Leptin levels and luteinizing hormone pulsatility in normal cycling women and their relationship to daily changes in metabolic rate

Transcription

1 REPRODUCTIVE ENDOCRINOLOGY Leptin levels and luteinizing hormone pulsatility in normal cycling women and their relationship to daily changes in metabolic rate Rebecca M. Fenichel, M.D., a Jennifer E. Dominguez, B.A., b Laurel Mayer, M.D., c B. Timothy Walsh, M.D., c Carol Boozer, M.D., d and Michelle P. Warren, M.D. a Departments of a Obstetrics and Gynecology; c Psychiatry, Columbia College of Physicians and Surgeons, New York, New York; b Yale University School of Medicine, New Haven, Connecticut; and d Department of Medicine, St. Luke s Roosevelt Hospital Center, New York, New York Objective: To determine whether leptin and LH secretion in normal women is related to changes in metabolic rate. Setting: Academic medical center. Patient(s) and Design: Ten young women with normal weight and menses were studied during the early follicular phase. Leptin and LH levels were sampled every 15 minutes over a 24-hour period. Metabolic rate was frequently sampled using indirect calorimetry. Luteinizing hormone pulsatility was analyzed using a Cluster Program analysis. Intervention(s): None. Main Outcome Measure(s): Leptin, LH, and metabolic rate levels. Result(s): All subjects demonstrated a diurnal leptin curve. Luteinizing hormone pulses were increased in amplitude and slower after the leptin peak. The average (SE) number of LH pulses per 6 hours slowed from to pulses after the leptin peak, whereas pulse amplitude increased from to miu/ml after the leptin peak. The LH interpulse interval increased from minutes to minutes after the leptin peak. Metabolic rate began to drop approximately 4 6 hours before leptin levels peaked, going from to kcal/min after the leptin peak. Conclusion(s): There is a significant association between the timing of the leptin peak, the nightly slowing of LH pulses, and the fall in metabolic rate, suggesting a metabolic cycle in normal individuals. (Fertil Steril Ò 2008; 90: Ó2008 by American Society for Reproductive Medicine.) Key Words: Leptin, luteinizing hormone, metabolic rate, premenopausal women A metabolic signal affecting reproduction has long been sought. Research suggests that LH pulses, which are required for normal ovulation, are intricately associated with secretion of leptin (1), a hormone that is intimately involved with nutritional status and energy homeostasis. This most likely occurs at the level of the hypothalamus, where leptin receptors are present (2). The role of leptin in determining normal LH pulsatility, however, is unclear. This relationship is important to define, both as a comparison for abnormal physiologic states, such as hypothalamic amenorrhea, and to further understand Received January 26, 2007; revised July 16, 2007; accepted July 18, B. T. Walsh receives research support from Eli Lilly & Company, Abbott Laboratories, and Ortho McNeil, Inc. M. P. Warren receives research support from Solvay Pharmaceuticals and is a consultant for Wyeth Pharmaceuticals and Amylin Pharmaceuticals. Reprint requests: Rebecca M. Fenichel, M.D., Department of Obstetrics and Gynecology, Columbia University, 622 West 168th Street, PH , New York, NY (FAX: ; rf2013@ columbia.edu). leptin LH interaction. Recently, leptin has been found to be a mediator of LH pulsatility in both normal and abnormal sates (1, 3). Leptin is synthesized in adipose tissue, and levels are known to be roughly proportional to fat mass, and as such leptin is thought to be a signal to the reproductive axis of adequate fat stores. However, leptin levels can change rapidly and thus are influenced by factors other than body composition. Women with disordered eating and decreased energy consumption have been shown to have lower leptin levels, with both a lower baseline and a blunted diurnal curve (4, 5). In these cases, both leptin levels and resting metabolic rate have been shown to be lower than in control subjects (6). At the same time, minute-to-minute changes in leptin levels have been observed to vary according to the availability of oxidative substrates in vitro and in vivo, because leptin levels fall in states of acute starvation (7 10). Similarly, leptin levels in young, non-obese female subjects decrease sharply in response to fasting, out of proportion with any weight loss achieved (11). This suggests that both leptin secretion and consequently LH pulsatility may be influenced /08/$34.00 Fertility and Sterility â Vol. 90, No. 4, October doi: /j.fertnstert Copyright ª2008 American Society for Reproductive Medicine, Published by Elsevier Inc.

