|Home | About | Journals | Submit | Contact Us | Français|
There is debate over the independent and combined effects of dieting and increased physical activity on improving metabolic risk factors (body composition and fat distribution).
The objective of the study was to conduct a randomized, controlled trial (CALERIE) to test the effect of a 25% energy deficit by diet alone or diet plus exercise for 6 months on body composition and fat distribution.
This was a randomized, controlled trial.
The study was conducted at an institutional research center.
Thirty-five of 36 overweight but otherwise healthy participants (16 males, 19 females) completed the study.
Participants were randomized to either control (healthy weight maintenance diet, n = 11), caloric restriction (CR; 25% reduction in energy intake, n = 12), or caloric restriction plus exercise (CR+EX; 12.5% reduction in energy intake + 12.5% increase in exercise energy expenditure, n = 12) for 6 months.
Changes in body composition by dual-energy x-ray absorptiometry and changes in abdominal fat distribution by multislice computed tomography were measured.
The calculated energy deficit across the intervention was not different between CR and CR+EX. Participants lost approximately 10% of body weight (CR: - 8.3 ± 0.8, CR+EX: - 8.1 ± 0.8 kg, P = 1.00), approximately 24% of fat mass (CR: - 5.8 ± 0.6, CR+EX: - 6.4 ± 0.6 kg, P = 0.99), and 27% of abdominal visceral fat (CR: 0.9 ± 0.2, CR+EX: 0.8 ± 0.2 kg, P = 1.00). Both whole-body and abdominal fat distribution were not altered by the intervention.
Exercise plays an equivalent role to CR in terms of energy balance; however, it can also improve aerobic fitness, which has other important cardiovascular and metabolic implications.
A GROWING BODY of literature demonstrates that in comparison with a dietary restriction intervention alone, exercise, accompanied with or without weight loss, can lead to favorable changes in body composition including a reduction in abdominal adiposity (1–4). It is therefore reasonable to hypothesize, when exercise is included in a weight loss therapy, greater improvements in body composition and metabolic outcomes may be evident. Few randomized, controlled trials, however, have specifically tested this hypothesis and compared the metabolic responses of a dietary restriction intervention to a dietary restriction plus exercise intervention. Collectively these reports indicate that exercise, when combined with dietary restriction, leads to similar reductions in weight (4–8) but more substantial improvements in glucose tolerance (9, 10), lipoprotein profiles (6–8, 11, 12), and the risks associated with coronary heart disease (8, 11). There is debate, however, regarding the change in fat mass. One study using hydrostatic weighing (8) reported an additional 80% reduction in fat mass when exercise was added to dietary restriction, whereas others have reported no difference (5, 6, 10). Furthermore, only one study used dual x-ray absorptiometry (7) and to our knowledge none have assessed total or abdominal fat distribution by computed tomography or magnetic resonance imaging. Therefore, the important role of these interventions on fat depots and their relationship to metabolic outcomes cannot be explained in these studies.
A combined exercise and dietary restriction intervention could further enhance the metabolic effects of a diet-only intervention through exercise-mediated lipolysis in adipose tissue and mitochondrial biogenesis and improved glucose uptake in skeletal muscle. In most of the randomized, controlled trials mentioned above (5–12), the degree of dietary restriction applied to the treatment arms (diet only or diet + exercise) was carefully matched, with the exercise component added on top. However, the exercise, although supervised (in most cases), was not designed to achieve a predetermined energy expenditure nor was it quantified throughout the intervention in terms of energy cost. Therefore, the total energy deficit applied to the diet + exercise groups was larger than that of the diet-only group, explaining, at least in part, the observed enhanced metabolic responses in the exercise groups. There is a need therefore to clarify whether dietary restriction when combined with exercise leads to greater improvements in body composition and fat distribution than calorie restriction alone when the total energy deficit is carefully matched between groups. Therefore, a key secondary aim of CALERIE, a randomized, controlled trial designed to study the effects of calorie restriction on metabolic adaptation in overweight men and women (13) was to investigate and compare the changes in body composition and fat distribution. In this study, a 25% energy restriction was prescribed by diet only or diet combined with exercise for 6 months and both energy intake and energy expenditure were rigorously controlled and monitored. We hypothesized that changes in body composition and abdominal fat would be enhanced during a caloric restriction intervention that combined dieting and exercise.
