Table 2 Multivariable associations between BMI and waist-to-height ratio with the presence of diabetic retinopathy in the studied population and stratified by sex.

From: Imaging-based body fat distribution and diabetic retinopathy in general US population with diabetes: an NHANES analysis (2003–2006 and 2011–2018)

 

Prevalence

Overall OR (95% CI)

Male OR (95% CI)

Female OR (95% CI)

BMI level

<25

23.6%

1 [Reference]

1 [Reference]

1 [Reference]

25–<30

20.9%

0.908 (0.506, 1.630)

1.270 (0.586, 2.751)

0.447 (0.212, 0.943)

30

18.9%

0.957 (0.537, 1.705)

1.113 (0.527, 2.352)

0.671 (0.299, 1.507)

P for trend

 

0.970

0.954

0.742

BMI tertiles by ethnicity

Tertile 1

24.2%

1 [Reference]

1 [Reference]

1 [Reference]

Tertile 2

18.0%

0.817 (0.535, 1.247)

0.774 (0.426, 1.403)

0.897 (0.451, 1.782)

Tertile 3

18.0%

0.836 (0.520, 1.342)

0.734 (0.388, 1.389)

0.903 (0.456, 1.785)

P for trend

 

0.445

0.314

0.779

BMI (Per unit increase)

 

1.001 (0.973, 1.030)

0.995 (0.956, 1.036)

1.004 (0.965, 1.044)

WHtR level

Tertile 1 (<0.61)

21.5%

1 [Reference]

1 [Reference]

1 [Reference]

Tertile 2 (0.61–0.69)

19.1%

1.116 (0.713, 1.748)

1.109 (0.628, 1.958)

1.153 (0.477, 2.785)

Tertile 3 (0.69)

18.6%

1.037 (0.650, 1.654)

0.949 (0.468, 1.925)

1.161 (0.510, 2.643)

P for trend

 

0.859

0.951

0.728

WHtR tertiles by ethnicity

Tertile 1

21.4%

1 [Reference]

1 [Reference]

1 [Reference]

Tertile 2

18.8%

1.078 (0.702, 1.654)

1.090 (0.625, 1.899)

1.144 (0.512, 2.554)

Tertile 3

19.1%

1.030 (0.649, 1.634)

0.887 (0.447, 1.757)

1.217 (0.540, 2.740)

P for trend

 

0.892

0.792

0.634

WHtR (Per 0.1-unit increase)

 

1.003 (0.813, 1.239)

1.000 (0.748, 1.337)

1.009 (0.730, 1.393)

  1. Adjusted for sex, age, race/ethnicity, diabetes duration, hemoglobin A1c level, blood pressure level, non-high-density lipoprotein cholesterol level.
  2. Stratified models are adjusted for covariates not stratified on.
  3. BMI body mass index, WHtR waist-to-height ratio, OR odds ratio, CI confidence interval.