Contemporary Clinical Trials 53 (2017) 151–161

Contents lists available at ScienceDirect

Contemporary Clinical Trials

journal homepage: www.elsevier.com/locate/conclintrial

DIETFITS study (diet intervention examining the factors interacting with

treatment success) – Study design and methods
Michael V. Stanton a,b, Jennifer L. Robinson a, Susan M. Kirkpatrick a, Sarah Farzinkhou a, Erin C. Avery a,
Joseph Rigdon d, Lisa C. Offringa a, John F. Trepanowski a, Michelle E. Hauser a, Jennifer C. Hartle a, Rise J. Cherin a,
Abby C. King a,c, John P.A. Ioannidis a,c,d, Manisha Desai c,d, Christopher D. Gardner a,⁎
a
Stanford Prevention Research Center, Department of Medicine, Stanford University Medical School, Stanford, CA, United States
b
War-Related Injury and Illness Study Center, Veterans Administration Palo Alto Health Care System, United States
c
Department of Health Research and Policy, Stanford University School of Medicine, Stanford, CA, United States
d
Department of Statistics, Stanford University School of Humanities and Sciences, Stanford, CA, United States

a r t i c l e i n f o a b s t r a c t

Article history: Numerous studies have attempted to identify successful dietary strategies for weight loss, and many have fo-
Received 12 August 2016 cused on Low-Fat vs. Low-Carbohydrate comparisons. Despite relatively small between-group differences in
Received in revised form 20 December 2016 weight loss found in most previous studies, researchers have consistently observed relatively large between-sub-
Accepted 22 December 2016
ject differences in weight loss within any given diet group (e.g., ~25 kg weight loss to ~5 kg weight gain). The
Available online 24 December 2016
primary objective of this study was to identify predisposing individual factors at baseline that help explain differ-
Keywords:
ential weight loss achieved by individuals assigned to the same diet, particularly a pre-determined multi-locus
Obesity genotype pattern and insulin resistance status. Secondary objectives included discovery strategies for further
Low fat identifying potential genetic risk scores. Exploratory objectives included investigation of an extensive set of phys-
Low carbohydrate iological, psychosocial, dietary, and behavioral variables as moderating and/or mediating variables and/or sec-
Nutrition ondary outcomes. The target population was generally healthy, free-living adults with BMI 28–40 kg/m2 (n =
Diet 600). The intervention consisted of a 12-month protocol of 22 one-hour evening instructional sessions led by
Weight loss registered dietitians, with ~15–20 participants/class. Key objectives of dietary instruction included focusing on
maximizing the dietary quality of both Low-Fat and Low-Carbohydrate diets (i.e., Healthy Low-Fat vs. Healthy
Low-Carbohydrate), and maximally differentiating the two diets from one another. Rather than seeking to deter-
mine if one dietary approach was better than the other for the general population, this study sought to examine
whether greater overall weight loss success could be achieved by matching different people to different diets.
Here we present the design and methods of the study.
© 2016 Elsevier Inc. All rights reserved.

1. Introduction substantial, ranging from highly successful to highly unsuccessful

(~25 kg weight loss to ~5 kg weight gain) [5–7].
Obesity has become one of the most significant public health chal- In 2013, our research group received NIH funding (R01 DK 91831) to
lenges of the 21st century [1,2]. Numerous studies have attempted to examine weight loss in overweight/obese adults after randomizing
identify successful dietary strategies for weight loss, with a particular them to Healthy Low-Fat vs. Healthy Low-Carb diets for 12 months.
emphasis on contrasting Low-Fat to Low-Carbohydrate (Low-Carb) The objective was to identify predisposing and measureable individual
diets. As summarized in multiple meta-analyses, the majority of these differences at baseline that would explain significant amounts of the dif-
diet trials have reported only modest mean weight loss (i.e., b 5% initial ferential weight loss achieved by individuals assigned to the same
body weight) after 12 months or longer, with limited influence of mac- weight loss diet. The study was designed to primarily address potential
ronutrient differences on average weight loss (i.e., average between- differences in genotype and insulin-glucose dynamics in particular,
group differences of 2–3 kg) [3,4]. Notably, despite these relatively while also collecting data on an extensive set of potentially relevant
small between-group average differences, weight loss variability be- physiological, psychosocial, dietary, and behavioral variables. Notably,
tween subjects within any given diet group in these studies has been the study was not designed to identify which of the two study diets
was the one best for weight loss, but rather, which diet was best for
which individuals for weight loss.- (i.e., the “whiches conundrum”) [8].
⁎ Corresponding author at: 1265 Welch Road, Stanford, CA 94305-5411, United States. Preliminary data suggested there were three distinct multi-locus geno-
E-mail address: [email protected] (C.D. Gardner). type patterns representing differential weight loss responses to

http://dx.doi.org/10.1016/j.cct.2016.12.021
1551-7144/© 2016 Elsevier Inc. All rights reserved.
152 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161

different diets: a Low-Fat Genotype (LFG), a Low-Carbohydrate Geno- eligibility. Those who remained eligible and interested after the survey
type (LCG), and a Neither Genotype. The three single nucleotide poly- were invited to attend a clinical screening that required written in-
morphisms (SNPs) considered to be components of the multi-locus formed consent. Potential participants who did not respond to the first
genotype included FABP2 (rs1799883), PPARG (rs1801282), and invitation received just one additional phone call or email invitation; if
ADRB2 (rs1042714). An additional goal was to determine whether there was no response after the second attempt, no additional efforts
other genotypic profiles had a differential impact on the effects of a par- were made to schedule the clinical screening visit. The 30-min in-per-
ticular diet on weight loss. son screening visit included measurements of height, weight, blood
Several additional significant objectives were considered of critical pressure, and a fasting fingerstick blood sample to measure glucose, tri-
importance to the study design. First, the intervention approach was de- glycerides, LDL-, HDL-, and total-cholesterol. If ineligible, contact was
signed to achieve maximal differentiation in intakes of dietary fat and discontinued after screening results were provided. Table 1 details the
carbohydrate in the free-living individuals randomized to each of the inclusion and exclusion criteria used in the screening process.
two diet groups. Both groups were asked to make large initial changes Eligible participants were required to attend an in-person, group-
from their baseline habitual diets, such that even after anticipated die- based, 60–75 min orientation, where study details were explained and
tary recidivism over the duration of the protocol, the 12-month differ- full study consent was obtained. Those who attended the orientation
entiation of fat vs. carbohydrate intake would be substantial. and provided written consent were then required to attend a pre-
Second, the intervention approach was designed to be comparably study workshop on the study's dietary assessment methodology using
challenging for the two groups. There are no standard definitions of the Nutrition Data System for Research (NDSR). During the hour-long
“Low-Fat” or “Low-Carb” in terms of grams/day or percent energy in- workshop, members of the dietary assessment team reviewed the
take. Some studies comparing the two have had ambitious goals for NDSR program, methods for collecting the 24-h recalls, and the instruc-
one group compared to modest goals for the other, making the compar- tions for using an accompanying Food Amounts Booklet. The workshop
ison unbalanced [9]. This study was designed so that both groups re- also included asking participants to fill out a phone call preference sheet
ceived equally demanding assignments. indicating their preferred phone numbers and time of day (i.e., morning,
Third, both diet approaches emphasized equally high dietary quality lunch, mid-day, or evening) for their interviewer-administered dietary
in terms of nutrient density (i.e., nutrients/Kcal). In this case, our objec- recalls.
tive was to avoid employing a study design that favored one diet over
the other in terms of overall dietary quality. For example, both diet 2.2. Run-in
groups were instructed to incorporate significant variety and quantity
of vegetables into their daily diets and to minimize added sugars and re- The intervention protocol began with a run-in period – between ori-
fined grains. entation date and the start date of the participants' first class – of ap-
In summary, for the three-diet design objectives described above, proximately one month (mean ± SD = 32 ± 10 days). During this
the protocol was designed to compare two dietary approaches that period potential participants were instructed to maintain their habitual
were maximally differentiated, equally demanding, and equally focused diet, exercise, and body weight so as to maximize the stability of their
on high quality nutrition. baseline measures.

2. Study design 2.3. Baseline data collection

Generally healthy women and men, 18–50 years of age, with a BMI Baseline data collection (see Table 2) included the following clinical
between 28 and 40 kg/m2, were randomized to a Healthy Low-Fat or measures: body weight, height, waist circumference, blood pressure,
Healthy Low-Carb weight loss diet for 12 months. The target sample blood sampling (to later assess such biomarkers as insulin, glucose,
size was n = 600. The intervention involved a series of 22 evening in- lipids, inflammatory markers, genotype) including an oral glucose toler-
structional sessions in groups of 12–22 participants per class. Partici- ance test (OGTT), resting energy expenditure (REE), and dual-energy x-
pants attended classes with the same group of individuals over time. ray absorptiometry (DXA). To assess dietary composition, unannounced
Classes were led by health educators who were all registered dietitians 24-h dietary recalls were conducted. Physical activity was assessed by
(Fig. 1). To accommodate the large sample size, enrollment was spread interviewer-administered 7-day recall [10]. Participants also completed
out across five cohorts between the spring of 2013 and the spring of a number of psychosocial questionnaires. There were optional sub-stud-
2015. Target enrollment for the five cohorts was n = 80 for Cohort 1, ies for which stool samples were collected and fat biopsies were obtain-
and n = 130 for Cohorts 2 through 5. All health educators led a similar ed. More detailed information about each of these measures is provided
number of Low-Fat and Low-Carb classes (i.e., for every cohort, each below.
health educator was assigned one Low-Fat and one Low-Carb class).
2.4. Randomization
2.1. Screening
Participants who completed their baseline clinic visit and data col-
Participants were recruited primarily through media advertisements lection were randomized to the Healthy Low-Fat or Healthy Low-Carb
(e.g., radio, online), and e-mail lists. Interested participants were re- diet group using an allocation sequence set by a computerized random
quired to complete a 10-min online survey to determine initial number generator as carried out by the study statistician. Once

Fig. 1. Summary of study protocol flow.

