months. A small group of age-matched males (n 5 13) (mean [6 SD] age 28.85 6
TABLE 1 Correlation coefficients, r 2, and P for comparing body mass index and leptin concentrations in the different groups. Group
Type of therapy received
Statistic
Young men
Young women
Older women
HRT
No HRT
OC
r r2 P
0.76 0.58 .0024
0.84 0.71 .0001
0.72 0.52 .0001
0.47 0.22 .0116
0.88 0.77 .0001
0.78 0.61 .0001
Note: HRT 5 hormone replacement therapy; OC 5 oral contraceptives.
1.13 years) was obtained to compare with the young group of normal females. All subjects gave their informed, written consent under different protocols approved by the Texas Tech University Health Sciences Center-Amarillo Institutional Review Board and Committee for the Use of Humans and Animals as Research Subjects, Southeastern Louisiana University Institutional Review Board. All serum samples were analyzed for serum leptin.
Assay methods Serum leptin was assayed with use of a second antibody RIA system available from Linco Diagnostics Incorporated (St. Louis, MO). The interassay coefficient of variation (CV)
FIGURE 1 Relationship of serum leptin concentration with BMI for male (n 5 13) and female (n 5 58) subjects between the ages of 20 and 35 years.
FERTILITY & STERILITYt
473
FIGURE 2 Relationship of serum leptin concentration with BMI for females 20 –35 years of age (n 5 58) with normal cycles and no form of contraception and postmenopausal women between the ages of 50 – 65 years of age (n 5 46). Postmenopausal subjects in the graph include both women on HRT and without HRT.
for low and high pools was 9.3% and 4.7%, respectively. The intra-assay CV averaged ,5% for all assays.
Statistics Pearson’s correlation coefficients were determined to assess the relationship between BMI and leptin for subgroups (young men, young women, older women, subjects receiving HRT, subjects not receiving HRT, and subjects receiving OC). Between-group analyses (independent t-tests) were conducted comparing group differences in gender, age, and HRT. Finally, to determine whether BMI affected the differences in leptin concentrations between groups (young males versus young females, young women versus older women, subjects receiving HRT versus those not receiving HRT, and subjects receiving OCs versus those not receiving OCs), four analyses of covariance (ANCOVA) were conducted with BMI as a covariate and leptin as the dependent variable.
RESULTS In all groups, leptin concentrations correlated significantly with BMI (Table 1; Figs. 1– 4). The t-tests indicated that leptin levels were significantly higher in young women versus young men (P 5 .004), older women versus younger women (P 5 .029), subjects who were receiving HRT versus those who were not (P 5 .0409), and subjects who were not receiving OCs versus those who were receiving them (P 5 474
Castracane et al.
.0026; Table 2). Using BMI as a covariate, ANCOVAs revealed that independent of BMI, leptin concentrations were significantly higher in younger women compared with young men (P 5 .0001), but no other statistically significant differences were found for the other three comparisons (Table 2). A post hoc decision was made to conduct power analyses to determine the extent to which we could detect statistically significant differences. We estimated power to detect moderate (0.50) and large (0.80) effect sizes for each analysis using the procedures recommended by Kraemer and Thiemann (1987) (11) for unequal group designs. Power to detect moderate effect sizes ranged from 0.40 to 0.70. To detect large effect sizes, the power of all analyses exceeded 0.70 (2 analyses were greater than 0.95) with the exception of the comparison between those receiving and not receiving HRT, which was 0.70. Acceptable range for power is considered to be 0.70 – 0.90 (12), and all analyses reached this level to detect significant differences for large effect sizes.
DISCUSSION The relationship of serum leptin concentration with some measure of adiposity (e.g., BMI, percent body fat) has been demonstrated in numerous reports (3–5). We have used the same approach to study serum leptin concentration with different physiologic states in women as well as comparison
Serum leptin in women: age and estrogen
Vol. 70, No. 3, September 1998
FIGURE 3 Relationship of serum leptin concentration with BMI for normal cycling females (20 –35 years) (n 5 58) and an age-matched group of women on low-dose oral contraceptives (n 5 23).
