Trial Information

The ACT-OUT Trial: ACTivity OUTcomes Based on the Consumption of a High Carbohydrate or High Fat Diet in Patients With Metabolic Syndrome

1. BACKGROUND AND RATIONALE:

Current obesity prevention emphasizes increasing physical activity and a low-fat,
calorie restricted diet to produce a negative caloric balance. Caloric restriction is
associated with decreases in energy expenditure. The other mainstay of obesity
prevention increasing levels of physical activity is associated with greater caloric
intake. Failure of the eat less-exercise more strategy is evident with epidemic of
obesity, metabolic syndrome (MetS) and type II diabetes. Moreover, the displacement of
dietary fat in the diet has led to compensatory increases in carbohydrate intake that
is associated with untoward metabolic effects.

Carbohydrate restricted diets are controversial despite benefits of this approach. The
controversy stems from a compensatory increase in dietary fat intake associated with
carbohydrate restriction and fears that greater fat intake increases risk of
cardiovascular disease. Carbohydrate restriction decreases insulin resistance and
ameliorates MetS. This protocol aims to establish a greater evidence base for
carbohydrate restricted diets as a therapeutic option for MetS. The investigators will
review current knowledge with respect to: (i) obesity and insulin resistance in Canada,
(ii) positions of major medical organizations (iii) shortcomings of the current obesity
prevention paradigm (iv) why carbohydrate restricted diets work.

1.1) Obesity and insulin resistance in Canada In the 1980s low fat, caloric restricted
diets became the cornerstone of dietary advice to reduce obesity. This approach
successfully reduced the percent of calories from fat with a consequential increase in
carbohydrate consumption. In the ensuing years, obesity become epidemic. In Canada,
the prevalence of obesity (BMI ≥ 30 kg/m2) increased from 10% in 1970s to 23% in 2004
with an additional 36% considered overweight (BMI 25-29.9). As obesity is a risk factor
for MetS and type II diabetes, the prevalence of these disorders has also reached
epidemic proportions. Approximately 2 million Canadians were living with diabetes in
2007 and by 2012 it is estimated that 2.8 million Canadians will be living with
diabetes. First Nations communities have been especially hard hit, with a prevalence
3-4 times higher and a younger age of onset than non-First Nations individuals. Novel
chronic disease prevention strategies are needed to address this crisis.

Insulin resistance underpins the pathophysiology of obesity, Insulin resistance is a
condition in which normal amounts of insulin are inadequate to produce a normal insulin
response from fat, muscle and liver cells. As a result, insulin resistant individuals
have both elevated blood glucose, fasting insulin and secrete greater amounts of
insulin in response to dietary carbohydrate. This compensatory hyperinsulinemia is
associated with adverse effects in tissues which retain their sensitivity to insulin.
Elevated blood glucose levels are associated with advanced glycation end products,
reactive oxygen species and low grade inflammation. These exposures are associated with
increased risk of cardiovascular disease & diabetes. Despite the evidence associating
elevated insulin levels to atherogenic changes in the vasculature, it remains a matter
of controversy as to whether persistently elevated insulin levels are pathogenic.

The diagnosis of MetS represents an entry-point for disease prevention efforts in
individuals with diabetes and cardiovascular disease. MetS is defined as a high waist
circumference in addition to two of the following: high triglycerides, low HDL and
elevated blood pressure. Other pathophysiologic changes are also observed including
low grade inflammation, increased uric acid, prothrombotic state, elevated ApoB, small
dense atherogenic LDL and endothelial dysfunction.

First line treatment for MetS is weight loss leading to reductions in insulin
resistance. Intensive lifestyle interventions reduce diabetes in high risk subjects.
Most lifestyle interventions for MetS are predicated on the hypothesis that individuals
can initiate and maintain weight loss by restricting calories and increasing physical
activity. The Diabetes Prevention Program showed that intensive intervention including
a low-fat calorie reduced diet and increase in physical activity was associated 1.5
kg/m2 decrease in BMI and a 58% reduction in incidence of diabetes in individuals at
high risk of diabetes. Less intensive interventions are associated with more modest
changes. A study of nutritional counselling was associated with a 0.4 kg/m2 decrease in
BMI. Thus, lifestyle interventions which emphasize 'eating less or exercising more have
limited value. The optimal diet for subjects with MetS has not been examined with hard
end-points. Small trials have compared carbohydrate restricted diets and low fat diets
using surrogate markers. A trial of severely obese subjects found greater weight loss
and greater improvements in lipid levels and hemoglobin A1c on carbohydrate restricted
diet than low fat/calorie restricted diets. In a study of adults with MetS, those
randomized to the carbohydrate restriction had greater weight loss, lower triglycerides
and lower insulin levels than individuals on a calorie restricted, low fat diet. These
results are consistent with a meta-analysis where carbohydrate restricted diets were
associated with improvements in atherogenic dyslipidemia, greater weight loss and
better compliance than low fat, calorie restricted diets. These studies highlight the
advantages of carbohydrate restriction.

