Higher risk of incident diabetes among patients with primary hyperparathyroidism.

OBJECTIVES: There is relatively scarce data regarding the association between primary hyperparathyroidism (PHPT) and incident diabetes in large population-based longitudinal studies. We aimed to evaluate the risk of incident diabetes in individuals with and without PHPT and investigate the association between serum calcium concentrations and the risk of incident diabetes in patients with PHPT. METHODS: We included 2749 PHPT patients and 13,745 age, sex and index year matched non-PHPT individuals during 2000-2019. We used Cox regression models to compare the risk of incident diabetes in individuals with and without PHPT, and the risk of incident diabetes in PHPT patients with serum calcium concentration above and below the median value. The association between serum calcium concentrations and the risk of incident diabetes was examined by restricted cubic spline analyses in patients with PHPT. RESULTS: During a median follow-up time of 5.17 years (IQR 2.17, 9.58), 433 patients (15.75%) with PHPT and 2110 individuals (15.35%) without PHPT developed diabetes, respectively. Patients with PHPT had a higher incidence rate of diabetes compared to non-PHPT individuals (27.60 [95% CI 25.00, 30.30] vs. 23.90 [95% CI 22.80, 24.90] per 1000 person-years, log-rank test p = .007]. Crude Cox regression model showed PHPT was associated with a 15% higher risk of incident diabetes (HR 1.15, 95%CI 1.04, 1.28). In patients with PHPT, a 44% higher risk of incident diabetes was found in patients with serum calcium concentrations above the median value (2.63 mmol/L), compared to those below the median value (HR 1.44, 95%CI 1.08, 1.90). Restricted cubic spline analyses confirmed a positive linear association between serum calcium concentrations and the risk of incident diabetes in those with PHPT (p-value for nonlinear = .751) CONCLUSIONS: Patients with PHPT had a higher risk of incident diabetes compared to non-PHPT individuals. A positive linear association was found between serum calcium concentrations and the risk of incident diabetes in patients with PHPT.

Precision Medicine to Redefine Insulin Secretion and Monogenic Diabetes-Randomized Controlled Trial (PRISM-RCT) in Chinese patients with young-onset diabetes: design, methods and baseline characteristics.

INTRODUCTION: We designed and implemented a patient-centered, data-driven, holistic care model with evaluation of its impacts on clinical outcomes in patients with young-onset type 2 diabetes (T2D) for which there is a lack of evidence-based practice guidelines. RESEARCH DESIGN AND METHODS: In this 3-year Precision Medicine to Redefine Insulin Secretion and Monogenic Diabetes-Randomized Controlled Trial, we evaluate the effects of a multicomponent care model integrating use of information and communication technology (Joint Asia Diabetes Evaluation (JADE) platform), biogenetic markers and patient-reported outcome measures in patients with T2D diagnosed at ≤40 years of age and aged ≤50 years. The JADE-PRISM group received 1 year of specialist-led team-based management using treatment algorithms guided by biogenetic markers (genome-wide single-nucleotide polymorphism arrays, exome-sequencing of 34 monogenic diabetes genes, C-peptide, autoantibodies) to achieve multiple treatment goals (glycated hemoglobin (HbA1c) <6.2%, blood pressure <120/75 mm Hg, low-density lipoprotein-cholesterol <1.2 mmol/L, waist circumference <80 cm (women) or <85 cm (men)) in a diabetes center setting versus usual care (JADE-only). The primary outcome is incidence of all diabetes-related complications. RESULTS: In 2020-2021, 884 patients (56.6% men, median (IQR) diabetes duration: 7 (3-12) years, current/ex-smokers: 32.5%, body mass index: 28.40±5.77 kg/m2, HbA1c: 7.52%±1.66%, insulin-treated: 27.7%) were assigned to JADE-only (n=443) or JADE-PRISM group (n=441). The profiles of the whole group included positive family history (74.7%), general obesity (51.4%), central obesity (79.2%), hypertension (66.7%), dyslipidemia (76.4%), albuminuria (35.4%), estimated glomerular filtration rate <60 mL/min/1.73 m2 (4.0%), retinopathy (13.8%), atherosclerotic cardiovascular disease (5.2%), cancer (3.1%), emotional distress (26%-38%) and suboptimal adherence (54%) with 5-item EuroQol for Quality of Life index of 0.88 (0.87-0.96). Overall, 13.7% attained ≥3 metabolic targets defined in secondary outcomes. In the JADE-PRISM group, 4.5% had pathogenic/likely pathogenic variants of monogenic diabetes genes; 5% had autoantibodies and 8.4% had fasting C-peptide <0.2 nmol/L. Other significant events included low/large birth weight (33.4%), childhood obesity (50.7%), mental illness (10.3%) and previous suicide attempts (3.6%). Among the women, 17.3% had polycystic ovary syndrome, 44.8% required insulin treatment during pregnancy and 17.3% experienced adverse pregnancy outcomes. CONCLUSIONS: Young-onset diabetes is characterized by complex etiologies with comorbidities including mental illness and lifecourse events. TRIAL REGISTRATION NUMBER: NCT04049149.

