Subject Selection

Patients with CHD, BMI >27 kg/m², and waist circumference >102 cm (men) or >88 cm (women) were eligible for participation. We chose these criteria instead of a BMI >25 kg/m² to avoid individuals with borderline excess adiposity with only a minimal need for weight reduction. Patients were enrolled from 2002 to 2006. A flow diagram of the recruiting process is given in Figure 1. We evaluated 114 subjects for eligibility. Subjects with diabetes mellitus were excluded, although 25 patients with a fasting glucose of 100 to 126 mg/dL were enrolled. Patients with severe deconditioning (peak aerobic capacity <14 mL O₂ · kg⁻¹ · min⁻¹) were excluded because in a preliminary nonrandomized trial such subjects were unable to markedly increase their exercise-related calorie expenditure and were less successful at accomplishing weight loss.²⁸ All subjects were hospitalized ≥3 months previously. The study population ultimately consisted of 74 individuals, 60 men and 14 women. Cardiac diagnoses included past myocardial infarction (n=34), coronary bypass surgery (n=34), percutaneous coronary intervention (n=28), or chronic stable angina (n=6). This protocol was approved by the University of Vermont Committee on Human Research and was registered as a clinical trial (NCT00628277). After providing informed consent, each patient underwent a clinical review to ascertain that evidence-based pharmacological therapies were being taken.²⁹ Baseline measures were taken at the University of Vermont General Clinical Research Center. Exercise was restricted for a minimum of 36 hours before all metabolic and exercise testing procedures.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Flow diagram of the recruiting process.

Body Composition and Physical Activity

Body composition measures included body weight, BMI, waist circumference, fat mass, and fat-free mass by dual energy x-ray absorptiometry (General Electric Lunar Prodigy, Madison, Wis). Abdominal visceral and subcutaneous fat areas were measured by CT scanning (General Electric Medical Systems, Milwaukee, Wis, and Philips Electronics N.V., Eindhoven, the Netherlands).³⁰ Physical activity energy expenditure was measured over 7 days with the doubly labeled water technique providing a noninvasive assessment of free-living activity energy expenditure after subtracting resting metabolic rate and thermic effect of food.³¹

Coronary risk factor assessments included insulin-stimulated glucose disposal, insulin, glucose, lipid profiles, resting blood pressure (BP), high sensitivity C-reactive protein, and plasminogen activator inhibitor-1 (PAI-1). Medications were withheld the morning of testing. BP was measured after 5 minutes in the seated position with a Dinamap (GE Medical Systems, Tampa, Fla) automated BP cuff. A mean of 3 measurements constituted the resting BP. Components of the metabolic syndrome were assessed, including waist circumference >102 cm (men) or >88 cm (women), systolic BP ≥135 mm Hg or diastolic BP ≥85 mm Hg, fasting triglycerides >150 mg/dL, high-density lipoprotein (HDL) cholesterol <40 mg/dL (men) or <50 mg/dL (women), and fasting blood glucose >100 mg/dL.³² The presence of ≥3 components defined the metabolic syndrome.

Insulin-stimulated glucose disposal was determined after an overnight fast with the euglycemic hyperinsulinemic clamp technique.^30,33 Testing was preceded by 3 days of standardized meals consisting of 200 to 250 g carbohydrate and 12 g fiber per 1000 kcal per day. Insulin-stimulated glucose uptake, an index of insulin sensitivity, was the average dextrose infusion rate during the final 30 minutes of the 3-hour clamp plus residual endogenous glucose production³⁰ expressed relative to fat-free mass.

The concentration of PAI-1 was determined by ELISA (Tintelize, Biopool, Umea, Sweden).³⁴ High-sensitivity C-reactive protein was measured with a colorimetric ELISA.³⁵ Isolated values of ≥10 mg/dL (3 of 74 at baseline) were eliminated from analysis unless C-reactive protein was elevated both before and after the 5-month intervention (1 of 71).

