| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Human Biology, Nutrition and Toxicology Research Centre NUTRIM (E.E.B., G.H., W.H.M.S.), Maastricht University, Maastricht, The Netherlands; Department of Health and Nutrition (E.F.M.F.), National Institute of Public Health and the Environment, Bilthoven, The Netherlands; Institute of Preventive Medicine (C.V., C.H., T.I.S.), Danish Epidemiology Science Centre, Copenhagen University Hospital, Copenhagen, Denmark; Department of Sports Medicine (V.S., J.P.), Centre of Preventive Medicine, Third Faculty of Medicine, Charles University, Prague, Czech Republic; Department of Physiology and Nutrition (A.M.), University of Navarra, Pamplona, Spain; Institute of Human Nutrition (M.P., S.T., A.A.), The Royal Veterinary and Agricultural University, Copenhagen, Denmark; School of Biomedical Sciences (K.P., I.A.M.), Queens Medical Centre, University of Nottingham Medical School, Nottingham, United Kingdom; Department of Nutrition (J.M.O.), Hôtel-Dieu Hospital, Paris, France; Obesity Research Unit of the French Institute of Health and Medical Research U586 (P.B., D.L.), Louis Bugnard Institute and Clinical Investigation Centre, Toulouse University Hospitals, Paul Sabatier University, Toulouse, France; and Department of Medicine (I.A.), Karolinska Institute, Huddinge University Hospital, Stockholm, Sweden
Address all correspondence and requests for reprints to: Dr. E. E. Blaak, Department of Human Biology, Nutrition Research Centre, Maastricht University, P.O. Box 616, 6200 MD Maastricht, The Netherlands. E-mail: e.blaak{at}hb.unimaas.nl.
| Abstract |
|---|
|
|
|---|
Aim: The aim of the present study was to study fatty acid use in the fasting state and in response to a high fat load in a large cohort of obese subjects (n = 701) and a lean reference group (n = 113).
Methods: Subjects from eight European centers underwent a test meal challenge containing 95 en% fat [energy content 50% of estimated resting energy expenditure (EE)]. Fasting and postprandial fat oxidation and circulating metabolites and hormones were determined over a 3-h period.
Results: Postprandial fat oxidation (as percent of postprandial EE, adjusted for fat mass, age, gender, center, and energy content of the meal) decreased with increasing body mass index (BMI) category (P < 0.01), an effect present only in those obese subjects with a relatively low fasting fat oxidation (below median, interaction BMI category x fasting fat oxidation, P < 0.001). Fasting fat oxidation increased with increasing BMI category (P < 0.001), which was normalized after adjustment for fat-free mass and fat mass. Furthermore, insulin resistance was positively associated with postprandial fat oxidation (P < 0.05) and negatively associated with fasting fat oxidation (expressed as percent of EE), independent of body composition.
Conclusions: The present data indicate an impaired capacity to regulate fat oxidation in the obese insulin-resistant state, which is hypothesized to play a role in the etiology of both obesity and insulin resistance.
| Introduction |
|---|
|
|
|---|
There are indications that obese subjects may have a diminished capacity to use fat as a fuel and adapt more slowly to a high dietary fat intake, compared with lean subjects (10). Additionally, postobese women (11) or women predisposed to the development of obesity (12) have lower postprandial and 24-h fat oxidation as compared with never-obese women. Also, longitudinal studies in Pima Indians of Arizona have shown that low-fat oxidizers show a greater risk of gaining body weight over time, compared with high-fat oxidizers (13). Thus, there is accumulating evidence that there may be a primary, possibly genetically, determined component of fat oxidation, which may result in a low fat oxidation rate predisposing subjects to the development of obesity.
