Type 2 diabetes mellitus (T2DM) has escalated into a global health crisis, which stands as the 11th leading cause of death worldwide due to chronic complications (1). Among the myriad microvascular complications associated with T2DM, diabetic nephropathy (DN) emerges as one of the most prevalent and severe, often culminating in end-stage kidney disease (ESKD). Current evidence suggests that DN is detected in approximately 33.6% of diabetic patients (2). It is generally characterized by an initial elevation in microalbuminuria excretion, a substantial increase in albuminuria, and a decline in glomerular filtration rate (GFR) (3). Research has underscored that diabetic patients exhibiting albuminuria are at a heightened risk of cardiovascular disease, mortality, and renal deterioration (4). Therefore, albuminuria serves as an early indicator of DN. Once DN manifests, its progression is challenging to reverse. Importantly, identifying diabetic patients prone to developing albuminuria could significantly aid in preventing the onset of DN.

Advanced glycation end products (AGEs) arise from the nonenzymatic glycosylation of proteins and lipids (5). This glycosylation process is intricate and slow. However, in a prolonged state of elevated glucose levels, glycosylation rates significantly accelerate, increasing AGEs. Studies have demonstrated a clear correlation between AGE accumulation in tissues and blood glucose levels (6). Furthermore, even after correcting hyperglycemia, AGE levels in diabetic tissues often fail to return to normal, leading to the concept of “metabolic memory” (7). Unlike HbA_1c, AGEs are not merely byproducts of hyperglycemia but are also implicated in the development of diabetes (8). It is now understood that AGEs can crosslink with proteins, altering their structure, interfering with their functional properties, and binding to the receptor for advanced glycation end products (RAGE), thereby activating proinflammatory signaling pathways (9). These processes are also thought to contribute to the development of diabetic microvascular complications (10). Therefore, AGEs are gaining increasing attention, especially concerning their potential role as markers of DN. However, current methods for measuring AGEs are often complex and costly, making the need for cost-effective, portable, and stable measurement methods paramount. The non-invasive diabetes detector (DM scan), developed using optical detection technology for AGEs, offers the advantage of rapid, non-invasive measurements without the risk of cross-infection. Nevertheless, few studies have explored the relationship between DN and AGEs measured by skin autofluorescence.

While the significance of glycemic control in DN management has been established, it is imperative to consider other metabolism-associated risk factors. Obesity, a burgeoning global public health concern (11), has also been linked to kidney disease (12), with body mass index (BMI) serving as a common measure of obesity. A study in the UK revealed a positive correlation between higher BMI and an increased incidence of microalbuminuria, with this association particularly pronounced among individuals with higher BMI (13). Beyond BMI, various metabolic metrics are employed to assess their relationship with kidney disease. One such metric, the triglyceride-glucose-BMI (TyG-BMI) index, is a product of fasting blood glucose and triglyceride levels combined with BMI. It is currently used to evaluate the association with diabetes (14) and is considered an alternative surrogate marker for insulin resistance (IR), which itself is linked to kidney disease (15). However, few studies have investigated the association between TyG-BMI and DN.

As the prevalence of diabetes continues to surge, the burden of diabetes-associated nephropathy is also poised to increase. Accordingly, there is a pressing need for enhanced clinical prevention strategies to mitigate modifiable DN risk factors. Most existing studies have predominantly focused on the relationship between individual risk factors and DN, with few examining potential synergistic effects among these risk factors. Acknowledging the influence of glycemic management on DN, this study incorporates HbA_1c into the model. Accordingly, we put forth the following hypotheses: 1) AGEs are associated with DN, 2) Obesity can modulate this relationship, and 3) An interaction exists between AGEs, obesity, HbA_1c, and DN. The outcomes of this study are anticipated to provide vital insights for healthcare providers and decision-makers, facilitating informed clinical decisions in the realm of healthcare.

