Twenty-four-hour Ghrelin Is Elevated after Calorie Restriction and Exercise Training in Non-obese Women


108 Noll Laboratory, Department of Kinesiology, Penn State University, University Park, PA 16802. E-mail:


Objective: The purpose of this study was to determine whether chronic energy deficiency achieved with caloric restriction combined with exercise is associated with changes in the 24-hour profile of ghrelin in non-obese, pre-menopausal women.

Research Methods and Procedures: Twelve non-obese (BMI = 18 to 25 kg/m2), non-exercising women (age, 18 to 24 years) were randomly assigned to a non-exercising control group or a diet and exercise group. The 3-month diet and exercise intervention yielded a daily energy deficit of −45.7 ± 12.4%. Serial measurements were made of body composition, energy balance, and feelings of fullness. Repeated blood sampling over 24 hours to measure ghrelin occurred before and after the study.

Results: Significant decreases in body weight, body fat, and feelings of fullness were observed in only the energy-deficit group (p < 0.05); significant changes in the following ghrelin features were found in only the deficit group (p < 0.05): elevations in baseline (+353 ± 118 pg/mL), lunch peak (+370 ± 102 pg/mL), dinner peak (+438 ± 149 pg/mL), nocturnal rise (+269 ± 77 pg/mL), and nocturnal peak (+510 ± 143 pg/mL). In addition, we found a larger dinner decline (−197 ± 52 pg/mL) and negative correlations between changes in the ghrelin dinner profile and changes in body weight (R = 0.784), 24-hour intake (R = 0.67), energy deficiency (R = 0.762), and feelings of fullness (R = 0.648; p < 0.05).

Discussion: Changes in ghrelin concentrations across the day after weight loss are closely associated with other physiological adaptations to energy deficiency, further supporting the role of ghrelin as a countermeasure to restore energy balance.


Much attention has been directed toward ghrelin because of its orexigenic (1, 2, 3) and adipogenic (4, 5) properties. Ghrelin and ratings of hunger have been shown to rise coincident with voluntary meal requests in men and women (6). Ghrelin infusions have led to significant increases in food intake and hunger and/or appetite (1, 2, 3). Negative associations between fasting ghrelin and body weight, percent body fat, and BMI have been observed in studies examining obese and anorexic individuals (7, 8). Diet and exercise interventions designed to restore normal body weight in clinical conditions have produced significant changes in ghrelin in association with changes in body weight and BMI (9, 10, 11). These findings lend support for the role of circulating ghrelin in the acute and chronic regulation of energy balance.

Ghrelin is at trough levels at ∼6:00 am, exhibits meal-related rises and declines throughout the day, and reaches nocturnal peak levels between 1:00 and 2:00 am (9, 12, 13). However, in women with anorexia nervosa, the daily and nocturnal ghrelin levels are significantly elevated compared with healthy controls, and the meal responses are abolished (14). This elevation along with the lack of post-prandial suppression may be a mechanism for energy balance restoration. In another study of anorexic patients, the ghrelin nocturnal peak was initially higher but returned to normal after weight gain (9). On the other end of the energy spectrum, obese individuals display either lower (11) or abolished (15) ghrelin meal-related profiles. After a 6-month weight loss program, ghrelin concentrations significantly increased at every time-point throughout the 24-hour period, as well as over each meal-related pattern, and resembled levels of normal-weight, healthy controls (11). Although these studies show that ghrelin concentrations across the day may change in response to chronic changes in energy balance, studies of obese and anorexic individuals represent models of the end-points of adaptation to these chronic changes. The role of ghrelin in energy balance regulation under normal physiological conditions may differ from what is observed in obese and anorexic individuals.

In normal-weight individuals, both the meal response and nocturnal pattern are altered after acute energy deficit and surplus conditions (16, 17), which is consistent with the idea of a role for ghrelin in energy balance restoration. Whether alterations in the meal-related and nocturnal profiles of ghrelin are associated with long-term regulation of energy balance in normal-weight individuals remains unanswered.

