Machine learning models for prediction of HF and CKD development in early-stage type 2 diabetes patients.

Chronic kidney disease (CKD) and heart failure (HF) are the first and most frequent comorbidities associated with mortality risks in early-stage type 2 diabetes mellitus (T2DM). However, efficient screening and risk assessment strategies for identifying T2DM patients at high risk of developing CKD and/or HF (CKD/HF) remains to be established. This study aimed to generate a novel machine learning (ML) model to predict the risk of developing CKD/HF in early-stage T2DM patients. The models were derived from a retrospective cohort of 217,054 T2DM patients without a history of cardiovascular and renal diseases extracted from a Japanese claims database. Among algorithms used for the ML, extreme gradient boosting exhibited the best performance for CKD/HF diagnosis and hospitalization after internal validation and was further validated using another dataset including 16,822 patients. In the external validation, 5-years prediction area under the receiver operating characteristic curves for CKD/HF diagnosis and hospitalization were 0.718 and 0.837, respectively. In Kaplan-Meier curves analysis, patients predicted to be at high risk showed significant increase in CKD/HF diagnosis and hospitalization compared with those at low risk. Thus, the developed model predicted the risk of developing CKD/HF in T2DM patients with reasonable probability in the external validation cohort. Clinical approach identifying T2DM at high risk of developing CKD/HF using ML models may contribute to improved prognosis by promoting early diagnosis and intervention.

Interpretation Method for Continuous Glucose Monitoring with Subsequence Time-Series Clustering.

We propose mini-batch top-n k-medoids to sequential pattern mining to improve CGM interpretation. Mecical workers can treat specific patient groups better by understanding the time series variation of blood glucose results. For 10 years, continuous glucose monitoring (CGM) has provided time-series data of blood glucose thanks to the invention of devices with low measurement errors. We conducted two experiments. In the first experiment, we evaluated the proposed method with a manually created dataset and confirmed that the method provides more accurate patterns than other clustering methods. In the second experiment, we applied the proposed method to a CGM dataset consisting of real data from 163 patients. We created two labels based on blood glucose (BG) statistics and found patterns that correlated with a specific label in each case.

Feature Set for a Prediction Model of Diabetic Kidney Disease Progression.

In this paper, we propose feature extraction method for prediction model for at the early stage of diabetic kidney disease (DKD) progression. DKD needs continuous treatment; however, a hospital visit interval of a patient at the early stage of DKD is normally from one month to three months, and this is not a short time period. Therefore it makes difficult to apply sophisticated approaches such as using convolutional neural networks because of the data limitation. The propose method uses with hierarchical clustering that can estimate a suitable interval for grouping inputted sequences. We evaluate the proposed method with a real-EMR dataset that consists of 30,810 patient records and conclude that the proposed method outperforms the baseline methods derived from related work.

Changes in Serum Immunoglobulin G4 Levels in Patients with Newly Diagnosed Graves' Disease.

OBJECTIVE: Immunoglobulin G4 (IgG4)-related disease (IgG4-RD) is an immune-mediated condition that can affect almost any organ. We investigated the association between IgG4-RD and the main characteristics of Graves' disease (GD) at the time of diagnosis. Additionally, we evaluated whether serum IgG4 levels change during treatment. DESIGN AND PATIENTS: Twenty-eight patients with newly diagnosed GD were enrolled into this longitudinal follow-up study. Serum IgG4 levels and thyroid function were measured in all the participants at the time of diagnosis. Further, the serum IgG4 levels of nine of 28 patients with untreated GD were measured after the achievement of euthyroid state (through the use of methimazole). RESULTS: Two (7.1%) of 28 patients with untreated GD had elevated serum IgG4 levels of >135 mg/dL. There was no significant difference in the average IgG4 levels before and after the achievement of euthyroid state (66.2±74.0 mg/dL vs. 50.5±47.3 mg/dL). In two patients, the elevated serum IgG4 levels returned to normal after treatment. However, one patient had an elevated serum IgG4 level of 136.6 mg/dL after treatment. CONCLUSIONS: This study showed that serum IgG4 levels varied with treatment in patients with GD, independent of thyroid function, suggesting that IgG4 might be indirectly related to GD.

Low-carbohydrate diet by staple change attenuates postprandial GIP and CPR levels in type 2 diabetes patients.

