Silent 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.

Modified arteriosclerosis score predicts the outcomes of diabetic kidney disease

BACKGROUND

The significance of renal arteriosclerosis in the prediction of the renal outcomes of diabetic kidney disease (DKD) remains undetermined.

METHODS

We enrolled 174 patients with DKD from three centres from January 2010 to July 2017. The severity and extent of arteriosclerosis were analysed on sections based on dual immunohistochemical staining of CD31 and α-smooth muscle actin. An X-tile plot was used to determine the optimal cut-off value. The primary endpoint was renal survival (RS), defined as the duration from renal biopsy to end-stage renal disease or death.

RESULTS

The baseline estimated glomerular filtration rate (eGFR) of 135 qualified patients was 45 (29 ~ 70) ml/min per 1.73 m², and the average 24-h urine protein was 4.52 (2.45 ~ 7.66) g/24 h. The number of glomeruli in the biopsy specimens was 21.07 ± 9.7. The proportion of severe arteriosclerosis in the kidney positively correlated with the Renal Pathology Society glomerular classification (r = 0.28, P < 0.012), interstitial fibrosis and tubular atrophy (IFTA) (r = 0.39, P < 0.001), urine protein (r = 0.213, P = 0.013), systolic BP (r = 0.305, P = 0.000), and age (r = 0.220, P = 0.010) and significantly negatively correlated with baseline eGFR (r = - 0.285, P = 0.001). In the multivariable model, the primary outcomes were significantly correlated with glomerular class (HR: 1.72, CI: 1.15 ~ 2.57), IFTA (HR: 1.96, CI: 1.26 ~ 3.06) and the modified arteriosclerosis score (HR: 2.21, CI: 1.18 ~ 4.13). After risk adjustment, RS was independently associated with the baseline eGFR (HR: 0.97, CI: 0.96 ~ 0.98), urine proteinuria (HR: 1.10, CI: 1.04 ~ 1.17) and the modified arteriosclerosis score (HR: 2.01, CI: 1.10 ~ 3.67), and the nomogram exhibited good calibration and acceptable discrimination (C-index = 0.82, CI: 0.75 ~ 0.87).

CONCLUSIONS

The severity and proportion of arteriosclerosis may be helpful prognostic indicators for DKD.

Diabetic Nephropathy without Diabetes

Diabetic nephropathy without diabetes (DNND), previously known as idiopathic nodular glomerulosclerosis, is an uncommon entity and thus rarely suspected; diagnosis is histological once diabetes is discarded. In this study we describe two new cases of DNND and review the literature. We analyzed all the individualized data of previous publications except one series of attached data. DNND appears to be favored by recognized cardiovascular risk factors. However, in contrast with diabetes, apparently no factor alone has been demonstrated to be sufficient to develop DNND. Other factors not considered as genetic and environmental factors could play a role or interact. The most plausible hypothesis for the occurrence of DNND would be a special form of atherosclerotic or metabolic glomerulopathy than can occur with or without diabetes. The clinical spectrum of cardiovascular risk factors and histological findings support this theory, with hypertension as one of the characteristic clinical features.