The objective of this part of the Nutrient-Gene Interactions in Human ObesityImplications for Dietary Guidelines multicenter project was to investigate fat oxidation rate in obese subjects with a wide range of adiposity (n = 701) and a lean reference group (n = 113) both in the fasting state as well as after a high-fat challenge.
| Subjects and Methods |
|---|
|
|
|---|
Eight hundred fourteen subjects of Caucasian origin were included (607 women). The basic selection criteria for both obese and reference subjects were age 2050 yr, lean body mass index (BMI) ranging between 18.5 and 25 kg/m2, and obese BMI 30 kg/m2 or greater. Exclusion criteria were weight change greater than 3 kg within the 3 months before the study start; drug-treated hypertension, diabetes, or hyperlipidemia; untreated thyroid disease; surgically treated obesity; pregnancy, alcohol, or drug abuse; and participation in other simultaneous ongoing trials. Additional exclusion criteria for lean subjects were a history of BMI greater than 25 kg/m2, medication except contraceptives, and chronic disease.
Subjects were recruited through the media, from waiting lists, from other ongoing population studies, and by self-referral or referral from a general physician or other clinical units and local obesity organizations. Before entering the study, all subjects were in good health as assessed by medical history and physical examination when needed. In total, six subjects were treated for thyroid disease (hypothyroidism). The study protocol was approved by the ethics committee of each center/country, and all subjects gave written informed consent before participating in the study.
Experimental design
The study was conducted in eight different centers in seven European countries; Denmark, The Netherlands, Sweden, United Kingdom (England), Czech Republic, France (two centers), and Spain. All subjects underwent a 1-d clinical investigation protocol. Subjects arrived at the research center at 0800 h after a 12-h overnight fast and a preceding 3-d dietary run-in period, in which they had to keep their habitual diet and avoid excessive physical activity and alcohol consumption. After the subjects voided their bladder, they underwent anthropometric and body composition assessments. Thereafter subjects lay on the bed for 3.5 h during which thermogenesis and substrate use and circulating hormones and metabolites were measured. Before the study, a 3-d weighed food record of 2 weekdays and 1 weekend day was performed. The dietary records were analyzed using the database routinely used in each center. Additionally, an estimate of habitual physical activity was obtained by means of the Baecke questionnaire using the sum of work, sport, and leisure scores of the questionnaire. This index of total physical activity was previously validated against average daily energy expenditure measured by the doubly labeled water method (14, 15).
Procedures
Body weight was measured with a calibrated scale with subjects in their underwear. A calibrated stadiometer was used to determine body height. Waist and hip circumferences were measured with the participant wearing only nonrestrictive underwear. All measurements were performed three times and the mean was recorded. Body composition was determined in all centers by means of the same brand of multifrequency bioimpedance meter (Quadscan 4000; Bodystat, Isle of Man, British Isles) with the subjects in supine position.
Test meal
To study nutrient partitioning, the response to a saturated fat load was measured. The liquid test meal (double cream with 40 g fat per 100 g adjusted with butter in three centers) consisted of 95 en% (percent of energy content fat load) fat, of which 60% was saturated fat, 3 en% carbohydrate, and 2 en% protein. Based on the predicted metabolic rate (World Health Organization, technical report series 724, Geneva, 1985) the energy content was fixed at 50% of predicted basal metabolic rate. Subjects were asked to drink the test meal within 10 min.
Metabolic rate
The experimental room was kept thermoneutral at 25 C. Energy expenditure (EE) and respiratory quotient (RQ) were measured using open circuit ventilated hood systems used routinely at each center during 30 min preprandial and during 3 postprandial hours. All equipment and procedures were standardized for the different centers using a standardized validation program. Before the start of the study, validation was assessed by 10 alcohol-burning tests per center. Within-subject variation was assessed by running repeated measurements on the same day from 10 lean and/or obese fasting subjects per center. The mean and SD variation in RQ were 0.668 ± 0.006. Likewise the within-subject coefficient of variation was 2.73 ± 1.10 and 2.89 ± 1.19% for RQ and EE, respectively.