Healthy, overweight [25 ≤ body mass index (BMI) < 30] men (aged 25 to <50 yr) and women (aged 25 to < 45 yr) were recruited from the local community by advertisement. Participants were excluded if they smoked; exercised more than twice per week; were pregnant, lactating, or postmenopausal; had a history of obesity (BMI > 32), diabetes, cardiovascular disease, eating disorders, psychological disorders, or substance abuse; or regularly used medications (except birth control). Details of the screening process have been previously described (13, 14). The study was approved by the Pennington Biomedical Research Center Institutional Review Board and the Data Safety Monitoring Board of CALERIE, and subjects provided written informed consent. The physical characteristics of the participants at baseline are summarized in Table 1.
After baseline testing, participants were randomized into one of three groups for 24 wk: control, healthy weight maintenance diet (five males, six females); CR group, 25% caloric restriction from baseline energy requirements (six males, six females); and CR+EX, 12.5% caloric restriction and 12.5% increase in energy expenditure through structured exercise from baseline energy requirements (five males, seven females). This study (13) also included a fourth group (low calorie diet). However, because the goal of this treatment group was to achieve a specific weight loss followed by weight maintenance (and not to sustain a defined level of CR), this group was excluded from this analysis. Subjects were stratified according to sex and BMI before randomization.
The baseline period was conducted over 5 wk to carefully establish individual energy requirements and thereafter to perform baseline testing. The energy intake required for weight maintenance during baseline and the subsequent energy deficit necessary to achieve the desired caloric restriction during the intervention were calculated from total daily energy expenditure assessed during two 14-d periods by doubly labeled water and from a 14-d period when participants consumed all meals prepared by the metabolic kitchen with adjustments for weight maintenance (13). During the last week of baseline, participants were admitted to the inpatient unit for 5 d during which body composition and metabolic assessments were conducted. The same inpatient stay was repeated at wk 12 (month 3) and 24 (month 6).
All participants received a diet based on the American Heart Association guidelines, 30% calories from fat, 15% from protein, and 55% from carbohydrate as previously described (13). During wk 1–12 and 22–24 of the intervention, participants were provided with all meals that were prepared by the metabolic kitchen at the center. On weekdays both breakfast and dinner were consumed at the center, whereas lunch, snacks, and weekend meals were packaged for take-out. During wk 13–22 participants self-selected a diet based on their individual calorie target. During the self-selected feeding, individual compliance to the prescribed dietary intervention was monitored from self-reported food records and changes in body weight collected and reviewed weekly during behavioral group or individual sessions.
Participants randomized to the CR+EX group increased their energy expenditure by 12.5% above baseline by undergoing a structured exercise regimen 5 d/wk. The exercise program was implemented gradually, and by wk 6 all participants were expending the required 12.5% of baseline energy expenditure. During the first 6 wk, energy intake was adjusted to maintain a total daily energy deficit of 25%. The amount of time necessary to expend the 12.5% calorie target was determined on an individual basis. Briefly, participants self-selected three exercise work-loads on a treadmill, stationary cycle, or stairmaster, and the oxygen uptake of these activities was measured by indirect calorimetry (V-max; Sensormedics, Yorba Linda, CA). The required exercise duration was then calculated from the average energy expenditure maintained during the exercise bout. Heart rate (Polar S-610; Polar Beat, Port Washington, NY) was also measured during the assessment of exercise energy expenditure and thereafter was used to inform the participants of the target heart rate required to be maintained during all in-patient and outpatient exercise sessions. The portable heart rate monitors were worn during all exercise session, and the data were downloaded and checked for compliance including average heart rate and number of minutes of exercise. At least three of the five weekly exercise sessions were performed under supervision at the center. The target energy cost was maintained at 403±63 kcal per session for women and 569 ± 118 kcal per session for men throughout the entire intervention, resulting in an average exercise duration of 53 ± 11 and 45 ± 14 min per session for women and men, respectively. To avoid any potential interference between acute exercise and clinical measures, no exercise was allowed for 2 d preceding admission to the clinic and during the in-patient stays.