Peak aerobic capacity (peak V̇o₂) was measured during a symptom-limited treadmill test using a graded modified Balke protocol until volitional fatigue, cardiovascular symptoms, or ≥2-mm ECG ST-segment depression. Expired gas was analyzed with a SensorMedics Vmax 29c (Yorba Linda, Calif) with measurement of peak V̇o₂ in milliliters of O₂ per kilogram per minute. Dietary macronutrient intake was estimated at baseline and weeks 15 and 51 with 3-day dietary diaries (Food Intake Analysis System, Houston, Tex).

Study Framework

Subjects were randomized to high-calorie-expenditure exercise or standard CR using a distance function method in sets of 4, balancing gender and BMI. After the 4-month intervention, subjects entered a 1-month weight stabilization phase, continuing to exercise but maintaining weight at <1-kg variation from the 4-month weight. We present data measured at 5 months as the primary outcome time point. Subjects underwent testing identical to that at baseline. Subsequently, subjects entered a maintenance phase, performing most but not all of their recommended exercise offsite and keeping exercise records. Each group was instructed to adhere to the exercise protocol established during the 5-month exercise intervention. Monthly onsite reviews of dietary and exercise records were conducted with the dietitian and exercise physiologist, as well as biweekly calls from the dietitian. Data at 1 year also are presented. Preventive medications were kept steady throughout the 12-month protocol.

Behavioral Weight Loss

The behavioral weight loss program included 16 hourly group counseling sessions led by a dietitian³⁶ emphasizing dietary records, itemization of food, and caloric content. The daily caloric goal was 500 kcal less than predicted maintenance calories,³⁶ which was independent of the exercise program to which they were assigned. No specific recommendations were provided on macronutrient intake. Meal replacement plans were not used. Features of behavior modification included self-monitoring, stimulus control, problem solving, social assertion, goal setting, feedback, relapse prevention, and family involvement.³⁶

Exercise Protocols

The exercise prescription for the high-calorie-expenditure CR group emphasized longer-duration (45 to 60 versus 25 to 40 minutes per session), lower-intensity (50% to 60% versus 65% to 70% peak V̇o₂), and more frequent (5 to 7 versus 3 times a week) exercise than the standard CR group. Walking was the preferred exercise modality to maximize caloric expenditure versus weight-supported exercises (cycling or rowing), which burns fewer calories.³⁷ Simplistically, the high-calorie-expenditure exercise training protocol was to “walk often and walk far.” The high-calorie-expenditure group had an exercise expenditure goal of ≥3000 to 3500 kcal/wk, attained after 2 to 4 weeks of gradually lengthening the exercise bouts. After performing all sessions onsite for 2 weeks, high-calorie-expenditure exercise subjects performed 2 to 4 sessions a week in the home environment using a heart rate monitor, subsequently downloaded at the CR center to ascertain duration of exercise. Both groups eventually performed 1 to 3 sessions a week onsite with home exercise logs. Exercise logs were reviewed weekly with the exercise physiologist to estimate caloric expenditure and to ascertain compliance. The standard CR group protocol included 25 minutes of treadmill walking and 8 minutes on 2 of 3 ergometers: cycle, rowing, or arm. After 5 months, subjects transitioned to a primarily home-based exercise program continuing the same exercise prescription. Subjects were allowed to perform up to 1 session a week at the CR facility.

The authors had full access and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.

Statistical Analyses

Results are presented as mean±SD. Repeated-measures ANOVA was used to determine overall treatment efficacy and to examine secondary outcomes between groups. When results indicated a differential treatment effect, the analysis was followed by an F test for simple effects to examine changes over time in each group separately. A 2-sided value of P<0.05 indicated statistical significance. Analyses were based on all subjects who remained in the study at 5 and 12 months, without imputation of missing values. Regression analysis was used to explore factors associated with fat mass loss and metabolic syndrome components during the intervention period, controlling for group assignment. Multivariate ANOVA was conducted on the change from baseline for a panel of cardiometabolic risk factors associated with obesity. The majority of the variables were approximately normally distributed. Log transformations for several variables were considered but did not change the conclusions. Thus, all results presented are based on untransformed data. The presence of the metabolic syndrome and the number of components present were computed both at baseline and at the end of treatment, with a logistic regression used to examine whether the presence of this syndrome changed differentially in each group after treatment. Differences in changes in the number of components of the metabolic syndrome were examined with the Friedman test, and generalized estimating equations were used to examine changes across time in the prevalence of components of the metabolic syndrome. We estimated that a sample size of 34 subjects a group was required to detect a differential group fat mass loss of 1.9 kg, assuming a between-subjects SD for the weight change of 2.5 kg with α=0.05 and a power of 80% based on a preliminary study of high-calorie-expenditure exercise.²⁸