Blood sampling
At least 30 min before the start of the resting measurement, a Teflon catheter was inserted in an antecubital forearm vein for blood sampling. Blood was drawn in the fasting state and every 60 min after the test meal. Concentrations of glucose, free fatty acids (FFAs), insulin, cortisol, and triacylglycerols (TAGs) were determined pre- and postprandially. IGF-I and leptin were determined only in baseline fasting samples.
Biochemical analysis
All blood analyses were performed in the laboratory of one of the centers or at a subcontracted commercial laboratory (for cortisol and IGF-I). Plasma glucose concentrations (ABX Diagnostics, Montpellier, France) and TAG (Sigma, St. Louis, MO; ABX Diagnostics) were measured on a COBAS MIRA automated spectrophotometric analyzer (Roche Diagnostica, Basel, Switzerland). FFAs (NEFA C kit; Wako Chemicals, Neuss Germany) were measured on a COBAS FARAH centrifugal spectrophotometer (Roche Diagnostica). Standard samples with known concentrations were included in each run for quality control. Plasma insulin and serum leptin concentrations were measured with a double-antibody RIA (Insulin RIA 100; Kabi-Pharmacia, Uppsala, Sweden; human leptin RIA kit, Linco Research, Inc., St. Charles, MO). Cortisol and IGF-I were both determined with ELISAs (IGF-I: Diagnostics Systems Laboratories, Inc., Webster, TX; cortisol: ADVIA Centaur, Bayer Health Care LLC, Tarrytown, NY).
Calculations
EE was calculated according the equation of Weir (16). Fat oxidation was calculated according to the equations of Frayn (17). In these calculations, nitrogen excretion was assumed to be similar to daily nitrogen intake. The homeostasis model assessment index for insulin resistance (HOMAIR) was calculated from fasting glucose and insulin according to the equation of Matthews et al. (18). For comparing postprandial responses, the area under the curve (AUC) was calculated according to the trapezium rule. Respiratory quotient, fat oxidation, and blood parameters were evaluated with two types of variables: the fasting value and the AUC during the postprandial period. Fasting fat oxidation was expressed both in absolute values (grams per minute) and relative to EE, and postprandial fat oxidation was expressed relative to EE AUC (percent EE or fat oxidative percentage).
Statistical analysis
Data are expressed as mean ± SD or mean and 95% confidence intervals. Statistical analysis was performed with SPSS 10 for Macintosh (SPSS Inc., Chicago, IL). All variables were checked for normal distribution and variables with a skewed distribution were ln-transformed to satisfy conditions of normality.
Fasting and postprandial RQ and fat oxidation were compared among predefined categories of BMI (World Health Organization, 2000) using analysis of covariance with adjustment for body composition (when appropriate) and in case of postprandial fat oxidation for fasting fat oxidation (as percent EE) and energy content of the meal [expressed as percent of measured resting EE (REE)].
Thereafter fasting and postprandial fat oxidation (expressed as percent of EE) were investigated as dependent variables in analysis of covariance. In the first analysis on fasting fat oxidation, the following parameters were entered: FM, BMI category, age, gender, center, HOMAIR, waist to hip ratio (WHR), physical activity, dietary fat intake, and fasting levels of FFAs (model 1). The fasting levels of the hormones IGF-I, leptin, and cortisol were then included (model 2). The impact of these factors is described as beta or regression coefficients.
In the analysis on postprandial fat oxidation, the model included in addition to FM, BMI category, age, gender, center, the energy content of the test meal, HOMA IR, WHR, physical and habitual dietary fat content, and fasting fat oxidation (model 1). Thereafter postprandial FFAs, TAG glucose, and insulin were entered (model 2). Finally, fasting IGF-I, leptin, and postprandial cortisol were entered (model 3). Interactions between main effects were taken into account and were retained within the models when significant. To avoid multicollinearity in the regression model, independent variables with a correlation greater than 0.8 were not simultaneously included into the model.
