Machine learning-based reproducible prediction of type 2 diabetes subtypes.

AIMS/HYPOTHESIS: Clustering-based subclassification of type 2 diabetes, which reflects pathophysiology and genetic predisposition, is a promising approach for providing personalised and effective therapeutic strategies. Ahlqvist's classification is currently the most vigorously validated method because of its superior ability to predict diabetes complications but it does not have strong consistency over time and requires HOMA2 indices, which are not routinely available in clinical practice and standard cohort studies. We developed a machine learning (ML) model to classify individuals with type 2 diabetes into Ahlqvist's subtypes consistently over time. METHODS: Cohort 1 dataset comprised 619 Japanese individuals with type 2 diabetes who were divided into training and test sets for ML models in a 7:3 ratio. Cohort 2 dataset, comprising 597 individuals with type 2 diabetes, was used for external validation. Participants were pre-labelled (T2Dkmeans) by unsupervised k-means clustering based on Ahlqvist's variables (age at diagnosis, BMI, HbA1c, HOMA2-B and HOMA2-IR) to four subtypes: severe insulin-deficient diabetes (SIDD), severe insulin-resistant diabetes (SIRD), mild obesity-related diabetes (MOD) and mild age-related diabetes (MARD). We adopted 15 variables for a multiclass classification random forest (RF) algorithm to predict type 2 diabetes subtypes (T2DRF15). The proximity matrix computed by RF was visualised using a uniform manifold approximation and projection. Finally, we used a putative subset with missing insulin-related variables to test the predictive performance of the validation cohort, consistency of subtypes over time and prediction ability of diabetes complications. RESULTS: T2DRF15 demonstrated a 94% accuracy for predicting T2Dkmeans type 2 diabetes subtypes (AUCs ≥0.99 and F1 score [an indicator calculated by harmonic mean from precision and recall] ≥0.9) and retained the predictive performance in the external validation cohort (86.3%). T2DRF15 showed an accuracy of 82.9% for detecting T2Dkmeans, also in a putative subset with missing insulin-related variables, when used with an imputation algorithm. In Kaplan-Meier analysis, the diabetes clusters of T2DRF15 demonstrated distinct accumulation risks of diabetic retinopathy in SIDD and that of chronic kidney disease in SIRD during a median observation period of 11.6 (4.5-18.3) years, similarly to the subtypes using T2Dkmeans. The predictive accuracy was improved after excluding individuals with low predictive probability, who were categorised as an 'undecidable' cluster. T2DRF15, after excluding undecidable individuals, showed higher consistency (100% for SIDD, 68.6% for SIRD, 94.4% for MOD and 97.9% for MARD) than T2Dkmeans. CONCLUSIONS/INTERPRETATION: The new ML model for predicting Ahlqvist's subtypes of type 2 diabetes has great potential for application in clinical practice and cohort studies because it can classify individuals with missing HOMA2 indices and predict glycaemic control, diabetic complications and treatment outcomes with long-term consistency by using readily available variables. Future studies are needed to assess whether our approach is applicable to research and/or clinical practice in multiethnic populations.

Young-onset type 2 diabetes mellitus enhances proteinuria, but not glomerular filtration rate decline: A Japanese cohort study.

AIMS/INTRODUCTION: The time course of chronic kidney disease in young-onset type 2 diabetes mellitus remains unclear. We compared the trajectories of proteinuria and estimated glomerular filtration rate (eGFR) decline between young-onset (aged ≤40 years) and late-onset (aged >40 years) type 2 diabetes mellitus in a Japanese multicenter cohort. MATERIALS AND METHODS: Participants without diabetic kidney disease were divided into two groups according to age at diagnosis: young- and late-onset. The primary endpoint was eGFR <60 mL/min/1.73 m2, proteinuria or both. Multivariable Cox proportional hazards were calculated to estimate incidence. RESULTS: Among 626 participants with type 2 diabetes mellitus, 78 (12.4%) had young-onset and 548 (87.6%) had late-onset diabetes. The incidence of eGFR <60 mL/min/1.73 m2 was lower (16.7% vs 33.5%, P = 0.003), but that of proteinuria was higher (46.2% vs 28.9%, P = 0.002) in the young-onset type 2 diabetes mellitus group. The Kaplan-Meyer curve showed that young-onset type 2 diabetes mellitus was associated with a decreased hazard ratio (HR) for eGFR <60 mL/min/1.73 m2 and an increased HR for proteinuria compared with late-onset type 2 diabetes mellitus. In the multivariate Cox analysis, young-onset type 2 diabetes mellitus increased the HR (95% confidence interval) of proteinuria (1.53, 95% confidence interval 1.03-2.26), but did not change the eGFR <60 mL/min/1.73 m2 HR. CONCLUSIONS: Young-onset type 2 diabetes mellitus has a lower HR of eGFR <60 mL/min/1.73 m2 and an increased HR of proteinuria compared with late-onset type 2 diabetes mellitus, indicating that young-onset type 2 diabetes mellitus has a different time course for the development of proteinuria and subsequent eGFR decline.