 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161 153

Table 1 diet group assignment at the start of the first class. The original random-
Inclusion and exclusion criteria (self-reported, unless otherwise indicated). ization plan included stratification based on preliminary data on geno-
Inclusion type predisposition. However, as will be described below in the
Age: ≥18 years of age
section describing the analysis plan, a decision was made, prior to en-
Women: pre-menopausal (self-report) and ≤50 years of age rollment of the first participant, to simplify this as a straight randomiza-
Men: ≤50 years of age tion to one of the two study arms.
Race/ethnicity: all
BMI (body mass index): 28–40 kg/m2 (measured in clinic)
2.5. Single blind
Body weight stable for the last two months, and not actively on a weight loss plan
No plans to move from the area over the next 12 months from start of study cohort
Available and able to participate in the evaluations and intervention for the study The study was single-blinded. It was not feasible to blind partici-
period pants to Healthy Low-Fat vs. Healthy Low-Carb dietary assignment.
Willing to accept random assignment However, for all staff collecting data (e.g., dietary assessment, DXA)
To enhance study generalizability, people on medications not noted below as
specific exclusions can participate if they have been stable on such medications
and for all laboratory personnel assaying samples (e.g., insulin, glucose),
for at least three months diet group assignments were masked. Only a limited number of staff not
Ability and willingness to give written informed consent involved in data collection or analysis, including the study coordinator
No known active psychiatric illness and health educators, knew the diet assignments. Subjects were explic-
itly instructed to not divulge their intervention assignment with assess-
Exclusion
ment staff.
Pregnant, lactating, within 6 months post-partum, or planning to become pregnant
in the next year
Diabetes (type 1 and 2) or history of gestational diabetes or on hypoglycemic
2.6. Staggered cohorts
medications for any other indication
Prevalent diseases: malabsorption, renal or liver disease, active neoplasms, recent The relatively large sample size of this single-site study required de-
myocardial infarction (b6 months) livery of the intervention to five staggered cohorts so as to maximize ef-
Currently smoking
ficiency of study staff and space requirements of intervention
History of serious arrhythmias, or cerebrovascular disease
History of bariatric surgery implementation. The start date of each cohort was staggered to mini-
Uncontrolled hyper- or hypothyroidism (TSH not within normal limits, mize overlap of major data collection time points (baseline and 3, 6,
self-report) and 12 months; see details in the Assessment Protocol section). Cohorts
Medications: Lipid lowering, antihypertensive medications, and those known to 1, 2, 3, 4 and 5 started in March 2013, September 2013, April 2014, Au-
affect weight/energy expenditure
Excessive alcohol intake (self-reported, ≥ to 2 drinks/day for men or ≥1 drink/day
gust 2014 and March 2015, respectively.
for women)
Musculoskeletal disorders precluding regular physical activity 3. Intervention protocol
Unable to follow either of the two study diets for reasons of food restrictions (e.g.,
vegan)
3.1. Class based education program
Currently under psychiatric care, or taking psychiatric medications
Inability to communicate effectively with study personnel
A team of health educators led the class-based education interven-
tion that was delivered over the 12-month protocol. Altogether there
randomized, participants were informed by e-mail of the date and time were 36 different class groups across all five cohorts – 18 Healthy
of their first intervention class. This communication did not inform Low-Fat class groups and 18 Healthy Low-Carb class groups. Each health
them of their diet group assignment. They were first informed of their educator taught one Healthy Low-Fat and one Healthy Low-Carb group
in each cohort. In Cohort 1 there were two health educators and in Co-
hort 2 an additional two health educators were hired to lead the classes.
Table 2
Data collection chart.
Before Cohort 3 got started, one of the original health educators left the
study, and a new health educator was hired. These four health educators
Assessmenta Screening Baseline 3 6 9 12 then completed Cohorts 3, 4, and 5. All five health educators were reg-
M M M M
istered dietitians (RDs), four of the five held Masters degrees, and two
Demographics X of the five were certified diabetes educators (CDE). Intervention fidelity
Screening survey X across health educators was established through weekly staff meetings,
Weight and waist circumference X X X X
Height X X X
during which time health educators shared information and class expe-
Blood pressure X X X X riences and engaged in group problem-solving around any issues that
Blood (i.e., OGTT insulin, OGTT glucose, X X X X came to light in the course of teaching the classes. Any behavioral issues
lipids, inflammatory markers)b with participants that arose were discussed with the study's senior be-
Diet composition – NDSRc X X X X
havioral scientist.
Physical activity 7-day recall X X X X
Medication/supplements taken X X X X All class sessions were held in the evenings, Monday through Thurs-
Questionnaires X X X X day. For the first eight weeks of each cohort, the sessions were held
Resting energy expenditure (REE)d X X X weekly. Class sessions then became less frequent – meeting once
Body composition (DXA)d X X X every two weeks for two months, then once every three weeks until
Fat biopsy (subset of ~n = 100)d X X
the six-month mark, and finally once every month for the remaining
Stool samples – microbiome sub-Study 1e X X
Stool samples – microbiome sub-Study 2f X X X X X six months. Overall, there were 22 instructional sessions throughout
a the year for each class group. The size of the class groups ranged from
Participant in-class weight and attendance were recorded at each of the 22 classes.
b
OGTT was not measured at 3 M. 12 to 22 participants (17 ± 2, mean ± SD). Once assigned to a particular
c
Diet composition data was collected on two weekdays and one weekend day at each diet assignment and class group, on a specific night of the week and, at a
time point. specific time, participants were not allowed to switch and sit in on any
d
Collected beginning in Cohort 2. other classes (i.e., not allowed to make-up for a missed class by sitting in
e
Stool samples in Sub-Study 1 were only collected in Cohort 2 and were also collected
at 10 weeks.
on another group's class); the class they attended was always their orig-
f
Stool samples in Sub-Study 2 were only collected in Cohort 3, and included whole stool inally assigned class day and time throughout the 22 sessions. Keeping
samples for 22 individuals at baseline. the classes exclusive in this manner was done with the intention of
154 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161

maximizing overall engagement and retention through the promotion carbohydrate. It was explained to participants that this was acceptable
of social cohesion, comfort and privacy, and the minimization of vulner- to the study researchers and was even to be expected due to the central
ability (i.e., avoiding newcomers in the class) among class members. For study hypothesis that each of the study diets would be easier for some
participants who missed a class session, they were provided with the participants and more difficult for others based on some combination
specific information from that session via written materials, as well as of genetic and/or metabolic predispositions. It was for these reasons
through brief email or phone contacts with their health educator. that “One Diet Does Not Fit All” was the name of the study used by the
research team for communication purposes with participants.
3.2. Dietary strategy – limbo-titrate-quality The third component of the overall dietary strategy was promoting
high dietary quality (Quality) for both groups for the full 12-month in-
There were three central components to the dietary strategy that tervention period. Optimizing diet quality was emphasized by giving
were repeatedly and consistently communicated to the participants re- both diet groups similar instructions to focus on whole, real foods that
gardless of diet assignment. The first was the “go as low as you can go” were mostly prepared at home when possible, and specifically included
(Limbo) strategy for the first eight weeks. Participants were instructed as many vegetables as possible, every day, however they liked them -
to progressively cut back on fat or carbohydrate intake until they had grilled, stir-fried, roasted, etc. They were also encouraged to choose
achieved a daily intake of no N 20 g of fat or carbohydrate per day, de- lean grass-fed and pasture-raised animal foods as well as sustainable
pending on their group assignment. Participants received explicit in- fish. With a focus on mostly consuming whole, real foods, both groups
structions that the rate of restriction was not critical to the study, and were likewise instructed to eliminate, as much as possible, processed
that reaching the 20 g per day in two vs. four vs. six vs. eight weeks food products, including those with added sugars, refined white flour
was not considered to carry any advantage or disadvantage. Therefore, products, or trans-fats. Participants were encouraged to prepare as
their rate could be variable and individually tailored. The instructions much of their own food as possible, and to optimize the inclusion of
also included a clear statement that, even though 20 g per day was the fresh, seasonal foods. When eating out or traveling, they were encour-
objective, any individuals who were unable to reach those low levels aged to ask for modifications to standard menu items that would help
would not be dropped from the study or considered to be non-compli- them adhere to their diet assignment (e.g., ordering salad dressing on
ant; rather, the expectation was more consistent with the concept of the the side for the Healthy Low-Fat group or a side of greens instead of
party game Limbo – go as low as you can go. Once participants reached mashed potatoes for the Healthy Low-Carb group).
their lowest level of fat or carbohydrate intake, they were encouraged Several of the topics related to Quality were specific to each of the
to maintain that level for at least a few weeks. There was no specific two different diet assignments. Those assigned to Healthy Low-Fat
set time for maintaining the lowest level. Rather, it was explained to were instructed to choose whole-grain foods (e.g., rather than whole
participants that the goal was to provide them with the personal expe- wheat flour products), including steel cut oats, farro, barley, quinoa,
rience of being anchored at the lowest level they could achieve and brown rice, and wild rice. Healthy Low-Fat participants were also en-
maintain, at least for a week or two. couraged to explore and consume a wide range of legumes and beans,
The second component of the dietary strategy (Titrate) involved fresh fruit, low-fat dairy products, and lean meats. Those assigned to
instructing participants to slowly add fat or carbohydrates back to Healthy Low-Carb were instructed to choose high quality oils and fats,
their diet in increments of 5–15 g/day, for periods of a week at a time, avocados, hard cheeses, nut butters, and nuts & seeds. During the Ti-
with no set endpoint goal for a specific level of fat or carbohydrate. trate phase, and throughout the remainder of the 12-month protocol,
For example, for participants who achieved 20 g/day within the first as the Healthy Low-Fat group added small amounts of fat back to the
eight weeks, and then maintained that level for at least a few weeks, diet, and as the Healthy Low-Carb group added small amounts of carbo-
they were encouraged to shift their daily goal to 25–35 g/day for a hydrate back to the diet, they were instructed to do so with these same
week or for possibly more than one week. During this process they quality foods.
were instructed to assess how the increased level of fat or carbohydrate Given that high quality foods can be more expensive than foods that
affected both their satisfaction with their daily intake (e.g., satiety, pal- are similar in type but lower in quality, the encouragement to choose
atability, and enjoyment) and their weight loss progress. If satisfaction quality was framed as a continuum as opposed to an either/or (e.g.,
and weight loss progress were acceptable, they had the option of main- for the Healthy Low-Fat participants, organic wheat berries was at the
taining that level of fat or carbohydrate intake for another week or highest level of quality, followed by conventional wheat berries, then
adding an incremental 5–15 g/day. Importantly, while participants whole wheat bread made with a minimal number of ingredients and
were encouraged to slowly add fat or carbohydrate to their diets in no additives, then a more conventional whole wheat bread with many
this manner, it was also made clear that they should not add back any ingredients including additives, and, finally, refined white flour bread
more than would be necessary to keep them at the lowest possible with many ingredients and additives was considered the lowest end
level over the long term while simultaneously addressing any concerns of the quality continuum). In other words, participants were encour-
about long-term satisfaction in areas related to satiety, palatability, and aged to choose the highest quality foods that they could reasonably
enjoyment. After adding back the designated grams/day, they could also find, realistically afford, and enjoy.
consider reversing that decision and instead reduce their intake based In summary, the diet strategy for both the Healthy Low-Fat and
on the factors mentioned above. At this point they could maintain that Healthy Low-Carb groups was a Limbo-Titrate-Quality approach, with
level for the remainder of the study, or try to add back small amounts the goal of having participants achieve an individualized, lowest-possi-
of fat or carbohydrate later in the study. Thus, for the purpose of overall ble level of fat or carbohydrate intake of maximal dietary quality, and
guidance, this Titrate component of the dietary strategy was described one that could conceivably be maintained long-term beyond the end
to participants as having the ultimate objective of having each one of of the 12-month protocol. Notably, there were no specific caloric restric-
them eventually find their individualized level of fat or carbohydrate tion goals for either diet and no single specific percentage of fat or car-
that was both: bohydrate to which they were told to strive as the final goal.
1. The lowest they could achieve, and
3.3. Similarities in intervention strategies for the two diet groups
2. The lowest they could conceivably maintain for many years to come
after the 12-month protocol ended.
Overall, there were four main foci of the instructional sessions: nu-
Inherent in this approach was the idea that the final level of fat or trition, behavior, emotions, and physical activity. While all of these
carbohydrate intake achieved among individuals within both diet were usually touched upon in each class, nutrition was typically the pri-
arms would vary substantially, with no single set target goal for fat or mary focus, and one other component per class was highlighted and
 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161 155

