| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Clinical Research Centers |
Reproductive Endocrine Unit and National Center for Infertility Research (A.E.T.), and the Mallinkrodt General Clinical Research Center (J.H., E.J.A.), Massachusetts General Hospital, Boston, Massachusetts 02114
Address all correspondence and requests for reprints to: Ann E. Taylor, M.D., Reproductive Endocrine Unit, Massachusetts General Hospital, BHX-5, 55 Fruit Street, Boston, Massachusetts 02114. E-mail aetaylor{at}partners.org
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Seven lean women without a history of eating disorders were studied on two occasions separated by one or two menstrual cycles. During one admission, they ate three regular meals plus a snack on each of 3 days. On the other admission, they ate the same number of calories, macronutrient matched to the normal diet, in a single evening meal. Glucose, insulin, and leptin were measured frequently for 12–14 h beginning at 0800 h on the third day of each diet, and an insulin tolerance test was performed while the subjects were fasting on the fourth day. Daily blood samples were obtained until ovulation was documented to assess any impact on menstrual function.
Ingestion of an entire day’s calories at dinner resulted in a significant increase in fasting glucose levels and a dramatic increase in insulin responses to the evening meal. The diurnal pattern of leptin secretion was altered, such that the gradual rise in leptin from 0800 h observed during the normal diet was abolished, and leptin did not begin to rise during the binge diet until at least 2 h after the evening meal. No changes were demonstrated in insulin sensitivity, follicular growth, or ovulation between the two diets.
We conclude that 1) ingestion of a large number of calories at one time (binge eating) impacts metabolic parameters even when total calories and macronutrients are appropriate for weight; 2) the timing of energy intake is an independent determinant of the diurnal rhythm of leptin secretion, indicating a relatively acute affect of energy balance on leptin dynamics; 3) the mechanism of exaggerated insulin secretion after a binge meal remains to be determined, but may be related to the altered diurnal pattern of leptin secretion; and 4) as most binge eating episodes in the population are associated with the ingestion of excess calories, it is hypothesized that binge eating behavior is associated with even greater metabolic dysfunction than that described herein.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
There are several potential mechanisms by which binge eating may cause or contribute to obesity. First, frequent binge meals would result in excess caloric ingestion, but a compensatory change in basal energy expenditure is not observed (11). Second, even if total calories were normal, the ingestion of fewer large meals may be metabolically worse than the ingestion of frequent small meals. Although the ingestion of six or nine small meals in a day has previously been demonstrated to reduce insulin secretion compared to the ingestion of three calorie- and nutrient-matched meals (12, 13, 14), the effects of a single daily meal, as is frequently observed in bingers and in many allegedly healthy Americans, has not previously been determined. If insulin levels are further increased by bingeing, they may contribute to increased hunger (15).
Finally, recent studies suggest that the adipocyte hormone leptin could modify the response to binge eating. Serum leptin levels demonstrate a diurnal rhythm with a peak in the middle of the night (16, 17) when most humans are not eating and a nadir first thing in the morning. Fasting results in a gradual decline in leptin levels, first observed after 6–8 h (18). Thus, modulation of the diurnal pattern of leptin secretion by the timing of meals might have important implications for modulation of hunger.
In addition to potential metabolic effects, binge eating is hypothesized to have effects on the hypothalamic-pituitary-gonadal axis. For example, an increase in serum insulin levels may increase ovarian androgen secretion (19, 20) and/or modify pituitary gonadotropin secretion (21), thereby disrupting menstrual cyclicity. In addition, abolition of the diurnal pattern of leptin secretion is associated with amenorrhea in women athletes (22), suggesting that the leptin pattern is either a marker for or a response to a normal reproductive hormone axis.
This study was designed to determine whether binge eating has adverse hormonal and metabolic effects. To study the effects of large meals per se independent of other potentially negative behaviors associated with bingeing, we studied lean normal women with regular ovulatory menstrual cycles and no history of eating disorders, who received meals that were matched for calorie and fat content.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Seven lean normal Caucasian women with regular menstrual cycles participated in this study. Subjects were healthy with no medical illnesses and were taking no medications. Each had a history of regular 26- to 32-day menstrual cycles. Ovulation was documented in the luteal phase before the study by a serum progesterone level greater than 6 ng/dL. Subjects had no history of eating disorders and no evidence of a current eating disorder based on the BITE Edinburgh Bulimic Investigation Test questionnaire (23, 24). All subjects had normal TSH and PRL levels. The protocol was approved by the Massachusetts General Hospital Subcommittee on Human Studies, and all subjects gave their informed consent.