Commencing with the induction of the weight maintenance diet at baseline, participants attended weekly meetings. Cognitive-behavioral techniques were used not only to teach subjects how to adhere to their meal and exercise plans but also to boost motivation and morale to the study interventions. Emphasis was placed on teaching participants the skills necessary to modify eating behavior and comply with the inventions during the outpatient phase of the study when participants were responsible for their own food preparation. To estimate the calorie content of foods while eating the self-selected diet, participants learned to use the health management resources calorie system (HMR, Boston, MA). Emphasis was also put on promoting perfect adherence to the exercise program.
Body weight was determined by the mean of two consecutive measurements obtained in the morning after a 12-h fast and corrected for the weight of a hospital gown. Whole-body percent body fat was measured using dual-energy x-ray absorptiometry (DXA; QDR 4500A; Hologic, Bedford, MA). Computed tomography (CT) was used for cross-sectional fat distribution in the abdomen, liver, and spleen (not reported), as previously described (15). Briefly eight cross-sectional, 1-cm-width scans were bench-marked to L4-L5 (GE Light Speed; General Electric, Milwaukee, WI), and scans were analyzed using commercially available software (Analyze; Analyze Direct, Lenexa, KS). Areas of bone, adipose tissue, and skeletal muscle were measured electronically by selecting regions of interest and defining adipose tissue and muscle for each subject as previously described (15). The interindividual coefficient of variation for percent fat (DXA) and sc and visceral adipose tissue mass (multislice CT) were determined in 16 (12 females, four males) individuals 28 d apart. The test-retest reliability was 1.5 ± 1.3% for percent fat and 3.0 ± 2.7 and 1.9 ± 1.7% for abdominal visceral and sc fat mass, respectively.
The energy balance of each participant was calculated at months 3 and 6 by the intake-balance method, i.e. from changes in body energy stores (fat mass and fat-free mass) vs. the energy intake calculated for weight maintenance at baseline. To convert the changes in fat mass (FM) and fat-free mass (FFM) to energy, the following energy coefficients were used; for weight loss, 1 g of FM = 9.3 kcal and 1 g of FFM = 1.1 kcal; for weight gain, 1 g FM = 13.1 kcal and 1 g FFM = 2.2 kcal (16). Therefore, the daily change in body energy stores from baseline to month 3 and baseline to month 6 is equal to the change in FM (multiplied by the energy coefficient) plus the change in FFM (multiplied by the energy coefficient), divided by the number of days between the body composition assessments by DXA.
Energy balance is then determined by dividing the daily change in energy stores by the energy intake required for weight maintenance and expressed as a percentage of baseline energy intake.
A positive number indicates a positive energy balance or weight gain, whereas a negative number indicates a negative energy balance or weight loss.
Data are expressed as means ± SEM and the level of significance for all statistical tests was set at P < 0.05. SAS (version 9.1; SAS Institute, Cary, NC) was used for analysis, and all analyses were performed by a biostatistician (A.A.) from the Biostatistics Core. The change and percent change from baseline to months 3 and 6 were computed for all variables, and ANOVA of the changes was used to determine differences. The factors tested in the model were treatment (CR, CR+EX, control), time (month 3, month 6), and sex and their interactions. Baseline values were included in the models as covariates. The statistical significance for all multiple comparisons was adjusted with respect to the Tukey-Kramer method to control for the type I error rate. One participant in the control group withdrew during the study (before month 3) for personal reasons. Data are therefore presented for 35 subjects.