explored in more detail. The nutrition focus was strongest in the first the remaining class sessions focused on how to put this knowledge
two months of the 12-month protocol, after which more and more em- into practice. The behavioral modification strategies were based on a So-
phasis was given to the non-dietary components of the sessions, and cial Cognitive Model, which views behavior, including health behaviors,
these were relatively similar in many regards for the two intervention as acquired and maintained through a complex set of behavioral, cogni-
groups (see Appendix for class topics). Specific class topics taught to tive, and environmental conditions [15,16]. Social cognitive interven-
both diet groups included mindful eating, food and mood, sleep and tion strategies have been found to be effective in promoting adoption
weight, food addiction, exercise, as well as tips and demonstrations on and retention in a number of lifestyle intervention studies with a
shopping, preparing, and cooking vegetables. All health educators range of adult populations [17–21]. Health educators exposed partici-
taught their classes through the lens of helping participants to focus pants to empirically-supported principles of self-regulatory behavior
on making sustainable lifestyle changes, not simply following a tempo- change (e.g., goal setting, self-efficacy building, supportive environ-
rary ‘diet.’ Beyond the classroom instruction, each health educator was ments, healthful self-reinforcement/rewards, relapse prevention) [18,
available to offer individual contact with their class members via 20,22]. Class themes that addressed behavior included ‘the power of
email and phone (e.g., ranging from several times a week to very rarely) habit,’ ‘the practice of mindful eating,’ ‘how emotions drive food deci-
to address specific dietary questions and review food logs. sions,’ ‘the concept of bulk cooking,’ ‘grocery shopping,’ ‘meal planning,’
The Quality principle was reinforced similarly for the two diet ‘meal timing,’ ‘dining out,’ and more. The impact of participants' sur-
groups via class food demonstrations highlighting vegetables and the rounding environments and contexts (e.g., at work, at home, with
use of simple cooking equipment and techniques to inspire and encour- friends) and how to deal with the pressures that might derail them
age healthy eating at home. Recipes were provided by health educators were also discussed in detail.
via newsletters. Other recipes, selected by the participants and relevant A key concept of behavior change was reinforcing the concept of
to their dietary assignments, were shared via email and printed booklet. “One Diet Does Not Fit All” as it applies to behavior change and motiva-
Participants were also encouraged to find and purchase food from local tion, as well as Healthy Low-Fat vs. Healthy Low-Carb diets. That is,
community supported agriculture (CSA) groups or take advantage of what works for one individual, may not work for someone else. The
home delivery food services that promoted high quality food values 12-month protocol involved helping participants find the right foods
(e.g., Good Eggs). to meet individual goals and the personal behaviors and habits needed
Efforts to maximize similar retention between the two groups in- to allow for sustainable change.
cluded the health educators consistently sending reminder e-mails be-
fore class, summary e-mails after class with class materials and 3.6. Ongoing intervention adaptation and development
pertinent links to other items that came up during class, as well as
reaching out via e-mail to individuals directly, as needed. Weekly emails As the study progressed from Cohort 1 to Cohort 5, the intervention
were sent to the groups even during the months when classes were delivery was refined in several minor ways with the intent of maximiz-
meeting less often. Participants were able to post questions or com- ing treatment fidelity. Importantly, each of these minor refinements
ments to the whole group as well as privately to their health educator. was implemented equally for both the Healthy Low-Fat and the Healthy
Low-Carb groups. For example, based on participant feedback in the
3.4. Physical activity first and second cohorts, health educators determined that the in-class
content could be improved by making it less dense, less didactic, and
The promotion of regular, moderate- or more vigorous-intensity more interactive. Therefore, the scope of the in-class content was re-
levels of physical activity was identical for the two diet arms and consis- duced and the interactivity of the material was increased, while
tent with national guidelines related to weight control. Health educa- retaining the guiding principles. Part of this transition was achieved by
tors recommended 60–90 min/day of moderate-intensity physical creating a set of videos for participants and instructing them to view
activity as a target goal [11–14] and encouraged any participants who the videos prior to class (e.g., a “flipped” classroom concept). As another
were not meeting these guidelines at baseline to work up to this level example, in-class cooking activities were made more participatory by
in the first three months of the study. Those who were already exercis- including more potlucks and demonstrations of participants' own
ing at this or a higher level were encouraged to add variety, increase in- recipes.
tensity, time, or frequency. Consistent with evidence from recent Additionally, in Cohort 5, an SMS (text messaging) accountability
national reports, health educators discussed the particular importance tool was developed for both groups that was introduced after the 6-
of physical activity for weight loss maintenance. In Cohorts 1 and 2, par- month data collection time point. Briefly, the tool involved three or
ticipants were provided with pedometers. For subsequent cohorts, four text messages per week with brief queries about adherence to
while participants were not given pedometers, since health educators their eating plan, emotions related to their adherence, general well-
realized that they were largely unused due to widespread use among being, and intentions for any actionable steps to address any lapses in
participants of other activity monitors (e.g., Fit Bit, Jawbone), they adherence or challenges with well-being. This addition was based on
were encouraged to use these types of wearable devices to regularly feedback from participants in earlier cohorts that had reported chal-
track their activity levels. Health educators emphasized a mix of cardio- lenges remaining fully engaged in the study in the latter six months of
vascular and strength training and proper form to prevent injury. They the protocol when the frequency of class meetings was reduced to
encouraged participants to start with moderate levels of physical activ- once per month.
ity (e.g., 10 min/day added to what they were currently doing) and in-
crease activity according to their lifestyle. Lastly, health educators 4. Assessment protocol
encouraged participants to find ways to be active that were truly enjoy-
able and practical for the participants, so that their physical activity As presented in Table 2, data were collected across a broad range of
could be sustainable for the long-term. domains. The primary data collected for all participants across all five
cohorts included demographics, personal and family health history,
3.5. Psychology and behavior modification clinical measures (e.g., weight, waist circumference, and blood work),
dietary intake, physical activity, and psychosocial variables. Resting en-
Health educators emphasized emotional awareness and behavior ergy expenditure and percent body fat (DXA) were assessed for Cohorts
modification to support the diet and weight loss program. While the ini- 2–5. Fat biopsies were obtained from a subset of participants who
tial eight class sessions were focused mostly on nutritional knowledge volunteered in Cohorts 2–5, and stool samples were collected from a
and understanding what the food changes would be, the majority of subset of participants who volunteered from Cohorts 2 and 3. With
156 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161

