Relation of luteinizing hormone levels to body mass index in premenopausal women

Transcription

1 FERTILITY AND STERILITY VOL. 69, NO. 3, MARCH 1998 Copyright 1998 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A. Relation of luteinizing hormone levels to body mass index in premenopausal women Kari Bohlke, M.S.,* Daniel W. Cramer, M.D., Sc.D.,* and Robert L. Barbieri, M.D.* Brigham and Women s Hospital, Harvard Medical School, Boston, Massachusetts Objective: To assess the relationship between body mass index (BMI) and basal LH and the LH-FSH ratio in normally menstruating women. Design: Cross-sectional analysis. Setting: A teaching hospital clinic. Patient(s): Premenopausal women without cancer, not currently using oral contraceptives, selected from a familial ovarian cancer clinic or the general population. Intervention(s): None. Main Outcome Measure(s): Early follicular phase plasma LH and FSH. Result(s): Luteinizing hormone increased slightly and nonsignificantly (P 0.44) from the first to the second quintile of BMI and decreased over all subsequent quintiles. Women in the highest quintile of BMI ( 27.1) had significantly lower LH levels than women in the lowest quintile of BMI ( 20.4; P 0.003). Compared with women in the second quintile of BMI who had the highest LH levels, women in the highest quintile of BMI had LH levels that were 40% lower. The relationship between BMI and the LH-FSH ratio was similar, though not as strong. Conclusion(s): Over most of the range of BMIs observed in this study, BMI was inversely associated with LH. These results suggest that the upper limits of normal for LH may need to be shifted downward for heavier women. (Fertil Steril 98;69: by American Society for Reproductive Medicine.) Key Words: Body mass index, LH, FSH, premenopausal Received July 31, 1997; revised and accepted October 30, This work was supported by grant 90BW25 from the American Institute for Cancer Research. Reprint requests: Daniel W. Cramer, M.D., Ob-Gyn Epidemiology Center, Brigham and Women s Hospital, 221 Longwood Avenue, Boston, Massachusetts (FAX: ). * Department of Obstetrics and Gynecology. Obstetrics and Gynecology Epidemiology Center /98/$19.00 PII S (97) An elevated LH level is an important biochemical feature of polycystic ovary syndrome (PCOS), but the degree of elevation appears to be inversely related to body mass index (BMI) (1). In obese women with PCOS, the absolute level of LH is less elevated and the ratio of LH to FSH is lower compared with women of normal weight with PCOS. However, the relationship between LH and body mass in normal women has not been as thoroughly characterized. This study of basal gonadotropin levels in reproductive-aged and normally menstruating women, who were not selected for obesity or PCOS, allows an examination of factors affecting LH, including body mass. MATERIALS AND METHODS Subjects in this study were healthy women with a family history of ovarian cancer who were recruited from a Familial Ovarian Cancer clinic established at the Brigham and Women s Hospital and women without such a family history who were recruited from the general population of greater Boston. All subjects provided informed consent in a protocol approved by the parent institution. Subjects ranged in age from 26 to 50 years, were predominantly white, were premenopausal, and currently were not using oral contraceptives (OCs) or cyclic progestins. Additional factors pertaining to the selection of the 106 women with a family history of ovarian cancer and the 116 women without this history previously were discussed (2). Data collected from subjects included information on parity, OC use, smoking history, and selfreported weight and height. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. 500

