The association of inflammatory biomarkers with clinical outcomes in diabetic retinopathy participants: data from NHANES 2009-2018.

Objective: The aim of this study was to assess the association of neutrophil lymphocyte ratio (NLR), monocyte to lymphocyte ratio (MLR), and system inflammation response index (SIRI) with the all-cause mortality and diabetes-cardiovascular mortality in participants with diabetic retinopathy (DR).

Methods: A total of 572 participants with DR from NHANES were included, and divided into survival group (n = 440) and all-cause death group (n = 132). NLR = neutrophil count/lymphocyte count, MLR = monocyte count/lymphocyte count, SIRI = (neutrophil count × monocyte count)/lymphocyte count. We utilized the NHANES Public-Use Linked Mortality File through April 26, 2022, to determine mortality status. Diabetes-cardiovascular death was defined as death resulting from heart disease, cerebrovascular disease, or diabetes mellitus. The Spearson Correlation Analysis, Kaplan-Meier curves, Cox proportional hazards regression models, Restricted cubic spline plots and Decision Curve Analysis were used.

Results: The all-cause mortality and diabetes-cardiovascular mortality were significantly higher in NLR ≥ 1.516, MLR ≥ 0.309, SIRI ≥ 0.756, and NLR + MLR + SIRI subgroups than NLR < 1.516, MLR < 0.309, SIRI < 0.756 subgroups, and other participants except NLR + MLR + SIRI (all P < 0.05). The HR of NLR, MLR, SIRI, NLR + MLR + SIRI for all-cause mortality were 1.979(1.13-3.468), 1.850(1.279-2.676), 1.821(1.096-3.025), 1.871(1.296-2.703), respectively. The hazard ratio of NLR, MLR, SIRI, NLR + MLR + SIRI for diabetes-cardiovascular mortality were 2.602(1.028-6.591), 2.673(1.483-4.818), 2.001(0.898-4.459), 2.554(1.426-4.575), respectively. In the restricted cubic spline plots, the relationship between NLR, MLR, SIRI and HR of all-cause mortality and diabetes-cardiovascular mortality was overall as "J" shaped. In both age < 60 and age > 60 years participants, the all-cause mortality and diabetes-cardiovascular mortality were significantly higher in NLR ≥ 1.516, MLR ≥ 0.309, SIRI ≥ 0.756, and NLR + MLR + SIRI subgroups than NLR < 1.516, MLR < 0.309, SIRI < 0.756 subgroups, and other participants except NLR + MLR + SIRI (all P < 0.05).

Conclusion: NLR, MLR, and SIRI may be three independent prognostic predictors for all-cause mortality and diabetes-cardiovascular mortality among individuals with DR. In practical clinical applications, combining NLR, MLR, and SIRI may enhance the prediction of all-cause mortality and diabetes-cardiovascular mortality in DR.