| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Family and Preventive Medicine, School of Medicine, University of California-San Diego (C.L., J.B., G.A.L., E.B.-C.), La Jolla, California 92093-0607; and Department of Epidemiology and Public Health, University College London Medical School (C.L.), London WC1E 6BT, United Kingdom
Address all correspondence and requests for reprints to: Dr. Elizabeth Barrett-Connor, Department of Family and Preventive Medicine, School of Medicine, University of California-San Diego, 9500 Gilman Drive, La Jolla, California 92093-0607. E-mail: ebarrettconnor{at}ucsd.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of this work was to study the association between ghrelin and the metabolic syndrome and identify social and behavioral determinants of ghrelin.
Design: This was a cross-sectional study.
Setting: The setting of this work was the Rancho Bernardo Study.
Participants: Study subjects included 1513 men and women, aged 51–90 yr in 1984–1987.
Outcomes: Total ghrelin, measured by RIA, and the metabolic syndrome, defined using Adult Treatment Panel III diagnostic criteria, were the outcome measures.
Results: Levels of ghrelin (mean ± SD) did not differ between the 848 men (1451 ± 532 pg/ml) and the 665 women (1459 ± 672 pg/ml) or by age. Education, alcohol intake, and smoking history were each significantly and positively associated with ghrelin in a dose-related manner, independent of body mass index (BMI). Compared with participants with the lowest third of ghrelin levels, the age- and sex-adjusted odds of having the metabolic syndrome were 18% lower in the middle third and 53% lower in the highest third. This corresponds to a 21% decrease per SD increase in ghrelin (odds ratio, 0.79; 95% confidence interval, 0.69, 0.89; P 
0.001); this was attenuated to 13% (odds ratio, 0.87; 95% confidence interval, 0.75, 1.01; P = 0.07), after adjustment for BMI. Of the five metabolic syndrome components, only the association between ghrelin and high-density lipoprotein cholesterol was independent of BMI. A significant association independent of BMI was also observed between insulin and ghrelin.
Conclusions: Ghrelin levels are influenced by lifestyle factors. The inverse association between endogenous ghrelin and the metabolic syndrome is largely explained by the strong ghrelin-BMI association.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Although the function of ghrelin as a short-duration, meal-related hunger signal in humans has been supported by evidence of a preprandial increase and postprandial decrease in ghrelin (2, 3, 4), a discernible diurnal rhythm has not been confirmed (5), and the role of ghrelin in the stimulation of appetite and food intake in humans remains controversial (6, 7, 8). Weight loss and fasting increase ghrelin levels, potentially as a compensatory response to an energy deficit (9, 10). Since the initial report (11), several studies of different populations have shown that levels of ghrelin are inversely related to body size. Although this points toward ghrelin as a marker of nutrient excess or depletion, its mode of action as a regulator of body fat stores remains unclear.
Obesity induces a number of metabolic disturbances, including decreased insulin sensitivity and dyslipidemia, and is associated with an excess risk of insulin resistance, diabetes, and cardiovascular disease (12, 13, 14). Obesity and insulin resistance are key components of the clustering of risk factors known as the metabolic syndrome (15, 16); the specific mechanisms for this co-occurrence of adverse risk factor levels are still a matter of debate (17). Recent reports have linked the adiposity signaling hormone ghrelin to insulin resistance and type II diabetes (18, 19, 20, 21, 22, 23, 24, 25); thus, ghrelin may be one of the pathophysiological mechanisms underlying the risk factor clustering observed in the metabolic syndrome. To our knowledge, the association between ghrelin and the prevalence of the metabolic syndrome and its components has not been examined.
The aims of this study were 1) to identify adiposity and lifestyle determinants of total serum ghrelin, and 2) to evaluate the association of ghrelin with the metabolic syndrome and its components in older community-dwelling adults, investigating the role of body size and lifestyle factors in these associations.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The Rancho Bernardo Study is a prospective cohort study of 82% of all adult residents of a southern California residential community, established from 1972–1974 (26). Participants aged 40 yr or older were invited to a follow-up examination between 1984 and 1987; 1056 men and 1300 women, aged 50 yr or older, participated at this visit, when blood samples were obtained for later hormone assays. The 1984–1987 visit included a standardized questionnaire on medical and medication history, smoking status, alcohol consumption, and physical activity. All prescriptions and/or pills taken were brought to the clinic for confirmation of medication use (27). Participants provided informed consent at each clinic visit, and research protocols were approved by the university institutional review board.