Protocol
Each subject was admitted to the Massachusetts General Hospital twice between 0800–1000 h in the early to midfollicular phase (cycle days 1–7) during two different menstrual cycles separated by less than 3 months. Each admission lasted for 3 days, during which all caloric ingestion was monitored and controlled.
During one admission, subjects received three meals a day and an afternoon snack, with 20% of calories at 0800 h for breakfast, 30% of calories at noon for lunch, 10% of calories at 1500 h for a snack, and the remaining 40% of calories at 1800 h for dinner. During the other admission, in random order, subjects received 100% of calories at 1800 h, with nothing but water and noncaloric, noncaffeinated beverages the rest of the day. For both diets, total calories were individually adjusted to maintain weight and were matched between studies of the same individual, and macronutrients were distributed as 50–55% carbohydrate, 35–40% fat, and 10–15% protein.
Twenty-four-hour collections of urine were obtained each day for measurements of creatinine, cortisol, and catecholamines (epinephrine, norepinephrine, and dopamine) as monitors of stress during the admissions.
On the third day (after 48 h of the test diet), each subject had blood sampled from 0800 h every 15 min for insulin and leptin and every 30 min for glucose. The first three women underwent 12 h of frequent sampling until 2000 h, but the sampling window was extended to 2200 h (14 h) for the remainder of the subjects, as glucose levels had not fallen to baseline by 2000 h after the 1800 h meal. At 0800 h on the fourth day, fasting blood was obtained for serum insulin, glucose, lipid, and androgen determinations before an insulin tolerance test. Regular human insulin (0.1 U/kg) was administered iv, and blood samples were obtained at -5, 0, 3, 5, 7, 10, and 15 min for measurement of glucose (25, 26). The insulin tolerance test was terminated by administration of iv glucose and/or the immediate ingestion of breakfast before discharge.
After each discharge, each subject obtained a daily blood sample until ovulation was documented by urinary LH kits (Sure Step Ovulation Predictor, Applied Biotech, Inc., San Diego, CA). The ovulation date of the first cycle was used to predict menses and schedule the second admission on a similar cycle day as the first. Subjects were requested to take oral iron replacement (ferrous gluconate, 300 mg daily or twice daily) from the time of acceptance into the study until completion of the menstrual cycle of the second study.
Assays
Glucose was measured by the glucose hexokinase method. Insulin was measured by an automated immunometric assay with a sensitivity of 1.0 µIU/L and intra- and interassay coefficients of variation of 3% and 6–12%, respectively. Leptin was measured by a RIA (Linco Research, Inc. St. Charles, MO) with a sensitivity of 0.5 ng/mL and intra- and interassay coefficients of variation of 4.5–6% and 6.1–8.5%, respectively. Urinary free cortisol was measured in a coated tube RIA (Incstar Corp., Stillwater, MN) after methylene chloride extraction. Urinary catecholamines were measured by Nichols Institute Diagnostics (San Juan Capistrano, CA). LH, FSH, estradiol, progesterone, testosterone, androstenedione, 17-hydroxyprogesterone, and dehydroepiandrosterone sulfate were measured by in-house RIAs as previously described (27). Cholesterol was measured by peroxidase-phenol-4-amino phenazone modified Trinder method, triglycerides by peroxidase enzymatic reaction, and high density lipoprotein and low density lipoprotein were measured by an enzymatic, colorimetric, nonprecipitate method. All lipids were measured using Boehringer Mannheim reagents (Indianapolis, IN) on a Hitachi Scientific Instruments 917 machine (Hialeah, FL).
Statistical analysis
Fasting glucose, insulin, and leptin on day 3 and fasting glucose, insulin, androgens, and lipids on day 4 were compared between diets by paired t testing. Mean levels of glucose, insulin, and leptin for the frequent sampling sessions were determined by averaging the individual samples and were compared by paired t testing. Similarly, mean levels of glucose and insulin were determined and compared for the window after the evening meal (1815 to 2000 or 2000 h). Lastly, glucose, insulin, and leptin levels across the day were compared by ANOVA for repeated measures.