Baseline energy requirements (determined from two 14-d doubly labeled water assessments and 14-d in-house feeding) were 2800 ± 158 kcal/d for CR, 2633 ± 134 for CR+EX, and 2873 ± 151 for control groups. Evaluation of energy deficit from the intake-balance method indicated a similar degree of energy restriction between CR and CR+EX. At month 3, the estimated energy deficit from baseline was -17.8 ± 1.8, -19.5 ± 1.7, and -2.1 ± 2.3% for CR, CR+EX, and control (P = 0.49 for CR vs. CR+EX). From baseline to month 6, the estimated energy deficit was -13.0 ± 1.4, -15.5 ± 1.6, and -0.3 ± 2.0% for CR, CR+EX, and control (P = 0.25 for CR vs. CR+EX). Whereas these data indicate a degree of energy restriction less than the prescribed 25% for CR and CR+EX, the calculations of energy balance do not take into account a progressive decrease in energy requirements that occur in parallel to weight loss and therefore underestimate the true energy deficit. Estimates of these adaptations provide values very close to the 25% energy deficit for the CR and CR+EX groups.
As expected, there was a significant reduction in body weight in both sexes (Fig. 1). The ANOVA identified a significant time by treatment interaction (P < 0.0001) but no effect of sex (P = 0.47) on the weight responses. Weight had declined -7.4 ± 0.5% for CR and -5.8 ± 0.3% for CR+EX at month 3 (CR vs. CR+EX, P = 0.71) and -10.4 ± 0.9% for CR and -10.1 ± 0.9% for CR+EX at month 6. The change in weight at both time points was significantly different from the control group (both P < 0.001). For the change in fat mass (Fig. 1), there was a significant time by treatment interaction (P < 0.001) and no effect of sex (P = 0.56). There was a similar decline in FM in CR and CR+EX at months 3 and 6 (P = 0.99). At month 6 the CR group lost an average 23.9 ± 3.0% (women: 21 ± 1%; men: 27 ± 6%) and CR+EX 24.8 ± 2.7% (women: 23 ± 3%; men: 27 ± 5%). With regard to FFM (Fig. 1), there was also a significant time by treatment interaction (P < 0.01) and a significant effect of sex (P = 0.02). Post hoc analyses indicate that for females, FFM was significantly reduced from baseline at each time point in all three groups, but the changes for CR and CR+EX were not different from controls or each other. For males, FFM was reduced from baseline in CR and CR+EX at months 3 and 6, and the reduction in FFM at month 3 in CR+EX was not different from baseline. The changes in FFM between CR and CR+EX were comparable at each time point.
The changes in abdominal and regional fat at month 6 are summarized in Table 2.
VAT was significantly reduced from baseline to months 3 (data not shown) and 6 in both sexes; however, the loss of fat in this compartment was substantially greater for men at both study time points (P = 0.05). There was a significant sex by treatment interaction (P < 0.001) with the post hoc analyses, indicating that for both intervention groups, the reduction in VAT was significantly greater in men, compared with women (P = 0.03). However, the changes between CR and CR+EX groups were not different in either sex (P > 0.9 for both).
For the sc compartment, however, there was a significant time-by-treatment interaction (P < 0.001) and no effect of sex (P = 0.36). At both time points, the decrease in SAT was not different between CR and CR+EX (month 3: P = 1.00, month 6: P = 0.98), and both groups were different from controls (P < 0.001 for all). These findings were true for both the superficial and deep portions of the SAT depot (Table 2).
To determine whether either intervention had a preferential change in any abdominal fat depots, we computed the changes in the ratios of visceral, superficial sc, and deep sc to each other and to the total fat mass in the abdomen. In addition, the ratio of visceral fat loss to total fat mass loss (by DXA) was also computed. As shown in Table 2, there was no change in the VAT to SAT in any of the groups at either time point. Similarly the ratio of fat loss in any abdominal compartment to the total abdominal fat depot [e.g. deep sc adipose tissue (DSAT) to total abdominal adipose tissue (TAT), superficial sc adipose tissue (SSAT) to TAT, or VAT to TAT] was also not different between groups (Table 2) in the women; however, men showed a preferential reduction in VAT, compared with total FM and deep sc, fat compared with superficial sc fat (Table 2). Interestingly changes in all fat depots occurred in a similar magnitude in the CR and CR+EX groups for both sexes (Fig. 2). For males the average reduction in fat from any depot was approximately 30% and approximately 25% for women. After the intervention the proportion of the visceral or sc depots to the total abdominal fat content did not change from the baseline in either intervention group (P = 1.00, for all) nor did the ratio of any abdominal fat depot to total fat mass (Fig. 2).