Anthropometry and blood pressure
Weight, height, and waist and hip circumferences were measured with participants wearing light clothing and no shoes. The body mass index (BMI) was calculated as weight (kilograms) divided by height (meters) squared. Waist circumference was measured over bare skin at the bending point (point marked when participant naturally bends forward and measured after participant has realigned to an upright position). In addition, minimum waist circumference was measured at the narrowest circumference; these measurements correlated by 98%, and the bending point measurement was used in all analyses (28). Hip circumference was measured at the iliac crest, and waist/hip ratio was calculated by dividing waist by hip circumference and was expressed as a percentage. Systolic and diastolic blood pressures were measured twice in seated resting subjects using the hypertension detection and follow-up program protocol (29); the mean of two readings was used in all analyses.
Laboratory measurements
Blood samples were obtained by venipuncture between 0730 and 1100 h after a requested 12-h fast. Lipid and lipoprotein levels were measured in a Center for Disease Control-certified lipid research clinics laboratory. Total cholesterol and triglyceride levels were measured by enzymatic techniques using an ABA-200 biochromatic analyzer (Abbott Laboratories, Inc., Irving, TX). High-density lipoprotein (HDL) cholesterol was measured by precipitating the other lipoproteins with heparin and manganese chloride according to the standard lipid research clinics protocol. Low-density lipoprotein (LDL) cholesterol was estimated using the Friedewald formula (30). Fasting and 2-h postchallenge glucose after a 75-g oral glucose tolerance test were measured by the glucose oxidase method. Fasting and 2-h insulin were measured by double-antibody RIA (31). Standardized insulin assays were not available for the first year of the study (29% of participants).
Blood samples used for ghrelin assays were obtained by venipuncture between 0730 and 1100 h after a requested 12-h fast in 1984–1987; sera were separated and frozen at –70 C until 2004, when they were thawed (for the second time) for measurement of ghrelin by RIA (Linco Research, Inc., St. Louis, MO). The ghrelin assay has a sensitivity of 95 pg/ml and intra- and interassay coefficients of variation of 8 and 15%, respectively. The laboratory reports no problem during two freeze-thaw cycles with this assay, and our levels are similar to those reported in the literature using the same assay (22). Log transformation of ghrelin was performed, resulting in the normalization of its skewed distribution.
Of the 2356 participants attending the 1984–1987 clinic visits, ghrelin was not measured in 351 women who were current estrogen users; 1535 of the remaining participants (860 men and 675 women) had sufficient stored sera for hormone measurement. Of these, 10 men and 12 women were excluded for missing height and/or weight data. Compared with those without ghrelin assays, participants included in this study were slightly older and more likely to be male, but did not differ in terms of weight, BMI, or the proportion of smokers or daily consumers of alcohol.
Metabolic syndrome and its components
The prevalence of the metabolic syndrome and its components were defined using Adult Treatment Panel III recommended diagnostic criteria (15). People were classified as having the metabolic syndrome if any three of the following were present: fasting plasma glucose of 6.1 mmol/liter or more (110 mg/dl), waist circumference greater than 102 cm (40 in.) in men or greater than 88 cm (35 in.) in women, triglycerides of 1.7 mmol/liter or higher (150 mg/dl), HDL cholesterol level less than 1.036 mmol/liter (40 mg/dl) in men or less than 1.295 mmol/liter (50 mg/dl) in women, or blood pressure of 130/85 mm Hg or higher. The use of antihypertensive medication was included as high blood pressure in this analysis. Type 2 diabetes was determined according to World Health Organization criteria (16). A person was regarded as diabetic if the fasting plasma glucose was 7 mmol/liter or more (126 mg/dl), if the 2-h glucose during an oral glucose tolerance test was 11.1 mmol/liter or more (200 mg/dl), or if he/she was using diabetes medication (oral or insulin). Whole body insulin resistance in the fasting state was assessed using the homeostasis model assessment for insulin resistance (HOMA-IR) and was calculated as fasting glucose (millimoles per liter) x fasting insulin (microunits per milliliter)/22.5 (32) after excluding participants with diabetes, because it has been recommended not to use HOMA as an index of insulin resistance in older individuals who have diabetes (33, 34). HOMA was log transformed, resulting in the normalization of its skewed distribution.
Statistical methods
Analyses were restricted to participants with complete information on ghrelin, height, and weight. For ease of interpretation, mean ghrelin levels are presented as untransformed values; all P values are based on tests for linear trend using log-transformed ghrelin. Analyses were performed in two steps; the first was adjusted for age and sex, and the second additionally included BMI in the model. Mean levels of ghrelin across social and behavioral categorical variables were obtained, and tests for trend across categories were carried out before and after adjustment for BMI.