Respiratory Displacement of the Right Adrenal Vein: Comparison of Inspiratory and Expiratory Computed Tomography With Catheter Venography.

OBJECTIVE: The aim of the study is to reveal the respiratory displacement of the right adrenal vein (RAV) to predict the exact location of the RAV during adrenal venous sampling (AVS). METHODS: Computed tomography (CT) scans obtained 45 seconds (breath-hold at inhalation) and 70 seconds (breath-hold at exhalation) after contrast material injection were compared to venograms of the RAV of patients with primary aldosteronism who underwent AVS between January 2016 and December 2020. The craniocaudal distance between the center of the Th11/12 disc and the RAV orifice was measured; the craniocaudal location of the RAV orifice was also specified relative to vertebral bodies and intervertebral discs on inspiratory phase CT (In-CT), expiratory phase CT (Ex-CT), and catheter venography. The transverse and vertical angles of the RAV and the position of the RAV orifice on the inferior vena cava (IVC) circumference were measured on In-CT and Ex-CT. RESULTS: In total, 51 patients (30 males, 21 females; mean age, 54.9 ± 11.1 years) were included. Craniocaudal distances between the center of the Th11/12 disc and RAV orifice were significantly different among the following 3 acquisitions: catheter venography versus In-CT (15.2 ± 8.4 mm); venography versus Ex-CT (5.6 ± 4.1 mm); and In-CT versus Ex-CT (19.6 ± 8.0 mm) (all, P < 0.001). The craniocaudal location of the RAV orifice on venography was significantly closer to that on Ex-CT than on In-CT (P < 0.001); measurements using venograms compared with In-CT and Ex-CT scans were within 1 level difference in 18 (35.3%) and 47 (92.2%) patients, respectively (P < 0.001). The vertical angle of the RAV was significantly more likely to be smaller on In-CT than on Ex-CT (P < 0.001). CONCLUSIONS: RAV locations and angles change with respiratory motion. It is crucial to consider the respiratory phase of CT because it can enable a more accurate prediction of the location of the RAV during AVS.

Role of perirenal adiposity in renal dysfunction among CKD individuals with or without diabetes: a Japanese cross-sectional study.

INTRODUCTION: It remains unclear whether increased perirenal fat (PRF) accumulation is equally related to renal involvement in patients with and without diabetes mellitus (DM). We evaluated the association between PRF volume (PRFV) and low glomerular filtration rate (GFR) and proteinuria in people with or without type 2 diabetes mellitus (T2DM). RESEARCH DESIGN AND METHODS: We performed a cross-sectional analysis of 473 individuals without T2DM (non-DM, n=202) and with T2DM (DM, n=271). PRFV (cm3), obtained from non-contrast CT, was indexed as PRF index (PRFV/body surface area, cm3/m2). Multivariate-adjusted models were used to determine the ORs of PRFV and PRFV index for detecting estimated GFR (eGFR) decrease of <60 mL/min/1.73 m2 proteinuria onset, or both. RESULTS: Although body mass index (BMI), visceral fat area, and waist circumference were comparable between the non-DM and DM groups, kidney volume, PRFV, and PRFV index were higher in individuals with T2DM than in those without T2DM. In the multivariate analysis, after adjusting for age, sex, BMI, hypertension, smoking history, and visceral fat area ≥100 cm2, the cut-off values of PRFV index were associated with an eGFR<60 in individuals with DM (OR 6.01, 95% CI 2.20 to 16.4, p<0.001) but not in those without DM. CONCLUSIONS: PRFV is associated with low eGFR in patients with T2DM but not in those without T2DM. This suggests that PRF accumulation is more closely related to the onset and progression of diabetic kidney disease (DKD) than non-DKD. Clarifying the mechanisms through which PRF influences DKD development could pave the way for novel prevention and treatment strategies.

