Factors Associated with Prevalent Diabetic Retinopathy in Chinese Americans: The Chinese American Eye Study.

Objective: To identify factors associated with prevalent diabetic retinopathy (DR) among Chinese American adults with type 2 diabetes mellitus (T2DM), and to compare these factors to ones previously described for a population-based sample of Latinos with a higher DR prevalence.

Design: Population-based cross-sectional study.

Participants: 4582 Chinese Americans aged 50 or older residing in Monterey Park, California.

Methods: Participants completed an in-home questionnaire on socio-demographic status and medical history, and a comprehensive clinical eye examination, using the same protocol implemented in the Los Angeles Latino Eye Study. Fundus photographs from 7 Early Treatment Diabetic Retinopathy fields were graded in a masked manner using a modified Airlie House grading system to assess presence and severity of DR. Logistic regression analyses based on a conceptual model of DR risk identified factors associated with prevalent DR.

Main Outcome Measures: Odds ratios (ORs) and 95% confidence intervals (CIs) for factors associated with DR and vision-threatening DR (VTDR).

Results: In total, 238 participants were diagnosed with any DR; 27 of these were classified as having VTDR. Both, any DR and VTDR showed statistically significant positive associations with T2DM duration (OR5-9 years = 1.24, OR10-14 years = 2.07, OR15+years = 3.99), glycosylated hemoglobin (HbA1c) (OR6.5-6.9% = 1.33, OR7-7.9% = 1.86, OR8%+ = 3.22), systolic blood pressure (SBP) (ORper 10mmHg+ = 1.19), and insulin treatment (ORinsulin+ = 2.44). For VTDR, we also found novel associations with antihypertensive drugs (OR: 0.18; 95% CI: 0.06-0.61) and statins (OR: 4.96; 95% CI: 1.60-16.41). Chinese Americans and Latinos had a nearly identical DR probability based on HbA1c and SBP. However, Latinos had a higher DR probability at every year of duration of T2DM (≥ 5 years).

Conclusions: While we observed an overall lower DR prevalence in Chinese Americans than in Latinos (35.8% of individuals with TD2M in Chinese Americans versus 42.0% in Latinos), our data indicate that the impact of increasing HbA1c and SBP on DR probability is incrementally the same in both populations. However, increasing T2DM duration is associated with higher DR probability in Latinos than Chinese Americans, even after controlling for other known predictors. Novel factors associated with VTDR include antihypertensive drugs and statins. However, to determine if these drugs impact VTDR susceptibility, we need longitudinal data and more cases.