Endocrinology and Metabolism

Calorie restriction or exercise: effects on coronary heart disease risk factors. A randomized, controlled trial

Luigi Fontana, Dennis T. Villareal, Edward P. Weiss, Susan B. Racette, Karen Steger-May, Samuel Klein, John O. Holloszy


Coronary heart disease (CHD) risk factors and the risk of CHD increase with increased adiposity. Fat loss induced by negative energy balance improves all metabolic CHD risk factors. To determine whether fat loss induced by long-term calorie restriction (CR) or increased energy expenditure induced by exercise (EX) has different effects on CHD risk factors in nonobese subjects, we conducted a 1-yr controlled trial involving 48 nonobese subjects who were randomly assigned to one of three groups: CR, 20% CR diet (n = 18); EX, 20% increase in energy expenditure through daily exercise with no increase in energy intake (n = 18); or HL, healthy lifestyle guidelines (n = 10). Subjects were 29 women and 17 men aged 57 ± 3 yr, with BMI 27.3 ± 2.0 kg/m2. Assessments included total body fat by DEXA, lipoproteins, blood pressure, HOMA-IR, C-reactive protein (CRP), and estimated 10-yr CHD risk score. Body fat decreased by 6.3 ± 3.8 kg in CR, 5.6 ± 4.4 kg in EX, and 0.4 ± 1.7 kg in HL, which corresponded to reductions of 24.9, 22.3, and 1.2% of baseline body fat mass, respectively. These CR- and EX-induced energy deficits were accompanied by reductions in most of the major CHD risk factors, including plasma LDL-cholesterol, total cholesterol/HDL ratio, HOMA-IR index, and CRP concentrations that were similar in the two intervention groups. Data from the present study provide evidence that CR- and EX-induced negative energy balance result in substantial and similar improvements in the major risk factors for CHD in normal-weight and overweight middle-aged adults.

  • cholesterol
  • C-reactive protein
  • insulin resistance

coronary heart disease (CHD) is the major cause of death in the United States (25). The relative risk of CHD and CHD risk factors increase with increased adiposity (2, 12). We (9) have found that surgical removal of large amounts of subcutaneous abdominal fat does not improve the metabolic CHD risk factors associated with obesity. Therefore, it is likely that fat loss induced by negative energy balance, which decreases visceral fat mass and fat cell size, is necessary to achieve metabolic benefits. Negative energy balance can be achieved by reducing energy intake or increasing energy expenditure. It is not known whether long-term negative energy balance induced by exercise alone or by caloric restriction alone is more beneficial in reducing CHD risk factors. However, data from epidemiological and physiological studies suggest that exercise has weight loss-independent benefits on CHD risk (1, 4, 5, 19, 21).

We conducted a 1-yr randomized, controlled trial in middle-aged lean and overweight men and women to evaluate the effect of body fat reduction induced by a 20% increase in energy expenditure alone or by a 20% decrease in energy intake alone on metabolic risk factors for CHD. We hypothesized that fat reduction induced by exercise would have greater beneficial effects on CHD risk factors than similar fat reduction induced by CR.



Forty-eight men and women aged 50–60 yr with a body mass index (BMI) in the 23.5–29.9 kg/m2 range, recruited from the St. Louis metropolitan area, participated in this study. Participants were weight stable for ≥3 mo and were nonsmokers; females were postmenopausal. Individuals who participated in regular exercise more than twice a week were excluded. This study was approved by the Human Studies Committee and the General Clinical Research Center Scientific Advisory Committee of Washington University School of Medicine. All subjects gave informed consent before their participation. The effects of CR and EX on body composition, insulin sensitivity, and bone and skeletal muscle mass in these subjects were reported recently (18, 2729).

Study design.

At baseline, all participants underwent a comprehensive medical examination, routine blood tests, and an electrocardiogram. Potential subjects were excluded if they had 1) a history of diabetes or a fasting blood glucose value ≥126 mg/dl, 2) a history or clinical evidence of CHD, stroke, or lung disease, 3) a resting blood pressure of ≥170 mmHg systolic and/or ≥100 mmHg diastolic, or 4) a recent history or evidence of malignancy.

Eligible participants were randomized, with stratification for sex, to one of three groups in a 2:2:1 sequence: caloric restriction (CR), exercise (EX), or healthy lifestyle (HL) for 1 yr. Forty-eight adults (30 females, 18 males) began the intervention.