A Collision Tumor of Pit-1/SF-1-positive Double Pituitary Adenoma and a Craniopharyngioma Coexisting with Graves' Disease.

Double or multiple pituitary adenomas expressing different types of transcription factors and collision tumors of pituitary adenomas and craniopharyngiomas are rare. In this report, we present a case of pituitary adenoma of two different cell populations, Pit-1 and SF-1, and an adenoma and craniopharyngioma collision tumor with coexisting Graves' disease. The patient had a 16-mm pituitary tumor with pituitary stalk calcification and optic chiasm compression but no visual dysfunction. Based on hormonal profile results, the tumor in the sella was considered a nonfunctioning pituitary adenoma; nevertheless, the pituitary stalk was invaded by a different lesion, which was later confirmed to be a craniopharyngioma. Using an endoscopic endonasal approach, the pituitary adenoma was removed; however, a small remnant remained medial to the right cavernous sinus. Because the pituitary stalk lesion was isolated from the pituitary adenoma, it was preserved to maintain pituitary function. Three years after the initial surgery, the patient suffered from Graves' disease and was treated with antithyroid medications. However, the intrasellar residual and pituitary stalk lesions gradually increased in size. A second surgery was performed, and the residual intrasellar and stalk lesions were completely removed. As per the initial and second histopathologies, the pituitary adenoma comprised different cell groups positive for thyroid-(TSH) and follicle-stimulating hormones, and each cell group was positive for Pit-1 and SF-1. The pituitary stalk lesion was an adamantinomatous craniopharyngioma. We believe that TSH-producing adenoma was involved in the development of Graves' disease or that treatment for Graves' disease increased TSH-producing adenoma.

A nutritional assessment tool, GNRI, predicts sarcopenia and its components in type 2 diabetes mellitus: A Japanese cross-sectional study.

Background: There are few reports evaluating the relationship between undernutrition and the risk of sarcopenia in type 2 diabetes mellitus (T2DM) patients. Objective: We investigated whether undernutritional status assessed by the geriatric nutritional risk index (GNRI) and controlling nutritional status (CONUT) were associated with the diagnosis of sarcopenia. Methods: This was a cross-sectional study of Japanese individuals with T2DM. Univariate or multivariate logistic regression analysis was performed to assess the association of albumin, GNRI, and CONUT with the diagnosis of sarcopenia. The optimal cut-off values were determined by the receiver operating characteristic (ROC) curve to diagnose sarcopenia. Results: In 479 individuals with T2DM, the median age was 71 years [IQR 62, 77], including 264 (55.1%) men. The median duration of diabetes was 17 [11, 23] years. The prevalence of sarcopenia was 41 (8.6%) in all, 21/264 (8.0%) in men, and 20/215 (9.3%) in women. AUCs were ordered from largest to smallest as follows: GNRI > albumin > CONUT. The cut-off values of GNRI were associated with a diagnosis of sarcopenia in multiple logistic regression analysis (odds ratio 9.91, 95% confidential interval 5.72-17.2), P < 0.001. The superiority of GNRI as compared to albumin and CONUT for detecting sarcopenia was also observed in the subclasses of men, women, body mass index (BMI) < 22, and BMI ≥ 22. Conclusions: Results showed that GNRI shows a superior diagnostic power in the diagnosis of sarcopenia. Additionally, its optimal cut-off points were useful overall or in the subclasses. Future large and prospective studies will be required to confirm the utility of the GNRI cut-off for undernutrition individuals at risk for sarcopenia.

Ectopic adrenocorticotropic hormone-secreting carcinoid with solitary cryptococcosis in the lungs.