In non-obese pre-menopausal women, we have previously shown that increases in fasting ghrelin were significantly correlated with the magnitude of weight loss produced by a 3-month supervised diet and exercise intervention (18), indicating a role for ghrelin in long-term energy homeostasis. We also examined acute changes in ghrelin over a typical day of eating in normal-weight sedentary women (19), showing that meal responses of ghrelin are significantly correlated with acute calorie intake. The latter studies and this study were part of a larger study on the endocrine and metabolic responses to chronic exercise training combined with caloric restriction in women. In this study, we extend our previous findings by examining the impact of the aforementioned diet and exercise intervention on the features of the 24-hour pattern of ghrelin. The purpose of this study was to determine whether changes in the meal-related and nocturnal profiles of ghrelin are associated with changes in body weight and other indices of energy balance in non-obese, pre-menopausal women.

Research Methods and Procedures

Subject Recruitment and Screening

This study was part of a larger prospective study designed to assess changes in endocrine and reproductive function in response to a controlled feeding and exercise intervention. The intervention was implemented in non-obese women to emulate exercise and restrictive eating patterns in which many young women engage. Subjects were recruited through newspaper advertisements and fliers. All study procedures were approved by the Pennsylvania State University Biomedical Institutional Review Board, and informed consent was obtained from all subjects. Demographic, medical history, menstrual history, physical activity, and eating attitudes questionnaires were completed, and body weight and height were measured.

Study inclusion was based on the following criteria: 1) 18 to 30 years of age; 2) body weight of 50 to 70 kg; 3) BMI of 18 to 25 kg/m2; 4) non-smoking; 5) no significant weight loss/gain ±2.3 kg in the 12 months; 6) <1 hour of purposeful aerobic exercise per week; 7) no hormonal contraceptives for the past 6 months; 8) not currently or previously diagnosed with disordered eating; 9) no medications that would interfere with measurement of hormones; 10) normal complete blood count and basic chemistry panel; and 11) suitable candidate for a feeding and exercise study.

Experimental Design

After the screening period and baseline monitoring, the subjects completed a 3-month diet and exercise intervention. The subjects who are part of this substudy had been randomly assigned to either a non-exercising control group, who consumed enough calories to maintain weight, or an energy deficit group, who experienced reductions in food calories and exercised to achieve a negative energy balance ranging from −30% to −60% compared with baseline energy needs, The diet and exercise and energy balance methods have been described previously (18). Before and after the diet and exercise intervention, body weight, body composition, aerobic fitness, and satiety were assessed. Twenty-four-hour blood sampling was performed before and after the intervention on Days 2 to 7 of the follicular phase. Ghrelin was measured in these samples. Thirteen subjects underwent baseline 24-hour blood sampling, but one did not complete the intervention, so 12 subjects (controls: n = 4; energy deficit: n = 8) completed all procedures.

Energy Balance Measurements

Dietary intake was controlled throughout the intervention (18). At baseline, each subject's daily energy requirement to maintain a stable body weight was determined through estimates of energy expenditure followed by a week-long “calibration” period allowing for adjustments in the dietary prescription if body weight changed. To estimate total energy expenditure, resting metabolic rate was measured using indirect calorimetry (18), and this value was added to the total calories expended during a 24-hour period as assessed by a triaxial accelerometer (RT3 accelerometer; Stayhealthy, Monrovia, CA) worn on the hip 24 h/d for 7 days at a time. The sum of calories from the resting metabolic rate and the RT3 accelerometer was operationally defined as the weight maintenance calorie level. This amount of calories was provided to the subjects each day for 7 days, and minor adjustments were made if body weight fluctuated during this time.

After the baseline period, the control group continued to consume the weight maintenance calorie level estimated from the 7-day calibration period (with minor adjustments if necessary). The energy deficit group was provided with fewer calories (mean ± standard error = −25 ± 8%) than those required to maintain initial body weight, and these subjects began to undergo supervised exercise training. All meals for both groups were made and provided at the metabolic kitchen in the General Clinical Research Center during the 3 intervention months. Each subject was required to eat at least two of the three weekday meals at this facility. One meal per day and weekend meals could be packed out. The diet was comprised of 55% carbohydrate, 30% fat, and 15% protein. Subjects were instructed to eat all of the food provided to them and only the food provided to them by the study. Any uneaten food was re-weighed and recorded for later subtraction from the prescribed intake total. Eating food not provided by the study was highly discouraged, but if this occurred, extra food was recorded on a log sheet, and calories and macronutrient composition were calculated using Nutritionist Pro (First Data Bank, Indianapolis, IN).

