The purpose of the study is to determine the effect of dietary sugar reduction in obese children and examine whether there are differential effects based on genotype of a single amino acid substitution in the PNPLA3 gene that is highly prevalent in Hispanics and associated with significantly elevated liver fat.

This dietary intervention aims at developing a more personalized and targeted treatment for NAFLD in Hispanic children and adolescents who are GG for the PNPLA3 variant. The investigators previous publications have shown that this particular demographic has a greater than 2-fold higher liver fat compared to GC and CC individuals. They have also demonstrated a significant gene*dietary sugar interaction with a significant association between liver fat and dietary sugar intake in GG subjects with no such association in GC or CC individuals. These studies suggests that different dietary strategies may have differential effects on reducing liver fat, depending on PNPLA3 genotype. To confirm this, the investigators will complete a clinical trial in 120 overweight and obese Hispanic children (12 - 18 years) with clinically verified NAFLD who will be randomized to one of two 12-week interventions:

Group 1 (standard of care control group): Dietary intervention focused on healthy eating (n=60; 30GG + 30GC/CC)

Group 2 (standard of care + sugar reduction): Dietary intervention based on healthy eating and sugar reduction focused on reduction of sugary beverages and added sugar towards a goal of 10% of daily calories (n=60; 30GG + 30GC/CC)

The following outcomes will be measured before and after intervention: Total liver fat fraction, and visceral and subcutaneous abdominal adipose tissue volume by magnetic resonance imaging (MRI); liver fibrosis by magnetic resonance elastography (MRE); total body fat by DEXA; liver enzymes, fasting insulin, glucose, lipids, free fatty acids and inflammatory markers, gut microbiome, and insulin and glucose response to an oral glucose challenge. The investigators hypothesize that liver fat fraction, liver fibrosis, and metabolic outcomes, such as fasting and 2h-glucose and insulin, and inflammatory biomarkers, will show significantly greater improvements with sugar reduction relative to control. In addition, the investigators also hypothesize a treatment*genotype interaction whereby the reduction in liver fat will be significantly greater in GG relative to CC/CG subjects.

Inclusion Criteria:

• Ethnicity: This study is limited to Hispanics because of their higher risk of NAFLD, higher frequency (~50%) of the at-risk PNPLA3 allele (G), and because no prior studies have targeted improvement in liver fat and NAFLD in this high-risk population. As with all of our ongoing studies, Hispanic ethnicity will be based on self-identity for the participants as well as their parents and grandparents.
• Gender: Males and females will be eligible for this study.
• Age: Children 12 to 18 years of age will be eligible. In our experience, children younger than around 10 years of age and greater than 18 years would require different intervention/counseling strategies. Therefore, we can develop a more consistent "age-neutral" approach if we limit the age range to 12-18 years.
• Weight status: Subjects will be eligible if they are obese, defined by a BMI > 95th percentile for age and gender.

Exclusion Criteria:

• Diabetes: Presence of type 1 or 2 diabetes, as defined by fasting plasma glucose > 126 mg/dl, or positive for diabetes related antibodies including ICA512 and GAD. Participants testing positive for diabetes will be referred for treatment. Subjects with pre-diabetes will be eligible for the study
• Pregnancy: Women who self-report as pregnant or obtain a positive pregnancy test result during Visit 1 will be excluded. Furthermore, should a woman become pregnant during the course of the intervention, she will be withdrawn from the study at that time and asked to no longer participate. This is in order to protect the mother and child from radiation involved with the DEXA scan and potential complications associated with a low-sugar diet.
• Medication: Taking any medications known to influence liver function, insulin action or lipid levels
• Self-prescribed dietary supplements: Taking any non-prescription supplements that could potentially affect liver function and liver fat (eg vitamin E or fish oils)
• Other metabolic diseases: Diagnosis of other syndromes or diseases that may influence insulin action and secretion (e.g., maturity-onset diabetes of the young, lipoatrophic diabetes, cystic fibrosis), or body composition and fat distribution (e.g. Cushing syndrome, Down syndrome, lipodystrophy)
• Other medical condition: Previously diagnosed with any major illness since birth (e.g. severe intrauterine growth retardation, chronic birth asphyxia, cancer)
• Familial hyperlipidemia: Patients with a family history of hyperlipidemia will be excluded, due to the particular genetic background of this disease, which may bias our results. Familial hyperlipidemia will be defined as LDL/cholesterol > 160 mg/dL and/or triglycerides > 200 mg/dL in both the participant AND at least one family member (first degree: parents or siblings).
• Smoking or drinking: Self-reported current smoking participants (more than 1 cigarette in the past week) will be excluded due to the potential effect of smoking on weight control and inflammatory status. Consumption of alcohol on a regular basis (40g/day alcohol per day determined by questionnaire) will also be excluded due to its important role in liver disease. Use of recreational drugs will also be an exclusion criteria, due to the potent effect of cannabinoids receptors on weight status and food intake.
• Participation in a weight-loss or exercise program: participants who have participated to a weight-loss or exercise program in the past three months will be excluded due to its potential effect on weight status.