Study design and population
Between 2020 and 2022, this case-control study included adults over the age of 18 who had recently been diagnosed with NAFLD and healthy controls who had been admitted to Taleghani Hospital in Tehran, Iran, and academic clinics for liver diseases of the Shahid Sadoughi University of Medical Sciences in Yazd, Iran. The control group included 552 people without a history of NAFLD who were recruited from the same hospital, while the case group included 340 consecutive patients with NAFLD who were diagnosed by a gastroenterologist. The patient sample procedure was evaluated by two dieticians. The following criteria have been used to diagnose NAFLD28,29,30: chronic elevation of liver enzymes (liver enzymes > 19 U/L for women and > 30 U/L for men), liver ultrasound compatible with NAFLD, presence of NAFLD grade II, III based on liver biopsy, abstinence from consumption of alcohol and the elimination of other possible causes of liver disease. A gastroenterologist confirmed the diagnosis of NAFLD when the case group was referred to our facilities for evaluation by Fibroscan28, which indicated a controlled attenuation parameter score greater than 237 and a fibrosis score greater than 7. In addition, patients from other outpatient clinics in the same hospital were recruited for the control group, including dermatology, ophthalmology, and otolaryngology, who had no history of NAFLD. Healthy controls had no history of chronic or inflammatory disease, had eaten consistently for the past six months, and had been physically active (such as diabetes, gastrointestinal or cardiovascular disorders, cancer, etc.). Laboratory tests and liver ultrasound, which confirmed the absence of fatty liver disease at any stage, served as the basis for the inclusion criteria for the control group. Furthermore, the matching of persons in the case and control group (1:1) was performed on the basis of the variables of age (± 3 years) and body mass index (BMI) (± 1 kg/m2). The following conditions precluded patients from participating: long-term dietary changes, weight loss, a specific disease, a history of liver or kidney disease (such as non-alcoholic steatohepatitis (NASH), alcoholic fatty liver disease, Wilson’s disease, cirrhosis , autoimmune disease liver disease, hemochromatosis, viral infections), diabetes, cancer, thyroid disorders and autoimmune diseases. By completing demographic, economic, and social questionnaires, information was collected on age, education level, employment status, medical history, smoking history, use of certain medications (other than typical NAFLD medications), and dietary history over the past six months. Participants’ physical activity levels were assessed using General Practice Physical Activity Questionnaires (GPPAQ). The GPPAQ is a short survey that measures your current level of physical activity30 and occupation and is classified into active, moderately active, moderately inactive or inactive categories (see Fig. 1). In this study, nutritionists were the interviewers. As a result, each patient answered each survey item honestly. Furthermore, informed consent was obtained from all subjects and all methods were performed in accordance with relevant guidelines and regulations. This study was approved by Shahid Beheshti University of Medical Sciences, Tehran, Iran, and Shahid Sadoughi University of Medical Sciences, Yazd, Iran.
Sample size calculation
The minimum sample size required for the current work was calculated based on the assumption of 1.5-fold NAFLD probabilities by OBS. Therefore, considering the type I error of 5%, the study power of 90%, and the ratio of controls to cases of approximately 1.5, the minimum required sample size was calculated.
The researchers carried out an anthropometric analysis. Weight was recorded to the nearest 100 g using a standard SECA 700 digital scale (SECA, Hamburg, Germany) with little clothing and no shoes. The height requirements of the patient were assessed using a Seca portable height gauge with 0.1 cm accuracy. Furthermore, a Seca waist measurer was used to establish the waist circumference (WC) in the central area between the corona iliac and the last gear. Also, hip circumference was calculated in centimeters by placing the same tape measure parallel to the floor at the fullest part of the buttocks. Weight (kg)/Height2(m) was used to determine body mass index (BMI) after weight and height were measured by the above method. All anthropometric measurements were conducted by the researcher in order to reduce the observational variance.