EX intervention.

The goal of the EX intervention was to induce an energy deficit comparable to that of the CR intervention but by increasing daily energy expenditure through exercise without changing caloric intake. The EX prescription began at an ∼16% increase in energy expenditure above baseline total energy expenditure (TEE) during the first 3 mo and was increased to 20% for the subsequent 9 mo. Our exercise trainers worked with participants individually to establish and monitor their exercise routines, providing advice, encouragement, and weekly exercise prescription updates. To avoid an increase or decrease of energy intake, the study dietitians periodically monitored energy intake using 7-day food diaries and provided consultation as needed. Participants were given the options of utilizing our exercise facility (which contains treadmills, a running track, cycle ergometers, rowing ergometers, elliptical machines, and stair climbers), another health club of their choosing, equipment in their homes, or outdoor exercise to achieve their energy expenditure goals. Because this was not a training study, exercise intensity was not prescribed, and exercise sessions could be completed in one or several daily bouts. All the participants were required to use a wrist watch-type heart rate (HR) monitor (S610; Polar Electro Oy, Kempele, Finland) which stored exercise-specific data for gross energy expenditure, HR, exercise duration, and the number of exercise sessions performed. Moreover, the participants met with the study dietitians on a weekly basis during which body weight was measured, the data from their HR monitors were downloaded, food records were analyzed and consultation was provided to avoid an increase in food intake.

CR intervention.

The goal of the dietary weight loss intervention trial was to decrease daily energy intake without changes in energy expenditure for the duration of the 1-yr intervention. To match the negative energy balance of the exercise group, the CR prescription began at an ∼16% decrease in energy intake below baseline daily TEE during the first 3 mo and was increased to 20% for the subsequent 9 mo. Diet prescriptions were based on energy intake at baseline, which was assumed to be equal to TEE as determined by the doubly labeled water (DLW) method (18). Baseline TEE was assessed during two 2-wk periods (representing average TEE over 4 wk). Each subject was given a prescription of total calories to consume daily (and how many calories to remove relative to their baseline intake) to achieve the desired percent CR. The macronutrient composition of the diet was flexible to accommodate individual preferences; a multivitamin supplement was given to all participants. For 5 days during the first month of the CR intervention, participants received all meals from the General Clinical Research Center metabolic kitchen. The participants met individually with the study dietitians on a weekly basis for measurement of body weight and instruction regarding strategies for reducing energy intake. Participants also attended weekly group meetings led by a dietitian and a behavioral psychologist. Participants were encouraged to record their food and beverage consumption daily using either 7-day food diaries or the Balance Log program (HealtheTech, Golden, CO) on Palm Pilots (Palm, Sunnyvale, CA) as a means of self-monitoring and to enable the dietitians to provide specific recommendations each week.

HL group.

The HL group received general information about a healthy diet, were offered free yoga classes, and served as a control group. The healthy lifestyle advice was provided in a brief consultation with a study dietitian and was centered on eating less high-fat food and increasing intake of fruits, vegetables, and whole grains. They were also encouraged to increase physical activity in activities of daily living, like walking, dancing, and stair climbing. At 6 mo, they received a newsletter containing similar educational themes and lower-fat recipes. HL participants did not receive diet or exercise prescription.

Body composition and TEE.

Body weight was measured in duplicate in the morning following a 12-h fast with the subject wearing a hospital gown and no shoes. Baseline body weight was calculated as the mean of five weekly, gowned weights measured during the 4-wk baseline period. Twelve-month body weights represent the mean of three weekly weights obtained at the beginning, middle, and end of the 2-wk assessment periods. Height was measured without shoes to the nearest 0.1 cm. BMI was calculated as weight divided by the square of height (kg/m2). Whole body fat mass (FM), fat-free mass (FFM), and %FM were assessed by dual-energy X-ray absorptiometry (DEXA, Delphi W; Hologic, Waltham, MA, software version 11.2). The reported body composition values represent the mean of two to three DEXA scans at baseline and two DEXA scans at 12 mo. TEE was measured by the DLW method in the three groups as described previously (18).

Blood analyses.