Exercise was supervised throughout the entire study (18). While the control group performed no exercise, the energy deficit group performed aerobic exercise five times per week at 70% to 80% of maximum heart rate as determined from tests of maximal aerobic capacity. Exercise duration was the number of minutes required to achieve the prescribed exercise expenditure calorie level (20 ± 2% of baseline energy expenditure). The total amount of calories expended during each exercise session was measured using the OwnCal feature on the Polar S610 heart rate monitor (Polar Electro Oy, Kempele, Finland). All training was supervised by personal trainers in a research training room. During supervised training sessions, the RT3 was not worn so as not to duplicate the measurement of calories assessed with the Polar heart rate monitor. To estimate total energy expended each day, the sum of resting metabolic rate (kilocalories per day), activity calories determined with the RT3, and exercise training calories as determined during exercise sessions with the Polar heart rate monitor was used. Energy balance was estimated daily by subtracting total energy expenditure from energy intake. Percent energy deficit was determined each day by multiplying the ratio of energy intake to energy expenditure by 100. Weekly averages of energy balance and energy deficit were calculated and monitored throughout the study. Adjustments were made in exercise calorie expenditure or in food intake calories if necessary to keep subjects in an energy deficit.

Fitness, Body Composition, and Satiety

Body composition and maximal aerobic capacity were determined during the before and after time-points (18, 20, 21). Satiety was assessed through perceptions of fullness collected before and after the breakfast meal collected 2 days during the first week of the intervention and 2 days during the last week of the intervention using a linear scale from 1 to 10: 1 being not full/not satisfied and 10 being full/satisfied.

Twenty-four-hour Repeated Blood Sampling

Subjects reported to the General Clinical Research Center at 7:30 am after fasting overnight and not exercising the day before. An intravenous catheter was inserted into a forearm vein. Blood samples were obtained every 10 minutes for 24 hours. Subjects remained in a supine position with their upper body and head slightly elevated. All postural changes were recorded. Meals were provided at 9:00 am (breakfast), 12:00 pm (lunch), 6:00 pm (dinner), and 9:00 pm (snack), and subjects consumed these meals within 30 minutes. Total calories over the day represented 85% of each subject's weight maintenance intake to account for negligible physical activity during the procedure. Each subject consumed a 500-calorie dinner, and the rest of the daily calories were distributed over the remaining meals (43% at breakfast, 49% at lunch, and 4% for the snack). The macronutrient composition of the food over the entire day averaged 55% carbohydrate, 30% fat, and 15% protein and was not different among the three meals. The composition of the meals was as follows: breakfast = 48% carbohydrate, 32% fat, and 20% protein; lunch = 54% carbohydrate, 32% fat, and 14% protein; and dinner = 55% carbohydrate, 30% fat, and 15% protein. A snack was provided at 9:00 pm and was comprised of 95% carbohydrate, 2% fat, and 4% protein. The meals consisted of typical foods such as an English muffin, orange juice, turkey lunchmeat sandwich, grapes, pork stir-fry, etc. Although the meal times were known by the subjects, no other information regarding food intake was provided, i.e., the type of food, how much food each meal would provide, or the total calories to be supplied over the day. Meals provided during the post-trial 24-hour repeated blood sampling procedure were identical in macronutrient composition and caloric content to those of the pre-trial 24-hour repeated blood sampling procedure.

Total Ghrelin

Total ghrelin was measured in serum samples from the 24-hour repeated blood sampling procedure every 20 minutes from 8:00 am to 8:00 pm and hourly from 8:00 pm to 8:00 am. Total ghrelin was measured using the Linco Research radioimmunoassay kit (St. Charles, MO). Assay sensitivity was 100 pg/mL. The intra-assay and inter-assay coefficients of variation for the high control were 2.7% and 16.7%, respectively; the intra-assay and inter-assay coefficients of variation for the low control were 1.2% and 14.7%, respectively.

At the time of sample collection, only total ghrelin was measurable with either the Linco Research or Phoenix Pharmaceuticals (Belmont, CA) radioimmunoassay kit. Quality control studies show a significant correlation between total ghrelin measured using Linco vs. Phoenix kits (22). Recently, other assays have been developed to measure the active form of ghrelin (containing the octanyl group). A positive correlation has been found between active and total ghrelin (22). Several studies have found a significant relationship between total ghrelin and food intake (1, 13). Both active and total ghrelin have been found to be proportionally elevated during fasting and decreased on feeding (23). Thus, the measurement of total ghrelin seems to be valid as related to food intake and/or energy homeostasis.

