Variability in glycated albumin levels predicts the progression of diabetic nephropathy

Predicting diabetic kidney disease for type 2 diabetes mellitus by machine learning in the real world: a multicenter retrospective study

Objective

Diabetic kidney disease (DKD) has been reported as a main microvascular complication of diabetes mellitus. Although renal biopsy is capable of distinguishing DKD from Non Diabetic kidney disease(NDKD), no gold standard has been validated to assess the development of DKD.This study aimed to build an auxiliary diagnosis model for type 2 Diabetic kidney disease (T2DKD) based on machine learning algorithms.

Methods

Clinical data on 3624 individuals with type 2 diabetes (T2DM) was gathered from January 1, 2019 to December 31, 2019 using a multi-center retrospective database. The data fell into a training set and a validation set at random at a ratio of 8:2. To identify critical clinical variables, the absolute shrinkage and selection operator with the lowest number was employed. Fifteen machine learning models were built to support the diagnosis of T2DKD, and the optimal model was selected in accordance with the area under the receiver operating characteristic curve (AUC) and accuracy. The model was improved with the use of Bayesian Optimization methods. The Shapley Additive explanations (SHAP) approach was used to illustrate prediction findings.

Results

DKD was diagnosed in 1856 (51.2 percent) of the 3624 individuals within the final cohort. As revealed by the SHAP findings, the Categorical Boosting (CatBoost) model achieved the optimal performance 1in the prediction of the risk of T2DKD, with an AUC of 0.86 based on the top 38 characteristics. The SHAP findings suggested that a simplified CatBoost model with an AUC of 0.84 was built in accordance with the top 12 characteristics. The more basic model features consisted of systolic blood pressure (SBP), creatinine (CREA), length of stay (LOS), thrombin time (TT), Age, prothrombin time (PT), platelet large cell ratio (P-LCR), albumin (ALB), glucose (GLU), fibrinogen (FIB-C), red blood cell distribution width-standard deviation (RDW-SD), as well as hemoglobin A1C(HbA1C).

Conclusion

A machine learning-based model for the prediction of the risk of developing T2DKD was built, and its effectiveness was verified. The CatBoost model can contribute to the diagnosis of T2DKD. Clinicians could gain more insights into the outcomes if the ML model is made interpretable.

Serum glycated albumin as good biomarker for predicting type 2 diabetes: A retrospective cohort study of China National Diabetes and Metabolic Disorders Survey

AIMS

Glycated albumin (GA) is a biomarker for short-term (2-3 weeks) glycaemic control. However, the predictive utility of GA for diabetes and prediabetes is largely uncharacterised. We aimed to investigate the relationships of baseline serum GA levels with incident diabetes and prediabetes.

METHODS

This was a longitudinal cohort study involving 516 subjects without diabetes or prediabetes at baseline. Blood glucose levels were observed during follow-up. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated using COX proportional hazard models. Receiver operating characteristic curves and areas under the curves (AUCs) were used to evaluate the discriminating abilities of glycaemic biomarkers and prediction models.

RESULTS

During a 9-year follow-up, 51 individuals (9.88%) developed diabetes and 92 (17.83%) prediabetes. Unadjusted HRs (95% CI) for both diabetes and prediabetes increased proportionally with increasing GA levels in a dose-response manner. Multivariable-adjusted HRs (95% CI) for diabetes were significantly elevated from 1.0 (reference) to 5.58 (1.86-16.74). However, the trend was no longer significant for prediabetes after multivariable adjustment. AUCs for GA, fasting blood glucose (FBG) and 2-h postprandial blood glucose (2h-PBG) for predicting diabetes were 0.698, 0.655 and 0.725, respectively. The AUCs for GA had no significant differences compared with those for FBG (p = 0.376) and 2h-PBG (p = 0.552). Replacing FBG or 2h-PBG or both with GA in diabetes prediction models made no significant changes to the AUCs of the models.

CONCLUSIONS

GA is of good prognostic utility in predicting diabetes. However, GA may not be a useful biomarker for predicting prediabetes.

Association between visit-to-visit variability of glycated albumin and diabetic retinopathy among patients with type 2 diabetes - A prospective cohort study

AIM

There is a paucity of studies regarding the association between long-term glycemic variability with the risk of diabetic retinopathy (DR) in patients with type 2 diabetes. Therefore, the purpose of this study is to explore the association of glycated albumin (GA) variability and HbA1c variability with the risk of DR in patients with type 2 diabetes.