Baseline Characteristics

The mean age of study participants was 64±9 years (range, 44 to 84 year), mean BMI was 32±4 kg/m², (range 27 to 45 kg/m²), and mean waist circumference was 110±10 cm. Study groups were similar at baseline in terms of age, gender, body weight, and body fat distribution (Table 1). Groups exhibited a similar distribution of cardiac risk factor measures, including BP, glucose disposal, glucose, insulin, and cardiorespiratory fitness; however, total cholesterol, triglycerides, and ratio of total to HDL cholesterol were higher in the high-calorie-expenditure exercise group (Table 2). All subjects were taking preventive cardiovascular medications in accordance with evidence-based guidelines.²⁷ They included aspirin (99%), statins (84%), β-blockers (72%), angiotensin-converting enzyme inhibitors/blockers (36%), and clopidogrel (28%), with no differences in use between study groups. Study subjects had a low level of aerobic fitness comparable to patients entering clinical CR programs (Table 2)⁹; thus, they were not a physically elite subset of CHD patients. A total of 3 subjects dropped out of the study, all during the first month, 1 in the high-calorie-expenditure group and 2 in the standard CR group, with no dropouts from month 5 to 12.

Combined Group Responses

To evaluate the overall effects of exercise and weight loss, we combined study groups (n=71). Mean weight loss was 6.2±5.1 kg; fat mass loss was 4.6±3.9 kg; and peak aerobic capacity increased by 8% from 22±5 to 24±7 mL O₂ · kg⁻¹ · min⁻¹ (P=0.0002; Tables 1 and 2⇑). These changes were associated with pronounced favorable effects on cardiac risk factors, including reductions in insulin resistance and fasting concentrations of insulin, glucose, serum triglycerides, ratio of total to HDL cholesterol, high-sensitivity C-reactive-protein, and PAI-1 and an increased HDL cholesterol (all P<0.05; Table 2). The prevalence of metabolic syndrome was reduced from 59% to 31% (P<0.0001). The average number of metabolic syndrome components decreased from 2.76±1.02 to 1.96±1.05 (P<0.001). High-sensitivity C-reactive protein decreased after exercise and weight loss only in patients not taking a statin (n=9) (4.3±3.7 to 2.8±3.5 mg/dL; P=0.002) with no effect in patients taking a statin (3.0±3.5 to 2.7±3.0 mg/dL; P=0.25; n=57; P<0.05 between groups). Fat mass change was the only independent predictor of change in metabolic score (R=0.26, P<0.02, adjusted for group) compared with changes in other anthropometric variables, fitness, physical activity, and dietary intake.