Data Analysis

Specific meal-related characteristics of the 24-hour profile of ghrelin were quantified by operationally defining specific features (Table 3). Nocturnal events were described by assessing similar features (Table 3). Baseline ghrelin was determined for the periods before and after the intervention for each individual as the lowest concentration after the nocturnal rise and, thus, was unrelated to any meal-related changes. While the subjects were given a small snack (<80 kcal) at 9:00 pm, no discernible ghrelin features were evident in response to the snack; thus, no data were analyzed surrounding this event. Energy balance data were averaged over 7 days before the intervention (“Pre”) and averaged over the last 7 days of the intervention (“Post”). Because of the small sample size, non-parametric two-tailed tests appropriate for paired comparisons, i.e., Wilcoxon signed rank tests, were performed. For comparisons between groups, the Mann-Whitney U test was used. Pearson correlation coefficient analyses were performed to examine relationships between ghrelin and other variables. Power analyses were performed on results using the Power Calculator from University of California, Los Angeles, Department of Statistics (http:calculators.stat.ucla.edupowercalc). All other analyses were performed using SPSS software (version 13.0; SPSS, Inc., Chicago, IL). All data are reported as mean ± standard error.

Table 3.  Comparison of pre- and post-intervention ghrelin meal-related and nocturnal averages in the controls and energy deficit group
Ghrelin concentrationsControls (n = 4)Energy deficit (n = 8)
  • Data expressed as mean ± standard error.

  • *

    The average of all time-points 2 hours before the meal/event up to and including 2 hours after the meal.

  • Pre- vs. post-study in energy deficit group; two-tailed Wilcoxon rank test (p < 0.05); observed power: 0.90.

  • Controls vs. energy deficit group; two-tailed Mann-Whitney U test (p < 0.05); observed power: 0.50.

Area under the curve (pg/mL/24 h)37,663 ± 2,88537,953 ± 3,752289 ± 137337,656 ± 2,51945,390 ± 2,5897,734 ± 2,851
Overall 24 hour average (pg/mL)1,616 ± 1221,634 ± 17818 ± 741,634 ± 1081,949 ± 111316 ± 116
Event averages*      
 Breakfast (pg/mL)1,559 ± 1821,514 ± 238−46 ± 1201,615 ± 1291,856 ± 8895 ± 86
 Lunch (pg/mL)1,511 ± 1731,514 ± 1993 ± 631,512 ± 1001,697 ± 102185 ± 82
 Dinner (pg/mL)1,655 ± 1031,709 ± 19854 ± 1791,681 ± 1122,034 ± 120353 ± 133
 Nocturnal (pg/mL)1,816 ± 1451,789 ± 158−28 ± 571,614 ± 1072,046 ± 123432 ± 141


Initial body weight was significantly higher (p ≤ 0.05) in the energy deficit subjects (59.6 ± 1.8 kg) compared with the controls (52.9 ± 0.5 kg), likely because of the difference in height. No other differences, including BMI, existed at baseline (Table 1). The intervention led to a significant (p < 0.05) decrease in body weight (−2.5 ± 0.9 kg), BMI (−0.91 ± 0.3 kg/m2), percent body fat (−3.8 ± 1.0%), and fat mass (−2.8 ± 0.8 kg) in the energy deficit group, with no change in the controls. Maximal oxygen uptake increased significantly in the energy deficit group. Table 2 depicts changes in the indices of energy balance across the intervention, calculated as the average of 7 days before the intervention (“Pre”) and the average of 7 days during the last week of the intervention (“Post”). Exercise calories for the energy deficit group averaged 357 ± 46 kcal/session (p < 0.05); intake was reduced significantly by 453 ± 146 kcal/d in the energy deficit group (p < 0.05). The feeling of fullness after a meal was significantly reduced after the intervention in the energy deficit group (−1.4 ± 0.6; p < 0.05) with no significant change in the controls.

Table 1.  Subject characteristics during the pre- and post-intervention in the controls and energy deficit group
 Controls (n = 4)Energy deficit (n = 8)
  • N/A, not applicable; Vo2max, maximal oxygen uptake. Data expressed as mean ± standard error.