Insulin sensitivity was estimated from the glucose disappearance rate (Kitt), which was calculated from the slope of the fall in glucose after insulin administration (26). The day of the LH surge was determined by using three of the following four criteria: the day of the peak LH, the day of the peak FSH, the day of or the day after the estradiol peak, and the day that progesterone doubled or exceeded 0.6 ng/mL (28). The day of the LH surge and the day of onset of menses were used to determine follicular and luteal phase lengths for each subject. LH, FSH, and estradiol levels were centered to the day of ovulation for each subject and meaned between subjects.
In addition, classic statistical methods for cross-over study designs were employed to demonstrate that there was no effect of diet order on any of the variables tested (29).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The mean subject age was 28.1 ± 3.0 yr (range, 19–40 yr) and the mean body mass index was 22.1 ± 0.9 kg/m2 (range, 20.5–25.9). The mean height was 162.2 cm (range, 150–170.5). During the normal diet, the mean body weight was 57.6 ± 1.0 kg on day 1 and 56.6 ± 1.0 kg on day 4 (P < 0.01, by paired t test). On the binge diet, the mean weight was 57.1 ± 1.1 kg on day 1 and 56.6 ± 1.0 kg on day 4 (P = 0.18). However, there was no significant difference in weights on the last day of the two diets.
Metabolic characteristics
Mean glucose and insulin levels over the 12 h of frequent
sampling on the third day of the diets are shown for the normal and
binge diets in Fig. 1
. On the normal
diet, glucose and insulin rose as expected after each meal and
gradually declined after dinner. On the binge diet, glucose and insulin
levels fell gradually across the day, rising after the evening
meal.
|
Fasting insulin levels were not significantly different on either the third or fourth day of either diet (normal, 3.5 ± 0.8; binge, 3.6 ± 1.1 µU/mL; P = 0.89). Similarly, the mean insulin levels from the entire frequent sampling study did not differ between the diets (normal, 13.0 ± 2.1; binge, 13.1 ± 2.4 µU/mL; P = 0.94). However, the insulin response to the evening meal (sum of insulin measurements from 1815–2200 h) was strikingly increased after the binge dinner (normal, 258.2 ± 49; binge, 559.0 ± 122.4 µU/mL; P = 0.01) despite virtually identical glucose responses.
The glucose responses to the insulin tolerance test were similar on the fourth day of the two diets, suggesting that there was no difference in insulin sensitivity. The Kitt was -3.2 ± 0.2 U after the normal diet and -3.6 ± 0.4 after the binge diet (P = 0.145).
The binge diet did not result in any significant difference in fasting
lipid levels by the fourth day of the diet (Table 1).
|
The two diets resulted in strikingly different patterns of leptin
secretion across the third day (Fig. 2).
In each individual, serum leptin levels were consistently higher than
baseline by 8 h after the initiation of breakfast on the normal
diet and continued to rise at the end of the frequent sampling (Fig. 3). Conversely, during the binge diet,
leptin levels continued to fall throughout the day and had not begun to
rise by 2200 h. Thus, although there was no significant difference
in fasting levels (normal, 9.9 ± 1.7; binge, 11.6 ± 2.2
ng/mL; P = 0.29) between the two diets, the mean leptin
levels across the third day were significantly lower on the binge diet
(normal, 11.3 ± 1.8; binge, 8.9 ± 1.6 ng/mL;
P = 0.04). From 1800–2200 h, leptin averaged 45%
lower on the binge diet than on the normal diet (P <
0.005).
|
|
)
demonstrate a pulsatile pattern of leptin secretion in all
subjects. Menstrual cyclicity
There was no significant difference in follicular or luteal phase
length after 3 days of either diet (Table 1). After centering daily
hormone levels to the day of ovulation, the patterns of gonadotropin
and sex steroid secretion were indistinguishable after 3 days of either
diet (data not shown). Finally, there was no difference in day 4
testosterone, 17-hydroxyprogesterone, androstenedione, or
dehydroepiandrosterone sulfate (Table 1) between the two diets.