1.2) Fat, sugar and positions of major medical organizations One of the principal
barriers to the widespread implementation of carbohydrate restricted diets is the fear
that increase in saturated fat consumption with its association with increase in
LDL-chol, will result in greater risk of cardiovascular disease (CVD). This diet-heart
hypothesis has formed the cornerstone of dietary advice for prevention of chronic
disease. The avoidance of saturated fat has led to general avoidance of dietary fat.
There is weak and contradictory evidence that both saturated fat and dietary fat
restriction are associated with health benefits. The largest test of the diet-heart
hypothesis was the Women's Health Initiative (WHI). This study randomized women to
either a low saturated fat diet or a regular diet. After 8 years, women randomized to
the low fat diet had lower LDL-chol but no decrease in CVD mortality, incidence of
breast or colorectal cancer. Consistent with the results of the WHI, a meta-analysis
which included 347,747 subjects found no relationship between saturated fat intake and
CVD.

While most experts would agree that it's too early to vindicate saturated fat as the
foremost dietary evil, evidence is accumulating highlighting the obesigenic and
atherogenic nature of refined carbohydrates. A trial aiming to decrease sweetened
beverage consumption found the intervention group had 7.7% less overweight and obese
children than controls. In NHANES (2006), consumption of added sugar was associated
with higher triglycerides and lower HDL. A prospective study of Dutch adults found
that replacing 5% of calories from saturated fat with high glycemic carbohydrates was
associated with a 33% increase in risk of myocardial infarction.

The current state of scientific confusion over the health effects of (refined)
carbohydrates is reflected in the diversity of dietary recommendations. American
Diabetes Association (ADA) dietary guidelines have suggested that carbohydrate
restricted diets are an option for weight loss for up to 1 year. In contrast, Canadian
Diabetes Association recommends a diet with 45%-60% of calories from carbohydrate with
a sucrose intake up to 10% of daily intake. The contradictory nature of these positions
highlights the need for more research into the health effects of carbohydrate
restricted diets. The endorsement of carbohydrate restriction by ADA suggests that
these diets are valid therapeutic options.

1.3) Shortcomings of the traditional obesity prevention paradigm The current obesity
prevention paradigm emphasizes caloric reduction and/or increased physical activity to
produce negative caloric balance. While physical activity has established benefits,
increasing physical activity as a weight loss strategy has met with disappointing
results. Observational studies associating high levels of physical activity with lower
adiposity are difficult to interpret as individuals with higher physical activity also
benefit from other (unmeasured) health promoting behaviours. In randomized trials,
physical activity interventions have shown equivocal results with respect to weight
loss. The physiologic limitation of this approach is that it is difficult to augment
physical activity without a compensatory increase in dietary intake. In studies where
physical activity is directly supervised and in the absence of a compensatory increase
in caloric intake, physical activity has been associated with weight loss < 3% of
initial body weight. In physical activity interventions of lower intensity, results are
less promising. A meta-analysis of school based physical activity interventions in
youth found no effect on BMI. The ineffectiveness of physical activity as a means of
weight loss is reflected in the American Heart Association and American College of
Sports of Sports Medicine position statements where they recommend 60-90 minutes of
moderate intensity activity/day to lose or maintain weight loss. Not surprisingly,
there is a very low prevalence of individuals who are moderately active for 60-90
minutes/day in the general population. Ironically, insulin resistant individuals who
would reap the greatest health benefits from physical activity may have a reduced
physiologic tolerance for it as insulin resistance has been associated with reduced
exercise tolerance through reversible changes in muscle mitochondria, reduced VO2 max
and possibly other mechanisms.