2 Besides collection of personal data, measurement of FSH, LH, and E 2 was performed from a blood specimen taken during the early follicular phase, defined as the first 3 days of the menstrual cycle (day 1 being the 1st day of menses). Samples were taken during normal working hours, but otherwise, no attempt was made to standardize time of blood drawing. Blood was drawn in heparanized tubes and separated into plasma, red cells, and buffy coat aliquots and stored at 70 C before assay. Estradiol was measured in plasma by a solid-phase RIA using unextracted plasma and a highly specific E 2 antibody (Diagnostic Products Corp., Los Angeles, CA). Luteinizing hormone and FSH were measured in plasma with the use of an immunoradiometric assay with the World Health Organization s Second International Reference Preparation of hmg as the standard. All assays were run in duplicate and averaged. Intra- and interassay coefficients of variation were 10% and 15%, respectively, for all three assays. Previously published articles have addressed the effect of various factors on basal FSH, including age, smoking history, and prior OC use (3, 4). Univariate analyses, including t-tests or F-tests for independent samples, were performed to determine the effect of BMI and other potential variables affecting LH and the LH-FSH ratio. The hormonal values were log transformed to achieve more normal distributions. Multiple linear regression was applied to assess the effects of several study variables on LH and the LH-FSH ratio with PROC REG in Statistical Analysis Systems (SAS, Cary, NC). RESULTS Table 1 shows geometric means for early follicular LH levels by various characteristics of the study population. Variables that did not significantly affect basal LH included age, smoking history, average cycle length and regularity, parity, length of prior OC use, and E 2. One variable that we previously reported to exert a significant effect on the basal LH level was family history status (i.e., family history of ovarian cancer). A second variable identified in this analysis to affect LH was BMI so that higher BMI was associated with lower LH. This is depicted graphically in Figure 1 showing mean LH and the 95% confidence limits of the mean by quintile of BMI. Compared with the peak LH mean of 4.98 IU/mL at BMIs between 20.5 and 21.8, women with a BMI in the upper quintile of 27.1 had a mean LH of 2.96 IU/mL or approximately 40% less than the peak mean LH. In a multiple linear regression model controlling for family history status, BMI remained a significant predictor of LH (P for the highest quintile of BMI relative to the lowest quintile of BMI; data not shown). Addition to the model of variables not significant in the univariate analyses, such as age, smoking, length of prior OC use, cycle length, TABLE 1 Mean basal LH levels by various characteristics of the study population. Variable No. of patients Geometric mean (IU/L) t or F* P value Status Positive family history Negative family history Cycle day Age (y) Body mass index Smoking None Past ( 10 pack-years) Past ( 10 pack-years) Current Cycle regularity No Yes Cycle length (days) Parity None Prior oral contraceptive use (y) None Estradiol level (pg/ml) * t-tests were applied to variables with only two categories, otherwise F-tests were used. and E 2 level, did not affect the significance of the association between BMI and LH (data not shown). Table 2 shows the mean basal LH-FSH ratio by various characteristics of the study population. Family history status, age, and cycle length were significantly associated with the LH-FSH ratio. Greater BMI tended to be associated with a smaller LH-FSH ratio, but the association was of borderline significance (P 0.07). In a multivariate model including variables for family history status, age, smoking, length of FERTILITY & STERILITY 501

3 FIGURE 1 Geometric mean LH and 95% confidence limits of the mean by quintile of BMI among normally menstruating women. prior OC use, cycle length, and E 2, the difference in the LH-FSH ratio between those in the highest quintile of BMI and those in the lowest quintile of BMI was statistically significant (P 0.04). DISCUSSION Studies addressing BMI and LH have been performed mostly in women with PCOS (5 13). All but two of these studies (6, 12) have found that obese women with PCOS have low LH levels compared with lean or normal weight women with PCOS. Five of these studies also have included healthy women (9 13); these studies show that obese women without PCOS stigmata also have lower LH levels than normal weight women, although the