2 by both the sum of energy stores as dictated by fat composition and by the short-term availability of energy fuels (12, 13). Metabolic rate is a strong indicator of both activity and caloric intake and thus energy availability. Little is known about metabolic rate in normal women in relation to the LH pulses. This study was undertaken to determine whether the 24-hour secretion of leptin and LH in normal women was related to changes in metabolic rate, which reflects activity and caloric intake. More specifically, we hypothesized that fluctuations in caloric availability may be associated with a metabolic reproductive rhythm involving leptin and LH in normal women; and that the normal physiologic nightly fall in metabolic rate, which reflects a decrease in caloric availability and activity, is reflected in a fall in leptin levels and a resultant change in LH pulsatility. Defining the relationship between leptin, reproductive hormones, and basal metabolic rate may provide clues as to the regulation of the reproductive cycle and the key role of energy homeostasis. MATERIALS AND METHODS Subjects Ten women were included in this study. They were recruited through flyers and advertisements at one of multiple sites including physician s offices, local colleges in the New York metropolitan area, and Internet web sites. To be included in the study, subjects had to be between the ages of 18 and 45 years, have regular, monthly menses by menstrual history, and be at 90% 100% of their ideal body weight as calculated from the 1959 Metropolitan Life height and weight tables. Subjects were screened to rule out a current or past history of eating disorders with the Structured Clinical Interview for Diagnosis DSM-IV (14). Dietary information was collected with 4-day food intake diaries, which were analyzed using special software (Nutrition Data System for Research, developed by the Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN). All subjects exercised for <5 hours per week by report, and none were using any form of hormonal contraception. Informed consent was obtained from all patients, and this study was approved by the institutional review boards of the New York State Psychiatric Institute, Columbia University, and the St. Luke s-roosevelt Hospital Center. All subjects had normal thyrotropin and cortisol levels, as well as normal diurnal cortisol variability during the 24 hours studied. Hormonal Studies Data were collected on two separate occasions. Hormonal studies were performed at the Irving Center for Clinical Research, part of the General Clinical Research Centers of Columbia University. All subjects except 1 were studied between days 1 7 of the follicular cycle, with 1 subject studied on day 8 and most subjects studied between days 3 and 7. Sampling began between 8:15 AM and 9:15 AM for all subjects. Subjects were kept overnight, during which IV sampling for hormone analysis took place through a small indwelling catheter every 15 minutes over a 24-hour period. After centrifugation, sera were frozen at 20 C until assays were performed. Energy intake was not controlled (because variability in energy status is known to disrupt menstrual function), though specific meal times were encouraged, and all subjects ate meals within 1 hour of each other, leading to little variability in meal times. Energy intake and sleep time were recorded. Hormonal Assays Luteinizing hormone was measured by a commercial solidphase chemiluminescent immunoassay (Immulite; Diagnostic Products Co., Los Angeles, CA) with an assay sensitivity of 0.1 miu/ml. The intra-assay coefficient of variation (CV) for LH was 3.6%, and the inter-assay CV was 5.0%. Leptin levels were measured using a commercial ELISA kit (Diagnostic System Labs, Inc., Webster, TX). Assay sensitivity was 0.25 ng/ml. The intra-assay CV was 2.5%, and the inter-assay CV was 3.6%. Metabolic Studies Metabolic rate was measured every 2 minutes during a second 24-hour period at the Obesity Research Center at St. Lukes- Roosevelt Hospital (New York, NY). Sampling began between 8:50 AM and 10:20 AM, with all subjects beginning approximately within 1 hour of the same time they began hormonal sampling on the previous study day. Subjects exercised twice for 10 minutes during the day, once in the morning at 11 AM and once in the afternoon at 3 PM. Again, energy intake was not controlled because variability in energy status is known to affect menstrual function, though subjects were encouraged to eat at specific times. As a result, all subjects ate meals within 1 hour of each other, leading to little variability in meal times. Energy intake was recorded. Subjects were weight-stable (within 1 kg), by report, for days before this phase of