Urine collections
The data for urinary catecholamines, cortisol, creatinine, and
volume are shown in Table 2. Although
there was no difference in urine volume, creatinine, or catecholamine
excretion between the first and second admissions, there was a small
effect of admission order on total urinary free cortisol, such that
cortisol excretion was higher during the first admission (52 ± 9
vs. 43 ± 6 µg/24 h for admission 1 vs. 2,
respectively; P = 0.03). Although urine volume and
creatinine and cortisol excretion were similar during the two diets,
dopamine excretion was significantly greater during the binge diet
(normal, 258 ± 33; binge, 397 ± 63 µg/24 h;
P < 0.001).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Previous studies have demonstrated that the ingestion of three meals a day results in higher total insulin levels than does the ingestion of six or nine small meals across the day (12, 13, 14). The current study, even given the limitation of sampling only for 12–14 h, demonstrates that this trend extends to even a single daily meal. The mechanism of hyperinsulinemia after the binge meal in these normal subjects remains to be determined. Although the doubling of insulin levels in response to similar glucose levels raises the possibility that there is insulin resistance, the insulin tolerance test the next morning does not confirm this. However, we cannot exclude the possibility that there is acute insulin resistance after the evening meal that is resolved by the overnight fast. In addition, the insulin tolerance test may not be a sufficiently sensitive test to identify significant changes in insulin sensitivity in this normal population.
In these studies, the increased insulin response to the binge dinner occurs without an increased glucose response compared to the normal dinner, raising the possibility that the observed hyperinsulinemia does not represent insulin resistance but, rather, an increase in other insulin secretagogues or a decrease in insulin suppressers. Glucagon-like peptide-1 (GLP-1) is a physiologically important regulator of insulin secretion that is itself secreted from the pancreas in response to oral, but not iv, glucose in a dose-dependent manner (30). Administration of increasing doses of exogenous GLP-1 stimulates increased insulin secretion up to 40% over the baseline (31, 32). Thus, increased GLP-1 secretion may explain part of the increased insulin response to the binge meal.
Leptin may also influence insulin secretion. In ambient glucose conditions, physiological leptin administration suppresses insulin secretion from rat islet cells or a human islet cell line (33). During the binge diet, leptin levels were 45% lower at the time of the evening meal compared to those during the normal diet. Thus, decreased leptin may also contribute to the increased insulin response to the binge meal.
A diurnal rhythm of leptin has previously been described (16) and has been postulated by some to be related to the cortisol diurnal rhythm (34, 35, 36). Our studies demonstrate that the pattern of food ingestion itself is an important factor controlling the leptin secretion profile, even when total 24-h cortisol excretion is constant. The current study design could not detect any potential change in the cortisol diurnal rhythm, however. In this study in which meal timing has been carefully controlled for 3 days, leptin secretion begins to increase approximately 6 h after food ingestion, a more acute response than may have been appreciated in previous studies. Careful circadian studies in which sleep, position, and caloric intake are constant for at least 24 h will be required to determine whether the entire diurnal rhythm of leptin is merely secondary to the pattern of food ingestion or whether there is also an underlying circadian rhythm that might reflect other circadian neuroendocrine rhythms.
In addition, it will be critical to determine whether the leptin diurnal rhythm can be modulated by the timing of food intake, as opposed to the amount, as recent studies of women athletes demonstrate that a flattened leptin diurnal rhythm is one of the few parameters that distinguish amenorrheic from eumenorrheic subjects (22). Such studies suggest that the pattern of leptin secretion across the day may have profound reproductive implications. Preliminary data from other studies of similar subjects in whom blood sampling continues for a full 24 h (Taylor, A. E., unpublished) suggest that a leptin diurnal rhythm is still maintained, although phase shifted, during binge eating. This is consistent with recent studies demonstrating a change in the leptin diurnal rhythm in response to simulated jet lag, even before the cortisol rhythm changed significantly (37). Thus, it may not be surprising that 3 days of binge eating were insufficient to induce a significant change in follicular or luteal phase length in the lean healthy young women. Whether modification of the pattern of caloric intake, with or without caloric restriction, will induce a physiologically significant flattening of the leptin diurnal rhythm remains to be determined.