The aim of the current study was to compare the effects of 6 months of caloric restriction by dietary restriction only or dietary restriction in combination with exercise, on body composition and abdominal fat distribution in overweight men and women. Our data suggest that when the level of caloric restriction imposed is precisely matched and carefully controlled, the changes in body composition and abdominal fat distribution are not further enhanced by the addition of exercise, rejecting our hypothesis. A novel finding was that fat depots, regardless of their location, were reduced by approximately 30% in men and 25% in women in such a way that fat distribution throughout the whole body (DXA) and specifically within the abdominal compartment (CT) was not altered by caloric restriction.
A major strength of the study is that the prescribed energy deficits were carefully determined from two 14-d assessments of energy requirements by doubly labeled water and from a 14-d weight maintenance feeding period. Furthermore, the energy expended during exercise was clearly defined and quantified throughout the study for energy cost. Estimates of energy balance calculated from changes in energy stores vs. the energy intake provided for weight maintenance at baseline indicates that both CR and CR+EX had a similar degree of energy restriction throughout the study.
The losses in fat mass (-27% for men and -22% for women) and VAT mass (-31% for men and -24% for women) were similar in both treatment groups. However, due to the small sample size, it is conceivable that this result represents a type II statistical error. In other words, we could not reject the null hypothesis when a true difference exists between CR and CR+EX. One might then ask whether these treatments were different, what size difference could we detect with our methods and these sample sizes. For VAT mass, for example, the smallest difference we could detect between CR and CR+EX with 12 participants per group, alpha less than 0.05 and power 0.80 or greater is 210 g, an amount probably not clinically relevant.
The inability of caloric restriction to alter the distribution of fat suggests that individuals are genetically or epigenetically programmed for fat storage in a particular pattern and that this programing cannot be easily overcome by weight loss. In support of this contention, twin studies have shown that FM and regional fat distribution are largely determined by genetic factors (17, 18) and that genetic heritage can explain changes in body composition and fat distribution during positive and negative energy balance (19). In contrast to our results, many studies of caloric restriction or weight loss in obese and morbidly obese individuals report more profound reductions in visceral fat, compared with sc fat (2, 3). These findings have been explained by the larger VAT depots to begin with (20) and also because lipolysis is higher in the VAT vs. non-VAT depots (21). However, when results are appropriately adjusted for differences in VAT mass before treatment, the enhanced reduction in VAT was no longer apparent (22).
There is some debate whether the inclusion of exercise in weight loss interventions is protective against the loss of FFM (23). In our study, FFM was reduced with the 6-month intervention and was not different between the intervention groups. Our data suggest that FFM is reduced in parallel with the degree of caloric restriction and that regular aerobic exercise (5 d/wk), at least in nonobese individuals, does not preserve lean mass.
To assess fully the independent role of exercise on changes in body composition and fat distribution this study would need to be repeated with an exercise-only group in the study design. To produce a 25% energy deficit by exercise only, it would require approximately 120 min of exercise per day for women and approximately 90 min for men, a daunting task. Few randomized, controlled trials of this nature have therefore been attempted. A 12-wk weight loss intervention, induced by either diet or exercise in obese men (2) showed that despite equivalent weight loss (~7.5%) and changes in ab- dominal fat distribution, exercise produced a greater reduction in fat mass. When the same trial was repeated in obese women (3), a similar reduction in weight (~6.5%) was achieved by both groups, but the exercise group lost approximately 6% more fat mass and approximately 10% more fat from VAT. A possible explanation of these findings could be that exercise independently reduces fat loss in men and women and selectively targets VAT. Indeed, a recent randomized, controlled trial of exercise only (24, 25) supports this argument, showing that exercise reduces weight, FM, and VAT in a dose-dependent manner. However, retrospective analyses from both studies (2, 3) indicated that the energy deficit achieved by the exercise groups exceeded those of the diet groups. Therefore, it is difficult to conclude that exercise can independently lead to greater improvements in fat distribution when the energy deficit achieved is a confounding factor in the interpretation.