A venous blood sample was taken to determine lipid and hormone concentrations after subjects had fasted for ≥12 h. In the EX group, blood samples were obtained ≥48 h after the last exercise session. Measurement of serum lipid and lipoprotein-cholesterol concentrations was performed in the Barnes-Jewish Hospital Laboratory. Cholesterol and glycerol-blanked triglycerides (TG) were measured by automated enzymatic commercial kits (Miles-Technicon, Tarrytown, NY). High-density lipoprotein cholesterol (HDL-C) was measured in plasma after precipitation of apolipoprotein B-containing lipoproteins by dextran sulfate (50,000 MW) and magnesium. These methods are continuously standardized by the Lipid Standardization Program of the Centers for Disease Control and Prevention. Two-hour, 75-g oral glucose tolerance tests were performed at baseline and at the end of the intervention. Plasma glucose was measured by the glucose oxidase method (YSI Instruments, Fullerton, CA), and insulin was measured by radioimmunoassay. Insulin resistance was calculated using homeostasis model assessment of insulin resistance {HOMA-IR = [fasting glucose (mmol/l) × fasting insulin 59]/22.5} (14). C-reactive protein (CRP) was measured using a high-sensitivity ELISA kit (ALPCO Diagnostics, Windham, NH).

Blood pressure.

Blood pressure (BP) was measured with an oscillometric BP monitor (Dinamap Procare 200; GE Healthcare, Waukesha, WI) with the patient in the supine position in the morning after a 12-h fast. In the EX group, BP was measured ≥48 h after the last exercise session. BP was measured by a certified nurse, and the measurement process was regularly monitored to assure protocol adherence.

Calculation of the 10-yr CHD Framingham risk score.

The risk score was calculated for each subject by using the 10-yr CHD risk score of Wilson et al. (27) at baseline and at 1 yr. In this algorithm, subjects receive a point score based on categorical values of age, total cholesterol, HDL-C, BP, smoking, and diabetes. The scoring sheet is available in the study by Wilson et al. and online at www.nhlbi.nih.gov/about/framingham/riskabs.htm (30).

Statistical analyses.

All participants who provided both baseline and 1-yr data were included in the analyses. Baseline characteristics were compared between groups using χ2 tests or Fisher's exact test for categorical variables, and analysis of variance for continuous variables. Analysis of covariance was used for between-group comparisons of data collected at baseline and at 1 yr, with the 12-mo value as the dependent variable and the baseline value as the covariate. When the overall model was significant, Tukey's adjusted pairwise comparisons of groups were performed. Paired t-tests were used for within-group comparisons. Three subjects changed lipid-lowering drug status during the study (2 EX, 1 HL). These subjects were not included in the analysis of lipids, lipoproteins, (hs)CRP, and 10-yr CHD risk. All statistical tests were two tailed, and significance was accepted at P ≤ 0.05. Data are presented as means ± SD at each time point, and for the change between baseline and 12 mo. All analyses were performed using SAS software, version 9.1.3, of the SAS System for Linux (SAS Institute, Cary, NC).


Of the 379 volunteers assessed for eligibility, the screening procedures excluded 212 women and 109 men because of a history or clinical evidence of obesity, diabetes, CHD, stroke, hypertension, or a history of malignancy, and 63 refused to participate. Of the remaining 58 subjects who were eligible and interested, 10 withdrew before baseline testing due to study-related issues (n = 7) and personal issues (n = 3). Of the 48 participants who began the intervention, 46 completed the study; one female was dropped from the CR group at 6 mo due to inability or refusal of the subject to schedule tests, and one male dropped out of the EX group at 9 mo for medical reasons unrelated to the intervention. Table 1 presents sex, age, race, and BMI data for the three groups. Analysis revealed no significant differences among groups at baseline, except for age. The CR and HL participants were younger than the EX participants (P = 0.002; Table 1). However, inclusion of age as a covariate in the outcome analyses did not affect the study findings (data not shown); therefore, age was not included as a covariate in the reported data.

View this table:
Table 1.

Baseline study subject characteristics

Intervention adherence.

Detailed information about adherence to the interventions has been published previously (18, 29); however, a brief summary of these data is provided here to assist in the interpretation of the data in the present report. According to 7-day food diaries, energy intake decreased by ∼300 kcal/day in the CR group and did not change significantly in the EX and HL groups. Based on 7-day physical activity recall questionnaires, physical activity energy expenditure increased by ∼4 MET-h/day (∼36%) in the EX group but did not change significantly in the CR or HL groups. Furthermore, on the basis of the HR monitor data, the participants in the exercise group exercised 5.8 days/wk and 62 min/session at 71% of maximal HR and had an average gross exercise energy expenditure of 317 kcal/day.