The failure of low fat, calorie restricted diets to curb the prevalence of obesity is
exemplified by the WHI. After 7.5 years of follow-up, the women randomized to the low
fat group decreased their weight by 0.8 kg but had a 1 cm increase in waist
circumference despite a reported decrease in caloric intake of 360 calories. Weight
loss in the low fat group was not different than women in the control group, who were
consuming a typical American diet. Calorie restricted diets emphasize the avoidance of
energy dense foods high in dietary fat. While avoidance of dietary fat has intuitive
appeal, calorie restriction is associated with a corresponding decrease in energy
expenditure, in effect, stimulating a starvation response. The compensatory decrease in
energy expenditure is associated with a plateauing of weight loss, an increased desire
to eat and weight gain.

Low fat/high carbohydrate diets require weight loss to produce positive metabolic
effects. This dependence on weight loss for health benefits combined with the
inevitable rebound weight gain associated with this dietary strategy has the potential
to worsen metabolic markers over the long term. The untoward metabolic changes
associated with a high carbohydrate intake are particularly deleterious in individuals
with MetS. This contrasts with the physiologic effects of carbohydrate withdrawal in
insulin resistant subjects that mirrors reversal of MetS. In addition to improvements
of MetS, carbohydrate restricted diets are associated with less dense and therefore
less atherogenic LDL particles and decreases in inflammatory markers.

A final limitation of the current obesity prevention paradigm is that macronutrients
(fat, protein and carbohydrates) are assumed to have equivalent metabolic effects. The
investigators will review metabolic effects of macronutrients relating to the insulin
response and energy balance. Carbohydrates are potent secretagogues of insulin that is
the primary regulator of adipose tissue metabolism and partitions dietary energy into
either storage (higher insulin) or oxidation (lower insulin). Mice genetically
engineered to lack the insulin receptor on fat cells consume more calories per gram of
body weight but are resistant to obesity. In type II diabetics, administrations of
exogenous insulin or drugs which stimulate insulin release from the pancreas are
associated with weight gain. Dietary fat has no effect on insulin secretion but delays
gastric emptying. Dietary protein lowers the glucose response to insulin 2-3 times more
effectively than dietary fat and plays an important role in appetite suppression and
potentially weight regulation. The addition of dietary fat, protein and fiber to dampen
the untoward effects of dietary carbohydrate in a meal is the concept behind the
glycemic index. Low glycemic index diet may be of less relevance to type II diabetics,
however, as the effect of dietary fat and protein in blunting glucose responses is
attenuated or absent. If obesity prevention is indeed evidence base then there are only
two tools left in the public health arsenal, bariatric surgery or carbohydrate
restriction.

1.4) Why do carbohydrate restricted diets work? There are at least two physiologic
characteristics of carbohydrate restricted diets which contribute to weight loss.
First, ad lib carbohydrate restricted diets have been associated with spontaneous
reductions in caloric intake. Second, these diets are associated with greater weight
loss per calorie than low-fat diets with an unexplained caloric deficit of about 200
kcal/day. The preferential weight loss of carbohydrate restricted diets was termed 'the
metabolic advantage'. The concept of a metabolic advantage associated with carbohydrate
restricted diets has been controversial among experts. While there is evidence
supporting its existence in animal models and in humans, this phenomena remains poorly
understood.

Ad lib carbohydrate restricted diets are often associated with spontaneous decreases in
caloric intake. This is caused by the insulin lowering effects of carbohydrate
restricted diets as high insulin levels have been associated with partitioning of
metabolic fuels into adipose tissue and greater food intake. A trial of obese subjects
found that those consuming a carbohydrate restricted diet had nearly a 30% decrease in
fasting insulin levels and a 570 kcal decrease in energy intake. Thus lower insulin
levels are a key mediator between lower carbohydrate consumption and weight loss. A
study in men found that 70% of the variability in weight loss from a carbohydrate
restricted diet was explained by changes in fasting insulin.

A number of trials have shown a metabolic advantage of carbohydrate restricted diets
compared to low fat diets. There are three possible explanations. First, there might be
greater energetic costs associated with metabolic interconversions of nutrients on
carbohydrate restricted diets. Second, individuals on a lower carbohydrate diet may
have higher energy expenditure due to a greater resting metabolic rate and/or greater
levels of physical activity. A final possibility is that lower weight loss is due to
the selective under-reporting of dietary intake in individuals randomized to low fat
diets. This latter explanation is not likely as this finding has been observed in
numerous randomized controlled trials.

