Diabetic retinopathy (DR) is one of the most common and specific microvascular complications of diabetes (1). According to the report of a systematic review and meta-analysis, the global population with DR was estimated to be 103.12 million in 2020 and is projected to increase to 160.50 million in 2045 (2). DR not only significantly contributes to the development of blindness and visual impairment but also is associated with an increased risk of all-cause mortality (3). DR has imposed a significant disease burden worldwide (2).

Endothelial dysfunction is an important pathological mechanism for the occurrence and development of DR (4, 5). However, the accuracy and reproducibility of assessing endothelial function, specifically flow-mediated dilation, are susceptible to operator variability, and there is no universally recognized biomarker for evaluating endothelial function. Recently, a novel Endothelial Activation and Stress Index (EASIX) based on biochemical markers including lactate dehydrogenase (LDH), creatinine, and platelets has been proposed and utilized for investigating the prognostic implications in patients with severe diseases and cancer (6, 7). However, there is currently a lack of research exploring the association between the EASIX and DR.

Vitamin B₆ is a crucial micronutrient for the human body, and its anti-inflammatory properties have been progressively elucidated (8). Recent studies have indicated that reduced vitamin B₆ intake may be linked to an elevated risk of developing DR and an increased mortality rate in patients with DR (9, 10), potentially due to the suppression of superoxide production in microvascular endothelial cells and the inflammatory response (10, 11). It is well known that dietary intake may not accurately reflect the body’s vitamin B₆ levels and does not reflect the catabolic status of vitamin B₆ catabolism status. Several studies have found that vitamin B₆ catabolism, or vitamin B₆ turnover, was strongly correlated with the risk of mortality (12, 13) and cardiovascular disease (CVD) (14) in individuals with diabetes. This association is reflected by the ratio of 4-pyridoxine (4-PA) to pyridoxal 5′-phosphate (PLP) (4-PA/PLP), a blood marker that is linked to vitamin B₆ levels. This ratio integrates changes in vitamin B₆ metabolism mediated by inflammation and serves as a functional indicator of inflammatory status and dysregulation in vitamin B₆ metabolism (14). Chronic inflammation plays a crucial role in the pathogenesis of DR associated with endothelial dysfunction (4, 5). However, few studies have investigated the correlation between the 4-PA/PLP ratio and DR, as well as whether this ratio influences DR risk through its impact on endothelial function.

In light of this background, our study aimed to explore the association between 4-PA/PLP and DR and further assess the mediating effect of EASIX on this association. It is hoped that this study could contribute to the early identification of high-risk populations and provide some references for DR management.

In the present study, we explored the associations between 4-PA/PLP, EASIX, and DR among patients with diabetes using the NHANES data. The findings suggested positive associations between high levels of 4-PA/PLP and log EASIX, as well as between high levels of 4-PA/PLP and DR. Interestingly, we found that log EASIX mediates the association between 4-PA/PLP and DR, with a mediation proportion of 69.06%. A significant mediating effect of log EASIX was more likely to occur among individuals with diabetes who were aged ≥60 years or had a duration of diabetes ≥10 years.

Vitamin B₆ is a crucial nutrient. There was evidence linking the relationship between vitamin B₆ and DR risk. An analysis of the data from the Japan Diabetes Complications Study revealed a significant inverse association between high vitamin B6 intake and the incidence of DR in individuals with type 2 diabetes in Japan, suggesting a potential protective role of increased vitamin B₆ consumption (10). In a narrative review, a high dietary intake of vitamin B₆ was also found to significantly reduce the risk of DR following an 8-year follow-up period (17). The involvement of inflammation in the pathogenesis of DR is widely acknowledged (18). Vitamin B₆ has been recognized for its antioxidant and anti-inflammatory properties, as well as its ability to modulate immunity and gene expression (19). An anti-inflammatory diet intake may decrease the levels of inflammatory markers and consequently lower the likelihood of developing DR. In addition, the deficiency of vitamin B₆ leads to an increase in homocysteine levels (20). Elevated homocysteine is considered a risk factor for DR (21). Similar to previous research, our study found that a high level of 4-PA/PLP was related to increased DR risk. PLP serves as a coenzyme form of vitamin B₆, playing a pivotal role in catalyzing over 160 diverse functions (22). 4-PA represents the primary metabolite of vitamin B₆ (23). 4-PA/PLP was used as an indicator for assessing vitamin B₆ turnover, reflecting a state of low vitamin B₆ status due to altered tissue distribution or increased vitamin B₆ turnover (24).