few exceptions, which are noted below, all data were collected at four carbohydrate, depending on assignment, to 20 g/day. The most com-
major study time points: baseline, 3, 6, and 12 months. mon dietary monitoring method used was the on-line MyFitnessPal
tool [29]. Many of the participants indicated experience with this tool
4.1. Dietary assessment methods prior to starting the study, and most participants who used it for the
first time reported general satisfaction with its ease of use. Participants
Primary dietary intake was assessed using three unannounced 24-h were able to share access of their MyFitnessPal results with their health
dietary recalls within a two-week window at each of the four major data educators, who were then able to review entries, as needed, to help
collection time points. Each participant was expected to complete two guide participants in their diet adherence. Other methods used by
weekday dietary recalls and one weekend recall at each data collection those preferring an alternative approach included a paper food log pro-
time point for a total of 12 recalls. Data were collected using NDSR, a vided by the health educators or one of several web-based tracking tools
computer-based software application developed at the University of that are similar to MyFitnessPal, (e.g., MyNetDiary and Lose It!). As the
Minnesota Nutrition Coordinating Center (NCC). Dietary recalls were study progressed, and with the increasing popularity of wearable de-
collected in a standardized fashion using a multiple-pass interview ap- vices, many participants reported tracking via their FitBit and UP
proach consisting of five steps to ensure completeness and accuracy wristbands.
[23,24]. First, participants were asked to list all foods and beverages con-
sumed in the previous 24 h (i.e., midnight to midnight). Second, the in- 4.3. Physical activity assessment
terviewer reviewed the list with the participant. Third, the interviewer
collected detailed information about each reported food and beverage The Stanford Seven-Day Physical Activity Recall (PAR) was adminis-
(e.g. method of preparation and amount consumed). Fourth, the inter- tered by trained study staff at the same time as one of the dietary recalls
viewer probed for commonly forgotten foods. Fifth, the information at each major data collection time point to assess participants' self-re-
was reviewed for completeness, correctness, and marked as a typical ported level of physical activity [10]. Originally developed at Stanford
or atypical day. Throughout the recall, the NDSR software searched for University in the early 1980′s, the PAR is a semi-structured interview
foods and brand name products by name and prompted the data collec- that documents the time an individual engages in physical activity,
tors with requests for additional detailed information [25]. All data col- strength, and flexibility activities during the 7 days preceding the inter-
lectors were trained by NDSR certified lead staff and were blinded to the view. An interviewer guides the participant through the recall of daily
assigned diets. activities to determine the length and intensity of the physical activities.
The NCC Food and Nutrient Database serves as the source of food Physical activity is measured as total energy expenditure and time spent
composition information in NDSR [26]. This database includes over in moderate, hard, and very hard physical activity. Hours per day spent
18,000 foods including 8000 brand name products. Ingredients and in the various categories of physical activity intensity are then converted
preparation methods allow for N160.000 food variants. Values for 165 to a daily average of metabolic equivalents (METS) and then used to es-
nutrients, nutrient ratios, and other food groups were generated by timate total energy expenditure per day in units of Kcal/kg/day. The
the database. The USDA Nutrient Data Laboratory is the primary source questionnaire also captures sleeping time, time spent cooking, and sev-
for nutrient composition for the database. These values are supplement- eral other lifestyle behaviors.
ed by food manufacturers' information and data available in the scientif-
ic literature [27]. Standardized, published imputation procedures were 4.4. Weight, height and waist circumference
applied to minimize missing values [28]. In addition, to the extent pos-
sible, the interviewers entered recipes or ingredients for homemade, Body weight was recorded without shoes to the nearest 0.1 kg using
restaurant, and other items not included in the software. The lead die- a calibrated Scale-tronix clinical scale. Height was measured to the
tary assessment nutritionist conducted a quality check for each cohort nearest 0.1 cm using a Seca wall-mounted stadiometer. Waist circum-
after data collection at each study collection point. This involved an in- ference was measured on the skin at the umbilicus to the nearest
depth review of both individual and composite reports for complete- 0.1 cm. All measurements were taken by a nurse at the Stanford Clinical
ness and errors. & Translational Research Unit (CTRU) at each time point. All clinic visits
As mentioned previously, all participants attended a pre-randomiza- started between 7:00 and 9:30 am, with participants in a fasted state for
tion 60-min, in-person training to learn the procedure for dietary recalls at least 10–12 h.
and were provided with a Food Amounts Booklet to be used at the time
of each data collection to enhance estimating portion sizes. When an in- 4.5. Blood pressure
person meeting was not possible, this training was done via email and/
or telephone (b1% of participants). After 5 min of sitting/resting, CTRU nurses obtained three blood
During each time point data collection window, participants re- pressure readings on the right arm one minute apart. These were col-
ceived unannounced phone calls to complete dietary recalls. If partici- lected automatically using a WelchAllyn, Spot Vital Signs LXi. If a
pants could not be reached via phone by the interviewers after participant's blood pressure was over 160 systolic or 90 diastolic, they
approximately five to seven attempted telephone calls, including rested another five minutes before taking the blood pressure again. If
voicemail messages, then the health educator for that specific partici- the blood pressure remained high after several readings, the study coor-
pant was notified and assisted in communication by either bringing dinator and sometimes the study physician were notified. In all cases,
this to the participant's attention in the next class, or sending an the participant was able to continue with the remainder of the clinic
email. If the participant still did not respond, the study coordinator visit. For analysis purposes, the first measurement was disregarded,
and/or Principal Investigator attempted to contact the participant via and the second and third measurements were averaged according to
email or phone. This process continued until the data were collected, NHANES guidelines [30].
the participant communicated with the study staff his/her wish to dis-
continue participation, or the data collection window closed. 4.6. Blood samples

4.2. Participant self-monitoring of diet Blood was sampled by fingerstick during the screening clinic visit for
all potential participants to assess a fasting lipid profile and blood glu-
Study participants employed a variety of methods to self-monitor cose using the Cholestech LDX machine. For all participants determined
adherence to their diet group assignment throughout the study, partic- eligible after screening that consented to participate in the study, subse-
ularly in the first eight weeks when the goal was to lower dietary fat or quent blood samples were taken at baseline, 3, 6 and 12 months via
 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161 157

venipuncture at the Stanford CTRU by trained nurses or phlebotomists. three arrays – CVD II, CVD III and Inflammation I – were used to assess
Aliquots of plasma and serum were obtained at all time points; buffy samples. The method is based on a proximity extension assay (PEA) in
coats were collected at baseline, 6 and 12 months. which 92 oligonucleotide-labeled antibody probe pairs can bind to
their respective target, present in the sample. The PEA technique has
4.7. Lipids and lipoproteins an advantage over conventional multiplex immunoassays, since only
correctly matched antibody pairs give rise to a signal, yielding an ex-
Lipids were assessed at all four times points (i.e., baseline, 3, 6, and tremely high specificity. PEA is a homogeneous assay that uses pairs of
12 months) from a fasting blood sample. Blood was collected into pur- antibodies equipped with DNA reporter molecules. When binding to
ple top EDTA vacutainer tubes. Samples were processed, aliquoted, their correct targets, they give rise to new DNA amplicons, each ID-
and frozen directly by the CTRU lab after being drawn. Samples were barcoding their respective antigens. The amplicons are subsequently
stored in a − 80° freezer until the time of processing for analysis. Plasma quantified using a Fluidigm BioMark™ HD real-time PCR platform.
triglycerides, total- and HDL-cholesterol were measured by enzymatic This PCR-based technique offers a high-throughput of simultaneous
endpoint analysis on a clinical chemistry analyzer (Liasys 330) using measurements of 92 protein biomarkers in just one microliter of sam-
methodology previously described [31–33]. LDL-cholesterol was calcu- ple, without any cross-reactivity. Proseek Multiplex provides accurate
lated using the Friedewald equation. Triglyceride and cholesterol mea- quantification below picogram per milliliter levels, even in small sam-
surements are standardized through the CDC-NHLBI lipid ples [40].
standardization program. Apolipoproteins B and AI were analyzed by
immunoturbidimetric assay using the K-assay reagent kits (Kamiya Bio- 4.10. Genotyping
medical). Lipoprotein particle concentrations were measured by ion
mobility, a process that allows for direct particle quantification as a DNA extraction was carried out by the Stanford Cancer Institute
function of particle diameter [34]. Ion mobility is based on the principle (SCI) Biorepository. Briefly, high quality DNA samples were extracted
that particles of a given size behave in a predictable manner when car- from the buffy coat/red blood cell suspension using laborious phenol/
ried in a laminar flow of air and subjected to an electric field. Briefly, a chloroform purification method, a modified procedure of Baas et al.
solution of lipoproteins, depleted of other serum proteins, is introduced [41], Gustafson et al. [42] and Paul et al. [43]. Processing utilized lysis
into a flow of air by electrospray. In the electrospray chamber, the buffer (144 mM NH4Cl, 14 mM NH4HCO3), pellet buffer (10 mM Tris
desolvated, highly charged lipoprotein particles are nearly neutralized pH 8.0, 10 mM EDTA, 150 mM NaCl), Proteinase K, 10% SDS, and
by ionized air, resulting in a known fraction of singly-charged particles RNAse, incubated at 50 °C for ~16 h, protein-denaturant buffers, phenol,
exiting the electrospray chamber. The particles are then carried by air- and chloroform. Processing involved precipitation by NaCl (100–
flow to a differential mobility analyzer (DMA), where a variable electric 250 mM, 2 volumes of cold Ethanol) and resuspending the dry pellet
potential, perpendicular to the direction of the airflow, causes the parti- in TE (10 mM Tris, 0.1 mM EDTA, pH 8.0). A series of QA&QC assays
cles to drift toward a collection slit. The velocity of the particles across were carried out to determine the DNA purity, concentration, integrity,
the airflow is proportional to the particle diameter and to the electrical and digestibility after the DNA was completely dissolved. The working
potential. Only singly charged particles are detectable. Those with dif- DNA solution was stored at 4 °C and the original stock at −80 °C.
ferent diameters reach the collection slit at different electrical potentials The Affymetrix Axiom® Genotyping platform was used for analysis
and are then carried to a particle counter where they are detected by of single nucleotide polymorphisms (SNPs) and insertions/deletions
light scatter and are counted after transition through the DMA. As a re- (indels). The specific microarray design used was the UK Biobank
sult, lipoprotein particle concentrations are measured directly as a func- Axiom® Array [44]. There are 820,967 SNPs and indel markers on the
tion of their particle diameters. The final numerical ion mobility output array, which included the three SNPs from the original study design -
is reported in nanomoles per liter for combined bins of particles FABP2 (rs 1,799,883), PPARG (rs 1,801,282), and ADRB2 (rs
summed into commonly reported subclasses of lipoproteins. These in- 1,042,714). The array was designed with imputation-aware algorithms,
clude: HDL 3, HDL 2b, HDL 2a, LDL 4c, LDL 4b, LDL 4a, LDL 3b, LDL 3a, enabling characterization of millions of additional markers. Simulation
LDL 2b, LDL 2a, LDL I, IDL 2, IDL 1 and VLDL sm, VLDL med, and VLDL results show high concordance between imputed genotypes and geno-
large. The peak LDL particle diameter for each sample is reported as is types in the Phase I 1000GP release for between 6 and 9 million
the LDL phenotype associated with it: pattern A, I, or B. A full description markers, depending on the population.
of this method is published elsewhere [34]. All of the lipid assays were DNA extracted as described above was analyzed at the Affymetrix
performed by the Krauss Lab at the Children's Hospital Oakland Re- Research Services Laboratory facility using the Axiom® 2.0 Assay Auto-
search Institute (CHORI, Oakland, CA) for all cohorts. mated Workflow [45]. Total genomic DNA (200 ng) was amplified and
randomly fragmented into 25 to 125 base pair (bp) fragments. These
4.8. Glucose, insulin, oral glucose tolerance test (OGTT) fragments are purified, re-suspended, and hybridized to Axiom®
Array Plates. Following hybridization, each polymorphic nucleotide
Blood was collected to assess post-fasting plasma glucose and insu- was queried via a multi-color ligation event carried out on the array sur-
lin via phlebotomy at the Stanford CTRU. Insulin levels were assessed face. After ligation, the arrays were stained and imaged on the
by radioimmunoassay by the Core Laboratory for Clinical Studies Wash- GeneTitan MC Instrument. Data management was performed using
ington University School of Medicine, St. Louis, Missouri [35]. Glucose Axiom™ Analysis Suite, Affymetrix® Genotyping Console™ Software
levels were analyzed using a Beckman Glucose Analyzer II (BGA II) by (GTC) or Affymetrix® Power Tools (APT) and SNPolisher R package to
electrochemical technique [36]. For the OGTT, serial blood sampling perform quality control analysis (QC), for samples, plates, and SNPs fil-
was collected under fasting conditions and then at 30, 60 and 120 min tering prior to downstream analysis, and advanced genotyping methods
after consuming 75 g of glucose solution [37,38]. The Matsuda index [46].
was calculated to assess insulin sensitivity according to the methods
of Matsuda et al. [39]. 4.11. Psychosocial questionnaires