Our studies confirm the pulsatile nature of leptin secretion (34, 38),
which is visible even when serum is sampled as infrequently as every 15
min. However, recent reports that leptin pulses typically occur
approximately every 30 min (38) suggests that the sampling frequency
for detecting the true frequency of pulsatile leptin secretion should
be at least once every 5 min. This estimate is based on previous
evidence that sampling frequency should be 6 times the interpulse
interval (39). Because the sampling protocol employed in this study was
suboptimal, formal pulse analysis was not performed. However, by visual
inspection of Fig. 3, it appears that leptin pulsatile secretion is
modulated by dietary intake, with an increase in pulse amplitude as
overall leptin levels are also rising. The use of a more frequent
sampling regimen will be required to determine whether pulse frequency
changes with meals.
The urine measurements of volume, creatinine, cortisol, and catecholamines were intended as a control for stress responses and metabolic balance during the 3 days of hospitalization and dietary control. As the development of obesity has been correlated with decreased sympathetic outflow in many experimental models (40), it was critical to demonstrate that bingeing did not exert its effects through modulation of epinephrine and norepinephrine excretion. Although there was a statistically significant increase in cortisol excretion during the first admission, the absolute magnitude was small and well within the normal range, making it unlikely to be physiologically relevant. Importantly, there was no difference in cortisol excretion between the two diets.
However, there was a striking increase in urinary dopamine excretion during the binge diet. Increased dopamine excretion has been associated with insulin resistance (41) and hypertension (42) in different populations, suggesting that it may be a true marker of an adverse outcome of binge eating. As the urine collections were for the full 24 h, whereas blood sampling only lasted for 12–14 h, the increased urinary dopamine secretion with the binge diet suggests that the exaggerated insulin secretion persisted throughout the night.
In summary, the ingestion of a single large daily meal has metabolic consequences, including increased fasting glucose levels, increased insulin responses to the binge meal, and modulation of the leptin diurnal rhythm. These factors, demonstrated here in young healthy lean women, are likely to be exacerbated in individuals who are already obese or insulin resistant. Because hyperinsulinemia has been associated with many adverse health outcomes, including an increased risk of cardiovascular disease, hypertension, and dyslipidemia, further studies are required to determine the long term consequences of the common practice of eating large evening meals.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received September 14, 1998.
Revised November 4, 1998.
Accepted November 10, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. Geliebter, C. N. Ochner, and R. Aviram-Friedman Appetite-Related Gut Peptides in Obesity and Binge Eating Disorder American Journal of Lifestyle Medicine, July 1, 2008; 2(4): 305 - 314. [Abstract] [PDF]
|
|
|
|
 |
|
 P. Monteleone, M. Fabrazzo, V. Martiadis, A. Fuschino, C. Serritella, N. Milici, and M. Maj Opposite Changes in Circulating Adiponectin in Women With Bulimia Nervosa or Binge Eating Disorder J. Clin. Endocrinol. Metab., November 1, 2003; 88(11): 5387 - 5391. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Kilciler, M. Ozata, C. Oktenli, S.Y. Sanisoglu, E. Bolu, N. Bingol, M. Kilciler, I. C. Ozdemir, and M. Kutlu Diurnal Leptin Secretion Is Intact in Male Hypogonadotropic Hypogonadism and Is Not Influenced by Exogenous Gonadotropins J. Clin. Endocrinol. Metab., November 1, 2002; 87(11): 5023 - 5029. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Monteleone, V. Martiadis, B. Colurcio, and M. Maj Leptin Secretion Is Related to Chronicity and Severity of the Illness in Bulimia Nervosa Psychosom Med, November 1, 2002; 64(6): 874 - 879. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Gill, A. E. Taylor, K. A. Martin, C. K. Welt, J. M. Adams, and J. E. Hall Specific Factors Predict the Response to Pulsatile Gonadotropin-Releasing Hormone Therapy in Polycystic Ovarian Syndrome J. Clin. Endocrinol. Metab., June 1, 2001; 86(6): 2428 - 2436. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. ECOCHARD, H. MARRET, M. BARBATO, and H. BOEHRINGER Gonadotropin and Body Mass Index: High FSH Levels in Lean, Normally Cycling Women Obstet. Gynecol., July 1, 2000; 96(1): 8 - 12. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Monteleone, F. Bortolotti, M. Fabrazzo, A. La Rocca, A. Fuschino, and M. Maj Plasma Leptin Response to Acute Fasting and Refeeding in Untreated Women with Bulimia Nervosa J. Clin. Endocrinol. Metab., July 1, 2000; 85(7): 2499 - 2503. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
