High-Volume Exercise Program in Obese Bariatric Surgery Patients: A Randomized, Controlled Trial

Abstract

Weight regain is a problem among many bariatric surgery patients. Whether a high-volume exercise program (HVEP), a strategy to limit weight regain, is feasible in these patients is unknown. The feasibility of an HVEP in obese post-bariatric-surgery patients was determined by randomizing 33 Roux-en-Y gastric bypass (RYGB) and gastric banding (GB) surgery patients with a mean BMI of 41 ± 6 kg/m² to an HVEP or control group for 12 weeks. The HVEP group was instructed to expend ≥2,000 kcal/week in moderate-intensity exercise. All patients were counseled to limit energy intake. Treatment effect was assessed by repeated measures analysis. During the last 4 weeks of the study, 53% of the HVEP group expended ≥2,000 kcal/week and 82% expended ≥1,500 kcal/week. Step count, reported time spent and energy expended during moderate physical activity, maximal oxygen consumption relative to weight, and incremental area under the postprandial blood glucose curve were significantly improved over 12 weeks in the HVEP group compared to controls (group-by-week effect: P = 0.009–0.03). Both groups reported significant improvement in some quality-of-life scales. Changes in weight, energy and macronutrient intake, resting energy expenditure (REE), fasting lipids and glucose, and fasting and postprandial insulin concentrations were not different between the two groups. HVEP is feasible in about 50% of the patients and enhances physical fitness and reduces postprandial blood glucose in bariatric surgery patients.

Introduction

Severe obesity (defined as BMI of ≥40 kg/m²) is a significant public health problem in the United States. According to a national survey conducted in the year 2007–2008, nearly 6% of US adults are severely obese (1). In addition, from the year 2001 to 2005, the prevalence of severe obesity has increased twice as fast compared with the prevalence of a BMI ≥30 kg/m² (2). Severe obesity is associated with a number of major comorbidities (3) and markedly lessens life expectancy (4). It is also associated with a poor quality of life (QOL) (5).

Dietary therapy has been ineffective in treating severe obesity in the long term (6). This as well as the above issues and the advent of laparoscopic bariatric surgery procedures has led to an exponential increase in Roux-en-Y gastric bypass (RYGB) and gastric banding (GB) surgeries (7). Although these surgeries lead to substantial excess weight loss and complete resolution or improvement of comorbidities (8,9,10), the long-term results are more modest with considerable weight regain and attenuation in the recovery from comorbidities (11,12).

Studies in non-bariatric-surgery patients (13,14) and data from the National Weight Control Registry (15) have suggested that individuals may need to expend ≥2,000 kcal/week in moderate-intensity exercise in order to lose and/or maintain weight loss. Whether bariatric surgery patients, many of whom are severely obese even after weight loss, can exercise at this level is not clear. According to a case-control study (16) in which bariatric surgery patients were compared with sex- and weight-matched controls, only 30% of the former group reported expending ≥1,500 kcal/week in exercise compared to 70% of the latter group. Similarly, according to another case-control study (17) in which bariatric surgery patients who had lost a large amount of weight were compared with subjects who had lost a similar amount of weight through nonsurgical means, approximately 30% of the former group reported expending at least ≥2,000 kcal/week in physical activity compared to about 60% of the latter group. Several case series studies on the other hand, generally noted a higher reported participation rate in exercise in post-bariatric-surgery patients but the energy expended during exercise was not given (18,19,20,21,22,23) and the information on the type, duration, and intensity of exercise was not always provided (18,19,22). Also none of the above studies used an objective method of assessing physical fitness and the studies were not randomized, controlled trials thus preventing determination of whether an exercise program is feasible in this population. These limitations make it difficult to interpret whether bariatric surgery patients can exercise at a certain level. Our study was designed to assess the feasibility of a high-volume exercise program (HVEP) in obese bariatric surgery patients in a randomized, controlled trial. The level of exercise training, and thus physical fitness, was assessed using an objective measure. It was hypothesized that the exercise goals would be met by most of the subjects in the HVEP group and that it would lead to improvement in fitness in these subjects compared to the controls. The secondary outcome goals were weight loss, comorbidities, and health-related QOL.