4.9. Targeted proteomics: OLINK Proseek® technique Participants completed a number of psychosocial instruments ad-
ministered online at baseline, 3, 6 and 12 months (see Appendix B).
Candidate protein biomarkers were assessed using a high-through- The questionnaires covered a variety of topics including General Health
put technique, the OLINK Proseek® Multiplex kits (www.olink.com). [47], Sleep Quality [48], Outcome Expectations and Outcome Realiza-
Each kit measures 92 disease-related proteins in plasma samples, and tions [49], Diet Self-Efficacy [50], Social Support & Eating Habits [51],
158 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161

Stress [52], Dieting History [53], Eating Inventory [54], Reasons for b55 g of carbohydrates per day on average, or being in the Healthy
Dieting [55], Eating Attitudes [56], Food Attitudes [57], Food Choices Low-Fat group and consuming b 40 g of fat per day on average, as deter-
[58], Body Image [59], Depressive Symptoms [60], Food Addiction mined during the 3-month NDSR diet data collection, and (2) losing ≥10
[61], Neighborhood Environment Walkability [62], Social Cohesion of lbs. in the six months since baseline.
Neighborhood [63], Group Cohesion Scale-Revised [64], Self-Control A subset of those who were eligible for the six month follow-up
and Self-Management [65], and Emotional Eating [66]. and met the additional eligibility criteria of having a TG:HDL-C
ratio (at the time of their screening visit) of equal to or N1.5 for
5. Sub-studies women and 2.5 for men, were asked to participate in a more elabo-
rate 6 month follow-up consisting of a meal tolerance test with fat
In addition to the above-mentioned measurements, participants had biopsy. For those completing the meal tolerance test, similar to the
an opportunity to participate in several sub-studies. Some were only of- OGTT, eligible participants fasted for 12 h prior to the appointment.
fered to certain cohorts but were required of all in that cohort, while Researchers took an initial fat biopsy at the beginning of the appoint-
others were optional. ment, then provided participants with a Healthy Low-Fat or Healthy
Low-Carb meal, depending on group assignment. A second fat biopsy
5.1. Dual energy X-ray absorptiometry (DXA) was taken two hours after the meal. Blood was collected from each
participant every hour for four hours, after the initial fasting blood
Dual Energy X-ray Absorptiometry (DXA) scans were performed to draw was taking.
examine whole body adiposity, lean body mass, and bone density at
baseline, 6 and 12 months. Scans provided body composition measure- 5.4. Stool collection for microbiome analyses
ments for six specific body areas (i.e., left arm, right arm, left leg, right
leg, trunk, and head) as well as whole body composition. Each individ- Stool samples were collected from a subset of volunteers from Co-
ual underwent DXA scans using a Hologic QDR-4500 W fan-beam scan- horts 2 and 3 to examine the microbiome. In Cohort 2, two samples of
ner (Bedford, MA) based on the manufacturer's guidelines. Quality stool taken one week apart were collected from participants at baseline,
control procedures were carried out regularly based on the and then subsequent samples were collected at 10 weeks and 6 months.
manufacturer's recommendations and the instrument was calibrated Participants were instructed to, “use the spoon attached to the cap to
weekly using appropriate phantoms supplied by the manufacturer. put several scoopfuls of stool into the collection tube until each tube
DXA data were collected for Cohorts 2 through 5 (i.e., resources were has a portion of the stool specimen equivalent to large marble or a
not available at the onset of the study for cohort 1). One technician com- walnut.”
pleted all scans for all participants at all time points to minimize poten- Stool sample collection in Cohort 3 involved five time points: base-
tial variability. line, 3, 6, 9 and 12 months. For the majority of the sampling, two
tubes of a walnut size sample of stool were taken from a single bowel
5.2. Resting energy expenditure (REE) movement. The baseline collection differed from the others in two
ways. First, for baseline, two samples were collected one week apart
Respiratory gas exchange was measured at the Stanford CTRU. The (i.e., each with two tubes). Second, for approximately one quarter of
assessment was performed at the same time (i.e., 7:00 am–9:30 am) the participants one of the two samples was a complete stool collection
and in the same room at each visit to minimize deviations in environ- rather than just a walnut sized portion.
mental conditions. Participants were measured in a fasting state after Participants from both cohorts were given ice packs and supplies to
lying supine for 5 min at their baseline, 6 and 12 month clinic visit. Spe- freeze the samples as soon as possible after collection and to keep the
cifically, measurements were taken after vitals but before the first blood samples on ice until they were delivered to the research unit. Once sam-
draw for the OGTT protocol. This was done using the PravoMedics ples were delivered to the research unit they were put into storage at
TrueOne 2400 metabolic cart [67,68]. Similar to the DXA data, REE –80 °C.
data were collected for Cohorts 2 through 5 (i.e., resources were not
available at the onset of the study for Cohort 1). 6. Analysis plan, including early modifications to the study design
Due to the high volume of our study population, two separate met-
abolic carts at the CTRU were purchased and used to measure REE An overview of the original analysis plan as described in the grant
[69]. The equipment was allowed to warm up for 30 min prior to cali- application to the NIH is provided below. However, modifications
bration and testing. Flow and gas calibration was performed every have been made to this original plan due to four important
morning the metabolic carts were to be used. Resting measurements developments.
were taken for a minimum of 20 min for each participant, and the first
5 min were discarded from analysis. Data were collected every 60 s. • First, after NIH funding was obtained in 2012, additional funding
Output files included the following variables: VO2, VO2/kg, METS, was acquired to augment the original study which enabled increasing
VCO2, VE, RQ, RR, Vt, FEO2, FECO2, and REE. Our analyses focused on the sample size from n = 400 to n = 600, as well as adding some
the average values for those variables from minute 6 to minute 20. of the metabolic and physiological assessments that are described
above.
5.3. Fat biopsies • Second, and related, the racial/ethnic composition of the study sample
was modified. The original study population was intended to be re-
Beginning in Cohort 2, fat biopsies were collected to assess fat cell bi- stricted to Caucasians because of the primary focus on a genetic inter-
ology from approximately 20% of participants. The procedure was of- action involving SNPs, for which allele frequencies have been
fered on a rolling basis to the first 20–30 volunteers per cohort. established for Caucasians but not other racial/ethnic groups. The
Participants were instructed to remain weight stable and to refrain three SNPs of original interest were considered to be components of
from using any over the counter anti-inflammatories for a week before a multi-locus genotype that included FABP2 (rs 1,799,883), PPARG
the procedure. Lower abdominal fat biopsies were taken at baseline. For (rs 1,801,282), and ADRB2 (rs 1,042,714). Preliminary data [70] sug-
all biopsies, participants were asked to fast from food and all non-water gested that there were three distinct multi-locus genotype patterns
liquids for 12 h prior to the procedure. Follow-up fat biopsies were with differential weight loss response to different diets: a Low-Fat Ge-
taken at 6 months from those meeting the following two eligibility notype (LFG), a Low-Carbohydrate Genotype (LCG), and a Neither
criteria: (1) being in the Healthy Low-Carb group and consuming Genotype, with rough distributions among Caucasians of 40:40:20.
 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161 159