Diet induced thermogenesis may have a small contribution to the metabolic advantage of
carbohydrate restricted diets. It accounts for 5-15% of total energy expenditure. The
energetic costs to metabolize macronutrients vary from 2.5% with dietary fat, 7-15%
with carbohydrate and 28-35% with protein. One study found that energy expenditure was
4% higher in a diet containing 30% compared to 10% of calories from protein. Thus the
modest increase in energy expenditure due to a higher protein intake would not likely
be enough to account for the metabolic advantage of carbohydrate restricted diets.

Physical activity is the greatest modifiable source of energy expenditure. The
hypothesis that the metabolic advantage of carbohydrate restricted diets could explain
lower insulin levels associated with spontaneous increases in physical activity has
received scant attention. This hypothesis is based on a number of well established
observations. First, type II diabetics taking exogenous insulin have decreases in
energy expenditure and weight gain. When insulin levels are lowered through
carbohydrate restriction or pharmacotherapy there is an attendant loss of fat mass and
some evidence for a modest increase in energy expenditure. A short study of obese
individuals found that a ketogenic, low carbohydrate diet with a 30% decrease in
insulin resistance was associated with a 20% increase in resting energy-expenditure,
though physical activity was not assessed in this study. Similarly, in a 6 month
randomized trial of obese women, those on the carbohydrate restricted diet had a
non-significant 5% increase in resting energy expenditure per kg of body weight
compared to baseline. This study also assessed physical activity using pedometers and
found no difference in steps/day between the carbohydrate restricted group and the
low-fat group despite a 3.7kg greater weight loss in the former and no difference in
reported caloric intake between groups. Measurement of physical activity in free-living
individuals is complex. Pedometers are an objective assessment of physical activity but
are unable to assess intensity, frequency and duration of activity. They are also
unsuitable for estimation of energy expenditure.

In summary, intensive lifestyle interventions reduce the incidence of type II diabetes
in individuals with metabolic syndrome. Central to the therapeutic effect of these
interventions is weight loss. There is a lack of consensus over what constitutes 'best
practices' for lifestyle interventions. Physical activity is associated with numerous
health benefits but is not an efficacious prescription for weight loss or weight
maintenance. Moreover, the optimal dietary prescription is controversial; some experts
suggest individuals with MetS should restrict their dietary fat intake while other
experts suggest dietary carbohydrate should be restricted. This controversy in dietary
prescription is reflected in the positions of major medical organizations. The American
Diabetes Association has endorsed carbohydrate restriction for weight loss since 2008.
The Canadian Diabetes Association, however, favours lower glycemic index foods and the
restriction of dietary fat. Despite the equivocal evidence, a low fat, calorie
restricted diet coupled with increased physical activity are considered the standard
lifestyle interventions for MetS.

2. OBJECTIVES This protocol has three aims. First, the investigators hypothesize that
adherence to a lower carbohydrate diet in individuals with MetS, resulting in lower
insulin resistance; will cause an increase in levels of spontaneous physical activity,
independent of changes in weight. Second, the investigators will examine changes in
cardiometabolic risk factors (ApoB, TG, HDL, blood pressure, CRP) when individuals are
randomized to either a carbohydrate restricted diet or a low fat/high carbohydrate
diet. Finally, the investigators will interview a sub-sample of participants from both
study arms (5 in each arm) to conduct open-ended interviews to better understand
quality of life issues with respect to the dietary assignment or lifestyle
intervention.

3. RESEARCH PLAN Our group has generated cross-sectional data showing a clinically
meaningful association between higher carbohydrate intake and lower physical activity,
assessed by accelerometer in glucose intolerant individuals. The limitation of these
findings relate to the inability to assess causality; that is, if carbohydrate
restriction causes increases in physical activity. The strength of the ACTout study is
its ability to delineate causal direction using a prospective design and randomization
to control for known and unknown confounders.

