Table 5.
Dietary intake of beverages, dietary patterns, 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 |
|---|---|---|---|---|---|
| Coffee | |||||
| Lee at al, 2022 Cross-sectional n = 1350 |
9 | Coffee Protective |
Age, sex, education, income, BMI, energy intake, hypertension, dyslipidemia, diabetes duration, HbA1c, smoking, alcohol, physical activity | Multivariable logistic regression models | Consumption ≥ 2 cups coffee/day vs. none for DR (OR: 0.53, 95%CI: 0.28–0.99, p for trend = 0.025) and VTDR (OR: 0.30, 95%CI: 0.10–0.91, p for trend = 0.005) |
| Kumari et al., 2014 Cross-sectional n = 353 |
9 | Coffee NS |
Sex, age, HbA1c, smoking, BMI, creatinine, education level, diabetes duration, family history of diabetes, hypertension, stroke, ischemic heart disease, dyslipidemia, and cancer | Multivariable logistic regression | Coffee drinker vs. never/rarely, OR: 1.36 (0.69–2.69) |
| Tea | |||||
| Ma et al., 2014 Case–control Case:100 Ctrl:100 |
8 | Green Tea Protective |
Diabetes duration, insulin treatment, family history of diabetes, fasting blood glucose, education, BMI, systolic blood pressure, smoking, alcohol, physical and, activity | Multivariable logistic regression | Regular Chinese green tea drinker vs. non-regular Chinese green tea drinker, OR: 0.48, CI: 0.24–0.97, p = 0.04 |
| Xu et al., 2020 Cross-sectional n = 5,281 |
7 | Tea Protective |
Age, sex, individual monthly income, fasting blood glucose, systolic blood pressure, occupation, educational level, smoking, alcohol | Multivariate logistic regression analyses | Tea consumers vs. non-tea consumers, OR: 0:29, 95%CI: 0.09–0.97, p = 0.04 |
| Milk | |||||
| Yan et al., 2019 Prospective n = 8122 |
6 | Milk NS |
Age, sex, income, educational level, BMI, hypertension, CVD, family history of diabetes, insulin treatment |
Cox regression model | No significant associations with DR (p = 0.74) |
| Diet soda | |||||
| Fenwick et al., 2018 Cross-sectional n = 609 |
8 | Diet soft drink Risk |
Age, sex, HbA1c, diabetes duration, insulin use, presence of at least one other diabetes complication, diabetes type, BMI, education antihypertensive medication, hyperlipidaemia, presence of comorbidity, smoking, alcohol energy intake, regular soft drink consumption |
Multinomial logistic regression | High-consumption (>4 cans [1.5 liters]/week) vs. no consumption for proliferative DR (OR = 2.62, 95%CI = 1.14–6.06, p = 0.024) |
| Mirghani et al., 2021 Cross-sectional n = 200 |
6 | Diet sugar-free carbonated soda beverage Risk |
NIL | Multiple regression analysis | Diet soda was associated with DR (p = 0.043) |
| Alcohol | |||||
| Fenwick et al., 2015 Cross-sectional n = 395 |
10 | Alcohol Protective |
Sex, gender, poorly controlled diabetes, diabetes duration, BMI, smoking, systolic blood pressure, insulin therapy, and presence of at least one other diabetic complication | Multivariable logistic regression |
Moderate vs. abstainers, OR: 0.47 (0.26–0.85), p = 0.013; moderate white wine vs. abstainers, OR: 0.48 (0.25–0.91), p = 0.024; moderate fortified wine vs. abstainers, OR: 0.15 (0.04–0.62), p = 0.009 |
| Beulens et al., 2008 Cross-sectional n = 1857 |
10 | Alcohol Protective |
Sex, Age, smoking, center, smoking, diabetes duration, physical activity, presence of CVD, systolic blood pressure, BMI, and HbA1C | Multivariable logistic regression |
Moderate vs. abstainers, OR: 0.60 (0.37–0.99), p = 0.023 |
| Lee et al., 2010 Prospective n = 1239 |
9 | Alcohol NS |
Sex, age, ethnicity, smoking, HbA1c, BMI, systolic blood pressure, and duration diabetes | Multivariable logistic regression | Moderate vs. none, OR: 1.08 (0.70–1.67) Heavy vs. none, OR: 1.07 (0.54–2.13), p = 0.8 |
| Moss et al., 1993 Prospective Younger: 439 Older: 478 |
9 | Alcohol NS |
Sex, age, HbA1c | Multivariable logistic regression |
Younger-onset diabetics per 1oz/day increase in alcohol consumption on DR incidence, OR: 2.09 (0.04–1.07);per 1oz/day increase in alcohol consumption on DR progression, OR: 1.25 (0.75–2.08). Older-onset diabetics per 1oz/day increase in alcohol consumption on DR incidence, OR: 0.75 (0.4–1.42); per 1oz/day increase in alcohol consumption on DR progression, OR: 0.73 (0.4–1.20) |
| Moss et al., 1992 Cross-sectional Younger: 891 Older: 987 |
9 | Alcohol Protective |
Diabetes duration, age, HbA1c, diastolic blood pressure, insulin therapy | Multivariable logistic regression |