BACKGROUND: Carcinoid tumors can on rare occasions ectopically produce adrenocorticotropic hormone (ACTH), causing Cushing's syndrome, and patients could become immunocompromised. Care must therefore be taken regarding infectious complications. In particular, ACTH-producing pulmonary carcinoid is not easy to diagnose by itself, and when combined with pulmonary nodules as infectious foci, each is very difficult to diagnose. CASE: The patient was a 71-year-old woman with refractory diabetes. She showed clinical symptoms of Cushing's syndrome during treatment for diabetes and ectopic ACTH production was suspected based on biochemical and imaging tests. Nodules were identified in the left lung apex and lingual segment. Examination of resected nodules revealed that the nodule in the apex was pulmonary cryptococcosis, while the nodule in the lingual segment represented typical carcinoid. After surgery, clinical symptoms, laboratory findings, and diabetes all improved. CONCLUSION: We present this very instructive case in terms of the difficulty of diagnosing ACTH-producing tumors, the possibility of infection complicating the immunodeficiency caused by ACTH-producing tumors, and the surgical strategy.

Detecting Sarcopenia Risk by Diabetes Clustering: A Japanese Prospective Cohort Study.

CONTEXT: Previous studies have assessed the usefulness of data-driven clustering for predicting complications in patients with diabetes mellitus. However, whether the diabetes clustering is useful in predicting sarcopenia remains unclear. OBJECTIVE: To evaluate the predictive power of diabetes clustering for the incidence of sarcopenia in a prospective Japanese cohort. DESIGN: Three-year prospective cohort study. SETTING AND PATIENTS: We recruited Japanese patients with type 1 or type 2 diabetes mellitus (n = 659) between January 2018 and February 2020 from the Fukushima Diabetes, Endocrinology, and Metabolism cohort. INTERVENTIONS: Kaplan-Meier and Cox proportional hazards models were used to measure the predictive values of the conventional and clustering-based classification of diabetes mellitus for the onset of sarcopenia. Sarcopenia was diagnosed according to the Asian Working Group for Sarcopenia (AWGS) 2019 consensus update. MAIN OUTCOME MEASURES: Onset of sarcopenia. RESULTS: Cluster analysis of a Japanese population revealed 5 diabetes clusters: cluster 1 [severe autoimmune diabetes (SAID)], cluster 2 [severe insulin-deficient diabetes (SIDD)], cluster 3 (severe insulin-resistant diabetes, cluster 4 (mild obesity-related diabetes), and cluster 5 (mild age-related diabetes). At baseline, 38 (6.5%) patients met the AWGS sarcopenia criteria, and 55 had newly developed sarcopenia within 3 years. The SAID and SIDD clusters were at high risk of developing sarcopenia after correction for known risk factors. CONCLUSIONS: This study reveals that among the 5 diabetes clusters, the SAID and SIDD clusters are at a high risk for developing sarcopenia. Clustering-based stratification may be beneficial for predicting and preventing sarcopenia in patients with diabetes.

Thoracoscopic resection of posterior mediastinal paraganglioma: perioperative management and surgical tips.

Posterior mediastinal paraganglioma (PM-PGL) is a rare disease that is difficult to diagnose. If PM-PGL is misdiagnosed preoperatively, surgeons may encounter severe tachycardia and hypertension and easy bleeding from the tumor during the operation. Therefore, it is essential to include PGL as a differential diagnosis for mediastinal tumors. We herein describe a 73-year-old Japanese man with a PM-PGL that was diagnosed preoperatively and resected safely by video-assisted thoracic surgery. Preoperative management of hypertension with doxazosin mesylate, soft coagulation of the peritumor area, and careful clipping of feeding arteries were effective for hemostasis. The patient's vital signs were stable during and after the operation.

Effect of a multitarget therapy with prednisolone, mycophenolate mofetil, and tacrolimus in a patient with type B insulin resistance syndrome complicated by lupus nephritis.

Type B insulin resistance syndrome (TBIR) is a rare autoimmune disease characterised by autoantibodies targeting insulin receptors. TBIR is often complicated by systemic lupus erythematosus (SLE). We describe the case of a 59-year-old Japanese man with TBIR complicated with lupus nephritis (LN), who presented with nephrotic syndrome and severe hypoglycaemia. Treatment with prednisolone (PSL), mycophenolate mofetil (MMF), and tacrolimus (TAC) resulted in improved SLE activity and glucose intolerance with the reduction of anti-insulin receptor autoantibodies. To the best of our knowledge, this is the first reported case of TBIR complicated with LN that was successfully treated using multitarget therapy with PSL, MMF, and TAC.

