ObjectiveTo preliminary investigate the association between systemic inflammatory biomarkers such as neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), and diabetic retinopathy (DR). MethodsA cross-sectional study. A total of 5 999 patients diagnosed with diabetes in the NHANES database during the survey periods from 1999 to 2018 were included in the study. They were divided into the DR group (1 331 cases) and the non-DR group (4 668 cases) based on the presence or absence of DR. Demographic and laboratory data (including complete blood count, blood glucose, and lipid profiles) were collected, and NLR, PLR, and SII were calculated. The survey weights were used to handle the complex sampling design. After adjusting for confounding factors such as C-reactive protein, a weighted multivariate logistic regression model was applied to analyze the association between the logarithmically transformed and quartile-categorized NLR, PLR, SII and DR, and the odds ratio (OR) and its 95% confidence interval (CI) were calculated. Restricted cubic spline (RCS) was used to analyze the dose-response relationship, and subgroup analysis and interaction tests were conducted through Bonferroni correction. ResultsAfter multivariate adjustment, logistic regression analysis showed in the fully adjusted model with log transformation, neither SII (OR=1.160, 95%CI 0.756-1.780), NLR (OR=0.834, 95%CI 0.669-1.040), nor PLR (OR=1.360, 95%CI 0.859-2.154) showed statistically significant linear associations with DR (P＞0.05). RCS analysis indicated that SII showed no statistically significant overall association with DR (Poverall=0.062), but a non-linear relationship was observed (Pnon-linear=0.045). There was a significant non-linear dose-response relationship between PLR and DR (Poverall=0.011, Pnon-linear=0.009); and there was also a significant non-linear dose-response relationship between NLR and DR (Poverall=0.017, Pnon-linear=0.035). Subgroup analysis revealed that the absence of significant associations between SII, NLR, PLR and DR remained stable across different subgroups. ConclusionBased on a large sample cross-sectional study using the NHANES database, no independent linear association is found between PLR, NLR or SII and DR, and the results remain consistent across various subgroups.