Despite its role in energy balance, exercise can produce health benefits independent of changes in body weight such as improvements in glucose tolerance (9) and aerobic fitness (26, 27). A low level of aerobic fitness has been identified as a stronger predictor of cardiovascular disease mortality than other risk factors including body fatness (28–30). Improved mitochondrial function and muscle oxidative capacity are believed to be important adaptations of exercise that link aerobic fitness to cardiovascular and metabolic disease (31). Participants in the CR+EX group significantly improved their peak aerobic capacity, whereas the CR and control groups did not (change in peak oxygen uptake; CR: 2.01 ± 1.76; CR+EX: 5.88 ± 1.27; control: -1.87 ± 1.05 ml·kg-1·min-1) and as detailed in a previous report (14), insulin sensitivity as measured by frequently sampled iv glucose tolerance test was increased at month 6 in both CR and CR+EX but only reached significance in the combined CR+EX intervention. We can speculate therefore that the combined CR+EX intervention may induce greater cardioprotective benefits through an improvement in aerobic fitness.
Participants in the CR+EX group self-selected their level of exercise intensity throughout the study because we believed compliance to the intervention would be enhanced with this strategy. Exercise intensity has been shown to influence body composition and cardiovascular and other metabolic outcomes in a dose-response manner (25, 32). Studies of exercise-induced weight loss suggest that high intensity exercise (65–80% maximal oxygen uptake) leads to greater improvements in visceral fat loss, insulin sensitivity, and lipoprotein profiles than moderate (40–55% maximal oxygen uptake) or low intensities. It might be argued that our approach may underestimate the role of exercise in the CR+EX intervention. Alternatively it could be argued that differences in body composition changes between these kinds of treatments are dependent on the resultant energy expenditure and energy deficit created by higher intensity exercise, rather than the exercise intensity itself (33, 34).
Previous randomized, controlled trials determining the role of dietary restriction in conjunction with exercise on body composition and metabolic risk factors have been confounded by the degree of energy restriction and therefore have produced conflicting results. Contrary to our hypothesis, our data indicate that when the degree of energy restriction is carefully matched, improvements in metabolic risk factors (body composition and fat distribution) in overweight men and women are dependent on the net energy deficit and that the inclusion of exercise does not contribute any added benefit in terms of changes body composition. Exercise therefore plays an equivalent role to caloric restriction in terms of energy balance; however, it can also improve aerobic fitness, which has other important cardiovascular and metabolic implications.
The authors thank the remaining members of Pennington CALERIE Research Team including: Donald Williamson, Walter Deutsch, Frank Greenway, Marlene Most, Jennifer Rood, James DeLany, Steven Anton, Emily York-Crowe, Enette Larson-Meyer, Catherine Champagne, Paula Geiselman, Michael Lefevre, Lilian de Jonge, Jennifer Howard, Jana Ihrig, Brenda Dahmer, Julia Volaufova, Darlene Marquis, Connie Murla, Aimee Stewart, Amanda Broussard, and Vanessa Tarver. Our thanks are extended to the excellent staffs of the Inpatient Clinic, Metabolic Kitchen, and Fitness Center. Finally, our profound gratitude goes to all the volunteers who dedicated so much time in participating in this very demanding research study.
This work was supported by U01 AG20478 (to E.R.). L.M.R. is supported by a Neil Hamilton-Fairley Training Fellowship awarded by the National Health and Medical Research Center of Australia (ID 349553). The CALERIE clinical trial registration number is NCT00099151 (clinicaltrials.gov).
Disclosure Summary: The authors have nothing to disclose.