Table 4.

Dietary intake of foods and diabetic retinopathy.

Study, Year Study Design Sample Size (n)	Quality score	Dietary Factor and Its Association with DR	Adjustment/Matched	Statistical Methods Analysis	Key Findings
Fruits, vegetables, and dietary fiber
Alcubierre et al., 2016 Case–control Case:146 Ctrl:148	10	Dietary fiber NS	Sex, age, diabetes duration, energy intake, systolic blood pressure, physical activity, waist circumference, HDL cholesterol, educational level and diabetes treatment	Multivariable logistic regression	Highest fiber intake tertile (T3) vs. lowest fiber intake tertile (T1), OR: 0.76 (0.33–0.76)
Tanaka et al., 2013 Prospective n = 978	10	Fruits, vegetables, and dietary fiber Protective	Sex, age, BMI, HbA1c, diabetes duration, insulin treatment, oral hypoglycaemic agents without insulin treatment, systolic blood pressure, LDL and HDL cholesterol, triglycerides, physical activity alcohol, smoking, total energy intake, proportions of dietary protein, fat, carbohydrate, saturated fatty acids, omega-6 PUFA and omega-3 PUFA and sodium	Multivariate Cox regression	Veg and fruit intake Q4 vs. Q1, HR: 0.59 (0.37–0.92), p < 0.01. Fruit intake Q4 vs. Q1, HR: 0.48(0.32–0.71), p = 0.01. Dietary fiber intake Q4 vs. Q1, HR: 0.63 (0.38–1.03), p = 0.07.
Ganesan et al., 2012 Cross-sectional n = 1261	10	Dietary fiber Protective	Sex, Age, diabetes duration, blood pressure, BMI, Hba1c, serum lipids, smoking, and, socioeconomic status.	Multivariable logistic regression	Low-fiber diet vs. healthy fiber diet for any DR, OR: 1.41 (1.02–1.94), p = 0.039. Low-fiber diet vs. healthy fiber diet for VTDR, OR: 2.24 (1.01–5.02), p = 0.049.
Roy et al., 2010 Prospective n = 469	9	Dietary fiber NS	Total fat, total caloric intake, oleic acid, linoleic acid, fiber, protein, sat fat, cholesterol, and sodium intakes	Multivariable logistic regression	No significant associations with DR (Data not shown)
Cundiff et al., 2005 Prospective n = 1412	8	Dietary fiber Protective	Intake of energy	Spearman correlation	Dietary fiber in g/1000kcal against DR progression rate, r = −0.10 (p = 0.002)
Yan et al., 2019 Prospective n = 8122	6	Fruits, vegetables, and dietary fiber NS	Age, sex, income, educational level, BMI, hypertension, CVD, family history of diabetes, insulin treatment	Cox regression model.	No significant associations with DR (p < 0.05)
Roy et al., 1989 Cross-sectional n = 34	5	Dietary fiber Protective	Diabetes duration	t test	Persons without retinopathy vs. persons with retinopathy, (p < 0.01)
Rice
Kadri et al., 2021 Prospective n = 261	8	Rice Risk	Age, sex, duration, antioxidants, pharmacological treatment, egg, fish, chapathi, rice	Multivariate regression analysis	Rice consumption yes vs. no, OR: 3.19, 95%CI: 1.17–8.69, p = 0.018
Cheese and wholemeal bread
Yan et al., 2019 Prospective n = 8122	6	Cheese and wholemeal bread Protective	Age, sex, income, educational level, BMI, hypertension, CVD, family history of diabetes, insulin treatment	Cox regression model.	Cheese intake highest quartiles vs. lowest HR: 0.58, 95%CI: 0.41–0.83, p = 0.007 and wholemeal bread HR: 0.64, CI: 0.4–0.89, p = 0.04
Fish
Sala-Vila et al., 2016 Prospective n = 3482	9	Oily fish Protective	Age, sex, BMI, intervention group, duration of diabetes, insulin treatment, oral hypoglycemic treatment, smoking, hypertension, systolic blood pressure, physical activity, and adherence to the Mediterranean diet	Cox proportional hazard model	>2 servings a week vs. <2 servings a week, HR: 0.41 (0.23–0.72), p = 0.002
Kadri et al., 2021 Prospective n = 261	8	Fish Protective	Age, sex, duration, antioxidants, pharmacological treatment, egg, fish, chapathi, rice	Multivariate regression analysis	Fish intake, more frequent vs. less frequent, OR: 0.42, 95%CI: 0.18–0.94, p < 0.05
Chua et al., 2018 Cross-sectional n = 357	8	Fish Protective	Age, sex, race, smoking diabetes duration, diabetic treatment, lipid-lowering medication use, systolic blood pressure, HbA1c, triglycerides	Ordered logistic and linear regression models	Per one serving increase in fish intake per week, OR: 0.91, 95%CI: 0.84–0.99, p = 0.038
Yan et al., 2019 Prospective n = 8122	6	Fish NS	Age, sex, income, educational level, BMI, hypertension, CVD, family history of diabetes, Insulin treatment	Cox regression model	No significant associations with DR (p = 0.22)
Alsbirk et al., 2021 Cross-sectional n = 510	6	Fish oil NS	Age, sex, diabetes type, diabetes duration, HbA1c, medication	Logistic regression	No significant association (p > 0.005)
Other types of food
Yan et al., 2019 Prospective n = 8122	6	Processed meat/breakfast cereal NS	Age, sex, income, educational level, BMI, hypertension, CVD, family history of diabetes, insulin treatment	Cox regression model.	No significant associations with DR (p > 0.05)

BMI—Body mass index, CVD—Cardiovascular disease, DR—Diabetic retinopathy, HDL—High-density lipoprotein, HbA1c—Glycated hemoglobin, PUFA—Polyunsaturated fatty acid, VTDR—Vision-threatening diabetic retinopathy.