A validated semi-quantitative food frequency questionnaire (FFQ) was used with 168 food items to collect data on food consumption during the previous year31. The FFQ consisted of a list of typical Iranian foods and their portions. Self-reports on the FFQ determined the average serving size and frequency of consumption for each food. The frequency of consumption of each food was as follows: never, 2-3 times a month, once a week, 2-4 times a week, 5-6 times a week and every day. Using standard Iranian household measurements, serving quantities have been given in grams32. Using the United States Department of Agriculture’s (USDA) National Nutrition Database, the daily intakes of nutrients for each individual were calculated33. The nutritional and calorie content of foods was analyzed using a customized version of Nutritionist 4 (First DatabankInc., Hearst Corp., San Bruno, CA, USA) for Iranian meals.
Calculation of Oxidative Balance (OBS) scores
In the present study, we used the method described by Goodman et al.34 to calculate the OBS of each participant. According to this method, a total of 13 dietary and non-dietary pro- and antioxidant components were selected, based on a priori knowledge of their association with oxidative stress. The components were divided into four groups: (1) dietary antioxidants (selenium, fiber, β-carotene, vitamin D, vitamin C, vitamin E, and folate); (2) dietary prooxidants (iron, saturated fatty acids (SFA) and polyunsaturated fatty acids (PUFA)); (3) non-dietary antioxidant (physical activity); and (4) nonfood prooxidants (smoking and obesity). Dietary factors were classified into quintiles. For dietary antioxidants and physical activity, the first to fifth quintiles were scored from 1 to 5. An inverse score was used for dietary prooxidants. For obesity, we assigned 1: BMI ≥ 30 kg/m2 and WC ≥ 102 cm in males and ≥ 88 cm in females, 3: o BMI ≥ 30 kg/m2 or WC ≥ 102 cm in males or ≥ 88 cm in females, and 5: BMI < 30 kg/m2 and WC < 102 in males or < 88 cm in females. For smoking, 1 was assigned: current smoker, 3: former smoker and 5: never smoker. The four component scores were then summed to calculate OBS for each participant. A higher OBS score indicates greater adherence to this diet-derived score, and a lower score indicates less adherence to this score. The minimum and maximum possible scores are 5 and 65, respectively.
The laboratory technician took 10 ml of venous blood from the individuals at the beginning and at the end of the test, after 10-12 hours of fasting. After clotting in the environment, the serum was rapidly separated by centrifugation and stored at -70°C until transported to the laboratory for testing. The concentrations of triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and fasting blood glucose (FBS) were determined using a kit from Pars Azmon Company (Tehran, Iran) using an enzymatic colorimetric technique. Enzymatic photometry was used with the Pars test kit (Parsazmun, Tehran, Iran) to determine the total cholesterol content. Using the Friedewald formula35, low-density lipoprotein cholesterol (LDL-C) concentration was also determined. LDL-C = TC (mg/dL) − HDL-C (mg/dL) − TG (mg/dL)/5. Based on a self-test (BT-3000), the enzymes alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were measured using commercially available enzyme reagents (Pars Azmoon, Tehran, Iran).
Statistical Package for Social Science v.21 (SPSS Inc., Chicago, IL, USA) software was used to conduct the statistical analysis. Data normality was examined using the Kolmogorov-Smirnov test and histogram plots. Baseline characteristics and dietary intakes were recorded as mean standard deviation (SD) for quantitative variables and for qualitative variables such as number and percentage. We used independent sample t-tests (or one-way ANOVA) and chi-square tests to compare data between two groups (or between OBS quartiles) for continuous and categorical variables, respectively. Nutrients were adjusted for total energy intake (kcal) using the residual method. Logistic regression was used to examine the relationship between OBS and NAFLD risk. Analyzes were adjusted for potential confounders such as age, gender, hip circumference, education, drug use, disease history, FBS, ALT, AST, lipid profiles, and energy intake. The odds ratio (OR) of NAFLD between score quartiles was estimated with a 95% confidence interval (CI). P-values <0.05 were considered statistically significant.
Ethics approval and consent to participate
This study was approved by the Research Council and Ethics Committee of Shahid Beheshti University of Medical Sciences, Tehran, Iran.