Methods and Procedures

Subjects

We recruited 33 RYGB and GB surgery patients for the study. Eligibility requirements included meeting the class 2 (BMI: 35.5–39.9 kg/m²) or class 3 (BMI: ≥40 kg/m²) obesity criteria, exercising <20 min/day within the previous 3 months, undergone bariatric surgery at least 3 months earlier, and being 18–65 years old. Patients were not eligible for the study if they weighed >180 kg, had functional limitations, such as not being able to climb 10 stairs or walk for 0.4 km (0.25 miles), due to arthritis or other musculoskeletal cause, were on weight loss medications, had serious cardiovascular disease, uncontrolled hypertension, hematocrit <30%, chronic kidney disease, untreated thyroid disorders, pulmonary disease severe enough to preclude participation in exercise training, or major neuropsychiatric illnesses impeding competence or compliance, or were pregnant, lactating, or taking recreational drugs.

The study was conducted at the Clinical and Translational Research Center and the St Paul Fitness Center at UT Southwestern Medical Center at Dallas. The protocol for this study was approved by the institutional review board, and all patients gave written informed consent.

Experimental design

The subjects were randomized to a HVEP (n = 21) or a control (n = 12) group for 12 weeks using a 2:1 randomization ratio and stratified by the type of surgery.

Exercise intervention

The exercise goal in the HVEP group was to expend ≥2,000 kcal/week in moderate-intensity aerobic exercise at 60–70% of maximal oxygen consumption (VO_2max). The subjects were instructed to achieve these goals gradually and were asked to expend 500 kcal during the first week and increase by 500 kcal every week until they achieved their goal of ≥2,000 kcal/week. Each subject was asked to exercise on the treadmill at a certain speed and grade and on the cycle ergometer or rowing machine at a particular wattage that would correspond to 60–70% of her/his measured VO_2max. When exercising elsewhere, they were asked to use similar equipment and follow the same individualized protocol used in the fitness center. Subjects who preferred to walk outside or on the walking track were asked to measure the distance that they walked and asked to complete this distance in a time period to achieve a pace associated with an intensity of 60–70% of the measured VO_2max. Once the intensity was achieved and the subjects became more physically fit, they were asked to increase their intensity in order to maintain the same level of perceived exertion. The subjects were asked to exercise at least 5 days a week. The exercise was partially supervised and the subjects were asked to come to the fitness center at least once or twice a week. About a third of the subjects came to the fitness unit once or twice a week and about a third performed all of their exercise under our supervision. The exercise supervision was led by one of the investigators (P.S.) and one or more of the other investigators were always present during the supervised sessions. Energy expenditure per week was calculated approximately from the work data shown on the exercise equipment and from the duration and distance of walking relative to body weight. Exercise away from the fitness center was monitored by asking the subjects to keep an exercise diary and/or using heart rate monitors (Polar Vantage XL monitors, Kempele, Finland). The subjects were asked to bring the diaries and heart rate monitors to the fitness center or mail the diaries every 2 weeks. The participants were provided feedback on how much energy they were expending through the logs that were kept at the fitness center and using the exercise diaries that they kept when exercising away from the fitness center.

Diet intervention

To optimize their diet and prevent nutritional deficiencies, the subjects in both the groups were instructed to follow the dietary guidelines developed for post-bariatric-surgery patients by the American Society for Metabolic and Bariatric Surgery and other groups (24,25). The instructions were provided through individual dietary counseling. They were asked to limit their energy intake to about 1,200–1,500 kcal/day. This was achieved by limiting portion size, energy-dense foods, and energy-containing drinks. They were also asked to chew foods thoroughly and take at least 20 min to eat the main meal, not to drink during and 30 min before and after a meal to prevent vomiting, diarrhea, and quick return to hunger, consume >60 g of protein per day to preserve lean body mass (LBM), consume the protein food sources before the fat and carbohydrate food sources, consume at least 5 servings of fruits and vegetables, and drink 2 liter/day (64 oz) of fluid in the form of frequent small amounts of water throughout the day (24,25). The RYGB surgery patients were asked to avoid food and drinks with added sugar and fruit juice to prevent the dumping syndrome (24,25). To improve compliance to the dietary recommendations, the subjects were asked to keep a daily diet diary, which was used by the investigators to provide individual feedback every 2 weeks. Lastly, all the subjects were asked to take multivitamin and mineral supplements as recommended for bariatric surgery patients (24,25) with the RYGB surgery patients requiring more extensive supplement therapy because of their greater risk for nutritional deficiencies.