the experiment. Metabolic rate was measured using a respiratory chamber indirect calorimeter, which has a net volume of 18.2 kl and was designed specifically for long-term experiments. The chamber is equipped with a high-precision gas (oxygen and carbon dioxide) exchange measurement system and a built-in motion detector to allow more precise measurement of the subject as she moves about the chamber. The subject was able to sleep, eat, watch television, and exercise as she normally would. A diary was kept by each subject and included food and calorie intake, as well as sleep and exercise. It was not possible to perform both the hormonal and metabolic rate sampling on the same day. Every effort was made to study subjects on the same day of the menstrual cycle for both metabolic rate and hormonal sampling; however, this was not always possible. Metabolic rate was measured as kilocalories per kilogram of lean body weight per unit time. Total body mass was differentiated into fat mass and fat-free mass. Body composition was evaluated by measuring weight and whole-body fat by dual x-ray absorptiometry (15). Dual x-ray absorptiometry is associated with a measurement precision of 1.2 % and 1162 Fenichel et al. Leptin, LH, and metabolic rate Vol. 90, No. 4, October 2008

3 a cross-subject accuracy in subjects of comparable fatness of better than 1%. The reports from the DPX-L scanner (GE Systems, Madison, WI) were analyzed using version 3.6 software and were used to deterine total body fat percentage. Statistics The leptin peak, which in most cases was the highest value of leptin recorded for each person, was selected as a point of reference. The leptin peak was chosen because of the known change in LH pulse patterns near the nocturnal peak in leptin levels (1). Two patients had the highest leptin value within the first few hours of the sampling period, 1 at first measurement and hours after the first measurement, and in these cases the second highest peak was taken. In both cases, the peak values were similar, and therefore the second peak was chosen because of known leptin physiology. The data for each subject were grouped into four 6-hour periods that corresponded to periods before and after the leptin peak. Sleep time was recorded by the research staff, and all parameters were compared during sleep time and awake time. Analysis of variance (Statistical Package for Social Sciences; SPSS, Chicago, IL) was used to evaluate differences in metabolic rate, leptin, and LH, with P<.05 considered significant. Nonparametric statistics were applied to the leptin data set because of 1 subject with very high leptin levels, and data reported for leptin are based on this analysis. If values increased significantly over time, we reported a significant linear relationship. Leptin and LH pulsatility characteristics were analyzed using Cluster Program version 8.0 (16). Luteinizing hormone pulse frequency and amplitude, as well as interpulse interval, were determined by the cluster analysis program (16). For peak detection, the algorithm was adjusted to require a threshold increase corresponding to a t statistic of 2.0 for the peak up-stroke and a t statistic of 2.0 for the downstroke, with nadir and peak test cluster sizes of 2 points and 1 point, respectively. These parameters permit both a reduced rate of false-positive peak detections while they preserve a minimum false-negative rate in vivo (17, 18). Cross-correlation analysis (SPSS) was used to detect significant correlations between time series of leptin and metabolic rate. RESULTS Baseline characteristics of the women included in this study are shown in Table 1. Eating Disorder Inventory scores revealed no symptoms of anorexia nervosa or bulimia. Leptin values were highly variable across subjects. However, all subjects demonstrated a diurnal leptin curve, with leptin levels usually peaking between 12 AM (midnight) and 4:30 AM (Fig. 1) and rising an average of 48% above baseline. One subject demonstrated much higher leptin levels, with an elevated baseline leptin level and a normal diurnal pattern. This subject had the highest percentage fat mass and a normal body mass index (BMI). All other subject characteristics, including nutritional intake and meal times, were similar. After assigning the time of leptin peak as 0 time, the 24-hour cycle was divided into four 6-hour periods three before and one after the leptin peak. The average leptin level was then calculated for each of these periods and is presented in Table 2. Average leptin levels significantly differed for segments 1 compared with segments 2 (P¼.044), segment 3 (P¼.015), and segment 4 (P¼.005), demonstrating the nightly leptin peak and morning nadir. Luteinizing hormone pulsatility was then measured using Cluster Program version 8.0 and was analyzed before and after the leptin peak. There was no significant correlation between leptin level and LH interpulse interval during the 24-hour period. However, as shown in prior studies, LH TABLE 1 Characteristics of 10 premenopausal women who participated in the study. Leptin (ng/ml) Subject no. Age (y) Age at menarche (y) Fat mass (%) BMI (kg/m 2 ) Nadir Peak Time of peak :15 PM :45 AM :15 AM :45 PM :20 AM :55 PM :05 PM :30 PM :45 AM :10 AM Note: BMI ¼ body mass index. Fenichel. Leptin, LH, and metabolic rate. Fertil Steril Fertility and Sterility â 1163

