Table 1.
Ultra-processed food consumption and NAFLD (n = 6).
| Author (year, location) | Source of data (mean/median follow-up time) | Population age-range and/or mean ± SD (%women) | Dietary questionnaire UPF assessment method | Adjustment | Main results |
|---|---|---|---|---|---|
| Intervention | |||||
| Hall KD. (2020, USA)53 | Inpatient, crossover RCT | n = 20 31.2±1.6 y (50%) n = 13 sub-sample with liver MRS |
UPF diet vs. unprocessed diet for 2 weeks, followed by the alternate diet for the next 2 weeks. All meals provided at an amount of 1.6∗EER, to consume ad libitum during 60 min. NOVA classification |
Randomization | Baseline liver fat (by MRS) was 1.2±0.1%. Liver fat was not significantly changed after the unprocessed diet (0.95±0.1%, p = 0.24) or the UPF diet (1.1±0.2%; p = 0.74) |
| Prospective | |||||
| Zhang S. (2022, China)48 | The Tianjin Chronic Low-grade Systemic Inflammation and Health (TCLSIH) (4.2 y) | n = 16,168 18-90 y 38.3±0.2 y (57.4%) | Quantitative 81/100-items FFQ (in the past month, previously validated) NOVA classification (nutrient density, g/1,000 kCal per day) |
Age, gender, BMI, smoking status, alcohol consumption, educational level, occupation, household income, physical activity, family history of diseases, depressive symptoms, energy intake, healthy diet score, hypertension, diabetes, and hyperlipidemia | Participants with the highest UPF consumption (4th quartile vs. 1st quartile) had 18% relatively higher risk of developing NAFLD (by AUS) (HR 1.18, 95% CI 1.07-1.30; p for trend <0.0001). HR (95% CI) for one standard deviation increment in UPF consumption, equivalent to 62.7 g/1,000kCal per day, was 1.06 (1.03-1.09) |
| Konieczna J. (2022, Spain)49 | Sub-sample from the Spanish Prevention with Mediterranean Diet (PREDIMED-Plus trial) Prospective analysis nested in RCT (1 y, first year) | n = 5,867 55-75 y 65.0±4.9 y (47.8%) | Semi-quantitative 143-items FFQ at baseline, 6- and 12-month follow-up NOVA classification (% of total food weight) (UPF coded as continuous and sex-specific quintiles) |
Age, gender, study arm, educational level, smoking status, height, physical activity, sedentary behavior, alcohol consumption, and follow-up time. (also sensitivity analysis for dietary factors, obesity measures, and related diseases) | A 10% increment in UPF consumption was associated with greater levels of NAFLD-related biomarkers; FLI score (β = 1.60, 95% CI 1.24-1.96) and HSI score (β = 0.43, 95% CI 0.29-0.57). FLI- estimates for Q5 vs. Q1; β = 3.73, 95% CI 3.10-4.35. HSI- estimates for Q5 vs. Q1; β = 0.93, 95% CI 0.67-1.18; p for trend <0.001) |
| Cross-sectional | |||||
| Liu Z. (2022, USA)50 | National Health and Nutrition Examination Survey (NHANES) | n = 6,545 >20 y mean 49.3 y (0.34ySE) (53.5%) | 24-hour dietary recall NOVA classification (% of total food weight) |
Age, gender, race/ethnicity, educational level, family income to poverty ratio, marital status, smoking status, BMI, serum ALT, fasting triglycerides, total cholesterol, and uric acid | Higher UPF consumption (4th quartile of >68.3% vs. 1st quartile of <41.6%) was associated with higher odds for probable NAFLD, as evaluated by FLI ≥30 (OR 1.83, 95% CI 1.33–2.53). A 10% increment in UPF consumption was associated with 15% higher odds for probable NAFLD (OR 1.15, 95% CI 1.09-1.22; p for trend <0.001) |
| Friden M. (2022, Sweden)52 | Prospective investigation of Obesity, Energy, and Metabolism (POEM) | n = 286 50 y (all participants) (53%) | Semi-quantitative 140-items FFQ (a shorter version of a previously validated FFQ) NOVA classification (% of total kCal) |
Gender, BMI, educational level, physical activity, smoking status, alcohol consumption, and dietary factors (protein, fiber, total sugar, saturated and polyunsaturated fat intake) | Intake of UPF was positively associated with liver fat (by MRI) in crude linear regression models (β = 0.02, p = 0.006). However, the association was attenuated after further adjustments. A 10% increment in UPF consumption was not associated with the prevalence of NAFLD (OR 1.32, 95% CI 0.84–2.09) |
| Ivancovsky-Wajcman D. (2021, Israel)51 | Hepatic screening study | n = 789 40-70 y 58.8±6.6 y (47.4%) |
Semi-quantitative 117-items FFQ NOVA classification (% of total kCal) |
Age, gender, BMI, saturated fat intake, protein intake, physical activity, coffee consumption, and fiber intake | UPF consumption (above median of 28% vs. under median) had no association with NAFLD (by AUS), NASH, and significant fibrosis biomarkers (FibroMax, BioPredictive). Higher UPF consumption among subjects with NAFLD was associated with higher odds for NASH (OR 1.89, 95% CI 1.07-3.38). Higher UPF consumption among ever smokers in the entire sample and those with NAFLD was associated with significant fibrosis (OR 1.89, 95% CI 1.03-3.45 and OR 2.85, 95% CI 1.14-7.14, respectively) |
ALT, alanine aminotransferase; AUS, abdominal ultrasound; FFQ, food frequency questionnaire; FLI, fatty liver index; HSI, hepatitis steatosis index; MRS, magnetic resonance spectroscopy; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; OR, odds ratio; RCT, randomized-controlled trial; UPF, ultra-processed food; WC, waist circumference.