To maintain the intent of the originally funded NIH study, it was de- A second primary hypothesis is that there will be a significant diet X
cided to complete the enrollment of n = 400 Caucasians. In order to insulin sensitivity interaction for weight loss success. Based on previous
broaden the generalizability of the study, it was decided to recruit findings [5,72–75], we predict that weight loss success will be greater
the additional n = 200 from participants of all non-Caucasian race/ for those on the Healthy Low-Carb diet who are more insulin resistant
ethnicities (which included Hispanics) that met all other study inclu- at baseline and greater for those on the Healthy Low-Fat diet who are
sion/exclusion criteria. more insulin sensitive.
• Third, the original plan to stratify randomization to the two diet Exploratory analyses will test whether various factors mediate the
groups by multi-locus genotype pattern was altered. Prior to ini- relationship between matched assignment and weight loss. We will
tiating the study we determined that the evidence supporting use contemporary mediation analyses techniques [76,77]. Selected me-
the proposed genotype patterns remained unclear, and we diators for these analyses include insulin sensitivity, energy intake, per-
wanted to explore other possible genotype interactions and asso- ceived appetite, satiety and hunger, resting energy expenditure, and
ciations. We still plan to test the original multi-locus genotype physical activity.
patterns, but we will also evaluate other SNPs secondarily in an
exploratory manner, as described in the original aims of the 6.2. Statistical analyses
grant application.
• Fourth, the original plan for primary analysis was to employ To include subjects with missing data, the primary hypothesis of
ANOVA to evaluate change in weight at 12 months as a function diet-genotype interaction will be assessed using a linear mixed
of study arm and genotype. This analysis – known as a com- model with weight as the outcome. The intention-to-treat (ITT)
plete-case analysis – includes only those subjects who contribute principle will be followed. All participants who were randomized
body weight data at both baseline and 12 months. To address this will be included in the analysis and analyzed according to their
issue of missing data, we initially proposed secondary analyses assigned treatment, irrespective of compliance. Weight change
that incorporated multiple imputation-based methods. This over time (i.e., baseline, 3, 6 and 12 months) between the interven-
would allow insight into the effect of assumptions regarding the tion groups will be assessed using a linear mixed effects model with
missing data on our primary findings. We currently propose main effect terms for time and whether or not a subject was a
using mixed effects regression methods to evaluate change in “match” (i.e., assigned to the diet implied by their genotype). The
body weight at 12 months as the primary analysis, keeping the model will include interaction terms for time and match and sub-
secondary analyses as proposed and adding the complete-case ject-specific intercepts to account for within-subject correlation
analysis as another secondary analysis. This allows us to better over time. Since participants were randomly assigned to diet
adhere to intent-to-treat principles as all subjects randomized groups, no baseline characteristics will be adjusted in the analysis.
will be included in the analysis, even if they do not contribute For the original genotypic classification of interest, statistical signif-
weight data at 12 months. We will use maximum likelihood tech- icance will be assessed at the 0.05 level. For all other SNPs, we will
niques for estimation, allowing subjects with any data points to assess significance after controlling the false discovery rate to be
contribute to the analysis, and borrowing strength from data con- no N0.05.
tributed across subjects.
6.3. Missing data, drop-outs, and intent-to-treat

Analyses that do not account for missing data can lead to biased
6.1. Hypotheses
and inefficient estimates. To address such issues, our original plan
was to perform both a complete-case analysis that excludes individ-
The first primary hypothesis of the study is that there will be a uals missing at least one variable in the model as well as a multiple
significant diet-genotype interaction for weight loss success. We imputation-based model. The latter allows adherence to intention-
predict that diet and genotype main effects will be non-significant to-treat principles in that all subjects randomized are included in
and that only by taking into account an interaction between the the analysis regardless of drop out. Importantly, multiple imputa-
two factors can we predict weight loss success. tion provides statistically valid results when the data are missing
We have planned a number of follow-up exploratory analyses as at random (i.e., the reason for missingness is related to observed
well. One hypothesis is that other obesity-relevant SNPs will predict variables only) [78]. Our current plan is to utilize mixed effects re-
weight loss success. This process includes evaluating the contribu- gression techniques that provide statistically valid results under
tion of the newly identified SNPs in predicting weight change. the same condition (i.e., missing at random) as those upon which
These SNPs will each be evaluated in turn and also expressed as a multiple imputation relies. In addition, these methods allow us to
weighted linear combination or score. The additional SNPs we will adhere to intention-to-treat principles as all subjects randomized
consider are those that have been previously and robustly docu- will be included in the analysis, regardless of attrition.
mented to have genome-wide significant associations with weight,
waist circumference, and/or metabolic phenotypes (e.g., lipid levels, 7. Conclusion
type 2 diabetes, and insulin resistance) in previous studies We will
systematically review these studies for these associated phenotypes At the core of the current study is a weight loss diet intervention
using the database of the Catalog of Genome-Wide Association stud- comparing a Healthy Low-Fat vs. a Healthy Low-Carb diet among
ies, which is continuously updated by the European Molecular Biolo- non-diabetic and generally healthy adults ages 18–50 years with a
gy Laboratory-European Bioinformatics Institute [71] and will BMI in the range of 28–40 kg/m2. However, the study was not de-
include all independent genetic variants with corresponding p- signed to simply test whether Healthy Low-Fat or Healthy Low-
values smaller than 5 × 10–8 for any of these phenotypes. These ex- Carb is better overall for weight loss success. The study was designed
ploratory analyses would serve to identify additional potential gene with the recognition from more than a dozen previous Low-Fat vs.
loci that may regulate response to specific diets. We expect these Low-Carb studies that the variability in weight loss within diet
exploratory associations to be studied in other larger studies in the groups typically ranges from highly successful to very disappointing,
future. There are currently almost 200 such independent genetic var- while the difference in average weight loss between diet groups is
iants that have been discovered, and we anticipate that approxi- typically negligible. Thus, this study was designed to examine inter-
mately 250 or more may be available in the next year or so. actions between diet group assignment and genotype and
160 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161