Age- and sex-adjusted means of each outcome across thirds of ghrelin were then calculated. Logged ghrelin was converted into internally derived SD scores (z-scores), and linear regression, including age and sex as covariates, was used to investigate changes in outcomes per 1-SD increment in ghrelin before and after adjustment for BMI. Changes in outcomes are calculated and expressed in their original units; however, all P values are based on tests for linear trend using log-transformed outcomes where appropriate.
Logistic regression was used to compare the odds of having the metabolic syndrome and its components across thirds of ghrelin, taking men and women with levels of ghrelin in the lowest third of the population as the comparison group and adjusting for age and sex. Changes in odds per 1-SD increment in ghrelin before and after adjustment for BMI were also calculated using logistic regression.
Mean levels of ghrelin according to the number (zero to five) of metabolic syndrome components present were also investigated, and a test for linear trend was based on log-transformed ghrelin adjusted for age and sex.
Analyses were initially stratified by sex to evaluate potential effect modification. Ghrelin associations with lifestyle variables and metabolic risk factors did not differ by sex, and tests for interaction were not statistically significant. Therefore, data are presented for men and women combined, adjusted for age and sex.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Educational attainment, alcohol intake, and smoking history were all associated with total serum ghrelin in a dose-related manner, with participants with higher education, greater alcohol intake, and current smoking exhibiting the highest levels of ghrelin, independent of age, sex, or BMI (Table 1
), notwithstanding a strong inverse relationship between ghrelin and BMI (Table 2). Participants who exercised at least three times a week had lower levels of ghrelin than those who did not; this association was strengthened and became statistically significant after adjustment for BMI (Table 1).
|
|
). Conversely, measures of central obesity (waist circumference and waist/hip ratio) did not account for the strong association between ghrelin and BMI (data not shown).
Systolic and diastolic blood pressures, fasting and postchallenge insulin, HOMA-IR, HDL cholesterol, and triglycerides were associated with ghrelin before and after adjustment for age and sex, such that participants with higher levels of ghrelin had more favorable levels of each of these cardiovascular risk factors (Table 2). In contrast, ghrelin was not associated with fasting or postchallenge glucose levels or with total or LDL cholesterol levels. Adjustment for BMI attenuated most associations; only fasting and postchallenge insulin and HDL cholesterol remained statistically associated with ghrelin. For each increase of 1 SD in ghrelin, fasting and postchallenge insulin levels decreased by 0.76 µU/ml (95% confidence interval, –1.38, –0.14; P = 0.01) and 6.67 µU/ml (95% confidence interval, –10.95, –2.38; P = 0.005), respectively, and HDL cholesterol increased by 1.12 mg/dl (95% confidence interval, 0.32, 1.92; P = 0.002), controlling for age, sex, and BMI (Table 2). These associations remained statistically significant after adjusting for education, alcohol intake and smoking or after excluding participants reporting insulin use (n = 12; data not shown). Ghrelin levels did not differ significantly between participants with or without diabetes (1462.6 vs. 1422.5 pg/ml, respectively; P = 0.5); similarly, no differences were observed when comparing those with or without use of drugs for diabetes, hypertension, thyroid function, or digitalis (data not shown).
Ghrelin and the metabolic syndrome
In age- and sex-adjusted analyses, the prevalence of each of the five metabolic syndrome components was lowest in the participants whose levels of ghrelin were in the highest third of the population (Table 3); these differences were statistically significant for high waist circumference, high triglycerides, and low HDL cholesterol, but not for hyperglycemia or high blood pressure, before and after adjustment for age and sex. After adjustment for BMI, only the association between ghrelin and HDL cholesterol remained statistically significant, with the odds of having low HDL cholesterol decreasing by 21%/SD increase in ghrelin (odds ratio, 0.79; 95% confidence interval, 0.69, 0.91; P 0.0001).
