Skip to main content
. 2023 Jul 18;14:4296. doi: 10.1038/s41467-023-39864-w

Table 7.

Multivariable-adjusted odds ratios of lifestyle factors with any DR and VTDR

Independent variables Any DR Any DR
OR (95% CI)a pa Non-VTDR VTDR
OR (95% CI)a pa OR (95% CI)a pa
Physical activity, MET min/week
  ≥600 vs <600 0.88 (0.77–0.99) 0.0375 0.88 (0.76–1.02) 0.0896 0.87 (0.73–1.03) 0.1023
Dietary intakeb
  Refined grains, g/day
    >450 vs ≤450 1.06 (0.97–1.16) 0.2040 1.11 (1.01–1.22) 0.0213 0.85 (0.69–1.06) 0.1405
  Whole grains, g/day
    >50 vs ≤50 1.03 (0.94–1.13) 0.4674 1.02 (0.93–1.13) 0.6167 1.08 (0.91–1.28) 0.3602
  Potatoes, g/day
    >31 vs ≤31 0.86 (0.79–0.94) 0.0003 0.84 (0.78–0.91) <0.0001 0.94 (0.77–1.14) 0.5201
  Soybean products, g/day
    >30 vs ≤30 1.00 (0.89–1.12) 0.9837 1.00 (0.90–1.12) 0.9440 0.99 (0.79–1.26) 0.9557
  Fresh vegetables, g/day
    >500 vs ≤500 1.12 (1.00–1.25) 0.0570 1.11 (0.98–1.26) 0.0966 1.14 (0.95–1.37) 0.1631
  Fresh fruits, g/day
    >100 vs ≤100 0.85 (0.77–0.95) 0.0039 0.87 (0.77–0.99) 0.0255 0.77 (0.63–0.95) 0.0113
  Dairy products, ml/day
    >100 vs ≤100 0.93 (0.86–1.00) 0.0594 0.90 (0.84–0.98) 0.0121 1.03 (0.85–1.26) 0.7447
  Red meat, g/day
    >100 vs ≤100 1.09 (0.97–1.21) 0.1345 1.10 (0.98–1.22) 0.0950 1.04 (0.85–1.27) 0.7089
  Poultry, g/day
    >15 vs ≤15 0.95 (0.86–1.05) 0.2885 0.98 (0.88–1.09) 0.6729 0.82 (0.64–1.05) 0.1133
  Seafood, g/day
   >26 vs ≤26 0.95 (0.85–1.06) 0.3175 0.96 (0.85–1.08) 0.4893 0.89 (0.75–1.05) 0.1679
  Eggs, g/day
    >50 vs ≤50 0.98 (0.89–1.07) 0.6114 0.96 (0.88–1.05) 0.3374 1.06 (0.85–1.30) 0.6100
  Nuts
    Yes vs no 0.94 (0.85–1.05) 0.2805 0.96 (0.86–1.07) 0.4297 0.89 (0.75–1.05) 0.1522
  Fresh juices
    Yes vs no 0.93 (0.76–1.12) 0.4158 0.89 (0.73–1.08) 0.2365 1.09 (0.73–1.62) 0.6614

Any DR any diabetic retinopathy. VTDR vision-threatening diabetic retinopathy. HbA1c hemoglobin A1c. BP blood pressure. LDL-C low-density lipoprotein cholesterol. MET metabolic equivalent. OR odds ratio. CI confidence interval.

aThe binary or multinomial surveylogistic regression (for two-sided test) was applied to calculate OR (95% CIs) and p value, respectively when the dependent variables were the presence or absence of any DR or the severity of DR (absence of any DR, presence of non-VTDR, and presence of VTDR). Reference category was no DR. In the models, the independent variables were simultaneously included with the entered method as follows: the independent variables presented in Tables 6 and 7, body-mass index, smoking and drinking status, use of glucose-lowering treatment, antihypertensive treatment, lipid-lowering treatment, and attainment of high-density lipoprotein cholesterol and triglycerides.

bThe consumption values of 13 food groups were categorical variables. The integer quartile cutoffs of consumption of the 11 food groups (refined grains, whole grains, potatoes, soybean products, fresh vegetables, fresh fruits, dairy products, red meat, poultry, seafood, and eggs) were selected to define high or low consumption. For nuts and fresh juices, consumption or non-consumption was defined.

Source data are provided as a Source Data file.