Higher RBP4 tracked with diabetic retinopathy across onset groups
Serum retinol-binding protein 4 rose with a higher burden of diabetic retinopathy in adults with type 2 diabetes, and the signal persisted after adjustment for age, sex, body mass index, blood pressure, lipids, hemoglobin A1c, diabetes duration, kidney function, and medication use. In this real-world cohort of 6,996 adults from Lishui People’s Hospital, elevated RBP4 was associated with greater odds of retinopathy in both age-at-onset strata, with the adjusted odds ratio reaching 1.28 in participants whose diabetes began before 65 years and 1.43 in those whose diabetes began at 65 years or later.
The pattern is biologically plausible. RBP4 has been tied to insulin resistance, inflammatory signaling, and endothelial dysfunction, all processes that can injure the retinal microvasculature and accelerate capillary dropout, leakage, and ischemic stress. When a biomarker shadows those pathways, it becomes more than a lab value; it starts to resemble a readout of vascular injury already in motion.
Late-onset diabetes carried a different clinical profile
Participants with diabetes onset at 65 years or older were older at baseline, had higher systolic blood pressure, and showed markedly lower estimated glomerular filtration rate than the earlier-onset group. They also had a shorter diabetes duration, averaging 5.73 years compared with 8.39 years in those diagnosed before 65 years. That shorter exposure window likely contributed to the lower overall burden of retinopathy in the later-onset group, even though the direction of the RBP4 association remained the same.
In practical terms, chronological age did not erase the relevance of cumulative disease time. Duration still mattered as a biologic clock, and the retinal vasculature appears to respond to that clock regardless of whether hyperglycemia began before or after the age of 65.
Higher RBP4 also clustered with a more adverse metabolic phenotype. Participants in the higher-RBP4 group had higher triglycerides, lower HDL cholesterol, and slightly worse glycemic control, while blood pressure and renal function were largely comparable across RBP4 strata. Sex distribution and use of glucose-lowering therapy were similar as well, which reduces the likelihood that the observed association was simply a reflection of treatment imbalance.
The age-at-onset split did not materially change the biomarker signal
The central question was whether the relationship between RBP4 and retinopathy differed by age at type 2 diabetes onset. It did not, at least not in a statistically convincing way. The formal interaction test between RBP4 category and onset age group was not significant, with a P value of 0.58. In other words, the excess retinopathy odds linked to higher RBP4 looked broadly similar whether diabetes began before 65 years or at 65 years and older.
That finding matters because age at diagnosis has sometimes been treated as a shorthand for disease biology. Here, the retinal association followed the biomarker rather than the age label. The retinal microcirculation seemed to care more about the metabolic and inflammatory burden captured by RBP4 and duration than about the threshold of 65 years itself.
A brief but important point stands out. The association was consistent, but not uniform in statistical strength.
Among later-onset participants, the adjusted odds ratio of 1.43 crossed the null with a wide confidence interval, which suggests fewer events or greater imprecision in that stratum. The direction still matched the younger-onset group, so the absence of statistical significance there should not be mistaken for a biologic reversal.
Diabetes duration emerged as the more relevant modifier
In subgroup analyses, diabetes duration was the only characteristic with a statistically significant interaction with RBP4, although the stratum-specific odds ratios were described as very similar. This is a useful reminder that interaction tests do not always imply a clinically dramatic separation; sometimes they simply indicate that the same biomarker behaves differently when measured against a background of longer or shorter exposure to diabetes.
Retinopathy develops through prolonged microvascular stress. Duration therefore functions as a surrogate for cumulative glycemic insult, oxidative injury, and hemodynamic strain, all of which can amplify the effect of a circulating biomarker such as RBP4. Even when two patients have the same current HbA1c, the one with longer disease exposure may have a very different retinal risk profile.
The investigators also performed generalized propensity score weighting to balance diabetes duration across exposure groups in a sensitivity analysis. The use of weighted logistic regression strengthens confidence that the RBP4-retinopathy link was not entirely driven by imbalance in duration, even though cross-sectional data cannot prove causality.
What this means for retinal risk assessment
The study positions RBP4 as a potential biomarker candidate for diabetic retinopathy risk stratification in routine clinical care. It does not justify screening changes on its own, but it does add a measurable signal that may help identify patients whose retinal risk is higher than glucose values alone would suggest. That is especially relevant in type 2 diabetes, where retinal disease often reflects a mixture of metabolic dysregulation, vascular dysfunction, renal impairment, and years of cumulative exposure.
Fundus photography and ETDRS-based grading gave the analysis a standardized ophthalmic anchor, which is important in observational work that relies on electronic medical records. The cohort also benefited from stable laboratory methods and consistent imaging protocols during the study window from 2018 through 2023, limiting some of the noise that can weaken real-world studies.
Still, the design sets firm limits. Cross-sectional data cannot determine whether elevated RBP4 precedes retinopathy, follows it, or simply travels alongside the same vascular injury. The hospital-based sample may also reflect local referral patterns and disease severity that do not fully mirror the broader population with type 2 diabetes.
The most defensible clinical reading is narrower and more actionable: higher serum RBP4 identified adults with type 2 diabetes who were more likely to have retinopathy, and this relationship did not meaningfully differ by whether diabetes began before or after age 65. Diabetes duration remained the more important contextual variable, which means retinal risk assessment still has to be anchored in time with disease, not age alone.