Association between albuminuria and prevalent diabetic retinopathy in type 2 diabetes: a cross-sectional study with exploratory analysis by carotid plaque status.
Diabetic retinopathy (DR) and diabetic kidney disease share common microvascular mechanisms, whereas carotid plaque reflects systemic atherosclerotic burden in type 2 diabetes (T2D). We examined the association between urinary albumin-to-creatinine ratio (ACR) and prevalent diabetic retinopathy (DR) in patients with type 2 diabetes and conducted exploratory analyses to assess whether this association differed according to carotid plaque status.
This cross-sectional study included 502 patients with T2D. DR was assessed using fundus photography, and carotid plaque was evaluated by ultrasonography. Because of its skewed distribution, ACR was natural log-transformed as lnACR. Multivariable logistic regression was used to examine the association between lnACR and DR. Subgroup and interaction analyses were performed according to carotid plaque status.
Among 502 participants, 95 (18.9%) had DR. In the extended logistic regression model adjusted for selected clinical covariates, lnACR was associated with prevalent DR (adjusted OR = 1.291, 95% CI: 1.045-1.595, P = 0.018). In a modified Poisson sensitivity analysis using the same covariates, the association was directionally similar but attenuated and did not reach statistical significance (adjusted PR = 1.160, 95% CI: 0.975-1.380, P = 0.094). The point estimate for the lnACR-DR association was numerically larger among participants with carotid plaque than among those without plaque, but the multiplicative interaction was not statistically significant (P for interaction = 0.220).
Higher lnACR was associated with prevalent DR in adjusted logistic regression, whereas the corresponding modified Poisson sensitivity analysis showed an attenuated, non-significant association. Carotid plaque-related findings were exploratory, and no statistically significant interaction was observed. Because of the cross-sectional complete-case design, these findings should not be interpreted as evidence of incident DR risk, prediction, or support for changes in ophthalmic screening practice.
This cross-sectional study included 502 patients with T2D. DR was assessed using fundus photography, and carotid plaque was evaluated by ultrasonography. Because of its skewed distribution, ACR was natural log-transformed as lnACR. Multivariable logistic regression was used to examine the association between lnACR and DR. Subgroup and interaction analyses were performed according to carotid plaque status.
Among 502 participants, 95 (18.9%) had DR. In the extended logistic regression model adjusted for selected clinical covariates, lnACR was associated with prevalent DR (adjusted OR = 1.291, 95% CI: 1.045-1.595, P = 0.018). In a modified Poisson sensitivity analysis using the same covariates, the association was directionally similar but attenuated and did not reach statistical significance (adjusted PR = 1.160, 95% CI: 0.975-1.380, P = 0.094). The point estimate for the lnACR-DR association was numerically larger among participants with carotid plaque than among those without plaque, but the multiplicative interaction was not statistically significant (P for interaction = 0.220).
Higher lnACR was associated with prevalent DR in adjusted logistic regression, whereas the corresponding modified Poisson sensitivity analysis showed an attenuated, non-significant association. Carotid plaque-related findings were exploratory, and no statistically significant interaction was observed. Because of the cross-sectional complete-case design, these findings should not be interpreted as evidence of incident DR risk, prediction, or support for changes in ophthalmic screening practice.