Diabetic kidney disease (DKD) is the most common microvascular complication of diabetes and the most common cause of end-stage renal disease (ESRD) (1). About 30-40% of people with diabetes develop DKD and half of the patients with DKD progress to ESRD (2), adding additional mortality and healthcare costs to patients (3, 4). Albuminuria is a marker of renal/glomerular disease and a key indicator of DKD (5). Albuminuria is a pathophysiological event triggered by several factors, including: damage to the endothelial glycocalyx, loss of endothelial cell function and rearrangement and injury of the cytoskeleton of podocytes (6). Accumulation of glucose and glycosylated proteins in endothelial cells leads to impaired glomerular filtration barrier function and changes in glomerular capillary permeability in patients with diabetic nephropathy (7, 8). Urinary albumin measurements are typically used to identify and monitor patients with kidney dysfunction (9) by measuring the ratio of albumin to creatinine (ACR) in urine (10, 11). Many studies have confirmed that high ACR indicates glomerular ultrafiltration and is related to renal failure, acute kidney injury, and progression of renal disease (12–14). An increase in urinary albumin excretion is widely recognized as a precursor of diabetic kidney disease, predicting an increased early risk and suggesting the need for early clinical intervention (15).

Hemoglobin A1c (HbA1c) represents the average fluctuation level of blood glucose in the body in recent three months. As a significant change in clinical practice, the American Diabetes Association’s 2010 guidelines recommended HbA1c as one of the diagnostic criteria for diabetes, and set a threshold of ≥ 6.5% (16). Since the DCCT (Diabetes Control and Complications Trial) and UKPDS (UK Prospective Diabetes Study) (17) showed that HbA1c level is a strong predictor of the risk of diabetic microvascular complications, glycosylated hemoglobin level has been the focus of diabetes management. Intensive glycemic therapy has been shown to reduce the incidence of albuminuria and delayed DKD progression to some extent (18). Currently, many studies have confirmed that HbA1c>7.0% is significantly associated with ACR and served as a risky indicator for severe DKD (19–21). However, the risk threshold of glycosylated hemoglobin observed in these studies had no obvious inflection point. Furthermore, there is evidence that the onset of microvascular complications may occur prior to the diagnosis of diabetes, so the above experiments may not accurately reflect the exact point at which the risk of DKD begins to increase.

The objective of our research is to explore the correlation between blood glucose and the continuous increase in ACR values and to reveal the adverse effects of hyperglycemia on renal function prior to the diagnostic criteria for diabetes being met. We attempt to determine the optimal blood glucose threshold to provide additional clinical evidence for the prevention and early screening of diabetic kidney injury, so as to prevent the occurrence and development of DKD.

Our study provides evidence that among blood glucose indicators, only HbA1c is independently associated with ACR. The risk threshold identified for HbA1c for the prevalence of albuminuria among a Chinese population was 5.5%. The increase in ACR was accompanied by a positive change in HbA1c, when HbA1c ≥ 5.5%. Furthermore, the correlation remained stable and significant after adjusting for sociodemographic characteristics and performing subgroup analyses based on gender, smoking habit, obesity, and hypertension. This result indicates the effect of blood sugar on renal damage may occur before the diagnosis of diabetes, emphasizing the necessity of early blood glucose control and renal function screening for DKD.

At present, many studies have shown that HbA1c levels affect albuminuria. For people with diabetes, data from the FinnDiane study and DMIDS project both suggest that HbA1c influences the development of albuminuria and renal disease (29, 30). Meanwhile, the AdDIT study, which was focused on adolescents, reached the same conclusion (31). The above results are consistent with those of our study. However, these analyses were performed only in a diabetic population with HbA1c >7% among enrolled subjects and showed a linear relationship without obvious inflection points. In addition, prediabetes is also an independent risk factor for glomerular ultrafiltration and increased ACR. The NHANES and the ARIC studies demonstrated that microvascular dysfunction is already present in prediabetes, and there was a significant trend toward an increased risk of kidney disease and ESRD with increased baseline HbA1c levels (32–34). In our analysis, we included the entire study population with levels of HbA1c ranging from 4% to greater than 9%. This result indicates that ACR accompanied by increased HbA1c reaches a turning point at 5.5%, suggesting this as the HbA1c threshold for developing kidney dysfunction.