  • *

    Controls vs. energy deficit group; two-tailed Mann-Whitney U test (p < 0.05); observed power: 0.50.

  • Pre- vs. Post-study in energy deficit group; two-tailed Wilcoxon test (p < 0.05); observed power: 0.40.

Age (years)20 ± 1N/AN/A20 ± 1N/AN/A
Height (cm)160 ± 1N/AN/A165 ± 1N/AN/A
Weight (kg)52.9 ± 0.551.6 ± 1.2−0.6 ± 0.359.6 ± 1.8*57.1 ± 1.9−2.5 ± 0.9
BMI (kg/m2)20.7 ± 0.620.1 ± 0.4−0.24 ± 0.121.9 ± 0.621.0 ± 0.7−0.91 ± 0.3
Body fat (%)24.2 ± 2.823.3 ± 2.3−0.09 ± 0.227.4 ± 2.023.6 ± 1.9−3.8 ± 1.0*
Fat mass (kg)12.8 ± 1.512.1 ± 1.1−0.04 ± 0.116.4 ± 1.413.6 ± 1.5−2.8 ± 0.8*
Fat-free mass (kg)40.1 ± 1.639.5 ± 1.6−0.45 ± 0.343.2 ± 1.543.5 ± 1.10.3 ± 0.7
Vo2max (mL/kg/min)36.2 ± 5.337.9 ± 0.81.8 ± 4.037.6 ± 1.546.0 ± 1.88.26 ± 1.8
Table 2.  Indices of energy balance before intervention and during the final week of the diet and exercise intervention in the controls and energy deficit group
 ControlsEnergy deficit
  • Data expressed as mean ± standard error.

  • *

    Numerical scale from 1 to 10.

  • Pre- vs. post-study in energy deficit group; two-tailed Wilcoxon rank test (p < 0.05); observed power: 0.90.

  • Controls vs. energy deficit group; two-tailed Mann-Whitney U test (p < 0.05); observed power: 0.80.

Intake (kcal/24 hours)1750 ± 1321850 ± 185100 ± 581800 ± 821347 ± 92−453 ± 146
Expenditure (kcal/24 hours)1774 ± 1091938 ± 174188 ± 881712 ± 982210 ± 163410 ± 118
Exercise (kcal/session)0000357 ± 46357 ± 46
Energy balance (kcal/24 h)0−88 ± 104−88 ± 1040−862 ± 225−862 ± 225
Energy deficit (%/d)0−5.3 ± 5.8−5.3 ± 5.80−45.7 ± 12.4−45.7 ± 2.4
Fullness before breakfast*3.0 ± 0.91.8 ± 1.2−1.2 ± 0.53.8 ± 0.42.9 ± 0.7−0.9 ± 0.6
Fullness after breakfast*9.2 ± 0.29.5 ± 0.30.2 ± 0.28.6 ± 0.47.2 ± 0.8−1.4 ± 0.6

The Pre and Post 24-hour ghrelin profiles for the controls and energy deficit subjects are shown in Figure 1A and B, respectively. Total ghrelin concentrations were found to rise before each meal, fall to trough levels after the meal, and reach peak levels during the evening hours in all subjects. There was a significant increase in both the 24-hour average and area under the curve of ghrelin (p < 0.01) from Pre to Post in the energy deficit group, whereas no changes occurred in the controls (Table 3; Figure 1A and B). Furthermore, the change in total ghrelin area under the curve from pre- to post-study was significantly greater in the energy deficit group (7734 ± 2851 pg/mL at 24 hours) compared with the controls (289 ± 1373 pg/mL at 24 hours; p < 0.05). The energy deficit group experienced a significant increase in the average dinner response from Pre to Post (Pre: 1681 ± 112 pg/mL vs. Post: 2034 ± 120 pg/mL; p < 0.01) and the nocturnal response (Pre: 1614 ± 107 pg/mL vs. Post 2046 ± 123 pg/mL; p < 0.01). No significant changes in overall meal-related or nocturnal profiles were observed in the controls. The change in the average nocturnal profile from pre- to post-study was significantly greater in the energy deficit group (432 ± 141 pg/mL) vs. controls (−28 ± 57 pg/mL; p < 0.05).

Figure 1.