Study design and participants: Study participants will be randomized to an ad lib
carbohydrate restricted group or a low fat, calorie reduced diet for 6 months. The
investigators will recruit individuals with metabolic syndrome (MetS) from St. Paul's
Metabolic Syndrome Program. Inclusion criteria include a diagnosis of MetS defined as an
increased waist circumference and two or more of the following: fasting blood sugar greater
than 5.6 mmol/L, fasting triglycerides greater than 1.7 mmol/L, high density lipoprotein
(HDL) less than 1.0 mmol/L in men and less than 1.3 mmol/L in women and blood pressure
greater than 135/85 or on antihypertensive medication. Potential participants will be
excluded if they are following a weight reducing diet, are abusing alcohol or other
psychoactive substances, are on psychiatric medication associated with weight gain or have
plans to travel during the study period.

Sample size calculation: To calculate sample size the investigators used cross-sectional
data with accelerometer measured physical activity and excellent quantitative measures of
diet. The investigators assumed a 15% difference in calories consumed as carbohydrates
between the dietary treatments. With a significance value of p < 0.05 (two sided test) the
investigators found that 27 participants would be required in each group to detect a
significant difference in accelerometer measured intensity of activity. To account for
possible attrition the investigators will overestimate the sample size by 10 in each arm.
The investigators will recruit 72 individuals with 36 in each arm. Based on our sample size
calculation, the investigators expect to detect time and diet based differences in addition
to a diet*time interaction.

Recruitment strategy: The investigators will recruit individuals with metabolic syndrome
from the Metabolic Syndrome Program. Flyers will be posted in the waiting room and the
examination rooms of the Program. Flyers will also be given to the administrative assistants
responsible for intake of participants into the Program who may choose to notify
participants of the study. If a potential participant expresses interest the coordinator
will explain the study in sufficient detail to enable informed consent. To assess diet prior
to inclusion into the study the investigators will ask potential participants to complete a
3 day diet history and wear the accelerometer for 7 days to ensure that they are not already
following a weight reducing diet and to assess baseline physical activity. Individuals
compliant with both the dietary recall and baseline assessment of physical activity will be
randomized into one of the two dietary assignments.

Randomization: The investigators will use block randomization to prevent the biasing of the
randomization process. After each block of subjects is enrolled, the investigators will use
a computerized randomization program to randomly assign participants to one of the two
dietary assignments. Due to the difficulty in concealing the dietary assignment, the
randomized trial will not be blinded; as both researchers and participants will be aware of
their dietary assignment. To minimize the contamination of study groups, individual and
group sessions will be scheduled at different times.

Carbohydrate restricted diet: Participants will be instructed to restrict carbohydrate
consumption to < 20 grams/day while not restricting their caloric intake. Participants will
be encouraged to consume vegetables with low carbohydrate content every day including 2 cups
of salad greens and 1 cup of vegetables 'that grow above the ground'. One of the physiologic
changes associated with a low carbohydrate intake is a loss of salt through the urine. If
salt is not replaced participants may experience headaches, nausea, dizziness, lethargy and
constipation. Participants will be counselled to increase their salt intake. This has been
shown to correct the natiuresis associated with carbohydrate restriction which can cause the
aforementioned side effects. Participants will be advised to continue their baseline level
of physical activity.

Control group: The control group will be randomized to the intensive lifestyle intervention
currently used by the Healthy Heart Program at St. Paul's Hospital to reduce symptoms
associated with metabolic syndrome. Participants will be instructed to replace high fat,
energy dense foods with foods rich in whole grains, fruits and vegetables. The macronutrient
distribution of this diet will be approximately 55% carbohydrate, 15% protein and 30% fat.
Individuals will also be instructed to reduce sodium intake to <2300 mg/day and in
individuals with hypertension, <1500 mg/day. At the end of the study, participants will have
the option to continue on either of the study diets.

Methods to ensure compliance to the dietary assignment: The study coordinator will
facilitate activities to support compliance to the dietary assignment, either one-on-one or
in a group setting. Group meetings will reinforce compliance by sharing cooking tips,
behaviour modification and relapse prevention strategies. In the one-on-one sessions, the
study coordinator will review 3 day food records and discuss strategies to increase
compliance. Participants from both groups will meet with their clinician after 3 months to
assess health indicators and to assure participants of the safety of their dietary
assignment.