METHODS

This prospective cohort study included 315 inpatients with type 2 diabetes (191 males and 124 females) with at least 3 measurements of GA and HbA1c within 2years prior to the baseline investigation. Different GA and HbA1c variability markers were calculated, including CV, variability independent of the mean (VIM), and the average real variability (ARV). Cox proportional hazard regression models were used to explore the association between visit-to-visit variability of GA and HbA1c and the risk of DR.

RESULTS

After an average follow-up of 3.42years, 81 patients developed incident DR. Multivariable-adjusted (diabetes duration, smoking status, systolic blood pressure, albumin to creatinine ratio, triglycerides, using fibrates, and mean HbA1c) hazard ratios of DR associated with each unit increase in GA-CV, GA-VIM, and GA-ARV were 1.05 (95% CI 1.02-1.09), 1.69 (95% CI 1.24-2.32), and 1.13 (95%CI 1.04-1.23), respectively. However, there was no significant association between visit-to-visit HbA1c variability and the risk of DR.

CONCLUSIONS

The present study indicated that visit-to-visit variability of GA can predict the risk of incident DR in patients with type 2 diabetes, and the prediction ability is independent of the average HbA1c levels.

Oxidized Albumin as a Mediator of Kidney Disease

Renal diseases are a global health concern, and nearly 24% of kidney disease patients are overweight or obese. Particularly, increased body mass index has been correlated with oxidative stress and urinary albumin excretion in kidney disease patients, also contributing to increased cardiovascular risk. Albumin is the main plasma protein and is able to partially cross the glomerular filtration barrier, being reabsorbed mainly by the proximal tubule through different mechanisms. However, it has been demonstrated that albumin suffers different posttranslational modifications, including oxidation, which appears to be tightly linked to kidney damage progression and is increased in obese patients. Plasma-oxidized albumin levels correlate with a decrease in estimated glomerular filtration rate and an increase in blood urea nitrogen in patients with chronic kidney disease. Moreover, oxidized albumin in kidney disease patients is independently correlated with higher plasma levels of transforming growth factor beta (TGF-β1), tumor necrosis factor (TNF-α), and interleukin (IL)-1β and IL-6. In addition, oxidized albumin exerts a direct effect on neutrophils by augmenting the levels of neutrophil gelatinase-associated lipocalin, a well-accepted biomarker for renal damage in patients and in different experimental settings. Moreover, it has been suggested that albumin oxidation occurs at early stages of chronic kidney disease, accelerating the patient requirements for dialytic treatment during disease progression. In this review, we summarize the evidence supporting the role of overweight- and obesity-induced oxidative stress as a critical factor for the progression of renal disease and cardiovascular morbimortality through albumin oxidation.

Visit-to-visit variability of glycated albumin was associated with incidence or progression of lower extremity atherosclerotic disease

BACKGROUND

The aim of this study was to investigate the association of visit-to-visit variability of hemoglobin A1c (HbA1c) and glycated albumin (GA) with the risk of lower extremity atherosclerotic disease (LEAD).

METHOD

We performed a prospective cohort study of 436 patients with type 2 diabetes (258 men and 178 women) with at least 3 measurements of HbA1c and GA prior to baseline investigation from the Department of Endocrinology and Metabolism, Shanghai Sixth People's Hospital. Different HbA1c and GA variability markers were calculated. Multivariable Cox proportional hazard regression models were used to demonstrate the association between visit-to-visit HbA1c and GA variability and the risk of incident or progressive LEAD.

RESULTS

During a mean follow-up period of 3.77 years, 112 participants developed LEAD. Multivariate-adjusted hazard ratios (HRs) of LEAD across tertiles of GA-CV values were 1.00, 1.06 (95% confidence interval [CI] 0.65-1.75), and 1.71 (95% CI 1.07-2.73) (P for trend = 0.042), respectively. When we used GA-VIM and GA-ARV values as exposures, similar positive associations with the risk of LEAD primary were found. Multivariate-adjusted HRs of LEAD for each 1 unit increase in GA-CV, GA-VIM and GA-ARV were 1.03 (95% CI 1.01-1.06), 1.32 (95% CI 1.03-1.69), and 1.07 (95%CI 1.01-1.15), respectively. However, there was no significant association between visit-to-visit variability of HbA1c and the risk of LEAD.

CONCLUSIONS

Visit-to-visit variability of GA may be an optimal biomarker in relation to LEAD risk among patients with type 2 diabetes.