with age-matched normal males. We attempted to determine whether estrogen would have an effect on serum leptin concentrations. It has been reported that estrogen receptors are present in adipose tissue (8 –10) and therefore might serve as a mechanism to regulate leptin production. In a recent study of rats and humans it was suggested that serum leptin levels may increase in response to estrogen treatment (13). Comparison with males represents an obvious distinction in which females would have elevated levels of estrogen over males. The other comparisons in this study involve normally cycling reproductive aged women compared with postmenopausal women who were or were not receiving estrogen replacement as well as a younger group who were receiving estrogen-progestin OCs. These groups were studied because of the dynamics of estrogen in these subjects. We have examined the relationship of females to males and with a small group of males have demonstrated, as have other investigators (5–7), that males consistently have lower serum leptin concentrations than do females (Fig. 1). In this study we have confined our subjects to a tighter age range of 20 –35 years than was used in these other studies. Because a large number of reports already have compared males and females, this aspect of our study was not of paramount importance but serves to demonstrate the validity of assay procedures and patient sampling. The factors that regulate leptin levels in males are not clear, although the suggestion that visceral versus subcutaFERTILITY & STERILITYt
neous adipose tissue, the distribution of which is different in males than females, may be a contributing factor (14). Similarly, in hypogonadal males, leptin levels are elevated and return to lower levels after testosterone treatment (15), suggesting that an endocrine involvement may be at the basis of the distinction between males and females. We were interested in the possibility that estrogens might represent a physiologic regulator of leptin production and/or secretion. Therefore, we obtained serum samples from a group of normally cycling younger reproductive aged women, 20 –35 years of age, which served as controls and compared serum leptin concentrations in this group with a group of older postmenopausal women, 50 – 65 years of age. In the original comparison (Fig. 2) we included women in the postmenopausal group, both with and without postmenopausal HRT. This comparison allows us to determine whether aging changes the relationship of serum leptin to BMI. This relationship was not different between the young and older group of women. Ostlund et al. (6) presented data on leptin concentration over a wide age range and suggested that circulating leptin levels are related inversely to age and were reduced by as much as 53% in subjects .60 years. Several factors make these data difficult to interpret. Data are presented not only as means 6 SE, which we think is inappropriate for such data, but in some cases (e.g., age-related data), leptin is plotted as the means 6 SE with apparently mixed male and female values. Because male leptin values are lower than 475
FIGURE 4 Relationship of serum leptin concentration to BMI for postmenopausal subjects who were receiving HRT (n 5 28) and who were not receiving HRT (n 5 18).
females, these data may be represented disproportionately in the lower BMI group. Moreover, it is not clear what the age distribution of the different genders might be. In our study, we have a balanced distribution of adiposity in each group to effectively scrutinize the difference between groups. The study by Kohrt et al. (16) conforms with our conclusion, that there was no effect of HRT on leptin concentration, although data presentation was not designed to evaluate the effects of HRT on leptin.
who were receiving OCs. We have used the same control group and have used an age-matched group of women on estrogen-progestin OCs. The comparison for these two groups shows that the regression lines for both sets of data (Fig. 3) are virtually superimposable. These results demonstrate no effect of OCs on serum leptin concentration compared with younger women and expressed on the basis of adiposity. This is the first study comparing the effect of OCs in normal subjects on leptin concentration.
Another approach to examine the effect of estrogen replacement as a possible physiologic regulator of leptin production and secretion was to study leptin levels in women
As a last comparison, we have subdivided subjects in the postmenopausal group to those receiving HRT and those with no postmenopausal treatment. In this comparison, there
TABLE 2 Body mass index and concentration of leptin in young men, young women, older women, subjects receiving HRT or OCs and subjects not receiving HRT or OCs. Group
Variable Mean (6 SD) BMI Mean (6 SD) leptin concentration (ng/mL) Adjusted mean leptin concentration (ng/mL)
Type of therapy received
Young men
Young women
Older women
HRT
No HRT
OC
28.09 6 3.28
29.19 6 7.94
29.33 6 5.47
30.16 6 4.11
28.01 6 7.00
23.66 6 4.016
9.67 6 5.14
25.12 6 15.20
27.76 6 14.75
30.80 6 12.86
23.04 6 16.57
15.43 6 8.63
11.09
25.24
27.61
29.21
25.51
21.83
Note: HRT 5 hormone replacement therapy; OC 5 oral contraceptives.
476
Castracane et al.