Body Composition by Group Assignment

High-calorie-expenditure exercise training yielded more than double the weight loss (8.2±4 versus 3.7±5 kg; P<0.001) and fat mass loss (5.9±4 versus 2.8±3 kg; P<0.001) and a greater reduction in waist circumference (7±5 versus 5±5 cm; P<0.02) than standard CR exercise (Table 1 and Figure 2). Additionally, high-calorie-expenditure exercise yielded a greater loss of total and intraabdominal fat, measured by computed tomography scan (Table 1). Our goal of increasing exercise energy expenditure was fulfilled, with physical activity energy expenditure increasing 615±427 kcal/d (P<0.001) in the high-calorie-expenditure exercise group versus 169±318 kcal/d (P<0.001) in the standard CR group (P<0.001 between groups; Table 3). The deficit in dietary caloric intake was similar between groups (−312±532 kcal/d in the high-calorie-expenditure group versus −254±424 cal/d in the standard CR group; P<0.01 within groups; P=0.69 between groups). Success with weight reduction did not differ by gender (results not shown). Using multiple regression analysis and controlling for group assignment, we found that exercise compliance (R=0.54, P<0.01) and physical activity energy expenditure (R=0.29, P=0.02) were the only independent predictors of fat mass change (cumulative R²=0.54). Adherence to the study interventions was good, with exercise attendance in the high-calorie-expenditure exercise and standard CR groups at 85% and 84%, respectively, and attendance at weight-loss sessions of 77% and 72%, respectively (both P=NS between groups). Five subjects in the high-calorie-expenditure group and 3 in the standard CR group missed ≥1 exercise session (baseline to month 5) because of an exercise-related medical problem (P=0.76). These were mostly aggravations of previously existing musculoskeletal conditions in addition to musculoskeletal overuse injuries rather than acute injuries. No exercise-related cardiac events occurred, although 1 individual in the high-calorie-expenditure group and 2 subjects in the standard CR group missed sessions for cardiac hospitalizations. Subjects did not experience greater difficulty accomplishing the high-calorie-expenditure training protocol. This was assessed with a study-specific physical activity satisfaction questionnaire (7-point Likert scale; 1=high level of satisfaction with physical activity, 7=low level of satisfaction with physical activity). Study groups had identical mean baseline scores of 2.7±1.2 with an overall increase in physical activity satisfaction at 5 months (combined group score, 2.3±1.7; P<0.01) and no difference in 5-month scores between groups (2.1±1.1 in the high-calorie-expenditure group versus 2.5±1.3 in the standard CR group; P=NS).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Weight change by group assignment. ▴ Indicate patients in standard CR exercise program; •, patients in high-calorie-expenditure exercise group. *P<0.05 vs preceding time point within group; †P<0.05 vs standard CR group at same time point.

Insulin Sensitivity and Cardiac Risk Factors by Group Status

The high-calorie-expenditure exercise group experienced a significantly greater increase in insulin-mediated glucose disposal, reflecting a decrease in insulin resistance, the hallmark of the insulin resistance/metabolic syndrome (Table 2). The high-calorie-expenditure exercise group also experienced a greater reduction in the total-to-HDL-cholesterol atherogenic ratio (Table 2) of −13% versus −3% (P<0.01). When effects of study interventions on a panel of cardiac risk factors (insulin sensitivity, waist circumference, systolic and diastolic BPs, plasma glucose, triglycerides, HDL cholesterol, low-density lipoprotein cholesterol, cardiorespiratory fitness, PAI-1, and high-sensitivity C-reactive protein) were entered into a model using a multivariate ANOVA, the high-calorie-expenditure group experienced a significantly greater improvement in overall cardiometabolic risk profile than the standard CR group (P=0.03). The high-calorie-expenditure exercise group also experienced a greater reduction in the number of components of the metabolic syndrome compared with patients in the standard CR group: 3.1±1.2 to 1.9±1.1 components versus 2.4±0.7 to 2.0±1.0 components (P=0.01 between groups; Figure 3). These differences in weight reduction and cardiac risk factors were related to the differential exercise protocol because group subjects restricted calories similarly. Cardiorespiratory fitness improved similarly by group (each P<0.01 within groups; P=NS between groups; Table 2) from an identical baseline of 22 to 24 mL O₂ · kg⁻¹ · min⁻¹.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. For each paired component of metabolic syndrome, the left bar represents the prevalence of the component at baseline, and the right bar is the prevalence at 5 months. Changes between groups were not statistically significant. Trig indicates triglycerides. *P<0.05, †P<0.01.

One-Year Results

Results at 1 year showed that both groups exhibited a weight regain of 1 kg (1.3 kg for the high-calorie-expenditure group versus 0.9 kg for the standard CR group; each P<0.05; P=NS between groups) compared with the 5-month measures, but body weight and body fat remained significantly lower than at baseline (each P<0.001) (Figure 2). Favorable alterations in the ratio of total to HDL cholesterol, HDL, peak aerobic capacity, and PAI-1, evident at 5 months in the high-calorie-expenditure exercise group, were maintained at 1 year (results not shown).