4 FIGURE 1 Frequently sampled LH, leptin, and metabolic rate (MR) levels for three representative subjects. Arrows represent meal times. Horizontal bars represent sleep time. Fenichel. Leptin, LH, and metabolic rate. Fertil Steril pulses were increased in amplitude and slowed after the leptin peak (Table 2). When averaged across all subjects, the number of LH pulses slowed from a peak of pulses during one 6-hour period ( 12 to 6 hours) to a low of pulses after the leptin peak (P¼.004), with a significant quadratic relationship (P<.001). At the same time, the LH pulse amplitude increased to a high of miu/ml after the leptin peak (P¼.015), with a significant linear trend (P¼.011). Changes in the LH interpulse interval were also analyzed before and after the leptin peak, with a significant increase in the interpulse interval from minutes to minutes (P¼.009) (Table 2). Therefore the leptin peak is followed by a dramatic increase in LH amplitude and a decrease in the number of LH pulses, as well as an increase in the LH interpulse interval. These changes are demonstrated in Fig. 1. Metabolic rate was measured every 2 minutes. Rates were averaged for each subject during each 6-hour period and then combined to produce the results seen in Table 2. A graph showing leptin and metabolic rate averages for each time period is shown in Fig. 2. Metabolic rate begins to drop significantly approximately 4 6 hours before leptin levels peak, from kcal/min to kcal/min after the leptin peak (P¼.004). Leptin, metabolic rate, and LH levels were compared 2 hours before and 2 hours after both dinner and lunch to assess for changes related to meals. Leptin levels 2 hours before and after lunch showed a small but significant increase (P<.001), whereas changes before and after dinner showed a trend toward significance (P¼.053). Luteinizing hormone interpulse interval showed an increase during the time periods before and after dinner (P<.001). Parameters were compared during sleep time and awake time. Although the number of LH pulses was lower during sleep time compared with awake time, no significant differences were found in LH interpulse interval during this time. Lastly, diurnal variation in leptin did not definitively 1164 Fenichel et al. Leptin, LH, and metabolic rate Vol. 90, No. 4, October 2008

5 TABLE 2 Results per 6-hour time segment calculated relative to each subject s individual leptin peak and then averaged across subjects. Time segment Variable One (L18 to L12 h) Two (L12 to L6 h) Three (L6 to0h) Four (0 to D6 h) Leptin (ng/ml) a Leptin change from baseline (%) LH (miu/ml) LH pulses per time segment b,c LH amplitude d,e LH interpulse interval (before and after leptin peak) (min) f Metabolic rate (kcal/min) g Note: Values are mean SE. Leptin peak occurs between times segments 3 and 4. a P¼.044 for segment 1 vs. 2, P¼.015 for segment 1 vs. 3, P¼.005 for segment 1 vs. 4. b P¼.011 for segment 1 vs. 2, P¼.004 for segment 2 vs. 4, P¼.001 for segment 3 vs. 4. c Significant quadratic relationship between all segments (P<.001). d P¼.011 for segment 2 vs. 4. e Significant linear trend between segments, P¼.011. f P¼.009 for before leptin peak vs. after leptin peak. g P¼.01 for segment 1 vs. 4, P¼.012 for segment 2 vs. 3, P<.001 for segment 2 vs. 4, P¼.001 for segment 3 vs. 4. Fenichel. Leptin, LH, and metabolic rate. Fertil Steril correlate with resting metabolic rate according to crosscorrelation analysis; however, 6 of 10 subjects showed a significant negative correlation for a time lag between 2 to 5 hours, with a drop in metabolic rate preceding a rise in leptin. This may represent a period of time between diurnal peaks of these two variables. DISCUSSION This study is the first to report a 24-hour study of leptin, LH, and metabolic rate in a group of normal, menstruating women. The nightly leptin peak is associated with a decrease in LH pulse frequency and an increase in amplitude of LH pulses, and these marked acute changes in the character of the LH pulses are accompanied by a drop in metabolic rate. The data we collected support the previous