Behavioral therapy

Behavioral therapy related to exercise was provided to the HVEP group on an individual basis, by one of the authors (M.S.), and included goal setting, self-monitoring, cognitive-behavioral strategies, and problem solving and relapse prevention. The behavioral therapy was provided every 2 weeks either in conjunction with the supervised exercise sessions or by telephone. Behavioral therapy related to diet was provided to subjects in both the groups on an individual basis and by the same author, and included all of the above strategies plus stimulus control, eating behavior, and stress management.

Measurements

All of the following measurements were collected at baseline, 6 and 12 weeks except VO_2max, dual-energy X-ray absorptiometry, and oral glucose tolerance test (OGTT), which were measured only at baseline and 12 weeks and demographic characteristics and health history, which were collected at baseline. Our primary outcome variable was VO_2max.

Physical fitness, physical activity, REE, and blood pressure

Physical fitness was assessed by measuring VO_2max (ml/kg/min) using a graded maximal exercise test on a treadmill and was supervised by a physician who monitored a 12-lead ECG. To establish the treadmill speed for the test, subjects performed a short walking warm-up at 0% grade and the speed was then increased until a steady-state heart rate of 60% of age-predicted maximum or a rating of perceived exertion of 11–13 (fairly light to somewhat hard physical activity) on the Borg scale (26) was maintained for 4 min. After a short recovery and stretching period, the test commenced at the previously determined speed and the grade was elevated 2% in 2-min stages until exhaustion. Blood pressure, heart rate, and rating of perceived exertion were measured during the last 30 s of each stage. The test continued until the participant reached volitional fatigue or met the other standard stopping criteria (27).

Physical activity levels were assessed by instructing the subjects to wear a pedometer (DIGI WALKER SW-200; New Lifestyles, Lees Summit, MO) for 7 days each at baseline and 6 and 12 weeks. It was also assessed by interviewer administered 7-day physical activity recall, a validated tool (28) that collects information on time spent in sleep, moderate, hard, and very hard activities. Data from the 7-day physical activity recall were used to calculate the energy expended (kcal/kg/day) in sleep, light, moderate, hard, and very hard activities by multiplying the reported average number of hours per day spent at each activity level by the metabolic equivalent (ratio of work metabolic rate and resting metabolic rate) for that activity (1, 1.5, 4, 6, and 10, respectively).

Resting energy expenditure (REE) was assessed by indirect calorimetry (Deltatrac II; Sensormedics, Yorba Linda, CA) during the fasting condition using a protocol similar to that described elsewhere (29). Resting blood pressure was assessed three times each at baseline and 6 and 12 weeks and prevalence of hypertension was diagnosed using the guidelines by the Joint National Committee 7 report on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (30).

Dietary intake

Dietary intake was assessed by 3 day (2 week days and one weekend day) food record, a validated technique (31). The food records were analyzed for nutrient intakes using the University of Minnesota Nutrition Data System for Research, version 5.0–3.5.

Anthropometry and dual-energy X-ray absorptiometry

Weight was measured in light clothing to the nearest 0.1 kg, and height without shoes and waist and hip circumferences were measured to the nearest 0.1 cm using standard procedures (32). Body composition was assessed using dual-energy X-ray absorptiometry (Hologic QDR4500, Bedford, MA). The dual-energy X-ray absorptiometry measurement was limited to subjects weighing <136 kg (18 HVEP subjects and 12 controls), the maximum weight supported by the table top.

Lipids, lipoproteins, glucose, and insulin concentrations

A blood sample was drawn after a 12-h overnight fast on 3 days each at baseline, 6, and 12 weeks for assessment of lipids, lipoproteins, glucose, and insulin. Serum total cholesterol, triglycerides, and high-density lipoprotein cholesterol were analyzed using enzymatic methods (33) and low-density lipoprotein cholesterol was calculated. Plasma insulin was measured using radioimmunoassay kits (Millipore, Billerica, MA).