Not all subjects fulfilled all the criteria for inclusion. An extra analysis excluding subjects with race other than white Caucasian (n = 4), women not being premenopausal (n = 8), subjects taking prescriptive medicine (n = 11), not meeting criteria for weight stability, and fasting glucose above 7 mM (n = 14) was performed afterward. Exclusion of these subjects did not affect the outcomes of the analyses.
| Results |
|---|
|
|
|---|
30 kg/m2) were slightly older, compared with the lean. As expected, percentage body fat, fat-free mass (FFM), FM, waist circumference, WHR, and habitual dietary fat intake and HOMAIR increased with increasing BMI, whereas level of habitual physical activity (as assessed by the Baecke questionnaire) decreased. Fasting concentrations of FFAs, glucose, insulin, and TAG increased with increasing BMI (Fig. 1
|
|
Figure 1
represents unadjusted circulating concentrations of metabolites before and after the high fat load according to BMI category. Postprandially there was an initial decrement (1 h) in circulating FFAs, followed by a slight increase from 1 to 3 h after the meal (Fig. 1A
). Postprandial FFA concentrations (Fig. 1A
) were significantly different between groups (P < 0.001), with the lowest overall values in the lean subjects, intermediate values in the group between BMI 3040 kg/m2, and the highest values in the most obese subjects. TAG concentrations increased postprandially and were higher in obese subjects, compared with the lean (Fig. 1B
, AUC, P < 0.001). Glucose was decreased at 1 h after the meal and increased again from 1 to 3 h postprandially (Fig. 1C
). Postprandial glucose (Fig. 1C
, AUC) was significantly lower in the lean group, compared with the obese groups. Insulin increased after the high-fat meal (Fig. 1D
), and postprandial insulin values increased with BMI category (AUC, P < 0.001). Postprandial cortisol concentrations decreased after the meal (P < 0.01), whereas postprandial cortisol (AUC) was not different between the BMI categories (data not shown).
Fat oxidation
Figure 1
, E and F, represents unadjusted values for EE and fat oxidation before and after the high fat load according to BMI category. Data on fat oxidation and RQ during fasting and in the postprandial state according to BMI categories are indicated in Table 2
.
|
REE increased with increasing BMI category. The fasting RQ decreased with BMI category, whereas adjustment for FM resulted in similar values across BMI categories. Fasting fat oxidation (milligrams per minute) increased with increasing BMI category, whether unadjusted, adjusted for FFM, or when expressed as percentage of REE (Table 2
). The beta coefficient for the effect of FFM on fat oxidation was 1.02E-03 (P < 0.01). FM significantly contributed to fat oxidation, adjusted for FFM (beta coefficient 4.82E-04, P < 0.001). Adjustment for both FFM and FM resulted in similar mean values for fat oxidation across the different BMI categories. Furthermore, adjustment of fat oxidative energy percentage for FM resulted in similar values among the different BMI categories (Table 2
).
Table 3
shows the results of an analysis of covariance with fasting fat oxidation (as percent of EE adjusted for FM and center) as dependent variable and multiple determinants.
|
Postprandial fat oxidation
Values for postprandial EE, RQ, and fat oxidation were adjusted for corresponding fasting values and energy content of the meal. The absolute AUC of postprandial EE was not different between groups (Table 2
). Postprandial RQ and postprandial fat oxidation (AUC) expressed as percentage of EE (AUC) were significantly different between BMI categories (Table 2
). However, there was a significant interaction between BMI category and fasting RQ (P < 0.05) or fasting fat oxidation (P < 0.001), indicating a reduced postprandial fat oxidation in those obese subjects with a fasting fat oxidation below median (for fat oxidation this is indicated in Fig. 2A
).