Endothelial cells play crucial roles in maintaining retinal vascular homeostasis (25). Retinal vascular dysfunction is induced by various factors, including advanced glycosylation end products and receptors, oxidative stress, pro-inflammatory cytokines and chemokines, proliferator-activated receptor-γ disruption, growth factors, and microRNA; these factors contribute to the impairment of retinal endothelial function, ultimately leading to the development of DR (4). The focus on markers of endothelial dysfunction as biomarkers of DR may be reasonable. EASIX, as a marker of endothelial injury, has been validated as an easily predictive biomarker for the prediction of survival outcomes in various medical conditions (26–28). However, there is a lack of research investigating the correlation between EASIX and the progression of DR. In this study, we observed that a high level of log EASIX was associated with an increased risk of DR. In addition, a positive association between 4-PA/PLP and log EASIX was found in this study. Oxidative stress is characterized by the imbalance between oxidation product levels and the body’s antioxidant capacity (29). Several studies revealed that nutrition rich in antioxidant properties possesses the ability to effectively scavenge free radicals and reactive oxygen species (ROS), thereby mitigating oxidative stress among individuals with diabetes (30, 31). A higher level of vitamin B₆ was found to normalize endothelial dysfunction in children with type 1 diabetes (32). This might explain the relationship between 4-PA/PLP and log EASIX.

To the best of our knowledge, this is the first study to explore the mediating effect of log EASIX for the association between 4-PA/PLP and DR among individuals with diabetes based on the NHANES data. The results showed that log EASIX partially mediated the relationship between 4-PA/PLP and DR risk, offering valuable insights into the potential role of endothelial function indicators in this relationship. In addition, the mediation proportion of log EASIX was 69.06%, indicating that it was largely involved in the process of 4-PA/PLP on the development of DR. In other words, the association between 4-PA/PLP and DR risk was serially mediated through the 4-PA/PLP-endothelial dysfunction path and the endothelial dysfunction-DR path, suggesting that 4-PA/PLP accelerated endothelium dysfunction and finally aggravated the DR progress.

The finding emphasizes the importance of reducing DR risk by improving endothelial dysfunction, especially in populations with high 4-PA/PLP. Simultaneously, we observed the specificity of age and diabetes duration in the mediating role of log EASIX between 4-PA/PLP and DR risk using stratified analysis. Log EASIX was not mediated in the individuals with diabetes who were aged <60 years, and it might be suggested that for the individuals with diabetes who were aged ≥60 years, reducing the EASIX would be the focus for preventing DR occurrence. Furthermore, among individuals with diabetes and a duration of diabetes ≥10 years, log EASIX partially mediated the association between 4-PA/PLP and DR risk (mediation proportion: 71.83%). However, among those with a duration of diabetes <10 years, the effect of 4-PA/PLP on the development of DR did not appear to be mediated by log EASIX. This finding may provide new insights into the reduced risk of DR in diabetes duration-specific populations.

Nevertheless, some limitations of our study should not be ignored. First, our study was cross-sectional; we cannot establish a causal relationship between 4-PA/PLP, EASIX, and DR. Second, we adjusted for many potential confounders, but we cannot entirely exclude the possibility of unmeasured confounders, such as medication therapy. Third, data on DR were collected by self-report, which might potentially introduce recall bias. In addition, due to the substantial amount of missing or unavailable data on vitamin B6, vitamin B12, and folate supplements in the NHANES database, the data collected for this study solely relied on dietary intake. Finally, similar to the majority of NHANES studies, the results of this study are limited to the US population, and the results need to be interpreted with caution. Further research is essential to confirm our findings in a prospective study.

This study identified a positive association between high levels of 4-PA/PLP and an increased risk of DR, with the relationship partially mediated by log EASIX. Further prospective studies are necessary to examine the effects of EASIX on the relationship between 4-PA/PLP and DR.