Diabetes mellitus in chronic kidney disease: Biomarkers beyond HbA1c to estimate glycemic control and diabetes-dependent morbidity and mortality

Diabetes mellitus (DM) is the leading cause of chronic kidney disease (CKD). Optimal glycemic control contributes to improved outcomes in patients with DM, particularly for microvascular damage, but blood glucose levels are too variable to provide an accurate assessment and instead markers averaging long-term glycemic load are used. The most established glycemic biomarker of long-term glycemic control is HbA1c. Nevertheless, HbA1c has pitfalls that limit its accuracy to estimate glycemic control, including the presence of altered red blood cell survival, hemoglobin glycation and suboptimal performance of HbA1c assays. Alternative methods to evaluate glycemic control in patients with DM include glycated albumin, fructosamine, 1-5 anhydroglucitol, continuous glucose measurement, self-monitoring of blood glucose and random blood glucose concentration measurements. Accordingly, our aim was to review the advantages and pitfalls of these methods in the context of CKD.

Anemia modifies the prognostic value of glycated hemoglobin in patients with diabetic chronic kidney disease

A common complication of chronic kidney disease (CKD), anemia can influence glycated hemoglobin (HbA1c) levels. In diabetic patients, anemia occurs earlier and with higher severity over the course of CKD stages. To elucidate the effect of hemoglobin (Hb) on the predictive value of HbA1c, we enrolled 1558 diabetic patients with stages 3-4 CKD, categorized according to baseline Hb and HbA1c quartiles. Linear regression revealed that higher HbA1c correlated significantly with higher Hb in the Hb < 10 g/dL group (β = 0.146, P = 0.004). A fully-adjusted Cox regression model revealed worse clinical outcomes in patients with higher HbA1c quartiles in the Hb ≥ 10 g/dL group. Hazard ratios for end-stage renal disease (ESRD), all-cause mortality, and composite endpoint (cardiovascular events and all-cause mortality) in patients with Hb ≥ 10 g/dL and the highest HbA1c quartile were 1.92 (95% confidence interval [CI], 1.17-3.15), 1.76 (95% CI, 1.02-3.03), and 1.54 (95% CI, 1.03-2.31), respectively. By contrast, HbA1c was not associated with clinical outcomes in the Hb < 10 g/dL group. In conclusion, in stages 3-4 diabetic CKD, higher HbA1c is associated with a higher risk of poor clinical outcomes in patients with Hb ≥ 10 g/dL.

Glycemic control of type 2 diabetes mellitus across stages of renal impairment: information for primary care providers

Chronic kidney disease (CKD) is a frequent complication of type 2 diabetes mellitus (T2DM) and elevates individuals' risk for cardiovascular disease, the leading cause of morbidity and mortality in T2DM. Achieving and maintaining tight glycemic control is key to preventing development or progression of CKD; however, improving glycemic control may be limited by effects of renal impairment on the efficacy and safety of T2DM treatments, necessitating dosing adjustments and careful evaluation of contraindications. Understanding the treatment considerations specific to each class of T2DM medication is important in individualizing therapy and improving glycemic, renal, and cardiovascular outcomes. Traditional glucose-lowering treatments include insulin, metformin, sulfonylureas, meglitinides, and thiazolidinediones. Each of these agents exhibits altered pharmacokinetics in patients with renal impairment except for the thiazolidinediones, which are metabolized by the liver and do not accumulate appreciably in patients with renal impairment. Newer glucose-lowering treatments include GLP-1 receptor agonists, DPP-4 inhibitors, and SGLT2 inhibitors. Of these, only the DPP-4 inhibitor linagliptin can be used across all stages of renal impairment without dosing restrictions or concerns regarding dose escalation, and all SGLT2 inhibitors are contraindicated when eGFR <45 mL/min/1.73m². Several of the newer treatments have also been investigated for effects on renal and cardiovascular outcomes, demonstrating potential benefits of the GLP-1 agonists liraglutide and semaglutide, as well as the SGLT2 inhibitors canagliflozin and empagliflozin, in reducing risk for some adverse renal and cardiovascular events. In addition, some DPP-4 inhibitors have been shown to reduce albuminuria, an indicator of glomerular dysfunction. Consideration of this information is useful in informing optimal management strategies for patients with T2DM and concomitant CKD. More clinical data from future and ongoing clinical trials, including data regarding potential renal and cardiovascular benefits, will be important in clarifying the safety and efficacy profiles of each of these agents in patients with CKD.