|
|
| Discussion |
|---|
|
|
|---|
Obesity and fat oxidation
Fasting fat oxidation (as percent of EE) increased with BMI category or FM. From the slope of the regression on FM and fat oxidative energy percentage, we estimated that a 10-kg increase in FM may result in a 1.9% increase in fat oxidative energy percentage. These data are consistent with previous studies (7, 8), both reporting a close correlation between FM and fasting or 24-h fat oxidation in obese subjects, whereas this relationship was not confirmed in another study (9). Data showing a significant contribution of FM to fat oxidation have been criticized due to the fact that data may be confounded by a greater range of FM vs. FFM (9) and due to differences in glycogen status in lean and obese subjects induced by a different impact of the recommended diet the days before the study between both groups (19). In our study, the range for both FM and FFM was large (FFM: 36.698.6 kg and FM: 4.8116 kg). Moreover, the correlation coefficient between FM and FFM was 0.17, indicating a low covariation, and it is valid to simultaneously include these parameters in the regression model. Furthermore, our subjects were requested to remain on their habitual dietary intake on the days preceding the measurement. One of our inclusion criteria was that subjects had to be weight stable (<3kg over 3 months) preceding inclusion into our study, indicating that fat oxidation in the present study is reflecting habitual fat oxidation. Thus, the present data support a significant contribution of FM to fasting fat oxidation.
Postprandial fat oxidation (as percent of EE) was diminished with increasing BMI, which was most pronounced in those obese subjects with a fasting fat oxidation below median (see Fig. 2A
). This seems consistent with a previous studies showing that obese subjects adapt more slowly to high-fat feeding in comparison with lean subjects (5, 10, 20). Additionally, a reduced postprandial and/or 24-h fat oxidation has already been reported in subjects predisposed to obesity (12), postobese women, compared with never-obese subjects (11), and normal-weight subjects with a strong family history of obesity (21). The latter studies suggest that interindividual differences in the capacity to adapt fat oxidation to fat intake may translate into differences in weight gain over time, when subjects are exposed to a high-fat diet. Flatt (1) has argued that one mechanism to adapt to a high-fat diet is an increase in FM, which may then increase fasting fat oxidation (by promoting FFA availability) until a new equilibrium is reached in which fat oxidation equals fat intake. Our data support this concept showing that both FM and fasting FFAs were strongly associated with fasting fat oxidation. From the finding that the reduced fat oxidative capacity postprandially was present only in those obese subjects with a relatively low fasting fat oxidation, it may be speculated that in these subjects the ability to increase fasting fat oxidation by increasing FFA availability is limited. Possibly other mechanisms like a reduction in dietary fat intake or an increase in physical activity may have contributed to the achievement of a new fat balance in these subjects.
Interestingly, previous research has shown a lowered oxidative capacity of skeletal muscle in obese subjects (22), suggesting that skeletal muscle may be the site of localization of the reduced relative postprandial fat oxidation in the present study. Additionally, a reduced 24-h fat/carbohydrate oxidation, which has been shown to be a significant predictor of weight gain (13), was related to a reduced skeletal muscle lipoprotein lipase activity (23), also supporting a role of skeletal muscle in the impaired regulation of fat oxidation.
Impact of insulin resistance
Independent of body composition, HOMAIR was negatively associated with fasting fat oxidation, indicating that insulin resistance per se is associated with a reduced fat oxidation. This seems consistent with data showing that skeletal muscle of insulin-resistant or type 2 diabetic subjects may be characterized by a lowered capacity to oxidize fatty acids during fasting conditions (24, 25), which may divert fatty acids toward an increased lipid accumulation in skeletal muscle, which is a strong marker of skeletal muscle insulin resistance (26).