Body weight and composition.

At baseline, 38 of 46 participants were overweight. Mean BMI was 27.3 ± 2.0 kg/m2 (range 23.2–29.9 kg/m2). Body fat averaged 39% in women and 26% in men. Weight loss averaged 6.6 ± 5.5 kg in the EX group, 8.2 ± 4.8 kg in the CR group, and 1.2 ± 2.1 kg in the HL group. The effects of the interventions on body composition were published in detail previously (18). Briefly, BMI (kg/m2), decreased significantly in both the exercise group (27.1 ± 1.9 to 24.8 ± 2.6) and the CR group (27.1 ± 2.5 to 24.2 ± 2.8). Total body fat also decreased significantly: 5.6 ± 4.9 kg in the EX group, and 6.3 ± 3.8 kg in the CR group.

CHD risk factors.

Total cholesterol decreased significantly only in the CR group (Table 2). LDL-C decreased significantly in response to both CR and EX: −19.3 and −16.8% of the baseline values. The 1-year LDL-C values for the CR and EX groups were significantly lower than that of the HL group (Table 2). HDL-C increased similarly in both the CR and EX groups, although the increases did not achieve statistical significance. The ratio of total cholesterol to HDL-C decreased significantly in both the EX and CR groups and was significantly lower than that of the HL group at 1 yr. Surprisingly, in light of the well-documented TG-lowering effect of exercise training, the plasma TG concentration was not decreased in the EX group, whereas both the CR and HL groups had significant reductions in serum TG concentrations (Table 2). In a subanalysis, we removed any subjects who were on lipid-lowering medications at the beginning and throughout the study (n = 3, all in HL group). None of the results changed except that the P value of 0.008 for TG change in the HL group became 0.08, and the among groups P value for 10-yr risk changed from 0.03 to 0.06. The HOMA-IR index was significantly reduced in both the CR and EX groups compared with the HL group at 1 yr (Table 3). Eight volunteers in the EX group and seven in the CR and HL groups had glucose intolerance (2-h glucose concentration >140 mg/dl) at baseline. After 1 yr of intervention, seven volunteers with impaired glucose tolerance at baseline in the EX group, five in the CR group, and three in the HL group had normal glucose tolerance (2-h glucose concentration <140 mg/dl) (29). Plasma CRP concentration decreased significantly only in the CR group (Table 3). Systolic BP and diastolic BP decreased only in the CR group; however, the decreases did not achieve statistical significance (Table 3). The estimated 10-yr CHD risk decreased significantly only in the CR group compared with the HL group (Table 3).

View this table:
Table 2.

Lipid and lipoproteins before and after intervention

View this table:
Table 3.

Major CVD risk factors before and after intervention


In this 1-yr randomized trial, we compared the effects of fat loss induced by reduced energy intake or by increased energy expenditure on several CHD risk factors in healthy normal-weight and overweight middle-aged men and women. Our results provide evidence that CR- and EX-induced negative energy balance improves CHD risk profile to a similar extent. Twelve months of the CR and EX intervention programs resulted in a similar (∼25%) reduction in body fat mass in both groups. Energy deficits induced by CR and EX were accompanied by reductions in most of the major CHD risk factors, including plasma LDL-C concentration, total cholesterol/HDL-C ratio, CRP concentrations, and HOMA-IR index, that were similar in the two groups.

Data from long-term clinical trials (≥6 mo) suggest that diet therapy is more effective than exercise in reducing body weight and body fat mass in obese subjects (10, 15). Two studies conducted in young obese subjects have shown that short-term exercise and CR are equally effective in improving insulin sensitivity (7, 20). However, few long-term studies have addressed the independent effect of CR and EX on CHD risk reduction in lean or overweight subjects (6, 8, 23). In those studies, exercise resulted in minimal or no weight loss, because of increased energy intake, noncompliance with exercise, or both (6, 8, 23). Therefore, it was not known whether similar reductions in energy bioavailability by CR and EX, resulting in similar fat mass losses, were equally effective in reducing the metabolic-related CHD risk factors. We have found that long-term major reductions in total body fat and abdominal visceral and subcutaneous fat depots similar to those obtained by CR are achievable with regular exercise if food intake is kept constant (18). The similar improvement in CHD risk in response to the EX-induced and CR-induced fat loss obtained in our trial was unexpected and counter to our initial hypothesis. Actually, it could be argued that CR was somewhat more effective than exercise, because it resulted in significant improvements in CRP, 10-yr CHD risk, and TG, whereas the changes in these variables did not attain statistical significance in the exercise group.

