The Personalized Nutrition Study (POINTS)

A person's genetic code is believed to affect how much weight he/she will lose during diets that vary in carbohydrate and dietary fat content. 'Carbohydrate responders' are hypothesized to lose more weight on diets that are high in carbohydrates, as compared to high in fats. 'Fat responders' are hypothesized to lose more weight on diets that are high in dietary fat, as compared to high in carbohydrates. The purpose of the proposed study is to test these hypotheses in a randomized controlled trial.

Obesity and its comorbidities are major public health challenges. To combat the obesity pandemic, many weight-loss strategies have been studied, often emphasizing either high carbohydrate (low fat) diets or high fat (low carbohydrate) diets. Mean weight loss differences between high-carbohydrate and high-fat diets that induce equal caloric deficits have been reported to be small; however, the individual weight loss response varies substantially within diet groups, suggesting that different individuals react differently to high-carbohydrate or high-fat diets. This assumption is supported by retrospective data showing that participants with carbohydrate-responsive polymorphisms lost 2-3 times more weight when assigned to a high-carbohydrate diet compared to a high-fat diet, and vice versa for those with dietary fat-responsive polymorphisms. Conversely, a recent randomized clinical trial aimed to determine the effect of a healthy high-fat diet (high in unsaturated fats) vs. a healthy high-carbohydrate diet (high in whole-grain foods) on 12-month weight change but did not find significant differences between the two groups and failed to find the hypothesized association between genotype patterns and weight loss induced by diets that varied in fat and carbohydrate content. However, an important caveat of their approach is that the single nucleotide polymorphisms selected by the investigators had not been previously associated with obesity or with dietary responses, which may explain their lack of predictive value in identifying differences in inter-individual responses. In addition, the fat composition of the diets was relatively high in both high- and low-fat groups. The inconsistent findings in the literature indicate a need for further research to determine if genetic factors affect weight loss when exposed to diets that vary in carbohydrates and dietary fats. The purpose of this randomized controlled parallel arm trial is to test the following hypotheses. Hypothesis 1 will test if participants assigned to the diet that corresponds to their genotype lose more weight than those assigned to a diet inconsistent with their genotype. Hypothesis 2 will analyze the fat responders and carbohydrate responders separately. Hypothesis 2a: Fat responders will lose more weight on the high-fat diet vs. the high-carbohydrate diet. Hypothesis 2b: Carbohydrate responders will lose more weight on the high-carbohydrate diet vs. the high-fat diet. Carbohydrate responders and fat responders will be randomized to one of the following two diets: A high-quality high-carbohydrate diet that is rich in whole-grain foods, or A high-quality high-fat diet that is rich in unsaturated fats and oils

Randomized controlled parallel arm trial with 4 groups over 12 weeks Fat responders receiving a high-fat diet (n=52) Fat responders receiving a high-carbohydrate diet (n=52) Carbohydrate responders receiving a high-fat diet (n=25) Carbohydrate responders receiving a high-carbohydrate diet (n=25) The total number per group is an estimate. We will not close cells to enroll this exact number per group, and the total number of people enrolled will not exceed 154.

Outcome assessors will be blind to diet assignment and genotype pattern. Interventionists will be blind to genotype pattern, but not diet type. To enhance external validity, participants will be told if they are carbohydrate or fat responders.

The high-fat diet will consist of ~40% energy from fat and ~45% from carbohydrates. Protein will be 15% of energy. All participants will be assigned an energy intake target that will result in a daily deficit of ~750 kcal, though no energy intake targets below 1,100 kcal/d (women) and 1,300 kcal/day (men) will be prescribed

The high-carbohydrate diet will consist of ~20% of energy from fat and ~65% from carbohydrates. Protein will be 15% of energy. All participants will be assigned an energy intake target that will result in a daily deficit of ~750 kcal, though no energy intake targets below 1,100 kcal/d (women) and 1,300 kcal/day (men) will be prescribed

Food cravings are assessed via the total score of the 33-item Food Craving Inventory (FCI). The FCI is scaled in a frequency format, assessing the frequency with which an individual experiences a craving for a particular food. All items are scored in the following manner: Never = 1, Rarely = 2, Sometimes = 3, Often = 4, & Always = 5. Change in food cravings is calculated as FCI total score at 12 weeks minus FCI total score at baseline.

Preference for high- versus low-fat foods is measured with the Food Preference Questionnaire (FPQ) and a fat preference index is calculated as the mean rating for high-fat foods divided by the mean rating for low-fat foods. Values greater than 1.0 reflect a higher fat preference, and values less than 1.0 reflect a lower fat preference. Items in the FPQ are rated on a 9-point Likert scale with the following anchors: 1 = dislike extremely; 5 = neutral, neither like nor dislike; 9 = like extremely. Change in Fat Preference Index is calculated as Fat Preference Index at 12 weeks minus Fat Preference Index at baseline.

Score on the Disinhibition subscale of the 36-item Three-Factor Eating Questionnaire. Items are scored as 'true' or 'false'. Change in Disinhibition is calculated as the score in the Disinhibition subscale at 12 weeks minus the score in the Disinhibition subscale at baseline.

Score on the Hunger subscale of the 36-item Three-Factor Eating Questionnaire. Items are scored as 'true' or 'false'. Change in Hunger is calculated as the score in the Hunger subscale at 12 weeks minus the score in the Hunger subscale at baseline.

Score on the Cognitive Restraint subscale of the 36-item Three-Factor Eating Questionnaire. Items are scored as 'true' or 'false'. Change in Cognitive Restraint is calculated as the score in the Cognitive Restraint subscale at 12 weeks minus the score in the Cognitive Restraint subscale at baseline

Inclusion Criteria: BMI ≥ 27.0 kg/m2 to < 47.5 kg/m2 Completed genealogy test and access to the raw data Fat responder or Carbohydrate responder, as determined by genetic risk score It is estimated that approximately 1/3 of people are fat responders, 1/3 are carbohydrate responders, and 1/3 are neither or will respond to either diet. Only carbohydrate and fat responders are eligible. Exclusion Criteria: Current smoker or has smoked in the previous year For females, pregnant or planned pregnancy during the study duration, or breast-feeding, based on self-report Conditions, diseases, or medications that affect body weight or metabolism (e.g., certain antipsychotic medications; type 2 diabetes mellitus; heart failure; cancer, excluding certain melanomas; etc.) Has gained or lost more than 10 pounds in the last 3 months Currently diagnosed with an eating disorder, major depression, or other condition that, in the judgment of the investigators, could affect the risk to the participant or study completion

Access to individual de-identified data is possible via Pennington Biomedical Research Center's Data Sharing policy.