metabolism (e.g., insulin resistance). Beyond the primary hypothe- Appendix B. Psychosocial questionnaires
ses about interactions with genotype and metabolism, the current
study will generate a rich data set to examine a wide range of phys-
iological and psychosocial factors that likely contribute to the het-
erogeneity of response to weight loss diets. The study is intended Number Questionnaire names
to test these hypotheses and then generate many more. It has been
1 Short Form Health Survey (SF-36)
designed to reframe a central question about diet and weight loss: 2 Pittsburgh Sleep Quality Index (PSQI)
Rather than searching for the one best diet to recommend to all, 3 Outcome Expectationsa
this study seeks to determine if overall success will be greater 4 Outcome Realizationsb
when different diets are matched to different people based on pre- 5 Self-Efficacy Scalec
6 Abbreviated Social Support for Eating Habits Survey
disposing individual differences. 7 Perceived Stress Scale (PSS-14)
8 Dieting Historya
Sources of support 9 Three-Factor Eating Questionnaire (TFEQ)
10 Treatment Self-Regulation Questionnaire (TSRQ-Diet)
11 Eating Attitudes Test (EAT)
National Institute of Diabetes and Digestive and Kidney Diseases NIH
12 Food Attitude Survey (FAS)
1R01DK091831, Nutrition Science Initiative, National Heart, Lung, and 13 Food Choice Questionnaire (FCS)
Blood Institute NIH T32HL007034, NIH 1 K12 GM088033, Stanford Clin- 14 Body Dissatisfaction subscale of the Eating Disorder Inventory (EDI)
ical and Translational Science Award (CTSA) to Spectrum NIH UL1 15 Beck Depression Inventory- (BDI-1A)
TR001085, and the War-Related Injury and Illness Study Center and 16 Yale Food Addiction Scale (YFAS)
17 Neighborhood Environment Walkability Scale (NEWS-A)a
VA Palo Alto Health Care System. The content is solely the responsibility 18 Social Cohesion and Trust Within Neighborhoods Scalea
of the authors and does not necessarily represent the official views of 19 Group Cohesiond
the NIH or other funders. 20 Self-Control and Self-Management Scale (SCMS)
21 Emotional Eating Scale (EES)
Acknowledgements a
Questionnaire only asked at Baseline
b
Questionnaire only asked at 3- and 12-month periods
c
We would like to acknowledge the many study team members Questionnaire skipped only at 12-month period
d
Questionnaire only asked at 3-month period.
who contributed to various aspects of the original design and/or im-
plementation the study, and for either contributing sections to the
References
manuscript or contributing helpful review. These team members in-
clude our health educators Jae Berman, Dalia Perelman, and Mandy [1] B.A. Swinburn, G. Sacks, K.D. Hall, K. McPherson, D.T. Finegood, M.L. Moodie, S.L.
Carroll, our diet assessment team of Molly Shimer-John, Valeria Gortmaker, The global obesity pandemic: shaped by global drivers and local envi-
ronments, Lancet 378 (9793) (2011) 804–814.
Alaimo and Diane Demis, Katherine Dotter who oversaw innova- [2] C.L. Ogden, M.D. Carroll, B.K. Kit, K.M. Flegal, Prevalence of childhood and adult obe-
tions and technology, Kyla Kent who provided expertise in DXA as- sity in the United States, 2011-2012, JAMA 311 (8) (2014) 806–814.
sessment, Alexandra Rossi who managed blood samples, and Dr. [3] B.C. Johnston, S. Kanters, K. Bandayrel, P. Wu, F. Naji, R.A. Siemieniuk, G.D. Ball, J.W.
Busse, K. Thorlund, G. Guyatt, J.P. Jansen, E.J. Mills, Comparison of weight loss among
Kenji Nagao who provided helpful review of the manuscript. We named diet programs in overweight and obese adults: a meta-analysis, JAMA 312
would like to acknowledge the work early in the study of Antonella (9) (2014) 923–933.
Dewell as the initial project coordinator and Josephine Lin as a re- [4] A.J. Nordmann, A. Nordmann, M. Briel, U. Keller, W.S. Yancy Jr., B.J. Brehm, H.C.
Bucher, Effects of low-carbohydrate vs low-fat diets on weight loss and cardiovascu-
search assistant. And, finally we would like to acknowledge the lar risk factors: a meta-analysis of randomized controlled trials, Arch. Intern. Med.
many participants who made this study possible. 166 (3) (2006) 285–293.
[5] C.D. Gardner, A. Kiazand, S. Alhassan, S. Kim, R.S. Stafford, R.R. Balise, H.C. Kraemer,
A.C. King, Comparison of the Atkins, Zone, Ornish, and LEARN diets for change in
Appendix A. Diet education class topics weight and related risk factors among overweight premenopausal women: the A
TO Z Weight Loss Study: a randomized trial, JAMA 297 (9) (2007) 969–977.
[6] I. Greenberg, M.J. Stampfer, D. Schwarzfuchs, I. Shai, D. Group, Adherence and suc-
cess in long-term weight loss diets: the dietary intervention randomized controlled
trial (DIRECT), J. Am. Coll. Nutr. 28 (2) (2009) 159–168.
Class number Class topicsa [7] W.S. Yancy Jr., E.C. Westman, J.R. McDuffie, S.C. Grambow, A.S. Jeffreys, J. Bolton, A.
Chalecki, E.Z. Oddone, A randomized trial of a low-carbohydrate diet vs orlistat
Class 1 Reveal Diet Assignment and Intro plus a low-fat diet for weight loss, Arch. Intern. Med. 170 (2) (2010) 136–145.
Class 2 Meal Planning [8] A.C. King, Behavioral medicine in the 21st century: transforming “the road less trav-
Class 3 Cooking Vegetables & Fat Acids Spices eled” into the “American way of life”, Ann. Behav. Med. 47 (1) (2014) 71–78.
Class 4 Real Food, Organic [9] L.A. Bazzano, T. Hu, K. Reynolds, L. Yao, C. Bunol, Y. Liu, C.S. Chen, M.J. Klag, P.K.
Class 5 Mindless Eating & Bulk Cooking Whelton, J. He, Effects of low-carbohydrate and low-fat diets: a randomized trial,
Class 6 Food & Mood Ann. Intern. Med. 161 (5) (2014) 309–318.
Class 7 Dining Out [10] J.F. Sallis, W.L. Haskell, P.D. Wood, S.P. Fortmann, T. Rogers, S.N. Blair, R.S.
Class 8 Beyond Activation Phase (Finding Diet Stability) Paffenbarger Jr., Physical activity assessment methodology in the Five-City project,
Class 9 Exercise Am. J. Epidemiol. 121 (1) (1985) 91–106.
Class 10 Behavior Change [11] Physical Activity Guidelines Advisory Committee, Physical Activity Guidelines Advi-
sory Committee Report, U.S. Department of Health and Human Services, Washing-
Class 11 Food Addiction
ton, DC, 2008.
Class 12 Mindful Eating
[12] A.H. Lichtenstein, L.J. Appel, M. Brands, M. Carnethon, S. Daniels, H.A. Franch, B.
Class 13 What's on Your Plate? (Food Recall Activity)
Franklin, P. Kris-Etherton, W.S. Harris, B. Howard, N. Karanja, M. Lefevre, L. Rudel,
Class 14 The Study (Explaining Available Study Data and Uses) F. Sacks, L. Van Horn, M. Winston, J. Wylie-Rosett, Diet and lifestyle recommenda-
Class 15 Sleep tions revision 2006: a scientific statement from the American Heart Association Nu-
Class 16 Accountability trition Committee, Circulation 114 (1) (2006) 82–96.
Class 17 Breaking the Plateau (Breaking Weight Plateau) [13] W.H.M. Saris, S.N. Blair, M.A. van Baak, S.B. Eaton, P.S.W. Davies, L. DiPierto, M.
Class 18 The Bucket List (Food, Behaviors, Exercise, & Emotions) Fogelholm, A. Rissanen, D. Schoeller, B. Swinburn, A. Temblay, K.R. Westerterp, H.
Class 19 Sharing Challenges and Solutions Wyatt, How much physical activity is enough to prevent unhealthy weight gain?
Class 20 Free from Diet Mentality Outcome of the IASO 1st stock conference and consensus statement, Obes. Rev. 4
Class 21 Healthy Weight for Life (2003) 101–104.
Class 22 End of Study (Christopher Leads) [14] Department of Health and Human Services, Dietary Guidelines, http://www.
healthierus.gov/dietaryguidelinesN2005.
a
Class topics were slightly updated throughout the study, although the content largely [15] A. Bandura, Social Foundations of Thought and Action: A Social Cognitive Theory,
remained the same. Prentice Hall, Englewood Cliffs, New Jersey, 1986.
 M.V. Stanton et al. / Contemporary Clinical Trials 53 (2017) 151–161 161