Composite pre- and post-intervention 24-hour profile of total ghrelin. (A) Controls (n = 4). (B) Energy deficit (n = 8).

Table 4 depicts results for specific features of the 24-hour ghrelin profile. There were no significant changes over time in specific meal profile features in the control group. In the energy deficit group, there were significant increases in baseline ghrelin (Pre: 1411 ± 107 pg/mL vs. Post: 1764 ± 113 pg/mL), pre-lunch ghrelin rise (Pre: 237 ± 35 pg/mL vs. Post: 608 ± 162 pg/mL), ghrelin lunch peak (Pre: 1670 ± 121 pg/mL vs. Post: 2040 ± 105 pg/mL), ghrelin dinner peak (Pre: 1977 ± 123 pg/mL vs. Post: 2414 ± 134 pg/mL), and ghrelin post-dinner trough (Pre: 1468 ± 120 pg/mL vs. Post: 1708 ± 93 pg/mL; p < 0.05). The ghrelin post-prandial dinner decline was larger at post-study (−706 ± 80 pg/mL) compared with pre-study (−509 ± 60 pg/mL; p < 0.01). However, when these changes were compared between the groups, there were no differences between the changes in the specific meal-related features of ghrelin in the controls vs. the energy deficit group. With regard to the nocturnal events, the energy deficit group exhibited a significant increase in the nocturnal pre-event rise (Pre: 394 ± 94 pg/mL vs. Post: 663 ± 116 pg/mL; p < 0.05; Table 4) and the nocturnal peak from Pre to Post (Pre: 1862 ± 96 pg/mL vs. Post: 2372 ± 148 pg/mL; p < 0.05; Table 4). The changes in the nocturnal peak, rise, and decline were all significantly greater in the energy deficit group compared with the controls (p < 0.05).

Table 4.  Comparison of individual meal-related features of pre- and post-intervention 24-hour profiles of ghrelin in the controls and energy deficit group
Meal-related and nocturnal profiles of ghrelinControls (n = 4)Energy deficit (n = 8)
  • Data expressed as mean ± standard error.

  • *

    Pre- vs. post-study in energy deficit group; two-tailed Wilcoxon rank tests (p < 0.05); observed power: 0.90.

Baseline (pg/mL) (nadir between 4:00 and 6:00 am)1409 ± 721604 ± 1681411 ± 1071764 ± 113*
  Pre-meal rise (pg/mL)371 ± 151356 ± 109493 ± 104325 ± 68
  Peak (pg/mL)1780 ± 1751908 ± 3031904 ± 1841866 ± 154
  Post-meal decline (pg/mL)−328 ± 54−653 ± 167−470 ± 110−461 ± 158
  Post-meal trough (pg/mL)1452 ± 1801255 ± 2191434 ± 971405 ± 133
  Pre-meal rise (pg/mL)376 ± 42519 ± 165237 ± 35608 ± 162*
  Peak (pg/mL)1763 ± 1701774 ± 2381670 ± 1212040 ± 105*
  Post-meal decline (pg/mL)−418 ± 81−378 ± 96−324 ± 41−529 ± 77
  Post-meal trough (pg/mL)1345 ± 911511 ± 1531346 ± 1081511 ± 153
  Pre-meal rise (pg/mL)583 ± 160591 ± 32630 ± 54925 ± 134
  Peak (pg/mL)1929 ± 1531953 ± 2501977 ± 1232414 ± 134*
  Post-meal decline (pg/mL)−528 ± 102−551 ± 126−509 ± 60−706 ± 80*
  Post-meal trough (pg/mL)1400 ± 1261351 ± 1421468 ± 1201708 ± 93*
 Nocturnal event    
  Pre-event rise (pg/mL)595 ± 102572 ± 50394 ± 94663 ± 116*
  Peak (pg/mL)1995 ± 1431924 ± 1601862 ± 962372 ± 148*
  Post-event decline (pg/mL)−636 ± 155−346 ± 35−451 ± 80−608 ± 139

To determine how tightly linked the observed changes in the 24-hour ghrelin profile were with changes in energy balance indices, correlation analyses were performed on all subjects. None of the changes in breakfast or lunch features was correlated with any of the changes in energy balance indices. The change in the dinner rise of ghrelin was negatively associated with the change in body weight (R = −0.70; p = 0.003), energy intake (R = −0.67; p = 0.048), calculated energy deficit (R = −0.76; p = 0.028), and the feeling of fullness after a meal (R = −0.65; p = 0.023). Negative correlations were also observed with the change in the nocturnal event average and the change in body weight (R = −0.601; p = 0.039), change in energy deficit (R = 0.735; p = 0.038), and change in feelings of fullness after a meal (R = −0.730; p = 0.007). There was a negative correlation between change in the average 24-hour ghrelin concentration and change in percentage body fat (R = −0.613; p = 0.045), such that the greater the increase in 24-hour ghrelin, the greater the decrease in percentage body fat.