Furthermore, HOMAIR was positively associated with postprandial fat oxidation This is consistent with the concept that skeletal muscle of insulin-resistant subjects may be characterized by an inability of insulin to suppress fat oxidation during postprandial conditions (24, 27), possibly related to an increased FFA supply as a result of the insulin resistance of lipolysis (postprandial FFA was significantly related to postprandial fat oxidation) and/ or an impaired suppression of muscle FFA uptake and oxidation. Thus, this inability to switch between carbohydrate and fat oxidation related to insulin resistance (metabolic inflexibility in substrate oxidation) represents an impaired ability to increase fat oxidation during fasting and a lowered ability of insulin to suppress fat oxidation during postprandial conditions (24, 27, 28). Interestingly, a recent study showed that the metabolic switching between substrates in vivo was preserved in human myotubes, separated from their neuroendocrine environment, supporting the hypothesis that the metabolic flexibility of substrate oxidation is an intrinsic property of skeletal muscle (29).
Impact of leptin
We found a negative independent association between fasting leptin and oxidative fat energy percentage. At first sight, this seems in contrast with previous findings showing that leptin may have a stimulatory effect on thermogenesis and fat oxidation (30, 31) and that leptin gene polymorphisms have been shown to be associated with a lower fat oxidation rate by 40% after a glucose load (32). However, it is possible that hyperleptinemia is a reflection of leptin resistance resulting in an unmatched secretion and clearance of leptin. Interestingly, in human muscle it has been shown that leptin resistance in obese subjects is associated with impaired muscle fat oxidation (33). Thus, the negative association between fasting leptin and fat oxidative energy percentage in our study may possibly reflect that some degree of leptin resistance in obesity may reduce fasting lipid oxidation.
Gender effects
Adjusted fasting fat oxidation was lower in females, compared with males, which is consistent with the previous literature (34). This reduced fasting fat oxidation may contribute to the increased fat storage in women, compared with men. On average, adjusted postprandial fat oxidation was higher in females. The underlying mechanism is not clear, at present, but may possibly be related to gender differences in postprandial lipolysis and fat storage (34). The interaction between fasting fat oxidation and gender in relation to postprandial fat oxidation remains to be elucidated.
In summary, postprandial fat oxidation (as percent EE) decreased with increasing BMI category, an effect present only in those obese subjects with a relatively low fasting fat oxidation. Independent of body composition, insulin resistance was associated with a reduced fasting fat oxidation but an increased postprandial fat oxidation (as percent EE), supporting the concept that insulin resistance may be associated with an inability to regulate fat metabolism (metabolic inflexibility). In conclusion, the present data indicate an impaired capacity to regulate fat oxidation in the obese insulin-resistant state, which is hypothesized to play a role in the etiology of both obesity and insulin resistance.
| Footnotes |
|---|
T.I.S. is a member of the Sanofi Aventis Advisory Board in Denmark. All other authors have nothing to declare.
First Published Online January 31, 2006
Abbreviations: AUC, Area under the curve; BMI, body mass index; EE, energy expenditure; 95 en%, percent of energy content load; FFA, free fatty acid; FFM, fat-free mass; FM, fat mass; HOMAIR, homeostasis model assessment index for insulin resistance; REE, resting EE; RQ, respiratory quotient; TAG, triacylglycerol; WHR, waist to hip ratio.
Received July 18, 2005.
Accepted January 23, 2006.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
J. E. Galgani, C. Moro, and E. Ravussin Metabolic flexibility and insulin resistance Am J Physiol Endocrinol Metab, November 1, 2008; 295(5): E1009 - E1017. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. A. Houmard Intramuscular lipid oxidation and obesity Am J Physiol Regulatory Integrative Comp Physiol, April 1, 2008; 294(4): R1111 - R1116. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Klimcakova, J. Polak, C. Moro, J. Hejnova, M. Majercik, N. Viguerie, M. Berlan, D. Langin, and V. Stich Dynamic Strength Training Improves Insulin Sensitivity without Altering Plasma Levels and Gene Expression of Adipokines in Subcutaneous Adipose Tissue in Obese Men J. Clin. Endocrinol. Metab., December 1, 2006; 91(12): 5107 - 5112. [Abstract] [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 |