work of Licinio et al. (1), showing that leptin has a diurnal rhythm in ovulatory women, peaking roughly between 12 AM and 4 AM. In all cases, as in our studies, a change in LH pulsatility followed. Data in the literature concerning the relationship between leptin and metabolic rate are unclear. Several studies have shown that leptin independently predicts resting metabolic rate (19, 20), even after controlling for fat-mass and fatfree mass, the two strongest correlates of leptin and resting metabolic rate, respectively. Some studies have shown an inverse relationship between leptin and resting metabolic rate; however, the majority of studies to date compare values in groups and across subjects (21, 22), and our data suggest a wide range of normal values in the subject population. In a study that looked at patients over time during weight loss or weight gain, neither absolute nor relative changes in leptin levels correlated with changes in energy expenditure (23). Leptin levels are strongly related to body fat mass, with increased levels in obese patients and decreased levels in patients in energy-deficient states (whether due to low energy consumption or high expenditure). However, leptin levels have also been shown to drop acutely during starvation, long before any change in body fat mass could be attained (24). It is interesting that, in our subjects, a decrease in metabolic rate clearly preceded both the leptin peak and any change in LH pulsatility. This suggests that a short-term metabolic signal may alter leptin secretion. In vitro studies have shown increased uptake and utilization of glucose by adipocytes to correlate with increased leptin messenger RNA expression, whereas blocking glucose oxidation lowers messenger RNA expression (25). Similarly, basal leptin secretion by adipocytes seems to depend on the availability and metabolism of glycolytic substrates, a process that is enhanced by glucose and insulin, with an effect seen as rapidly as 2 4 hours (26). Although resting metabolic rate reflects the metabolism of oxidative fuels in fat-free mass, it may be that the same phenomena is occurring in adipocytes and determining leptin levels. Thus, a short-term signal such as caloric deficit may affect leptin levels by lowering availability of oxidative fuels, which in turn may function as the signal to the reproductive axis of low energy states reflected in changes in LH pulsatility. Elucidating the specific signal molecules that Fertility and Sterility â 1165

6 FIGURE 2 Mean metabolic rate and leptin levels over 24 hours. Fenichel. Leptin, LH, and metabolic rate. Fertil Steril may directly link changes in metabolism with changes in reproductive function is an important area of future research. Finally, given that metabolic rate falls before leptin levels peak, we suggest that it is unlikely that changes in leptin levels lead to changes in metabolic rate, as has been proposed. It has long been suspected that there is an energy threshold required for reproduction. We suggest that in normal subjects baseline levels of leptin are most likely related to fat stores, whereas the 24-hour, minute-to-minute leptin secretion may depend on availability of oxidative fuels, which also determines metabolic rate. We also suggest that, in this system, leptin levels gradually peak and then fall, which may affect the GnRH pulse generator and in turn LH pulses. Together, these effects may create a daily metabolic cycle, as seen in our subjects. Studies have shown that leptin levels tend to rise approximately 3 6 hours after a given meal (27, 28). In addition, several studies in normal, nonfasting women show a consistent rise in leptin starting after the first meal of the day and increasing toward a nightly peak (1, 11, 29). Studies in normal, healthy men have also shown that delaying breakfast by several hours resulted in leptin levels that continued to decline until subjects were allowed to eat (30). After allowing the first meal, leptin levels began to rise again. The fall in leptin, which occurs consistently at night, likely follows a nadir in available oxidative fuels. These patterns may be altered in women with hypothalamic amenorrhea who have lower leptin levels due to low fat stores and show patterns of restrictive eating, thereby affecting the leptin peak and blunting its diurnal variation. Other clinical studies support this hypothesis. Body fat is a prime determinant of leptin levels. In women with low BMI, low leptin levels, and hypothalamic amenorrhea, leptin administration will restore LH pulses (3). Caloric restriction also decreases leptin secretion and suppresses leptin s diurnal rhythm (12). The nocturnal slowing of LH pulses has been described since 1973 (31), but a physiologic basis was recently proposed by Licinio relating this slowing to a nocturnal leptin peak (1). Multiple studies of the nutritionally related model suggest that the secretion of these two hormones is interrelated, and administration of leptin recently restored LH pulsatility in some subjects. We suggest that normal caloric intake and metabolic rate may also influence nocturnal leptin dynamics. Alternatively, metabolic rate or another determinant of metabolic rate could be independently affecting both leptin levels and LH pulsatility. There is evidence that sleep stage is also associated with changes in LH pulsatility (32); however, it is notable that these studies did not measure leptin levels or mealtimes. Our data do not support sleep stage as the only variable associated with changes in LH pulsatility because there were no significant differences in interpulse interval comparing sleep time with wake time. Our data support the idea that the slowing of LH pulses may be more related to the leptin peak. Licinio et al. (1) showed that LH pulse parameters including pulse amplitude, width, and interpulse interval were 1166 Fenichel et al. Leptin, LH, and metabolic rate Vol. 90, No. 4, October 2008

7 significantly different during daytime hours when leptin levels were lower than at night time when leptin levels were higher. In addition, they showed that the nocturnal rise in leptin was associated with significant changes in the profile of LH pulses. Our data suggest that the nightly leptin peak is most closely related to overnight changes in LH pulsatility. Our study has several limitations. Our sample size is small, and there was a large amount of variability in both leptin levels and metabolic rate. However, this suggests a great deal of variability within the normal population. Interestingly, our data support the observation that there is a great deal of variability in leptin levels between normal-weight subjects. Many factors contribute to leptin secretion, including acute caloric intake, fat mass, and BMI (28, 30, 33). Gender, adipocyte size, fat distribution, insulin levels, glucocorticoids, and cytokines (such as tumor necrosis factor-a and interleukin-1) may affect leptin levels as well (34). As described above, 1 of our subjects had very high baseline and peak leptin levels; therefore, there may be yet-undetermined factors that affect daily leptin secretion. Ideally, our data at 0 hours and 24 hours would be comparable; variation in this regard may be due to unavoidable but different conditions during data collection for hormonal samples and metabolic rate, and the experimental conditions. Although our patients were not recruited on the basis of BMI, most of our patients were at the lower end of the normal BMI range, which may affect interpretation or extrapolation of the data. However, it is important to note that all of our subjects had appropriate caloric intake and no history of disordered eating, indicating a normal energy balance. Although the exact nature of the signal linking energy homeostasis with reproduction is still unknown, it is possible that leptin plays a key role. We offer further evidence of a temporal relationship between metabolic rate, leptin levels, and LH pulsatility and propose that there is a daily metabolic cycle that reflects the availability of metabolic fuels and that may be disrupted when energy availability is low. REFERENCES 1. Licinio J, Negrao AB, Mantzoros C, Kaklamani V, Wong ML, Bongiorno PB, et al. Synchronicity of frequently sampled, 24-h concentrations of circulating leptin, luteinizing hormone, and estradiol in healthy women. Proc Natl Acad Sci U S A 1998;95: Zamorano PL, Mahesh VB, De Sevilla LM, Chorich LP, Bhat GK, Brann DW. Expression and localization of the leptin receptor in endocrine and neuroendocrine tissues of the rat. Neuroendocrinology 1997;65: Welt CK, Chan JL, Bullen J, Murphy R, Smith P, DePaoli AM, et al. Recombinant human leptin in women with hypothalamic amenorrhea. N Engl J Med 2004;351: Warren MP, Voussoughian F, Geer EB, Hyle EP, Adberg CL, Ramos RH. Functional hypothalamic amenorrhea: hypoleptinemia and disordered eating. J Clin Endocrinol Metab 1999;84: Laughlin GA, Yen SS. Hypoleptinemia in women athletes: absence of a diurnal rhythm with amenorrhea. J Clin Endocrinol Metab 1997;82: Polito A, Fabbri A, Ferro-Luzzi A, Cuzzolaro M, Censi L, Ciarapica D, et al. Basal metabolic rate in anorexia nervosa: relation to body composition and leptin concentrations. Am J Clin Nutr 2000;71: Elimam A, Marcus C. Meal timing, fasting and glucocorticoids interplay in serum leptin concentrations and diurnal profile. Eur J Endocrinol 2002;147: Kolaczynski JW, Considine RV, Ohannesian