Younger-onset diabetes population per 1oz/day increase in alcohol consumption for PDR, OR: 0.49, (0.27–0.92) Older-onset: no significant associations |
| Gupta et al., 2020 Prospective n = 656 |
8 | Alcohol Protective |
Age, sex, BMI, smoking, systolic blood pressure, income, HbA1c, diabetes duration, hyperlipidaemia, CKD, antidiabetic medication | Multivariable analyses | Alcohol consumption vs. non-drinkers, OR: 0.36 (0.13 to 0.98) p = 0.045; occasional drinker (≤2 days/week) vs. non-drinkers, OR:0.17, (0.04–0.69), p = 0.013) |
| Thapa et al., 2018 Cross-sectional n = 1860 |
8 | Alcohol Risk |
NIL | Multivariable logistic regression analysis | Alcohol consumption yes vs. no for DR (OR:4.3, 95%CI: 1.6–11.3, p = 0.004) and vision-threatening DR (OR: 8.6, 95%CI: 1.7–47.2, p = 0.010) |
| Harjutsalo et al., 2013 Cross-sectional n = 3608 |
8 | Alcohol Protective |
Sex, diabetes duration, age at onset of diabetes, triglycerides, HbA1C, HDL cholesterol, social class, BMI, smoking status, lipid-lowering agents and hypertension |
Multivariable logistic regression | Abstainers vs. light consumers, OR: 1.42 (1.11–1.82), p < 0.05; former users vs. light consumers, OR: 1.73 (1.07–2.79), p < 0.05 |
| Cundiff et al., 2005 Prospective n = 1412 |
8 | Alcohol NS |
Intake of energy | Spearman correlation | No significant association with DR (p = 0.26) |
| Young et al., 1984 Prospective n = 296 |
8 | Alcohol Risk |
Diabetes duration, impotence and glycemic control | Multivariable logistic regression |
Heavy consumption vs. none–moderate consumption, RR: 2.25 (1.15–4.42) |
| Giuffre et al., 2004 Case–control Case:45 Ctrl:87 |
7 | Alcohol NS |
Diabetes duration, duration of oral treatment and duration of insulin therapy |
Multivariable logistic regression |
No significant association with DR (data not shown) |
| Kawasaki et al., 2018 Cross-sectional n = 363 |
5 | Alcohol NS |
Age, sex, HbA1c, diabetes duration, medication, BMI, lifetime maximum body weight, systolic blood pressure, diastolic blood pressure, non-HDL cholesterol, HDL-cholesterol, LDL, estimated glomerular filtration rate, history of myocardial infarction, history of stroke, alcohol, smoking, number of oral hypoglycemic agents, number of antihypertensive agents | Multiple logistic regression model | No signification was seen (p = 0.759) |
| Acan et al., 2018 Cross-sectional n = 413 |
3 | Alcohol Risk |
NIL | t test | p = 0.010 |
| Mediterranean Diet | |||||
| Ghaemi et al., 2021 Prospective n = 22187 |
7 | Mediterranean diet Protective |
Age, sex, time, HbA1c, fasting plasma glucose, HDL-cholesterol, total cholesterol, total triglycerides, systolic blood pressure, obesity, smoking, diabetes duration | Pooled logistic regression models | Mediterranean diet against incident retinopathy in type 1 DM (OR: 0.32, 95%CI: 0.24–0.44, p = <0.001) and type 2 DM (OR: 0.68, 95%CI: 0.61–0.71, p = <0.001) |
| Diaz-Lopez et al., 2015 Interventional n = 3614 |
ModerateBias | Mediterranean diet Protective |
Sex, age, waist circumference, BMI, smoking, physical activity, hypertension, educational level, dyslipidemia, family history of premature coronary heart disease, and baseline adherence |
Multivariate Cox regression | Mediterranean diet vs. control diet, HR: 0.60 (0.37–0.96) |
| Caloric Intake | |||||
| Alcubierre et al., 2016 Case–control Case:146 Control:148 |
10 | Caloric intake 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 energy intake tertile (T3) vs. lowest energy intake tertile (T1), OR: 0.73 (0.37–1.46) |
| Roy et al., 2010 Prospective n = 469 |
10 | Caloric intake Risk |
Sex, age, total caloric intake, oleic acid intake, physical exercise, glycated hemoglobin, carbohydrate intake, protein intake, and hypertension | Multivariable logistic regression | Higher caloric intake, OR: 1.48 (1.15–1.92), p = 0.003 |
| Cundiff et al., 2005 Prospective n = 1412 |
8 | Caloric intake Risk |
NIL | Spearman correlation | Calories in kcal against DR progression rate, r = 0.07 (p = 0.007) |
BMI—Body mass index, CVD—Cardiovascular disease, CKD—Chronic kidney disease, DM—Diabetes Mellitus, DR—Diabetic retinopathy, HDL—High-density lipoprotein, HbA1c—Glycated hemoglobin, LDL—Low-density lipoprotein, OR—Odds ratio, PDR—Proliferative diabetic retinopathy, RR—Relative risk, VTDR—Vision-threatening diabetic retinopathy.