A standard OGTT with 75 g of oral dextrose (Glucose Drink; Azer Scientific, Morgantown, PA) was conducted after a 12-h overnight fast in nondiabetic GB surgery patients. The OGTT was limited to GB subjects (n = 23) and not conducted in RYGB surgery patients because of their propensity to develop the dumping syndrome following an oral glucose load. An intravenous catheter was placed in the forearm vein and blood was collected for determination of insulin and glucose concentrations at −30, −15, and 0 min before glucose ingestion, and at 30-min intervals thereafter for 120 min.

Prevalence of hypercholesterolemia, low high-density lipoprotein cholesterol, high non-high-density lipoprotein cholesterol, and hypertriglyceridemia was diagnosed using the National Cholesterol Education Program, Adult Treatment Panel III (NCEP ATP III) guidelines (34), and type 2 diabetes using the American Diabetes Association and World Health Organization Expert Committees guidelines (35).

Health-related QOL and current health and medication use

Health-related QOL was assessed using validated questionnaires: SF-36 (the Medical Outcomes Study 36-item Short-Form Health Survey questionnaire version 2.0; refs. 36,37) and the IWQOL-L (Impact of Weight on QOL-Lite; ref. 38). Current health and medication use was collected by questionnaire.

Statistical analysis

The effect of treatment on the continuous outcome variables was assessed by mixed-effects models for repeated measures with a between group factor (HVEP and control group) and a repeated factor (evaluation week). The difference in response between the two groups was assessed by using the group-by-week interaction factor and changes within groups were assessed by tests of the effect slices in the mixed-effects models for repeated measures analysis. To compare plasma glucose and insulin concentrations during the OGTT test, the incremental area under the curve for these variables was computed for each subject using the trapezoidal rule and the 12-week difference from baseline was compared between groups by the nonparametric Wilcoxon Rank Sum test.

The treatment outcomes reported in this paper are intention-to-treat analyses and excluded one subject in the HVEP group and four subjects in the control group who did not provide any follow-up data after baseline. Only nonadjusted results are reported in the paper because adjusting for age, baseline BMI, type of surgery, and banding adjustments did not change the results. Variables that did not satisfy the analysis assumptions (triglycerides, insulin, and glucose concentrations) were log-transformed before analyses. Group differences at baseline were assessed by two-sample t-tests for continuous variables and Fisher's exact tests for categorical variables. All data are reported as means and standard deviations unless otherwise noted. All analyses were carried out using the SAS statistical software, version 9.2 (SAS Institute, Cary, NC).

Results

Baseline participant characteristics

Baseline characteristics are shown in Table 1. Mean BMI was >40 kg/m² in both groups and mean age was 53.9 years in the control group and 47.3 years in the HVEP group. More than 90% of the subjects were female. Most of the subjects were non-Hispanic whites or African Americans and had undergone the GB surgery. Less than 20% of the subjects were current smokers. The average duration since surgery ranged from 3 months to 3.5 years in the control group and from 3 months to 8.5 years in the HVEP group. None of these variables were different between the two groups. Presence of comorbidities was also not different except for the prevalence of diabetes, which was significantly higher in the control group compared to the HVEP group.

Table 1.  Baseline characteristics of the subjects

Subject attrition and problems during exercise

Four out of 12 subjects in the control group did not provide any data after baseline. Three of these four subjects dropped out because they would have preferred to be in the exercise group and the fourth subject did not have time for the study. Five out of 21 subjects in the HVEP group also did not complete the study. Four of these five subjects provided data at both baseline and 6 weeks whereas one provided data at only baseline. The reason for the drop out in the HVEP group was because they did not have enough time to exercise. Excluding the four subjects in the control group and one subject in the HVEP group who did not provide any data after baseline from our data set (intention-to-treat analysis data set) did not affect the baseline BMI and prevalence of type of surgery but age was significantly higher in the control group compared to the HVEP group (P = 0.02). The subjects in the HVEP group did not encounter any major problems during exercise other than the occasional muscle or joint soreness.

Exercise energy expenditure, moderate physical activity METs, step count, VO_2max, and REE

At 4 weeks, 30% of the subjects in the HVEP expended ≥2,000 kcal/week and 55% expended ≥1,500 kcal/week. During the second 4-week period, 40% of the subjects in the HVEP expended ≥2,000 kcal/week and 65% expended ≥1,500 kcal/week and the respective percentages for the last 4 weeks were 53 and 82% (based on the mean of each 4-week period).