Novel strategies for glycaemic control and preventing diabetic complications applying the clustering-based classification of adult-onset diabetes mellitus: A perspective.

Diabetes mellitus results from an interplay between insulin resistance and ß-cell dysfunction. Since their relative contributions to its pathogenesis are difficult to quantify, therapeutic strategies for glycaemic control are determined primarily based on two limited metrics: plasma glucose and haemoglobin A1c. Recent attempts have been made to subclassify diabetes mellitus to better predict its associated pathology and plan appropriate therapeutic strategies. These classifications are based on data-driven cluster analysis using autoimmunity, age, obesity (metabolically unhealthy and healthy phenotypes), insulin secretory capacity and resistance, and ethnicity. This review addresses potential therapeutic strategies for the cluster-based classifications of adult-onset diabetes mellitus to achieve better glycaemic control and prevent or at least delay the concomitant complications.

High FIB4 index is an independent risk factor of diabetic kidney disease in type 2 diabetes.

Nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) may be linked to development of chronic kidney diseases (CKD). The FIB4 index, a noninvasive liver fibrosis score, has been reported to predict CKD in non-diabetic patients, but there are no reports yet in diabetic cases. Therefore, we evaluated the prognostic impact of FIB4 index on the risk of developing diabetic kidney disease (DKD) in Japanese patients with type 2 diabetes in a retrospective cohort study. We assessed patients with type 2 diabetes with an eGFR ≥ 60 mL/min/1.73 m2 and without dipstick positive proteinuria (≥ 1 +) at their first visit to our department. Participants were divided into two groups based on the FIB4 index at their first visit: FIB4 index > 1.3 and FIB4 index ≤ 1.3. The primary endpoint was defined as a decrease in eGFR < 60 mL/min/1.73 m2 or the onset of proteinuria during the course of treatment. The average age of all 584 type 2 diabetic participants (360 [61.6%] men) was 55 ± 11 years. There were 187 patients in the FIB4 index group > 1.3 (32.0%) and the median observation period was 6.0 (3.8-11.0) years. Kaplan-Meier survival analysis indicated that the risks of developing DKD, eGFR < 60 and proteinuria were all higher in FIB4 index > 1.3 patients than in FIB4 ≤ 1.3 patients. In the Cox regression analysis, an FIB4 index > 1.3 was a significant predictor for onset of DKD (HR 1.54, 95% CI 1.15-2.08) and proteinuria (HR 1.55, 95% CI 1.08-2.23), but not for an eGFR < 60 (HR 1.14, 95% CI 0.79-1.99). To the best of our knowledge, this is the first study to demonstrate that an FIB4 index > 1.3 has a prognostic impact on the development of CKD and proteinuria in type 2 diabetic patients. This warrants further investigation of the prognostic impact of the development of DKD or proteinuria.

Walking Speed is the Sole Determinant Criterion of Sarcopenia of Mild Cognitive Impairment in Japanese Elderly Patients with Type 2 Diabetes Mellitus.

Diabetes mellitus is a risk factor for mild cognitive impairment (MCI) and dementia. However, how the clinical characteristics of MCI patients with type 2 diabetes mellitus are linked to sarcopenia and/or its criteria remain to be elucidated. Japanese patients with type 2 diabetes mellitus were categorized into the MCI group for MoCA-J (the Japanese version of the Montreal cognitive assessment) score <26, and into the non-MCI group for MoCA-J ≥26. Sarcopenia was defined by a low skeletal mass index along with low muscle strength (handgrip strength) or low physical performance (walking speed <1.0 m/s). Univariate and multivariate-adjusted odds ratio models were used to determine the independent contributors for MoCA-J <26. Among 438 participants, 221 (50.5%) and 217 (49.5%) comprised the non-MCI and MCI groups, respectively. In the MCI group, age (61 ± 12 vs. 71 ± 10 years, p < 0.01) and duration of diabetes mellitus (14 ± 9 vs. 17 ± 9 years, p < 0.01) were higher than those in the non-MCI group. Patients in the MCI group exhibited lower hand grip strength, walking speed, and skeletal mass index, but higher prevalence of sarcopenia. Only walking speed (rather than muscle loss or muscle weakness) was found to be an independent determinant of MCI after adjusting for multiple factors, such as age, gender, body mass index (BMI), duration of diabetes mellitus, hypertension, dyslipidemia, smoking, drinking, estimated glomerular filtration rate (eGFR), HbA1c, and history of coronary heart diseases and stroke. In subgroup analysis, a group consisting of male patients aged ≥65 years, with BMI <25, showed a significant OR for walking speed. This study showed that slow walking speed is a sole determinant criterion of sarcopenia of MCI in patients with type 2 diabetes mellitus. It was suggested that walking speed is an important factor in the prediction and prevention of MCI development in patients with diabetes mellitus.