In this retrospective cross-sectional study, several key findings emerged. First, we observed a DN incidence of 37.2% among hospitalized T2DM patients, slightly higher than the rates reported in previous studies (2). Notably, among these DN patients, 58.4% had a BMI exceeding 24 kg/m², and only 11.5% had HbA_1c levels below 7%. This finding highlights the inadequacy of comprehensive T2DM management among this population. Second, our study revealed a significant correlation between AGEs and DN, with higher AGE levels indicating an increased risk of DN. Considering the influence of age on both AGEs and DN, we introduced the AGE_age index, which integrates AGEs and age. Lastly, we identified a three-way interaction among obesity, AGEs, and DN with HbA_1c in this regulatory relationship. These findings were supported by the results of sensitivity analyses, emphasizing their robustness.

Unlike diabetic macroangiopathy, diabetic microangiopathy is more closely associated with blood glucose, as evidenced in numerous large clinical studies (27, 28). Chronic hyperglycemia leads to increased oxidative stress, initiating the accumulation of AGEs in cells via activation through pathways such as the hexose pathway, polyol pathway, and protein kinase C, resulting in overexpression of RAGE and subsequent activation of various inflammatory cytokines (29). Studies on animals have indicated that inhibiting carboxymethyllysine (CML) may protect against DN progression (30), while young diabetic rats treated with AGEs precursors exhibit renal lesions similar to those seen in aged diabetic rats (31). AGEs are, therefore, crucial in DN development, and AGEs-generated markers can be harnessed to assess DN risk.

Recent studies have shown that non-invasive devices measuring skin AGE fluorescence can be used for diabetes screening, offering a simple and rapid approach (32). However, previous research on the association between non-invasive skin AGEs and diabetic complications has primarily focused on Caucasian populations, showing significant positive correlations between AGEs and diabetic vascular complications (33). Given the impact of skin tone on skin AGE levels, research on the relationship between AGEs and DN in Chinese diabetic populations remains limited. In this study, we employed UACR as a marker for DN to investigate the AGE-DN relationship. Given the significance of age in both AGEs and DN, we introduced the AGE_age index. We found that AGE_age levels were significantly elevated in DN, and after adjusting for factors including age, sex, and HbA_1c, AGE_age remained positively correlated with UACR levels. This finding indicates that AGE_age influences UACR independently of HbA_1c, underlining its value in assessing DN.

One of the management strategies for T2DM is lifestyle modification, including weight loss. A longitudinal study involving 369,362 participants aged 2-15 years indicated that a high percentage of T2DM patients were obese (47.1%), with only 4.33% having a normal BMI (34). This underscores the strong link between obesity and diabetes. Moreover, studies have independently identified BMI as a risk factor for DN (16). Large population-based investigations have corroborated the increased risk of nephropathy in individuals with both diabetes and obesity, and this risk remains elevated even after stringent glycemic control (35). This highlights the role of obesity in DN development, independently of blood glucose control. Overall, our study findings confirm the association of BMI with DN, emphasizing the importance of BMI control in T2DM management.

The interaction between BMI and AGEs has become a research hotspot. AGEs typically accumulate slowly through glycation processes, with hyperglycemia and hyperlipidemia accelerating AGE accumulation in vivo (36). Given that both hyperglycemia and hyperlipidemia are prevalent in obese individuals, it is reasonable to speculate that AGE levels are higher in obese patients, as supported by previous research (37). Our study consistently validated the association between BMI and AGE_age. In vitro and animal experiments further supported this relationship, demonstrating that RAGE overexpression induces adipocyte hypertrophy (38) and that mice fed a high-fat high-AGE diet exhibit greater weight gain and more visceral fat compared with mice fed a high-fat low AGE diet for 6 weeks (39). Additionally, obese individuals often have less healthy dietary habits, consuming highly processed Western-style foods rich in exogenous AGEs, which can be absorbed into the bloodstream and accumulate in the body (40). Considering this interaction, we propose that AGEs interact with BMI to facilitate DN development. Our study validated this hypothesis, with moderating analysis showing that AGE_age interacts with BMI to increase the UACR. In contrast, HbA_1c and BMI did not exhibit a synergistic effect on DN risk, underscoring the greater importance of AGE_age in DN, with BMI exacerbating the condition. Although HbA_1c did not exert a moderating effect on BMI, we identified a three-way interaction between AGE_age, HbA_1c, BMI, and UACR, suggesting that patients with T2DM, especially those with higher AGEs, obesity, and HbA_1c levels, are at a heightened risk of urinary proteinuria. Effective management of HbA_1c and weight reduction can mitigate the impact of AGEs on UACR, emphasizing the importance of a comprehensive approach. On one hand, it involves strict blood glucose control to reduce HbA_1c levels and minimize endogenous AGE production. On the other hand, it necessitates dietary control to reduce the consumption of high-AGE foods, thereby decreasing the absorption of exogenous AGEs and lowering the risk of obesity.