differences were usually not significant. Our study has identified a statistically significant inverse association between BMI and LH in normal women, none of whom were known to have PCOS. The association persisted after controlling for smoking, past OC use, and cycle length. Although some studies have shown that LH levels rise among women in their late 40s (14, 15), age was not associated with LH in this study of women with a mean age of 39, and age did not confound the association between BMI and LH. Greater BMI is also associated with a lower LH- FSH ratio, although the relationship is not as strong as that between BMI and LH. Thus, as BMI increases, the absolute and relative level of LH decreases in both women with PCOS and normal women, even though the LH levels will be greater in the PCOS group in any weight category. Most studies addressing LH and BMI have dichotomized PCOS and normal women according to BMI of 25. This type of analysis does not allow resolution of the precise relationship between LH and BMI. However, a recent analysis by Taylor et al. (13) showed a continuous effect of BMI on LH over the entire range of BMI in the women with PCOS that they studied. In our data, LH was slightly lower for women in the lowest quintile of BMI ( 20.4), increased for women in the second quintile ( ), and then decreased progressively. Although it is possible that the relationship between BMI 502 Bohlke et al. LH and body mass index Vol. 69, No. 3, March 1998

4 TABLE 2 Mean values of the LH-FSH ratio by various characteristics of the study population. Variable No. of patients Geometric mean t or F* P value Status Positive family history Negative family history Cycle day Age (y) Body mass index Smoking None Past ( 10 pack-years) Past ( 10 pack-years) Current Cycle regularity No Yes Cycle length (d) Parity None Prior oral contraceptive use (y) None Estradiol level (pg/ml) * t-tests were applied to variables with only two categories, otherwise F-tests were used. and LH is more complicated than a simple linear decrease with increasing BMI, the decrease in LH with increasing BMI is the relationship most apparent in our data and may be relevant for defining the upper limits of normal for LH in clinical evaluation. Given the 40% lower mean LH levels observed among women with BMIs of 27.1 relative to women with BMIs between 20.5 and 21.8, our data suggest that the upper limits of normal LH may need to be shifted downward as BMI increases. A limitation of the study is that BMI was based on self-reported height and weight. Although self-reports of both height and weight have been shown to be quite accurate (16 18), there is a tendency for people to underestimate weight and overestimate height. The result is a slight bias downward in BMI. If underreporting of weight is most pronounced among the heaviest women, as has been observed by a number of studies (16 18), it may attenuate the association between BMI and LH. Therefore, our results may be an underestimate of the true association between BMI and LH. Our study is unable to address the mechanism(s) underlying the association between LH and BMI. It seems unlikely that the association is mediated through E 2 levels because we observed no significant relationship between LH and E 2 in our data based on early follicular phase specimens. However, a key factor that we did not measure may be insulin. Fasting insulin levels are positively correlated with BMI (or body fat) and negatively correlated with LH, whereas insulin resistance is positively correlated with BMI (or body fat) and negatively correlated with LH (7, 9, 10, 19 21). In a study by Holte et al. (10) that included measurement of fasting insulin, the addition of the insulin level to a multiple regression model eliminated the significance of BMI as a determinant of LH in both normal subjects and those with PCOS. In summary, the inverse association between BMI and LH observed in women with PCOS also pertains to normal premenopausal women selected neither for obesity nor PCOS. Issues of clinical relevance to the treatment of PCOS, including weight loss or medications that improve insulin sensitivity, also may be relevant to obese women