The number of steps/day increased from about 5,500 steps at baseline to nearly 10,000 steps/day at 12 weeks in the HVEP group and increased only slightly in the control group (Figure 1a). The group-by-week interaction (P = 0.03) and within-group change in the HVEP group (P < 0.0001) were statistically significant.

Figure 1. Changes in step count and VO_2max during the study. The individual values and means (horizontal lines) are shown for (a) step count and (b) VO_2max in the control and HVEP groups. Signficant group-by-week interaction and within-group change in the HVEP group was seen for step count (P = 0.03 and P < 0.0001, respectively) and VO_2max (P = 0.009 and P = 0.001, respectively) as evaluated by repeated measures analysis. HVEP, high-volume exercise program; VO_2max, maximal oxygen consumption.

Based on the 7-day physical activity recall, the reported time spent and energy expended during moderate physical activity increased by more than three times over 12 weeks in the HVEP group but remained the same in the control group compared to respective baseline values (Table 2). There was a significant group-by-week interaction (P = 0.02) and within-group change in the HVEP group (P < 0.0001) but not in the control group (P = 0.99). There was no group-by-week interaction for the reported time spent and energy expended during light physical activity. The median values for time spent in vigorous (hard and very hard) physical activity and the energy expenditure related to these activities were zero for both groups and at all time points.

Table 2.  Reported energy expended and time spent during moderate and light physical activity from 7-day physical activity recall and resting energy expenditure during the study

VO_2max relative to body weight, the primary outcome variable, increased by 10% in the HVEP group (baseline: 17.4 ± 3.3 (mean ± SD) ml/kg/min; 12 weeks: 19.2 ± 4.2 ml/kg/min) and did not change much in the control group (baseline: 17.6 ± 1.4 ml/kg/min; 12 weeks: 17.1 ± 1.7 ml/kg/min) over 12 weeks compared to the corresponding baseline values (Figure 1b). There was a significant group-by-week interaction (P = 0.009) and within-group change in the HVEP group (P = 0.001) but not in the control group (P = 0.41).

REE decreased by 2.6% from baseline to 12 weeks in the HVEP group and by 5.5% over the same time period in the control group (Table 2). Neither the group-by-week interaction (P = 0.21) nor the within-group changes (P = 0.50 and P = 0.10 in HVEP and control groups, respectively) were significant, however.

Energy and macronutrient intake

Reported energy intake decreased by 358 kcal/day in the HVEP group and by 593 kcal/day in the control group over 12 weeks (Table 3). Protein intake was maintained at ≥60 g in the HVEP group over 12 weeks and tended to decrease by 12–16 g/day to about 50 g/day in the control group (Table 3). Percent energy from fat declined slightly in both the groups (Table 2). The group-by-week interaction or the within-group changes were not significant for any of the variables except energy intake that reduced significantly within the HVEP group (P = 0.02) but not in the control group (P = 0.07).

Table 3.  Energy and macronutrient intake during the study

Body weight, waist and hip circumferences, and body composition

Body weight, and waist and hip circumferences decreased similarly in the two groups over 12 weeks (Table 4). Percent total body fat, percent trunk fat, and LBM also decreased similarly and slightly in both groups (Table 4). The group-by-week interaction was not significant for any of these variables but the within-group changes were significant for body weight and hip circumference in both the HVEP (P = 0.009 and P = 0.02, respectively) and the control (P = 0.006 and P = 0.002, respectively) group.

Table 4.  Body weight, waist and hip circumference, and body composition during the study

Plasma glucose, insulin, and lipid concentrations and blood pressure

There was no significant group-by-week interaction or within-group changes for fasting glucose and insulin concentrations or 2-h postprandial glucose and insulin concentrations, although the 2-h postprandial blood glucose response decreased by 11% in the HVEP and remained the same in the control group (Table 5). There was a significant group-by-week interaction for the incremental area under the curve postprandial glucose response (P = 0.03) (Figure 2a) but not insulin response (Figure 2b).

Table 5.  Plasma glucose, insulin, and lipid concentrations, and blood pressure during the study

Figure 2. Changes in incremental area under the curve (iAUC) for postprandial blood glucose and insulin levels during the study. Individual values and geometrical means (horizontal lines) are shown for postprandial blood (a) glucose iAUC and (b) insulin iAUC in the control and HVEP groups. A significant group-by-week interaction (P = 0.03) was seen for glucose iAUC. HVEP, high-volume exercise program; iAUC, incremental area under the curve.