AIMS: The aim of this study is to investigate the effects of a low-carbohydrate staple food (i.e., low-carbohydrate bread) on glucose and lipid metabolism and pancreatic and enteroendocrine hormone secretion in comparison with meals containing normal-carbohydrate bread, without consideration of the carbohydrate content of the side dishes. METHODS: T2DM patients (n = 41) were provided meals containing low-carbohydrate bread (LB) together with side dishes or normal-carbohydrate bread (NB) together with side dishes every other day as a breakfast. Blood glucose levels were evaluated by using a continuous glucose monitoring system; blood samples were collected before and 1 and 2 h after the breakfast. RESULTS: Postprandial blood glucose levels, plasma insulin, plasma glucose-dependent insulinotropic polypeptide (GIP) and plasma triglyceride were significantly lower and plasma glucagon levels were significantly higher in LB compared with those in NB. Plasma glucagon-like peptide-1 (GLP-1) levels did not differ in the LB and NB groups. CONCLUSIONS: These results indicate that changing only the carbohydrate content of the staple food has benefits on glucose and lipid metabolism in T2DM patients concomitant with the decrease of insulin and GIP secretion, which ameliorate body weight gain and insulin resistance.

Artificial intelligence predicts the progression of diabetic kidney disease using big data machine learning.

Artificial intelligence (AI) is expected to support clinical judgement in medicine. We constructed a new predictive model for diabetic kidney diseases (DKD) using AI, processing natural language and longitudinal data with big data machine learning, based on the electronic medical records (EMR) of 64,059 diabetes patients. AI extracted raw features from the previous 6 months as the reference period and selected 24 factors to find time series patterns relating to 6-month DKD aggravation, using a convolutional autoencoder. AI constructed the predictive model with 3,073 features, including time series data using logistic regression analysis. AI could predict DKD aggravation with 71% accuracy. Furthermore, the group with DKD aggravation had a significantly higher incidence of hemodialysis than the non-aggravation group, over 10 years (N = 2,900). The new predictive model by AI could detect progression of DKD and may contribute to more effective and accurate intervention to reduce hemodialysis.

Risk Prediction of Diabetic Nephropathy via Interpretable Feature Extraction from EHR Using Convolutional Autoencoder.

This paper describes a technology for predicting the aggravation of diabetic nephropathy from electronic health record (EHR). For the prediction, we used features extracted from event sequence of lab tests in EHR with a stacked convolutional autoencoder which can extract both local and global temporal information. The extracted features can be interpreted as similarities to a small number of typical sequences of lab tests, that may help us to understand the disease courses and to provide detailed health guidance. In our experiments on real-world EHRs, we confirmed that our approach performed better than baseline methods and that the extracted features were promising for understanding the disease.

Simultaneous measurement of serum chemokines in autoimmune thyroid diseases: possible role of IP-10 in the inflammatory response.

Autoimmune thyroid diseases (AITDs), including Graves' diseases (GD) and Hashimoto's thyroiditis (HT), are the most common autoimmune diseases, and are mainly mediated by T cells that produce cytokines and chemokines in abnormal amounts. Few reports have described the circulating chemokines active in AITDs. Recently, we used a new multiplex immunobead assay to simultaneously measure cytokines and chemokines in small volume serum samples from patients with AITDs. We measured 23 selected serum chemokines in patients with GD (n=45) or HT (n=26), and healthy controls (n=9). GD patients were further classified as either untreated, intractable, or in remission, while HT patients were classified as either hypothyroid or euthyroid. Of the 23 serum chemokines assayed, only the serum level of IP-10 (CXCL10/interferon-γ-inducible protein 10) was elevated, depending on disease activity, in GD or HT compared with healthy controls. However, the serum level of IP-10 was also increased in both untreated GD patients and hypothyroid HT patients, suggesting that levels of this cytokine may not be affected by disease specificity. In conclusion, autoimmune inflammation in patients with AITD is closely related to the level of the serum chemokine, IP-10. Therefore, IP-10 might be a good biomarker for tissue inflammation in the thyroid, but not a useful biomarker for predicting disease specific activity, the progression of AITDs, or responsiveness to treatment because of its independence from thyroid function or disease specificity.

Relative levels of the inflammatory cytokine TNFα and the soluble CD40 ligand profile in serum correlate with the thyrotoxic activity of Graves' disease.