[16] A. Bandura, Self-efficacy: The Exercise of Control, W. H. Freeman and Company, New [49] K.L. Neff, A.C. King, Exercise program adherence in older adults: the importance of
York, New York, 1997. achieving one's expected benefits, medicine, exercise, Nutr. Health 4 (1995)
[17] K.D. Brownell, L.R. Cohen, Adherence to dietary regimens. 2: Components of effec- 355–362.
tive interventions, Behav. Med. 20 (4) (1995) 155–164. [50] A.W. Garcia, A.C. King, Predicting long-term adherence to aerobic exercise: a com-
[18] J.P. Foreyt, G.K. Goodrick, Impact on behavior therapy on weight loss, Am. J. Health parison of two models, J. Sport Exerc. Psychol. 13 (1991) 394–410.
Promot. 8 (6) (1994) 466–468. [51] J.F. Sallis, R.M. Grossman, R.B. Pinski, T.L. Patterson, P.R. Nader, The development of
[19] R.W. Jeffery, L.H. Epstein, G. Wilson, A. Drewnowski, A.J. Stunkard, R.R. Wing, Long- scales to measure social support for diet and exercise behaviors, Prev. Med. 16 (6)
term maintenance of weight loss: current status, Health Psychol. 19 (Suppl. 1) (1987) 825–836.
(2000) 5–16. [52] S. Cohen, T. Kamarck, R. Mermelstein, A global measure of perceived stress, J. Health
[20] A.C. King, B. Frey-Hewitt, D.M. Dreon, P.D. Wood, Diet vs exercise in weight mainte- Soc. Behav. 24 (4) (1983) 385–396.
nance. The effects of minimal intervention strategies on long-term outcomes in [53] S.A. French, R.W. Jeffery, Current dieting, weight loss history, and weight suppres-
men, Arch. Intern. Med. 149 (12) (1989) 2741–2746. sion: behavioral correlates of three dimensions of dieting, Addict. Behav. 22 (1)
[21] R.R. Wing, D.F. Tate, A.A. Gorin, H.A. Raynor, J.L. Fava, A self-regulation program for (1997) 31–44.
maintenance of weight loss, N. Engl. J. Med. 355 (15) (2006) 1563–1571. [54] A.J. Stunkard, S. Messick, The three-factor eating questionnaire to measure dietary
[22] K.D. Brownell, The LEARN Manual for Weight Management, 10th ed. American restraint, disinhibition and hunger, J. Psychosom. Res. 29 (1) (1985) 71–83.
Health Publishing Company, Dallas, TX, 2004. [55] G.C. Williams, E.L. Deci, D.H. Ryan, Building health-care partnerships by supporting
[23] D. Feskanich, B.H. Sielaff, K. Chong, I.M. Buzzard, Computerized collection and anal- autonomy: promoting maintainted health behavior change and positive health out-
ysis of dietary intake information, Comput. Methods Prog. Biomed. 30 (1) (1989) comes, in: A.L. Suchman, P. Hinton-Walker, R. Botelho (Eds.), Partnerships in
47–57. Healthcare: Transforming Relational Process, University of Rochester Press, Roches-
[24] R.K. Johnson, P. Driscoll, M.I. Goran, Comparison of multiple-pass 24-hour recall es- ter, NY 1998, pp. 67–87.
timates of energy intake with total energy expenditure determined by the doubly [56] D.M. Garner, P.E. Garfinkel, The eating attitudes test: an index of the symptoms of
labeled water method in young children, J. Am. Diet. Assoc. 96 (11) (1996) anorexia nervosa, Psychol. Med. 9 (2) (1979) 273–279.
1140–1144. [57] R.A. Frank, N.J. van der Klaauw, The contribution of chemosensory factors to individ-
[25] L. Harnack, M. Stevens, N. Van Heel, S. Schakel, J.T. Dwyer, J. Himes, A computer- ual differences in reported food preferences, Appetite 22 (2) (1994) 101–123.
based approach for assessing dietary supplement use in conjunction with dietary [58] J. Brug, L. Lechner, H. De Vries, Psychosocial determinants of fruit and vegetable con-
recalls, J. Food Compos. Anal. 21 (2008) S78–S82. sumption, Appetite 25 (3) (1995) 285–296.
[26] Y.A. Sievert, S.F. Schakel, I.M. Buzzard, Maintenance of a nutrient database for clini- [59] D.M. Garner, M.P. Olmstead, J. Polivy, Development and validation of a multidimen-
cal trials, Control. Clin. Trials 10 (4) (1989) 416–425. sional eating disorder inventory for anorexia-nervosa and bulimia, Int. J. Eat. Disor-
[27] S. Schakel, Y.A. Sievert, I.M. Buzzard, Sources of data for developing and maintaining der 2 (2) (1983) 15–34.
a nutrient database, J. Am. Diet. Assoc. 88 (10) (1988) 1268–1271. [60] A.T. Beck, C.H. Ward, M. Mendelson, J. Mock, J. Erbaugh, An inventory for measuring
[28] S. Schakel, I. Buzzard, S. Gebhardt, Procedures for estimating nutrient values for food depression, Arch. Gen. Psychiatry 4 (1961) 561–571.
composition databases, J. Food Compos. Anal. 10 (1997) 102–114. [61] A.N. Gearhardt, W.R. Corbin, K.D. Brownell, Preliminary validation of the Yale Food
[29] A. Lee, M. Lee, MyFitnessPal, LLC, www.myfitnesspal.com2015 accessed December Addiction Scale, Appetite 52 (2) (2009) 430–436.
1.2015. [62] B.E. Saelens, J.F. Sallis, J.B. Black, D. Chen, Neighborhood-based differences in physi-
[30] National Center for Health Statistics, in: C.f.D.C.a.P. US Department of Health and cal activity: An environment scale evaluation, Am. J. Public Health 93 (9) (2003)
Human Services (Ed.), Centers for Disease Control and Prevention, National Health 1552–1558.
and Nutrition Examination Survey 1999–2000. Data Documentation, Codebook, [63] R.J. Sampson, S.W. Raudenbush, F. Earls, Neighborhoods and violent crime: a multi-
and Frequencies: Blood Pressure (BPX), 2002 (Hyattsville, MD). level study of collective efficacy, Science 277 (5328) (1997) 918–924.
[31] C.C. Allain, L.S. Poon, C.S. Chan, W. Richmond, P.C. Fu, Enzymatic determination of [64] T. Treadwell, N. Lavertue, V.K. Kumar, V. Veeraraghavan, The group cohesion scale-
total serum cholesterol, Clin. Chem. 20 (4) (1974) 470–475. revised: Reliability and validity, Int. J. Action Methods 54 (1) (2001) 3.
[32] U. Nagele, E.O. Hagele, G. Sauer, E. Wiedemann, P. Lehmann, A.W. Wahlefeld, W. [65] P.G. Mezo, The self-control and self-management scale (SCMS): Development of an
Gruber, Reagent for the enzymatic determination of serum total triglycerides with adaptive self-regulatory coping skills instrument, J. Psychopathol. Behav. Assess. 31
improved lipolytic efficiency, J. Clin. Chem. Clin. Biochem. 22 (2) (1984) 165–174. (2008) 83–93.
[33] G.R. Warnick, T. Nguyen, A.A. Albers, Comparison of improved precipitation [66] B. Arnow, J. Kenardy, W.S. Agras, The Emotional Eating Scale: the development of a
methods for quantification of high-density lipoprotein cholesterol, Clin. Chem. 31 measure to assess coping with negative affect by eating, Int. J. Eat. Disord. 18 (1)
(2) (1985) 217–222. (1995) 79–90.
[34] M.P. Caulfield, S. Li, G. Lee, P.J. Blanche, W.A. Salameh, W.H. Benner, R.E. Reitz, R.M. [67] D.R. Bassett Jr., E.T. Howley, D.L. Thompson, G.A. King, S.J. Strath, J.E. McLaughlin, B.B.
Krauss, Direct determination of lipoprotein particle sizes and concentrations by ion Parr, Validity of inspiratory and expiratory methods of measuring gas exchange
mobility analysis, Clin. Chem. 54 (8) (2008) 1307–1316. with a computerized system, J. Appl. Physiol. 91 (1) (1985) 218–224 2001.
[35] C.R. Morgan, A. Lazarow, Immunoassay of insulin: two antibody system: plasma in- [68] S.E. Crouter, A. Antczak, J.R. Hudak, D.M. DellaValle, J.D. Haas, Accuracy and reliabil-
sulin levels in normal, subdiabetic and diabetic rats, Diabetes 12 (1963) 115–126. ity of the ParvoMedics TrueOne 2400 and MedGraphics VO2000 metabolic systems,
[36] A.H. Kadish, D.A. Hall, A new method for the continuous monitoring of blood glucose Eur. J. Appl. Physiol. 98 (2) (2006) 139–151.
by measurement of dissolved oxygen, Clin. Chem. 11 (9) (1965) 869–875. [69] D.J. Macfarlane, H.L. Wu, Inter-unit variability in two ParvoMedics TrueOne 2400 au-
[37] J.M. Mooy, P.A. Grootenhuis, H. de Vries, P.J. Kostense, C. Popp-Snijders, L.M. Bouter, tomated metabolic gas analysis systems, Eur. J. Appl. Physiol. 113 (3) (2013)
R.J. Heine, Intra-individual variation of glucose, specific insulin and proinsulin con- 753–762.
centrations measured by two oral glucose tolerance tests in a general Caucasian [70] C. Draper, L. Wilkins, G. Breton, L. Perusse, R. Debusk, S. Ramakrishnan, D. Krempin,
population: the Hoorn Study, Diabetologia 39 (3) (1996) 298–305. Genetic markers for weight management and methods of use thereof, Google
[38] J.K. Olijhoek, J.D. Banga, P.A. Doevendans, F.L. Visseren, Oral glucose tolerance test or Patents2012.
metabolic syndrome criteria to predict risk in patients with coronary heart disease, [71] L.A. Hindorff, J. MacArthur, J. Morales, H.A. Junkins, P.N. Hall, A.K. Klemm, T. Manolio,
Eur. Heart J. 26 (6) (2005) 623 (author reply 623-4). in: E.M.B.L.-E.B. Institute (Ed.), A Catalog of Published Genome-Wide Association
[39] M. Matsuda, R.A. DeFronzo, Insulin sensitivity indices obtained from oral glucose Studies, 2015.
tolerance testing: comparison with the euglycemic insulin clamp, Diabetes Care [72] M.A. Cornier, W.T. Donahoo, R. Pereira, I. Gurevich, R. Westergren, S. Enerback, P.J.
22 (9) (1999) 1462–1470. Eckel, M.L. Goalstone, J.O. Hill, R.H. Eckel, B. Draznin, Insulin sensitivity determines
[40] Olink Proteomics, Proseek Multiplex Product Site, http://www.olink.com/products/ the effectiveness of dietary macronutrient composition on weight loss in obese
proseek-multiplex/N2016. women, Obes. Res. 13 (4) (2005) 703–709.
[41] F. Baas, H. Bikker, G.J. van Ommen, J.J. de Vijlder, Unusual scarcity of restriction site [73] C.B. Ebbeling, M.M. Leidig, H.A. Feldman, M.M. Lovesky, D.S. Ludwig, Effects of a low-
polymorphism in the human thyroglobulin gene. A linkage study suggesting autoso- glycemic load vs low-fat diet in obese young adults: a randomized trial, JAMA 297
mal dominance of a defective thyroglobulin allele, Hum. Genet. 67 (3) (1984) (19) (2007) 2092–2102.
301–305. [74] C.D. Gardner, L.C. Offringa, J.C. Hartle, K. Kapphahn, R. Cherin, Weight loss on low-fat
[42] S. Gustafson, J.A. Proper, E.J. Bowie, S.S. Sommer, Parameters affecting the yield of vs. low-carbohydrate diets by insulin resistance status among overweight adults
DNA from human blood, Anal. Biochem. 165 (2) (1987) 294–299. and adults with obesity: a randomized pilot trial, Obesity, 24(1), 2016, pp. 79–86
[43] J. Paul, U. Froster-Iskenius, W. Moje, E. Schwinger, Heterozygous female carriers of Silver Spring.
the marker-X-chromosome: IQ estimation and replication status of fra(X)(q), [75] A.G. Pittas, S.K. Das, C.L. Hajduk, J. Golden, E. Saltzman, P.C. Stark, A.S. Greenberg, S.B.
Hum. Genet. 66 (4) (1984) 344–346. Roberts, A low-glycemic load diet facilitates greater weight loss in overweight
[44] UK Biobank Array Design Group, in: UK Biobank (Ed.), UK Biobank Array Content adults with high insulin secretion but not in overweight adults with low insulin se-
Summary Brochure, 2014. cretion in the CALERIE trial, Diabetes Care 28 (12) (2005) 2939–2941.
[45] Affymetrix, Axiom 2.0 Assay Automated Workflow User Guide, Santa Clara, CA, [76] B. Jo, Causal inference in randomized experiments with mediational processes,
2016. Psychol. Methods 13 (4) (2008) 314–336.
[46] Affymetrix, Analysis Guide: Axiom Genotyping Solution Data Analysis Guide Santa [77] H.C. Kraemer, G.T. Wilson, C.G. Fairburn, W.S. Agras, Mediators and moderators of
Clara, CA, 2015. treatment effects in randomized clinical trials, Arch. Gen. Psychiatry 59 (10)
[47] J.E. Ware Jr., C.D. Sherbourne, The MOS 36-item short-form health survey (SF-36). I. (2002) 877–883.
Conceptual framework and item selection, Med. Care 30 (6) (1992) 473–483. [78] R.J.A. Little, D.B. Rubin, Statistical Analysis with Missing Data, Wiley, New York,
[48] D.J. Buysse, C.F. Reynolds 3rd, T.H. Monk, S.R. Berman, D.J. Kupfer, The Pittsburgh 2002.
sleep quality index: a new instrument for psychiatric practice and research, Psychi-
atry Res. 28 (2) (1989) 193–213.