Qualitative interviews: The investigators will conduct hour long open-ended interviews with
8-10 participants from the study to understand the lived experience of a prudent diet and a
carbohydrate restricted diet. Interviews will be recorded and transcribed. The
investigators will use grounded theory to look for theories which emerge from the data as
opposed to having interpreting the data with an a priori analysis. The investigators will
then assign descriptive categories to themes which emerge from the data, such themes could
include: effect of their diet on diabetes, perceptions of hunger, mood etc. These
descriptive categories will then be placed into thematic categories and analyzed in light of
the existing of the literature.

Measurement of body composition and estimation of resting energy expenditure:

The investigators will use bioelectrical impedance to assess body composition at baseline
and at 6 months. Bioelectrical impedance has been shown to provide a reasonable assessment
of body composition compared to DXA, which is considered a gold standard. The investigators
will normalize fat mass and fat-free mass to height by using a fat-mass index and a fat-free
mass index (kg/m2), respectively. Height will be measured with a stadiometer at baseline.
Waist circumference will be assessed at baseline and at 6 months, after exhalation, using a
flexible tape measure half way between the hip bone and the lowest rib. The investigators
will estimate energy expenditure using previously published equations which take into
account fat-free mass, fat mass, gender and age.

Questionnaire: The investigators will use the Applied Health Indicators Questionnaire and
the Patient Health Questionnaire, to collect relevant health information from the patient.
These self-administered questionnaires will be completed at baseline, 3 and 6 months.

Physical activity assessment: Participants will wear an accelerometer to assess physical
activity at baseline and week of every month for 6 months. An accelerometer is a small
electronic monitor worn on the waist that measures vertical accelerations and is thus
considered an objective measure of physical activity. Patients will be instructed to wear
the accelerometer during waking hours for 1 week of every month, exclusive of time spent
bathing or when in water. Data will be downloaded in one-minute epochs and categorized as
light, moderate or vigorous activity. Days will be excluded when the accelerometer is worn
for less than 80% of the average time worn on the other days. The investigators will
estimate activity-based energy expenditure using previously validated equations. As
accelerometers do not capture activity associated with cycling, swimming or skiing the
investigators will ask participants about the number of times they did these activities in
the period that they wore the accelerometer.

Cardiometabolic indices: Blood samples will be drawn and placed in individually labelled
tubes with EDTA, centrifuged immediately and stored at -70°C. As changes in insulin action
are central to our hypothesis the investigators will assess two distinct aspects of insulin
metabolism in this study, insulin secretion and insulin resistance. The investigators will
assess insulin resistance using HOMA assessment of insulin resistance calculated from
fasting insulin and fasting glucose. The investigators will assess the concentration of
C-peptide, a cleavage product related to insulin synthesis that is synthesized in equimolar
concentrations to insulin. As C-peptide has a longer half-life than insulin, it can be used
a marker of insulin secretion from the pancreas. Blood Pressure will be assessed at
baseline, 3 and 6 months using BP-TRU monitors. Lipid, lipoprotein and C-reactive protein
levels will be measured at baseline, 3 and 6 months. Total cholesterol, triglyceride,
high-density lipoprotein cholesterol and apolipoprotein B will be measured using previously
described methods. LDL-chol levels will be calculated using the Friedewald formula for
patients whose plasma triglyceride level is less than 4 mmol/L. β-hydroxybutyrate will also
be assessed in both groups at the aforementioned time points and will be used as a biomarker
of adherence to a carbohydrate restricted diet. The investigators will also assess serum
Leptin and uric acid at baseline and 6 months.

Statistical analyses: The investigators will use two different approaches to analyze the
data. First the investigators will use an intention to treat analysis using baseline values
to impute missing data in participants who may drop-out of the study. The investigators will
also examine the effect of the dietary assignment on outcomes in participants who were
compliant with their dietary assignment. Baseline characteristics will be compared between
the two groups using t tests. The investigators will examine three inter-related dependent
variables, capturing different aspects of physical activity, in the analysis strategy.
First, the investigators will examine the association between dietary assignment and changes
in accelerometer variables. The investigators will also examine time spent in
moderate-to-vigorous physical activity and activity based energy expenditure. As activity
based energy will be estimated from accelerometer , the investigators will not consider it a
unique comparison. The investigators will thus set the level of statistical significance at
p<0.025 to adjust for multiple comparisons. The investigators will obtain differences in key
variables such as physical activity and change in insulin concentration. To assess the
effects of the dietary assignment on physical activity, the investigators will use three-way
repeated-measures ANOVA, including change in body weight in the model with time as the
repeated factor.