BACKGROUND: Interactions between CD40 and its ligand (CD40L) have important roles in T-cell-dependent activation of B cells, which may be related to the thyrotoxic activity of Graves' disease (GD). Soluble forms of CD40 ligand (sCD40L) are released from activated T cells and platelets, and several types of inflammatory cytokines are increased in patients with hyperthyroid GD. The aim of this study was to assess sCD40L and other cytokines as clinical indicators of disease activity or as possible markers of remission in GD. METHODS: Serum levels of sCD40L, interleukin 18 (IL-18), tumor necrosis factor-alpha (TNFα), and TNFα receptors 1 and 2 (TNFR1 and TNFR2) were investigated in patients with active GD (GD-A), intractable GD (GD-IT), inactive GD (GD-IA), GD in remission (GD-R), and Hashimoto's thyroiditis (HT), and in control subjects (CON). RESULTS: Serum concentrations of sCD40L were higher in the GD-A and GD-IT groups than in the HT and CON groups. Similarly, serum concentrations of IL-18, which induces Th1 cytokines, such as interferon-γ, were higher in the GD-A and GD-IT groups than in all other groups. Serum levels of TNFR1 and TNFR2 were also significantly higher in the GD-A than in all other groups. The mean serum concentration of TNFα was higher in the GD-R compared with the GD-A and GD-IT groups, although the difference was not significant. Serum sCD40L concentrations in the GD-R group were lower than in the GD-A and GD-IT groups. Finally, the ratio of serum TNFα to sCD40L was higher in the GD-R group than in the GD-A and GD-IT groups. This is the first report that serum sCD40L is increased in active GD, and that the serum TNFα:sCD40L ratio is a marker for remission in GD. CONCLUSIONS: Our results suggest that not only thyrotoxicosis, but also the activity of the immunoreaction presenting as anti-thyrotropin receptor antibodies (TRAb) titer in GD, affects inflammatory cytokine serum profiles. Serum profiles of cytokines vary in patients with GD depending on disease activity. An elevated serum TNFα:sCD40L ratio indicates declining disease activity and reflects a shift from Th2 to Th1 dominance, suggesting that suppression of sCD40L or increased production of TNFα is required to initiate or maintain remission of GD.

Genetic variations in calpain-10 gene are not a major factor in the occurrence of type 2 diabetes in Japanese.

The 112/121 haplotype combination defined by the UCSNP-43, -19, and -63 alleles in the calpain-10 gene is associated with type 2 diabetes in Mexican Americans. To determine whether this genetic variation constitutes risk of type 2 diabetes in Japanese, we investigated its frequency in 177 patients with type 2 diabetes and 172 controls. Though this variation occurs in Japanese more frequently than in Mexican Americans, there is no significant difference in frequency between diabetic (29.9%) and control (31.9%) subjects. We also screened all exons and the putative promoter of the calpain-10 gene for mutations in 96 of the genotyped patients, resulting in the identification of 7 coding variants, including 3 missense mutations and 5 nucleotide alterations in the promoter. However, their frequencies all are similar in patients and controls, suggesting that these genetic variations are not a major factor in the occurrence of type 2 diabetes in Japanese, although they could yet be associated with various phenotypes of the disease.

Distinctive response of thyroid-infiltrating mononuclear cells to B cell activation through CD40 and interleukin-4 in Graves' patients.

The possible role of abnormal T cell-dependent B-cell activation in Graves' disease was investigated by comparing lymphocyte subset distribution and the production of soluble CD8 (sCD8), sCD23, IL-10 and IL-12 by peripheral blood cells (PBMC) and thyroid-infiltrating lymphocytes (TL) in vitro. In TL, the percentage of CD8(+) cells was slightly higher and the sCD8 concentration was significantly higher than in PBMC. The ratio CD23(+) cells to CD20(+) cells (activated B/pan B cells) was increased in TL compared to PBMC from Graves' or normal controls, although the percentage of CD20(+) cells was decreased. Compared to PBMC in Graves' disease, the relative ratio of IL-10 to IL-12 release (IL-10/IL-12) by unstimulated TL was increased, despite a lack of significant difference between PBMC and TL in mean values for either IL-10 or IL-12 secretion. Incubating PBMC with a combination of anti-CD40 monoclonal antibodies and interleukin-4 (IL-4) resulted in B cell activation, reflected in an increase in the sCD23 level in both controls and Graves' patients, but especially prominent in the latter. Stimulation with anti-CD40 antibody and IL-4 also decreased the percentage of CD8(+) cells in PBMC but not TL from both Graves' disease and normal controls, and the percentage of CD8(+) cells in TL was higher than PBMC after the stimulation. The sCD23 concentration in TL was decreased compared to PBMC both in patients with Graves' disease and normal controls. However, in contrast to the increased responses observed in Graves' PBMC or normal controls after stimulation, sCD23 levels remained the same in stimulated TL from Graves' patients. This combination of B cell stimulants increased production of IL-10 in PBMC but not in TL obtained from patients with Graves' disease, and the increased IL-10/IL-12 ratio declined to a value no different from that in PBMC group after stimulation. Thus, T cell-dependent B-cell activation via a CD40 pathway may cause a shift in the Th(1)/Th(2) balance to Th(2) dominance in Graves' disease, while increased CD8(+) cells in TL may suppress sCD23 production and IL-10-producing Th(2) cells.