To our knowledge, this is the first study to show that meal-related and nocturnal profiles of ghrelin are elevated and significantly related to several metabolic adaptations after a prolonged period of caloric restriction combined with exercise resulting in weight loss in non-obese, pre-menopausal women. These findings are in line with the idea that ghrelin exhibits compensatory increases during chronic energy deficiency and extend our previous work showing significant increases in fasting ghrelin with weight loss after the same intervention (18). These results provide evidence that adjustments in meal-related and nocturnal profiles also occur.

Our previous study showed that increases in fasting ghrelin with weight loss were compensatory, in that the time-course of changes was such that significant decreases in resting metabolic rate, fat mass, and body weight occurred before changes in fasting ghrelin (18). In this study, we found tight associations between the meal-related and nocturnal features and the decreases in body weight, food intake, satiety, and the magnitude of energy deficiency that occurred over the course of the intervention. Although correlations do not prove cause and effect, these results, combined with our findings with fasting ghrelin, are in line with the suggestion that ghrelin secretory dynamics across the day are important for regulation of long-term energy homeostasis.

We found significant effects of the intervention on specific features of ghrelin meal-related profiles. The pre-meal ghrelin concentrations (both lunch and dinner meal peaks) were significantly elevated after the energy deficit. The elevation in the dinner peak was also accompanied by reduced feelings of fullness. It is intriguing to postulate that ghrelin may act in a compensatory manner to increase food intake, through decreases in satiety, and subsequently restore energy balance through elevations in the meal-related peaks throughout the day. While numerous studies in animals and humans (1, 4, 5) support a role for ghrelin in the stimulation of food intake per se, the mechanism for this effect remains unclear, and whether or not serum ghrelin levels relate to food intake in a dose–response fashion has been difficult to discern, because studies examining this have yielded mixed results (6, 13, 24, 25). Until a direct and dose-dependent role for pre-meal rises in ghrelin in the stimulation of food intake can be proven, our finding of weight loss–induced increases in pre-meal ghrelin peaks is only suggestive of the idea that ghrelin may play a role in the stimulation of compensatory increases in food intake in an attempt to restore baseline body weight.

Concerning the post-prandial response of ghrelin after a chronic energy deficit, smaller post-meal declines were expected to maximize additional food intake throughout the day in an effort to restore energy balance. This thought is in line with the abolition of the post-meal decline observed in anorexia (14). We actually observed larger post-prandial declines during the dinner meal after the intervention. In our previous study in normal-weight women, we observed significant associations between the amount of calories consumed in a meal and the magnitude of the post-prandial decline of ghrelin, such that the greater the calorie content of the meal, the greater the decline in ghrelin (19). However, in this study, the calorie content of the dinner was identical (500 kcal) during Pre and Post, as were the content and composition of all food eaten during the 24-hour procedures. One potential explanation for the larger post-prandial decline might be related to how the ingested calories were “sensed” by the body. Presumably, the significant loss of body weight (−2.5 ± 0.9 kg) experienced by the deficit group represented a reduction in energy requirements. Thus, the same 500-calorie dinner provided during the pre-intervention represented a larger percentage of post-intervention 24-hour calorie needs than it did before the intervention and, thus, was “sensed” as a relatively larger meal. This, perhaps, explains the larger post-meal decline in ghrelin after the study. Further studies are clearly needed to fully identify the potential mechanisms surrounding changes in ghrelin post-prandial responses that occur in response to chronic changes in energy balance.