|
0.0001 for trend; Fig. 1). The prevalence of the metabolic syndrome was 12% in those with ghrelin levels in the highest third of the population, 19% in those with ghrelin levels in the middle, and 23% in those with ghrelin levels in the lowest third (Table 3). In an age- and sex-adjusted comparison, with men and women in the lowest third of ghrelin as the reference group, the odds of having the metabolic syndrome was 18% lower in the middle and 53% lower in the highest third of ghrelin levels. This corresponds to a 21% decrease/SD increase in ghrelin (odds ratio, 0.79; 95% confidence interval, 0.69, 0.89; P 0.0001); this was attenuated to 13% (odds ratio, 0.87; 95% confidence interval, 0.75, 1.01; P = 0.07) after adjustment for BMI. The association between ghrelin and the metabolic syndrome did not differ significantly between participants who were middle-aged and those who were older (>65 yr; data not shown). Furthermore, there was no significant interaction in the association between ghrelin and the metabolic syndrome between diabetic and nondiabetic participants (P = 0.4 for interaction).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Few studies have investigated whether insulin and glucose metabolism explain physiological variations in ghrelin in a population-based sample. Although we are aware of no other report of circulating ghrelin and the metabolic syndrome, other observational studies have reported that low ghrelin levels are associated with insulin, insulin resistance, and type II diabetes (11, 19, 20, 25, 35). Meal-related changes in ghrelin are reciprocal to those in insulin (2), suggesting that they may be secondary to postprandial changes in plasma glucose and insulin (36, 37). Experimental studies in humans have shown that ghrelin levels are reduced during euglycemic hyperinsulinemic glucose clamp conditions (38, 39, 40, 41, 42) as well as after insulin-induced hypoglycemia (38, 41). One study, however, found that, unlike the suppression of ghrelin after a meal-induced rise in insulin, parenteral administration of glucose and insulin had little effect on ghrelin levels (43), and changes only occurred during sustained hyperinsulinemia (44). It remains unclear to what degree insulin and glucose each contribute to the suppression of ghrelin, but administration of natural ghrelin has been shown to reduce insulin levels in healthy human subjects (45). Thus, cross-sectional findings of an inverse correlation between ghrelin and insulin may also reflect a role for ghrelin in the etiology of insulin resistance or down-regulated levels in the hyperinsulinemic state.
An inverse association between ghrelin and body size in normal weight and obese populations is well established (9, 11, 19, 46). In this study, overall body size (BMI) was more strongly associated with ghrelin than central adiposity (waist circumference or waist/hip ratio). In support of this, a study using dual x-ray absorptiometry reported that ghrelin levels reflect body weight more strongly than percent body fat or body fat distribution (20). Central obesity is a key component of the metabolic syndrome, and adjustment for BMI in associations between ghrelin and central obesity or the metabolic syndrome may be inappropriate due to the strong colinearity between waist circumference and BMI (r = 0.76 in this cohort). Overadjustment may occur if obesity mediated the association between ghrelin and metabolic risk or insulin resistance; research into the direction of these associations is warranted.
Observational studies have reported a positive association between ghrelin and HDL cholesterol (20, 21). Our findings confirm and extend these results by showing that the ghrelin-HDL association is independent of body size and lifestyle factors known to be associated with both ghrelin and HDL levels.
To our knowledge this is the first study to report social and behavioral determinants of ghrelin. We found that levels of ghrelin were higher in men and women with greater educational attainment, more alcohol consumption, and smoking exposure. These associations with ghrelin showed a dose-response pattern and were independent of the potential confounding effects of BMI. Ghrelin levels were lower in participants who reported more frequent exercise, but this association was explained by their smoking status; exercisers were more likely to have never smoked or to have given up smoking, both of which were associated with lower ghrelin levels. Our results are in agreement with a small study of middle-aged men reporting higher ghrelin levels in current smokers compared with nonsmokers (21). A study investigating the influence of alcohol consumption on ghrelin levels in abstinent chronic alcoholics reported that fasting ghrelin levels were positively correlated with the duration of alcohol abstinence, and that daily alcohol consumption before abstinence was correlated inversely with ghrelin levels after abstinence (47). These findings are contrary to ours, which show a positive, dose-related association between ghrelin and the frequency of alcohol intake. Nutritional abnormalities, adverse health behaviors, and long-term health consequences associated with alcohol dependence may contribute to the different findings.
We identified novel social and behavioral determinants of ghrelin; however, these only explained a little more of the variation in ghrelin levels between individuals than age, sex, and BMI alone (5% vs. 2.6%; data not shown), and the determinants of ghrelin levels in the general population remain poorly understood.