Factors Associated with Risk of Diabetic Complications in Novel Cluster-Based Diabetes Subgroups: A Japanese Retrospective Cohort Study.

Diabetes is a complex and heterogeneous disease, making the prediction of the risks of diabetic complications challenging. Novel adult-onset diabetes subgroups have been studied using cluster analysis, but its application in East Asians remains unclear. We conducted a retrospective cohort study to elucidate the clinical utility of cluster-based subgroup analysis in the Japanese population. Cluster analysis based on anti-glutamate decarboxylase antibody (GAD antibody) levels, age at diagnosis, body mass index (BMI), hemoglobin A1c (A1c), and homeostatic model assessment 2 estimates of ß-cell function and insulin resistance was performed in 1520 diabetic patients. The risk of developing diabetic complications was analyzed using Kaplan-Meier analysis and the Cox proportional hazards model. By cluster analysis, we identified five distinct subgroups of adult-onset diabetes in the Japanese population. The risk of diabetic complications varied greatly among the clusters. Patients with severe autoimmune diabetes or severe insulin deficiency diabetes were at an increased risk of diabetic retinopathy, and those with severe insulin resistant diabetes (SIRD) had the highest risk of developing diabetic kidney disease (DKD). After adjusting for uncorrectable and correctable risk factors, SIRD was found to be an independent risk factor for DKD. In conclusion, we identified five subgroups of adult-onset diabetes and the risk factors for diabetic complications in the Japanese population. This new classification system can be effective in predicting the risk of diabetic complications and for providing optimal treatment.

Burden of Undiagnosed Type 2 Diabetes in Diabetic Kidney Disease: A Japanese Retrospective Cohort Study.

The risk of developing diabetic kidney disease (DKD) in patients with undiagnosed diabetes mellitus (UD) has never been evaluated. We studied the burden of UD on the risk of developing DKD in the Japanese population in a single-center retrospective cohort study. The patients with type 2 diabetes mellitus, but without DKD (estimated glomerular filtration rate (eGFR) < 60 mL/min/1.73 m2 or proteinuria), were recruited from January 2018 to January 2019; medical records were scrutinized retrospectively from January 2003 until May 2019. The individuals, with diabetes that could not be denied based on past and current records, comprised the undiagnosed diabetes (UD) group whereas those with confirmed diagnosis comprised the diagnosed diabetes (DD) group. The group differences were tested using a Kaplan-Meier curve and Cox proportional hazards model. Among the 408 participants, 164 (40.2%) and 244 (59.8%) comprised the DD and UD groups, respectively. The baseline parameters, including age, male gender, and BMI were comparable between the groups, but the plasma glucose, HbA1c levels, and diabetic retinopathy prevalence were higher in the UD group. The risk of developing DKD (log rank test, p < 0.001), an eGFR of < 60 mL/min/1.73 m2 (p = 0.001) and proteinuria (p = 0.007) were also higher in the UD group. The unadjusted and adjusted hazard ratios for DKD were 1.760 ((95% CI: 1.323-2.341), p < 0.001) and 1.566 ((95% CI: 1.159-2.115), p = 0.003), respectively, for the UD group. In conclusion, this is the first report showing that UD is a strong risk factor for DKD. The notion that a longer duration of untreated diabetes mellitus is involved strongly in the risk of developing DKD warrants the need for the identification and monitoring of UD patients.