There was a tendency for fasting lipid concentrations and blood pressure to decline slightly over 12 weeks in both the groups except high-density lipoprotein cholesterol, which tended to increase by 3 mg/dl in HVEP group and SBP and DBP, which tended to increase by 2–3 mm Hg in the control group compared to the corresponding baseline values (Table 5). The group-by-week interaction or within-group change was not statistically significant for any of these variables.

Health-related QOL

Data on health-related QOL are presented in Table 6. The HVEP group reported significant improvement in 4 of the 5 IWQOL-L scales including physical function (P = 0.049), self-esteem (P = 0.0002), sexual life (P = 0.02), public distress (P = 0.003), and the total score (P = 0.0004) over 12 weeks whereas the control group reported improvement in self-esteem (P = 0.004), sexual life (P = 0.04), and work (or daily activities) (P = 0.04) and the total score (P = 0.012). There was no group-by-week interaction for any of the scales except the interaction for self-esteem, which was close to significance (P = 0.05). According to the data from the SF-36 questionnaire, there was no group-by-week interaction for any of the scales but the HVEP group reported a significant improvement in emotional well being (P = 0.001), energy levels (P = 0.0002), and mental QOL total score (P = 0.006) over 12 weeks whereas QOL in the control group did not change.

Table 6.  Quality of life during the study

Discussion

Our study is the first study to examine the feasibility and efficacy of an HVEP in mostly severely obese bariatric surgery patients. Our results show that during the last 4 weeks of the 12-week study, about 50% of the subjects were performing a large volume (≥2,000 kcal/week) of moderate-intensity exercise and >80% were expending at least 1,500 kcal/week. The increased amount of time spent exercising was not compensated for by reducing physical activity levels at other times of the day because the step count increased substantially from about 4,500 steps/day to nearly 10,000 steps/day. The increase in moderate-intensity exercise was associated with a significant 10% increase in VO_2max relative to body weight in the HVEP group and this increase was not related to the number of months since bariatric surgery was performed. A recent small study by Stegen et al. (39) in which patients who underwent RYGB surgery were allowed to choose whether or not to undergo a 12-week endurance and resistance exercise training program a month after bariatric surgery, reported no group-by-time interaction effect, unlike our study, but a within-group increase (25–26%) in peak VO_2max relative to body weight in both the exercise and control groups. It is not clear why the control group in their study showed the same improvement in fitness levels as the exercise group despite adjusting VO_2max for weight loss. In a pre-post study by Sartorio et al. (40), VO_2max relative to body weight increased by 20% in severely obese non-bariatric-subjects after 3 weeks of endurance training. The smaller increase in VO_2max in our study may be due to the fact that not all the HVEP subjects in our study met the exercise goal unlike the subjects in the other studies who were completely supervised during the exercise training and may have been more motivated because they chose to exercise. Nonetheless even a moderate increase in VO_2max is important because physical fitness is inversely correlated with BMI (41) and associated with a reduced risk of all cause mortality (42).

Weight loss, a secondary outcome variable, was similar (∼4.5 kg) in both groups. According to a meta-analysis including studies that compared diet and exercise intervention with diet only intervention in non-bariatric-surgery subjects, weight loss was only slightly higher in the former group compared to the latter group (43). Stegen et al. (39) also reported no difference in weight loss in the RYGB surgery patients who chose to exercise compared to the patients who chose not to exercise. Despite similar weight changes, the control group in our study reported reducing their energy intake by 1.7 times as much as the HVEP group (593 kcal/day vs. 358 kcal/day, respectively) suggesting that the latter group may have partly compensated for the energy deficit caused by exercise by decreasing their energy intake to a smaller extent than the former group. Nevertheless, bariatric surgery patients who engage in an exercise training program may be able to achieve similar weight loss without markedly reducing their energy intake compared to patients who are just dieting.