While obesity is primarily linked to dietary factors, there are additional contributors to obesity, including IR. The development of IR is closely associated with obesity in a complex relationship, both being integral components of the metabolic syndrome. IR is a well-established risk factor for cardiovascular and cerebrovascular diseases and plays a significant role in DN. Animal studies have shown that mice fed a high-fat diet, resulting in obesity and IR, exhibit increased UACR levels and altered renal outcomes, indicating tubular dilation and interstitial vacuolation (41). Therefore, we examined another metabolic indicator, TyG-BMI, to represent IR. TyG-BMI, derived from the product of the TyG index and BMI, effectively reflects various metabolic processes in the body. Studies have previously established that elevated levels of TyG-BMI can heighten the risk of prediabetes, especially among non-obese individuals (42). Causality between TyG-BMI and the incidence of diabetes has been reported, particularly in non-obese populations (14). Nevertheless, the relationship between TyG-BMI and DN has received less attention. Our study provided hitherto undocumented evidence of a significant positive relationship between TyG-BMI and UACR, indicating that TyG-BMI is a potential risk factor for DN, possibly surpassing BMI’s significance. As a moderating variable, TyG-BMI exerts a distinct influence on the relationship between AGEs and UACR levels. Concurrently, in vitro and animal experiments suggest that AGEs can influence cellular insulin sensitivity and insulin secretion capacity (43, 44). This implies that non-obese type 2 diabetes patients, despite seemingly meeting BMI standards, should consider other metabolic factors since BMI fails to capture fat distribution, and abdominal obesity is more strongly associated with IR.

Herein, we established a retrospective model to assess the correlation between these metabolic indicators and DN. We unveiled the intricate interaction among AGEs, obesity-related metabolic metrics, and HbA_1c, all associated with UACR levels. This underscores the significance of comprehensive diabetes management. Given that albuminuria in diabetic patients is largely preventable, effective management and treatment strategies should persist even after the onset of DN, aiming to retard disease progression. Comprehensive management awareness is imperative for diabetic patients, and early, timely interventions can substantially reduce the incidence of DN.

This study boasts several strengths, including its multilevel design and the inclusion of a substantial sample size. Furthermore, our study uniquely investigates DN by exploring the relationship between obesity and non-invasive AGEs, offering compelling insights into preventing proteinuria in type 2 diabetes mellitus. However, certain limitations should be acknowledged. First, in recent years, a subtype of DN has been proposed with low estimated glomerular filtration rate but without albuminuria, accounting for about 10.1% of diabetes patients (45), thus this subtype therefore needs to be studied to adjust the management strategy. Second, the cross-sectional nature of this study makes it challenging to establish causal relationships or confirm long-term clinical outcomes.

In conclusion, our study highlights the higher incidence of DN within the hospitalized T2DM population. We propose multifaceted management strategies to prevent DN in T2DM patients. Additionally, we introduce AGE_age, a non-invasive measure of accumulated AGEs adjusted for age, as a promising approach for identifying patients at high risk of developing DN.