With respect to a glucose threshold in terms of HbA1c levels, current clinical studies have not consistently demonstrated a relationship for predicting the onset of DKD. In the diabetic population, the ADVANCE study showed no significant increase in the risk of retinopathy or renal complications with an HbA1c level below 6.5% (48 mmol/mol) (35). This result was indistinguishable from diagnostic criteria and only included for those with confirmed diabetes. The ESTHER study demonstrated that the risk for reduced kidney function with respect to eGFR linearly increased to more than three-fold when HbA1c levels were increased, with a turning point at HbA1c = 6.4% (36). In our study, the inflection point is the effect corresponding to ACR, speculating that the increase of ACR occurs earlier than the decline of eGFR in DKD (37, 38). Normal glomerular filtration barriers include porous glomerular endothelium, glomerular basement membrane and podocyte foot processes. For pathological reasons, endothelial dysfunction is the key factors for the damage of filtration membrane during the progression of DKD (38, 39). Chronic hyperglycemia induces oxidative stress that impairs endothelial function and podocytes, leading to impaired glomerular filtration barrier, glomerular hyperfiltration, and increased albumin excretion. In both type 1 and type 2 diabetes patients, a triad of endothelial cell glycocalyx damage, increased vascular permeability and albuminuria occurred (40, 41). The ORIGIN trial followed type 2 diabetes or prediabetes mellitus patients for 6.2 years, and findings suggested a gradual increase in the risk of adverse renal outcomes, from the lowest group (HbA1c< 5.7%) to the highest group (HbA1c > 7.4%) (42). However, the analytic method of that study divided HbA1c into five groups based on the hazard ratio, so only a range in blood glucose threshold was given. Evidence of continuous changes in blood sugar was lacking, and there was no precise inflection point. In our study, we identified an important cutoff at 5.5%, and to our surprise, it did not meet the diagnostic criteria for prediabetes, which is defined as HbA1c of 5.7–6.4% (39–47 mmol/mol) according to the 2019 ADA clinical standards (16). The classification of pre-diabetes diagnostic criteria is still controversial. Many studies have suggested the use of different HbA1c cut-off values to diagnose pre-diabetes for different populations due to genetic and biological differences (43–45). Our result places particular emphasis on people with HbA1c greater than 5.5% potentially having an increased risk for albuminuria and provides more clinical evidence for the diagnosis of prediabetes in Chinese population.

Currently, there is sufficient evidence that intensive glucose therapy significantly reduces albuminuria and improves the composite end point of nephropathy. As of 2019, the ADA recommended target for T2DM glycemic control is HbA1c < 6.5% after adequate consideration of the risk of hypoglycemia. The DCCT study of a T1DM population, the UKPDS study of a T2DM population, and the ADVANCE study all showed that intensive glycemic therapy significantly improved renal clinical benefits compared to conventional glycemic control (46, 47). These studies only focused on people who already had diabetes, the glycemic control threshold and risk of proteinuria were different from those of the general population. There is a lack of recommendations for glycemic control in the normal population who did not meet the diagnostic criteria for diabetes to avoid diabetic kidney injury. Our results provide strong evidence for predicting the starting point of renal function impairment and the glycemic control threshold. With the development of treatments and the popularization of physical examination, current strategies for managing blood glucose should also consider the risk of disease in potentially at-risk populations and evaluate the clinical benefits of early intervention.

The National Institute for Clinical Excellence (UK) and the Standards of Medical Care in Diabetes from ADA both recommend kidney care for all people with type 2 diabetes that includes measuring ACR or albumin levels annually (48, 49). Based on the results of our study, we recommend early glycemic management and kidney function screening of ACR levels for people with risk factors. In terms of the duration of DKD, the timing of intervention may be critical (50). Early and more focused interventions for at-risk populations have the potential to convey substantial, long-term improvements in public health. Therefore, it is particularly important that public health and health care policies place greater emphasis on achieving early glycemic control and complication screening in at-risk populations of DKD (51–53).

Our analysis had the following limitations. First, the design of this study was a cross-sectional survey, which can only identify the correlation but cannot conclude cause and effect between ACR and HbA1c. Second, although we used two regression models to estimate a single point of change, the choice of the point of change may also be skewed. However, this deviation, while having a slight effect on the regression coefficient, does not affect the nonlinear conclusion. Furthermore, while we adjusted for many of the covariables associated with ACR in the multiple regression analysis, other potential factors, such as social status, personal income level, and family lifestyle may also influence the regression results and should be included in the adjustment. Third, we only used morning urine samples to measure ACR levels, but levels of ACR fluctuate with time. Therefore, combining this measurement with 24-hour urine albumin quantitative results to assess the risk of albuminuria would be more scientific and rigorous. However, there was a good correlation between on-site urine ACR samples and urine samples collected 24 hours a day. Urine ACR assessment through field samples may be a reliable alternative to epidemiological specimen collection (54). Fourth, the population in this study primarily included community residents over 40 years old of Chinese subjects. Therefore, the current findings cannot be reliably extrapolated to other races or age groups. Fifth, our analysis lacked relevant information on lowering glucose treatment for diabetic patients and did not analyze the effect of lowering glucose treatment in the model for correction. However, studies have found that HbA1c appears to be an independent risk factor for the development of albuminuria even adjusted for use of antidiabetic drugs (55, 56). Therefore, we suggested that the missing information on glucose-lowering therapy in diabetic patients may not influence the relationship between HbA1c and ACR. However, the study will be more rigorous if detailed information of antidiabetic therapies could be gathered.

Our study provides evidence for an HbA1c threshold in which the risk of albuminuria begins to increase at a cutoff point of 5.5% in a Chinese population. We found that HbA1c≥5.5% was positively and independently associated with ACR. After adjusting for the influence of risk factors, the correlation remained stable and significant. This finding underlines the need for screening renal function in at-risk populations and the importance of early glucose management to delay the development of renal complications from diabetes.

Data are available upon reasonable request. Main document data and additional unpublished data from the study are available by sending Email to xumt@mail.sysu.edu.cn with proper purposes.

The studies involving human participants were reviewed and approved by Ruijin Hospital Ethics Committee Shanghai JiaoTong University School of Medicine. The patients/participants provided their written informed consent to participate in this study.

All authors contributed to the conception of the study, execution of data collection, data analysis or writing and revision of the paper. All authors contributed to the article and approved the submitted version.

This work was supported by National Key R & D Program of China (No. 2016YFC0901204), Medical Scientific Research Foundation of Guangdong Province of China (A2019040), and National Natural Science Foundation of Guangdong, China (2019A1515011110).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.