Exercise has been shown to favorably alter lipids and lipoprotein concentrations, especially to raise plasma HDL-C and lower TG concentration, BP, and inflammation and improve insulin sensitivity (26). However, in the present study we found that only plasma concentrations of LDL-C, HOMA-IR index, and total cholesterol/HDL-C ratio were significantly improved by EX-induced fat loss. Serum HDL-C concentration was also improved, but the increase did not reach statistical significance, whereas serum TG concentration was unchanged. Possible explanations for the smaller-than-expected responses include the low initial TG level of the EX group, the relatively mild intensity of the exercise for most of the participants, and the interval of at least 48 h between the last exercise session and the blood collection and BP measurements. Exercise generally lowers plasma TG concentrations when TG are moderately to markedly elevated, and high-intensity exercise is more effective than low-intensity exercise (4, 11). Most of our subjects who were assigned to the EX intervention participated in regular, long periods of low-intensity exercise, mostly walking and jogging. We purposely waited 48 h after the last bout of exercise before conducting our follow-up studies so that we could evaluate the long-term rather than the acute effects of the exercise (26). The TG-lowering effect of exercise is mediated by an increase in lipoprotein lipase in skeletal muscle, and this adaptation appears to require rather vigorous exercise and reverses rapidly (3, 22). The BP-lowering effect of exercise is largely a short-term response that is generally lost within 24 h (24).

Despite the relatively low baseline values for most of the measured variables, our findings provide evidence that a 12-mo negative energy balance achieved through CR or EX results in major reductions in many CHD risk factors. For example, on the basis of the data generated from the clinical trials involving statins, both the EX- and CR- induced LDL-C drops obtained in this study should have resulted in a ∼15% reduction in the relative risk of CHD (16). Moreover, on the basis of the HOMA-IR index, many of our volunteers with impaired glucose tolerance at baseline were normal after 12 mo of intervention (17). We think that our findings have public health relevance for the lowering of lifetime risk for CHD. In a recent study it was reported that people with no CHD risk factors at age 50 yr have a very low remaining lifetime risk for CHD and markedly longer survival, whereas people with more than two major risk factors have high lifetime risk and substantially shorter survival (13). Moreover, our findings support the notion that sustained and inexpensive lifestyle changes can have marked beneficial effects on many CHD risk factors.

The strengths and limitations of our study warrant comment. A major strength is the randomized controlled trial design, which minimized the potential for selection bias. To avoid the potentially confounding effects of disease on CHD risk factors, individuals with overt disease were excluded. In addition, we carefully measured energy intake and energy expenditure by using both nutrition assessment software and direct analysis through DLW (18). Our study had a high retention rate of enrolled participants and good adherence to the study interventions, as shown by the successful weight reduction over 12 mo in both groups (18). The HL group did show some beneficial changes, including slight decreases in body weight, TG, and CRP concentration. These changes were probably due to the fact that the volunteers enrolled in the HL group did make some changes in their diet and exercise regimens. The relatively small number of participants limits our ability to generalize some of the conclusions. However, this was an exploratory, technically challenging, and highly labor-intensive study, and a recruitment target of 48 participants was determined to be realistic and consistent with our resources and was based on the sample needed to address feasibility and body composition outcomes in the parent study (18). Despite this restriction, the CR and EX groups underwent a number of changes that were large enough to be statistically significant despite the small number of subjects.

In conclusion, the results of this randomized, controlled trial provide evidence that CR- and EX-induced negative energy balances result in substantial and similar improvements in the major risk factors for CHD. These data support the conclusion that sustained lifestyle changes can have marked beneficial effects on CHD risk.


This work was supported by National Institutes of Health (NIH) Cooperative Agreement AG-20487, NIH General Clinical Research Center RR-00036, Diabetes Research Training Center DK-20579, and NIH Clinical Nutrition Research Unit DK-56341. E. P. Weiss was supported by NIH Grant AG-00078.


We are grateful to the study participants for their cooperation and to the staff of the Applied Physiology Laboratory and the nurses of the General Clinical Research Center at Washington University School of Medicine for their skilled assistance.

Present address for E. P. Weiss: Department of Nutrition and Dietetics, St. Louis University, St. Louis, MO.


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


View Abstract