The finding that an elevation in the nocturnal pattern occurred in response to weight loss has been previously reported in obese subjects who lost weight (11), although the absolute concentrations of ghrelin remained lower than in our normal-weight subjects. Data from our laboratory examining the relation between meal energy content across a typical day of eating and features of the nocturnal profile showed no significant association with acute nutrient intake (19). Other data suggest that the nocturnal rise may be related to the onset of sleep (26, 27). Possible explanations for our finding might be that the nocturnal rise in ghrelin is associated with chronic, not acute, changes in energy balance or that the neuroendocrine factors associated with the onset of sleep also changed in proportion to the changes observed in our intervention.

When comparing these results to those in clinical populations, individuals with anorexia displayed constant 24-hour elevations in ghrelin with no breakfast, lunch, or dinner pre-meal rises and post-meal declines (8, 9, 14, 28). Although we did not observe an abolishment of the meal response in response to our intervention, the elevations in the meal-related and nocturnal features of ghrelin are in line with the idea that ghrelin may increase to facilitate compensatory changes to restore body weight, just as it can be reasoned that the insensitivity to nutrient ingestion in anorexics results in sustained elevations in ghrelin for the same purpose. Perhaps, our findings indicate a potential intermediate step in a progression to re-establish energy balance by initially increasing the nocturnal pattern and the meal response of ghrelin to override the energy deficiency by increasing the drive to eat. It is unknown whether an extension of our intervention accompanied by more weight loss might have led to an actual abolishment of the meal response, resulting in a sustained elevation. At the other end of the energy spectrum, obese individuals exhibit lower fasting ghrelin and blunted meal responses and nocturnal rises (11, 29). With body weight restoration in both obese and anorexic subjects, fasting ghrelin levels return to normal (10, 11, 28). Taken together, these findings indicate directional changes in fasting, meal-related, and nocturnal profiles similar to those observed in this study. This may indicate that the factors that modulate ghrelin in extreme states of energy balance are similar to those that are in place under normal physiological conditions vs. being a product of other changes related to these pathological states.

While all subjects were randomly assigned to either the control group or the energy deficit group at the start of the intervention, they were not purposely matched for baseline body weight. Our controls had significantly lower initial body weights (52.9 ± 0.5 kg) compared with the energy deficit group (59.6 ± 1.8 kg), likely because of the 5-cm difference in height. Accordingly, both groups were within the normal range for BMI. While several studies in overweight, obese, and anorexic patients and two studies in normal-weight individuals have shown a relationship between fasting ghrelin concentrations and body weight and/or BMI (7, 11, 14, 16, 30), we found no difference in baseline total ghrelin between our control and energy deficit groups (1409 ± 72 vs. 1411 ± 107 pg/mL; p = 0.933) even though they differed in body weight. Nor did we find that there was a relationship between baseline ghrelin concentrations and initial body weight (r = 0.117; p = 0.718). We cannot be sure why we did not observe this relationship in our group of normal-weight subjects. One difference between our study and others is that our fasting ghrelin samples represented the times between 4:00 and 6:00 am, whereas times closer to 8:00 am are typically used in other studies. Our subjects might have also spanned a relatively smaller range of body weights than that represented in other studies that included clinical populations. Irrespective of this difference, the lack of relationship between baseline body weight and fasting ghrelin in our study makes it unlikely that our baseline differences in body weight had an impact on the ghrelin responses to our intervention.

In conclusion, the observed elevations in the meal-related and nocturnal profiles of ghrelin are tightly linked to changes in indices of energy balance after a long-term energy deficit in non-obese, pre-menopausal women. Although a limitation of our study was the small number of control subjects and low power for some analyses, we believe our findings suggest, but do not prove, that ghrelin is involved with the long-term regulation of energy homeostasis and that meal-related and nocturnal secretory dynamics of ghrelin are of particular importance in this regard. Obviously, our study was not designed to directly test whether weight loss–induced increases in ghrelin are associated with an increase in voluntary food intake and a return to baseline body weight. Future studies examining similar interventions in free-living individuals whose eating is not controlled are needed to address whether fluctuations in these aspects of ghrelin secretion over time are associated with post-weight loss-induced increases in food intake and body weight. Experimental studies using pharmacological agents are obviously also necessary to confirm a direct role for ghrelin in long-term energy homeostasis.


We thank the members of the Exercise Endocrinology and Metabolism Laboratory, the subjects, and the General Clinical Research Center staff for efforts during this study. This study was supported by NIH Grants 1R01HD39245-01A1 and M01 RR 10732.


  1. The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.