Although this is the largest population-based study to report social and behavioral determinants of ghrelin and its association with the metabolic syndrome, some limitations should be noted. Participants are adult residents of Rancho Bernardo, a predominantly white, upper-middle class community. Thus, although our results are more generalizable than patient-based studies and/or small selected samples, they may not apply to other social classes or ethnic groups. A single morning sample of ghrelin may be inadequate to characterize an individual’s usual blood level throughout the day, although a single fasting 0800 h ghrelin level correlates well with the 24-h area under the curve value in lean and obese individuals (20). Long-term storage and previous freeze-thaws may potentially distort levels of ghrelin; however, one study reported no significant decrease in ghrelin values after repeated freezing and thawing (48). Furthermore, our levels are similar to those reported in the literature using the same assay (22, 49). Lastly, the correlations between ghrelin and BMI were virtually identical when comparing our study to one using fresh fasting samples for ghrelin measures (partial r adjusted for age and sex = –0.12 and –0.14, respectively) (25).
We conclude that the strong inverse associations between ghrelin and metabolic syndrome components observed in this study were largely explained by the greater BMI of individuals with lower levels of ghrelin. BMI did not explain associations between ghrelin and levels of insulin and HDL cholesterol or the influence of educational attainment, alcohol consumption, smoking, and exercise on ghrelin. Prospective studies clarifying whether ghrelin plays a role in the development of insulin resistance, diabetes, and cardiovascular disease are warranted and should account for the influence of lifestyle factors on both ghrelin and these outcomes.
|
Footnotes |
|---|
First Published Online October 4, 2005
Abbreviations: BMI, Body mass index; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment for insulin resistance; LDL, low-density lipoprotein.
Received June 17, 2005.
Accepted September 23, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  O. Sentissi, J. Epelbaum, J.-P. Olie, and M.-F. Poirier Leptin and Ghrelin Levels in Patients With Schizophrenia During Different Antipsychotics Treatment: A Review Schizophr Bull, November 1, 2008; 34(6): 1189 - 1199. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Ingelsson, M. G. Larson, X. Yin, T. J. Wang, J. B. Meigs, I. Lipinska, E. J. Benjamin, J. F. Keaney Jr., and R. S. Vasan Circulating Ghrelin, Leptin, and Soluble Leptin Receptor Concentrations and Cardiometabolic Risk Factors in a Community-Based Sample J. Clin. Endocrinol. Metab., August 1, 2008; 93(8): 3149 - 3157. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Mager, M. Kolehmainen, V. D F de Mello, U. Schwab, D. E Laaksonen, R. Rauramaa, H. Gylling, M. Atalay, L. Pulkkinen, and M. Uusitupa Expression of ghrelin gene in peripheral blood mononuclear cells and plasma ghrelin concentrations in patients with metabolic syndrome. Eur. J. Endocrinol., April 1, 2008; 158(4): 499 - 510. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. De Vriese, M. Hacquebard, F. Gregoire, Y. Carpentier, and C. Delporte Ghrelin Interacts with Human Plasma Lipoproteins Endocrinology, May 1, 2007; 148(5): 2355 - 2362. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-C. Chang, C.-H. Hung, H.-L. Chen, K.-L. Hwang, and C.-Y. Lin Peritoneal transport characteristics and dwelling time significantly impact ghrelin clearance in peritoneal dialysis patients Nephrol. Dial. Transplant., January 1, 2007; 22(1): 224 - 228. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. T. Vestergaard, R. Dall, K. H. W. Lange, M. Kjaer, J. S. Christiansen, and J. O. L. Jorgensen The Ghrelin Response to Exercise before and after Growth Hormone Administration J. Clin. Endocrinol. Metab., January 1, 2007; 92(1): 297 - 303. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. H. Park, Y. J. Oh, G. Y. Kim, S. E. Kim, K.-H. Paik, S. J. Han, A. H. Kim, S. H. Chu, E. K. Kwon, S. W. Kim, et al. Obestatin Is Not Elevated or Correlated with Insulin in Children with Prader-Willi Syndrome J. Clin. Endocrinol. Metab., January 1, 2007; 92(1): 229 - 234. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Langenberg, J. Bergstrom, C. Scheidt-Nave, J. Pfeilschifter, and E. Barrett-Connor Cardiovascular Death and the Metabolic Syndrome: Role of adiposity-signaling hormones and inflammatory markers. Diabetes Care, June 1, 2006; 29(6): 1363 - 1369. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Bouros, A. Tzouvelekis, S. Anevlavis, M. Doris, S. Tryfon, M. Froudarakis, V. Zournatzi, and A. Kukuvitis Smoking acutely increases plasma ghrelin concentrations. Clin. Chem., April 1, 2006; 52(4): 777 - 778. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
