Discovering Determinants of Food Intake by Application of Artificial Intelligence to Complex, High-Dimensional Data

Discovering Determinants of Food Intake by Application of Machine Learning to Complex, High-Dimensional Data

Background:

Many people in the United States are overweight or obese. Researchers want to learn why some people can overeat and not gain weight, whereas others who do not overeat still gain weight.

Objective:

To study factors related to food intake that can lead to weight gain over time.

Eligibility:

Healthy adults aged 18 to 60 years.

Design:

Participants will have 6 to 8 clinic visits over 2 years. The first 3 or 4 study visits will be 1 week apart.

Procedures during visits may include the following:

Collection of blood, hair, urine, and stool samples.

Measurement of the waist, neck, thighs, and other parts of the body.

Dual energy x-ray absorption (DXA) scan: Participants will lie still on a padded table while they are scanned to measure body fat.

Physical activity monitor: Participants will wear a monitor on the wrist for 2 weeks.

Cognitive tests: Participants will perform tasks to measure attention, memory, and brain function.

Continuous glucose monitor. Participants will wear a device that measures their blood glucose for 1 week.

Mixed meal test and stomach emptying test. Participants will drink a breakfast shake and swallow a dose of acetaminophen. Blood will be drawn over the next 4 hours.

Resting metabolic rate: Participants will wear a clear hood over their head while they rest for 20 minutes. The hood will measure the gases they breathe.

Breakfast and lunch test. Participants will eat a standard breakfast. They will be allowed to select from foods and eat as much as they like at lunch. They will be asked how hungry or full they are.

Questionnaires. Participants will answer questions about their health, sleep, physical activity, and eating.

Study Description:

The study will be an outpatient study with 6-7 total visits, 3-4 completed at baseline and 3 follow-up visits, occurring over 2 years. We will collect broad, multiscale data (e.g., clinical, genetic, microbiome, energy expenditure, physical activity, psychosocial). Then, we will apply artificial intelligence algorithms with the goal of identifying novel determinants of food intake.

Objectives:

Primary Objective:

To identify novel determinants of objectively measured ad libitum energy intake using broad, multiscale data.

Secondary Objectives:

1. To investigate whether determinants of ad libitum energy intake predict changes in weight, adiposity, and markers of metabolic health.
2. To determine associations of metabolic measures with psychosocial, hormonal, or other metabolic markers
3. To identify novel determinants of ad libitum self-reported energy and nutrient intake

Endpoints:

Primary Endpoint: Energy intake during ad libitum meal test, including total energy and macronutrient specific energy intake (e.g., carbohydrate, fat, protein) measured during the initial baseline period (visit 3). The potential determinants (broad multiscale data) will be collected during visits 1, 2, and 3.

Secondary Endpoints:

1. Measures at 0.5, 1, and 2 years of body weight, BMI, fat mass, body fat percentage, HbA1c, and fasting glucose, lipids, insulin, and c-peptide.
2. Psychosocial measures (visits 1-6), metabolic measures (e.g., fasting glucose, HbA1c, lipids, insulin, and c-peptide), resting metabolic rate, total energy expenditure, and physical activity
3. Energy and nutrient intake from ASA-24. The potential determinants (broad multiscale data) will be collected during visits 1, 2, and 3.

• INCLUSION CRITERIA:

To be eligible to participate in this study, an individual must meet all the following criteria:

1. Stated willingness to comply with all study procedures and availability for the duration of the study
2. Age 18 to 60
3. In good general health as evidenced by medical history and physical exam

EXCLUSION CRITERIA:

An individual who meets any of the following criteria will be excluded from participation in this study:

1. Weight >=204 kg (>=450 pounds, maximum weight of the iDXA machine as per manufacturer s manual), or weight <45.4 kg (<100 pounds, minimum weight allowed based on the NIH guidelines of blood drawing for research purposes)
2. Unstable weight (+/-5%) within past 6 months as determined by volunteer self-report.
3. Current use of medications, dietary supplements, or alternative therapies known to alter energy metabolism
4. Inability to consume provided food based on food allergy or intolerance or other food restrictions (e.g., kosher, vegan, or vegetarian)
5. Current pregnancy, or recent pregnancy within past 6 months or currently lactating
6. History or clinical manifestation of:

6i. Diabetes, including fasting glucose >= 126 mg/dl or HbA1c >= 6.5%, or history

6ii. Surgery for treatment of obesity

6iii. Endocrine disorders, such as Cushing s disease, pituitary disorders, or hypo- and hyperthyroidism (TSH <0.1 or >= 10 uIU/mL)

6iv. Pulmonary disorders including chronic obstructive pulmonary disease or other lung disease that which would limit ability to follow the protocol

6v. History of coronary artery disease, heart failure, arrhythmias, and peripheral artery disease that which would limit ability to follow the protocol

6vi. Liver disease, including cirrhosis, active hepatitis B or C based on history, and AST or ALT >= 3x normal

6vii. Gastrointestinal disease including Crohn's disease, ulcerative colitis, celiac disease, or other malabsorptive disorders by history

6viii. Hematologic disorders including significant anemia (male hemoglobin < 13.0 g/dL or female hemoglobin < 11.0 g/dL)

6ix. Renal disease including abnormal kidney function (eGFR <60 mL/min/1.73m^2)

6x. Central nervous system disease, including cerebrovascular accidents, dementia, and neurodegenerative disorders by history

6xi. Cancer requiring treatment in the past 5 years, except for nonmelanoma skin cancers or cancers that have clearly been cured

6xii. Infectious disease such as active tuberculosis, HIV (by self-report), chronic coccidiomycosis or other chronic infections that might influence appetite

6xiii. Diagnosis of binge eating disorder, anorexia, or major psychiatric disorders including depression, schizophrenia, and psychosis

6xiv. Menopausal transition (late) to early postmenopause (e.g., >= 2 skipped cycles and an interval of amenorrhea >= 60 days and <12 months)

7. Alcohol abuse as defined by >= 8-point score on the Alcohol Use Disorders Identification Test (AUDIT) questionnaire interview

8. Inability to provide informed consent

9. Any disorder, unwillingness, or inability not specifically mentioned above may serve as criteria for exclusion at the discretion of the investigators, such as those that jeopardize the safety of the participant or others, or would interfere with completion of study procedures