Serum leptin in women: age and estrogen
Vol. 70, No. 3, September 1998
was no statistically significant difference between groups, indicating that standard doses of postmenopausal estrogen replacement have no effect on production and secretory dynamics of leptin from adipose tissue in this age group. The power of this analysis to detect a large effect size was within the acceptable range (0.7– 0.9) (11). Some studies, e.g., Rosenbaum et al (5), have used means 6 SE of serum leptin without correction for BMI or fat mass to compare leptin levels in different groups. However, it is important to remember that the major determinant of leptin production is some measure of fat mass or adiposity and any group that is not balanced for distribution of body fat in its subjects or not corrected for body composition may be biased when expressed as means 6 SE. Our adjusted means clearly demonstrate that without using BMI as a covariate, all groups compared were significantly different; however, with BMI as a covariate, only the male-female comparison was significantly different. In summary, this study has demonstrated that serum leptin levels in young normally cycling females, when expressed on the basis of BMI, is greater than age-matched normal males as has been demonstrated by other investigators. More importantly, with an adequate number of subjects and a well-distributed BMI range, there is no effect of estrogen administration on this serum leptin distribution, either in subjects receiving OCs or HRT. This appears to be the first investigation to have studied OCs and leptin. We believe the advantage in this study is an adequate number of subjects as well as a better BMI distribution in our different populations.
Acknowledgments: We thank Edward Hebert, Ph.D. (Department of Kinesiology and Health Studies, Southeastern Louisiana University, Hammond, LA), for invaluable help in statistical interpretation. We also appreciate the
FERTILITY & STERILITYt
help of many individuals in recruiting subjects, in performing phlebotomy, and in manipulating data. We also thank Debbie Pallmeyer for help in preparing the manuscript.
References 1. Pelleymounter MA, Cullen MJ, Baker MB, Hecht R, Winters D, Boone T, et al. Effects of the obese gene product on body weight regulation in ob/ob mice. Science 1995;269:540 –3. 2. Chehab FF, Lim ME, Ronghua L. Correction of the sterility defect in homozygous obese female mice by treatment with the human recombinant leptin. Nat Genet 1996;12:318 –20. 3. Weigle DS, Bukowsk TR, Foster DC, Holderman S, Kramer JM, Lasser G, et al. Recombinant ob protein reduces feeding and body weight in the ob/ob mouse. J Clin Invest 1995;96:2065–70. 4. Considine RV, Sinha MK, Heiman ML, Kriauciunas A, Stephens TW, Nyce MR, et al. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med 1996;334:292–5. 5. Rosenbaum M, Nicolson M, Hirsch J, Heymsfield SB, Gallagher D, Chu F, et al. Effects of gender, body composition and menopause on plasma concentrations of leptin. J Clin Endocrinol Metab 1996;81: 3424 –7. 6. Ostlund RE Jr, Yang JW, Klein S, Gingerich R. Relation between plasma leptin concentration and body fat, gender, diet, age and metabolic covariates. J Clin Endocrinol Metab 1996;81:3909 –13. 7. Kennedy A, Gettys TW, Watson P, Wallace P, Ganaway E, Pan Q, et al. The metabolic significance of leptin in humans: gender-based differences in relationship to adiposity, insulin sensitivity, and energy expenditure. J Clin Endocrinol Metab 1997;82:1293–300. 8. Wade GN, Gray JM. Cytoplasmic 17b-[3H]estradiol binding in rat adipose tissues. Endocrinology 1978;103:1695–701. 9. Pedersen SB, Hansen PS, Lund S, Andersen PH, Odgaard A, Richelsen B. Identification of oestrogen receptors and oestrogen receptor mRNA in human adipose tissue. Eur J Clin Invest 1996;26:262–9. 10. Prins JB, O’Rahilly S, Chatterjee VKK. Steroid hormones and adipose tissue. Eur J Clin Invest 1996;26:259 – 61. 11. Kraemer HC, Thiemann S. How many subjects? Statistical Power Analysis in Research. Newbury Park (CA): Sage Publications, 1987. 12. Grimm LG. Statistical applications for the behavior sciences. New York: John Wiley and Sons, 1993. 13. Shimizu H, Shimomura Y, Nakanishi Y, Futawatari T, Ohtani K, Sato N, et al. Estrogen increases in vivo leptin production in rats and human subjects. J Endocrinol 1997;154:285–92. 14. Hube F, Lietz U, Igel M, Jensen PB, Tornqvist H, Joost H-G, et al. Differences in leptin on RNA levels between omental and subcutaneous abdominal adipose tissue from obese humans. Horm Metab Res 1996; 28:690 –3. 15. Jockenhovel F, Blum WF, Vogel E, Englaro P, Muller-Wieland D, Reinwein D, et al. Testosterone substitution normalizes elevated serum leptin levels in hypogonadal men. J Clin Endocrinol Metab 1997;82: 2510 –3. 16. Kohrt WM, Landt M, Birge SJ Jr. Serum leptin levels are reduced in response to exercise training, but not hormone replacement therapy, in older women. J Clin Endocrinol Metab 1996;81:3980 –5.
477