without PCOS stigmata. References 1. Laatikainen T, Tulenheimo A, Andersson B, Karkkainen J. Obesity, serum steroid levels, and pulsatile gonadotropin secretion in polycystic ovarian disease. Eur J Obstet Gynecol Reprod Biol 1983;15: Cramer DW, Barbieri RL, Muto MG, Kelly A, Brucks JP, Harlow BL. Characteristics of women with a family history of ovarian cancer. II. Follicular phase hormone levels. Cancer 1994;74: Barbieri RL, Gao X, Xu H, Cramer DW. Effects of previous use of oral contraceptives on early follicular phase follicle-stimulating hormone. Fertil Steril 1995;64: Cramer DW, Barbieri RL, Xu H, Reichardt JKY. Determinants of basal follicle-stimulating hormone levels. J Clin Endocrinol Metab 1194;79: Paradisi R, Venturoli S, Pasquali R, Capelli M, Porcu E, Fabbri R, et al. Effects of obesity on gonadotropin secretion in patients with polycystic ovarian disease. J Endocrinol Invest 1986;9: Dunaif A, Mandeli J, Fluhr H, Dobrjansky A. The impact of obesity and chronic hyperinsulinemia on gonadotropin release and gonadal steroid secretion in the polycystic ovary syndrome. J Clin Endocrinol Metab 1988;66: Conway GS, Jacobs HS, Holley JMP, Wass JAH. Effects of luteinizing hormone, insulin, insulin-like growth factor-1 and insulin-like growth factor small binding protein 1 in the polycystic ovary syndrome. Clin Endocrinol (Oxf) 1990;33: Anttila L, Ding YQ, Ruutiainen K, Erkkola R, Irjala K, Huhtaniemi I. Clinical features and circulating gonadotropin, insulin, and androgen interactions in women with polycystic ovarian disease. Fertil Steril 1991;55: Dale PO, Tanbo T, Vaaler S, Abyholm T. Body weight, hyperinsulinemia, and gonadotropin levels in the polycystic ovary syndrome: evidence of two distinct populations. Fertil Steril 1992;58: Holte J, Bergh T, Gennarelli G, Wide L. The independent effects of FERTILITY & STERILITY 503

5 polycystic ovary syndrome and obesity on serum concentrations of gonadotrophins and sex steroids in premenopausal women. Clin Endocrinol 1994;41: Morales AJ, Laughlin GA, Butzow T, Maheshwari H, Baumann G, Yen SSC. Insulin, somatotropic, and luteinizing hormone axes in non-obese and obese women with polycystic ovary syndrome: common and distinct features. J Clin Endocrinol Metab 1996;81: Tropeano G, Vuolo IP, Lucisano A, Liberale I, Barini A, Carfagna P, et al. Gondotropin levels in women with polycystic ovary syndrome: the relationship to body weight and insulin levels. J Endocrinol Invest 1996;19: Taylor AE, McCourt B, Martin KA, Anderson EJ, Adams JM, Schenfeld D, et al. Determinants of abnormal gonadotropin secretion in clinically defined women with polycystic ovary syndrome. J Clin Endocrinol Metab 1997;82: Reyes FI, Winter JS, Faiman C. Pituitary-ovarian relationships preceding the menopause. I. A cross-sectional study of serum follicle-stimulating hormone, luteinizing hormone, prolactin, estradiol, and progesterone levels. Am J Obstet Gynecol 1977;129: Lee SJ, Lenton EA, Sexton L, Cooke ID. The effect of age on the cyclical patterns of plasma LH, FSH, oestradiol and progesterone in women with regular menstrual cycles. Hum Reprod 1988;3: Stewart AL. The reliability and validity of self-reported weight and height. J Chron Dis 1982;35: Palta M, Prineas RJ, Berman R, Hannan P. Comparison of self-reported and measured height and weight. Am J Epidemiol 1982;115: Stewart AW, Jackson RT, Ford MA, Beaglehole R. Underestimation of relative weight by use of self-reported height and weight. Am J Epidemiol 1987;125: Grulet H, Hecart AC, Delmer B, Gross A, Sulmont V, Leutenegger M, et al. Roles of LH and insulin resistance in non-obese and obese polycystic ovary syndrome. Clin Endocrinol (Oxf) 1993;38: Pasquali R, Casimirri F, Venturoli S, Paradisi R, Mattoli L, Capelli M, et al. Insulin resistance in patients with polycystic ovaries: its relationship to body weight and androgen levels. Acta Endocrinol (Copenh) 1983;104: Buyalos RP, Geffner ME, Watanabe RM, Bergman RN, Gornbein JA, Judd HL. The influence of luteinizing hormone and insulin on sex steroids and sex hormone-binding globulin in the polycystic ovarian syndrome. Fertil Steril 1993;60: Bohlke et al. LH and body mass index Vol. 69, No. 3, March 1998