REE did not change significantly in either group over 12 weeks although the decrease tended to be slightly less in the HVEP group than in the control group. These results are probably not explained by changes in percent body fat loss and LBM which decreased slightly though not significantly in both groups. Tremblay et al. (44) have also reported that REE was not modified significantly in overweight males who underwent 100 days of endurance training despite weight loss. These results suggest that a strategy such as exercise that helps to maintain REE may be useful in preventing weight regain. To help preserve LBM and thus REE, >60 g of protein intake per day is recommended in bariatric surgery patients (24,25). Reported protein intake was maintained at or above 60 g/day in the HVEP group but tended to decrease to 55 g/day in the control group. Further emphasis on increasing protein intake to help to preserve LBM may be necessary in bariatric surgery patients.

The HVEP group also showed a reduction in the incremental area under the curve postprandial blood glucose response. These results are corroborated by other studies (45,46), which have reported improved postprandial blood glucose response following 12 (ref. 45) or 20 (ref. 46) weeks of endurance training in non-bariatric-surgery obese (45) and healthy sedentary (46) individuals.

The HVEP group tended to report greater improvement in health-related QOL especially physical function, self-esteem, sexual life, public distress, energy levels, and emotional and mental well being than the control group but there was no group-by-week interaction possibly because of the limited sample size. According to some case series, health-related QOL deteriorates in the long-term following bariatric surgery possibly due to weight regain (47,48) whereas becoming or continuing to be highly active after bariatric surgery is associated with improved mental health-related QOL (23). A 6-month randomized, controlled trial in non-bariatric-surgery overweight or obese subjects found an improvement in all mental and physical aspects of QOL except bodily pain following exercise training and this relationship was exercise dose dependent and independent of weight change (49).

A limitation of this study is that the exercise training was implemented for only 12 weeks and in a limited number of subjects. Some of the subjects who did not meet the exercise goal felt that they needed more than 12 weeks to reach the 2,000 kcal/week goal. A larger study with a longer duration may have provided a more comprehensive assessment of exercise feasibility. The drop-out rate was higher in the control group than in the HVEP group because most of the subjects in the control group would have preferred to be in the HVEP group. Future studies should include low volume of exercise or flexibility exercises such as yoga in the control group to improve retention rate. Another limitation is that we did not perform a meal tolerance test in the RYGB surgery patients given that they may experience the dumping syndrome in response to the OGTT test. The dietary and exercise counseling was provided at an individual level and not at the group level. The latter format would have provided group support. However, both groups received frequent counseling from the investigators regarding their exercise and/or dietary intervention. Lastly, we used an unsealed pedometer to assess physical activity and the subjects logged their step count every day during the measurement period. We do not expect this to have affected the comparison of change in physical activity between the HVEP and control group, however, because any reactivity to an unsealed pedometer in the HVEP group would have also occurred in the control group. In addition, whether an unsealed pedometer results in a different step count from a sealed pedometer is controversial (50,51).

A major strength of this study is that it is the first randomized, controlled exercise trial in bariatric surgery patients. It is also the first study to objectively assess exercise capacity and training effect in this population by measuring their VO_2max. In addition, both groups were also asked to improve their overall diet including reducing their energy intake. This is important as bariatric surgery patients increase their energy intake over time (11) following surgery possibly contributing to some of the weight regain. Lastly, each subject received behavioral therapy as discussed earlier.

In conclusion, a high-volume moderate-intensity exercise program is feasible in about 50% of severely obese bariatric surgery patients and improves physical fitness. It also improves postprandial blood glucose response. Whether a HVEP helps to maintain weight loss and improvement in comorbidities in these patients remains to be evaluated in long-term studies, however. The studies also need to assess how exercise over the long term affects factors that influence energy balance including energy intake, nonexercise activity levels, body composition, metabolic rate, and gastrointestinal hormones related to satiety and hunger.

ACKNOWLEDGEMENT

We would like to acknowledge Drs Savitha Shastry, Zahid Ahmad, and Vinaya Simha for providing clinical care to the patients during the study, Sarah Masood for assisting with data collection, Sheena Shah-Simpson for analyzing the food records, Rosemary Son and Drs David Provost, Nancy Puzziferri, and Nirmal S. Jayaseelan for helping to recruit the subjects, and Dr Joel Mitchell for consultation. The study was partly funded by NIH grants M01-RR00633 and UL1-RR-024982 and by the Southwest Medical Foundation.

DISCLOSURE

The authors declared no conflict of interest.