J, Marco C, Opentanova I, Nyce MR, et al. Responses of leptin to short-term fasting and refeeding in humans: a link with ketogenesis but not ketones themselves. Diabetes 1996;45: Mars M, de Graaf C, de Groot LC, Kok FJ. Decreases in fasting leptin and insulin concentrations after acute energy restriction and subsequent compensation in food intake. Am J Clin Nutr 2005;81: Mars M, de Graaf C, de Groot CP, van Rossum CT, Kok FJ. Fasting leptin and appetite responses induced by a 4-day 65%-energy-restricted diet. Int J Obes (Lond) 2006;30: Weigle DS, Duell PB, Connor WE, Steiner RA, Soules MR, Kuijper JL. Effect of fasting, refeeding, and dietary fat restriction on plasma leptin levels. J Clin Endocrinol Metab 1997;82: Hilton LK, Loucks AB. Low energy availability, not exercise stress, suppresses the diurnal rhythm of leptin in healthy young women. Am J Physiol Endocrinol Metab 2000;278:E Loucks AB, Thuma JR. Luteinizing hormone pulsatility is disrupted at a threshold of energy availability in regularly menstruating women. J Clin Endocrinol Metab 2003;88: Garner DM, Olmstead MP, Polivy J. Development and validation of a multidimensional eating disorder inventory for anorexia nervosa and bulimia. Int J Eating Disord 1983;2: Mazess RB, Barden HS, Bisek JP, Hanson J. Dual-energy x-ray absorptiometry for total-body and regional bone-mineral and soft-tissue composition. Am J Clin Nutr 1990;51: Veldhuis JD, Johnson ML. Cluster analysis: a simple, versatile, and robust algorithm for endocrine pulse detection. Am J Physiol 1986;250: E Urban RJ, Johnson ML, Veldhuis JD. Biophysical modeling of sensitivity and positive accuracy of detecting episodic endocrine signals. Am J Physiol 1989;257:E Urban RJ, Johnson ML, Veldhuis JD. In vivo biological validation and biophysical modeling of the sensitivity and positive accuracy of endocrine peak detection. I. The LH pulse signal. Endocrinology 1989;124: Toth MJ, Sites CK, Poehlman ET. Hormonal and physiological correlates of energy expenditure and substrate oxidation in middle-aged, premenopausal women. J Clin Endocrinol Metab 1999;84: Wauters M, Considine RV, Chagnon M, Mertens I, Rankinen T, Bouchard C, et al. Leptin levels, leptin receptor gene polymorphisms, and energy metabolism in women. Obes Res 2002;10: Soares MJ, Piers LS, O Dea K, Collier GR. Plasma leptin concentrations, basal metabolic rates and respiratory quotients in young and older adults. Int J Obes Relat Metab Disord 2000;24: Niskanen L, Haffner S, Karhunen LJ, Turpeinen AK, Miettinen H, Uusitupa MI. Serum leptin in relation to resting energy expenditure and fuel metabolism in obese subjects. Int J Obes Relat Metab Disord 1997;21: Rosenbaum M, Nicolson M, Hirsch J, Murphy E, Chu F, Leibel RL. Effects of weight change on plasma leptin concentrations and energy expenditure. J Clin Endocrinol Metab 1997;82: Chan JL, Heist K, DePaoli AM, Veldhuis JD, Mantzoros CS. The role of falling leptin levels in the neuroendocrine and metabolic adaptation to short-term starvation in healthy men. J Clin Invest 2003;111: Mueller WM, Gregoire FM, Stanhope KL, Mobbs CV, Mizuno TM, Warden CH, et al. Evidence that glucose metabolism regulates leptin secretion from cultured rat adipocytes. Endocrinology 1998;139: Cammisotto PG, Gelinas Y, Deshaies Y, Bukowiecki LJ. Regulation of leptin secretion from white adipocytes by insulin, glycolytic substrates, and amino acids. Am J Physiol Endocrinol Metab 2005;289: E Romon M, Lebel P, Fruchart JC, Dallongeville J. Postprandial leptin response to carbohydrate and fat meals in obese women. J Am Coll Nutr 2003;22: Fertility and Sterility â 1167

8 28. Saad MF, Riad-Gabriel MG, Khan A, Sharma A, Michael R, Jinagouda SD. Diurnal and ultradian rhythmicity of plasma leptin: effects of gender and adiposity. J Clin Endocrinol Metab 1998;83: Taylor AE, Hubbard J, Anderson EJ. Impact of binge eating on metabolic and leptin dynamics in normal young women. J Clin Endocrinol Metab 1999;84: Schoeller DA, Cella LK, Sinha MK, Caro JF. Entrainment of the diurnal rhythm of plasma leptin to meal timing. J Clin Invest 1997;100: Kapen S, Boyar R, Hellman L, Weitzman ED. Episodic release of luteinizing hormone at mid-menstrual cycle in normal adult women. J Clin Endocrinol Metab 1973;36: Hall JE, Sullivan JP, Richardson GS. Brief wake episodes modulate sleep-inhibited luteinizing hormone secretion in the early follicular phase. J Clin Endocrinol Metab 2005;90: 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: Chan JL, Mantzoros CS. Role of leptin in energy-deprivation states: normal human physiology and clinical implications for hypothalamic amenorrhoea and anorexia nervosa. Lancet 2005;366: Fenichel et al. Leptin, LH, and metabolic rate Vol. 90, No. 4, October 2008