Association of dipeptidyl peptidase-4 inhibitor initiation at glycated haemoglobin <7.5% with reduced major clinical events mediated by low glycated haemoglobin variability.

AIM: Therapeutic inertia, hypoglycaemia and poor treatment persistence can lead to glycaemic fluctuation and poor outcomes in type 2 diabetes (T2D). We compared glycated haemoglobin (HbA1c) variability, insulin initiation, severe hypoglycaemia and clinical events in patients with T2D initiated dipeptidyl peptidase-4 inhibitors (DPP4is) at low versus high HbA1c thresholds. METHODS: Using territory-wide electronic medical records in Hong Kong, we curated a propensity score-matched cohort of patients initiated DPP4i at HbA1c <7.5% versus ≥7.5% in 2007-2019. We expressed the HbA1c variability score (HVS) as a proportion of HbA1c varied by ≥0.5% compared with preceding values. We used the Cox model to compare the risks of insulin initiation and clinical outcomes, adjusted for time-varying variables between the two groups. Mediation analysis estimated the effects of HbA1c variability on outcomes. RESULTS: Among 6874 insulin-naïve patients who initiated DPP4i, 88.7% were treated with metformin and 79.6% with sulphonylureas at baseline (54.9% men; mean age 65.2 ± 11.4 years). After a median follow-up of 4.6 years, compared with the high-threshold plus high-HVS group (≥50%), the low-threshold plus low-HVS (<50%) group had reduced hazard ratios (95% confidence interval) of insulin initiation (0.35, 0.31-0.40), severe hypoglycaemia (0.38, 0.34-0.44), major adverse cardiovascular endpoints (0.76, 0.66-0.88), heart failure (0.42, 0.36-0.49), end-stage kidney disease (0.65, 0.36-0.49) and mortality (0.45, 0.35-0.57). Reduced HbA1c variability explained 31.1%-81.2% of the effect size of DPP4i initiation at HbA1c <7.5% versus ≥7.5% on outcomes. CONCLUSIONS: In Chinese patients with T2D, avoiding therapeutic inertia with intensified glycaemic control at HbA1c <7.5% using drugs with low risk of hypoglycaemia and good tolerability, such as DPP4i, delayed insulin treatment, reduced HbA1c variability and improved clinical events.

Monogenic diabetes in a Chinese population with young-onset diabetes: A 17-year prospective follow-up study in Hong Kong.

AIMS: Asians have a high prevalence of young-onset diabetes, but the pattern of monogenic diabetes is unknown. We aimed to determine the prevalence of monogenic diabetes in Chinese patients with young-onset diabetes and compare the clinical characteristics and outcome between patients with and without monogenic diabetes. MATERIALS AND METHODS: We sequenced a targeted panel of 33 genes related to monogenic diabetes in 1021 Chinese patients with non-type 1 diabetes diagnosed at age ≤40 years. Incident complications including cardiovascular disease (CVD), end-stage kidney disease (ESKD) and all-cause death were captured since enrolment (1995-2012) until 2019. RESULTS: In this cohort (mean ± SD age at diagnosis: 33.0 ± 6.0 years, median[IQR] diabetes duration 7.0[1.0-15.0] years at baseline, 44.9% men), 22(2.2%, 95% confidence interval[CI] 1.4%-3.2%) had monogenic diabetes. Pathogenic (P) or likely pathogenic (LP) variants were detected in GCK (n = 6), HNF1A (n = 9), HNF4A (n = 1), PLIN1 (n = 1) and PPARG (n = 2), together with copy number variations in HNF1B (n = 3). Over a median follow-up of 17.1 years, 5(22.7%) patients with monogenic diabetes (incidence rate 12.3[95% CI 5.1-29.4] per 1000 person-years) versus 254(25.4%) without monogenic diabetes (incidence rate 16.7[95% CI 14.8-18.9] per 1000 person-years) developed the composite outcome of CVD, ESKD and/or death (p = 0.490). The multivariable Cox model did not show any difference in hazards for composite events between groups. CONCLUSIONS: In Chinese with young-onset non-type 1 diabetes, at least 2% of cases were contributed by monogenic diabetes, over 80% of which were accounted for by P/LP variants in common MODY genes. The incidence of diabetes complications was similar between patients with and without monogenic diabetes.

A novel electronic health record-based, machine-learning model to predict severe hypoglycemia leading to hospitalizations in older adults with diabetes: A territory-wide cohort and modeling study.

BACKGROUND: Older adults with diabetes are at high risk of severe hypoglycemia (SH). Many machine-learning (ML) models predict short-term hypoglycemia are not specific for older adults and show poor precision-recall. We aimed to develop a multidimensional, electronic health record (EHR)-based ML model to predict one-year risk of SH requiring hospitalization in older adults with diabetes. METHODS AND FINDINGS: We adopted a case-control design for a retrospective territory-wide cohort of 1,456,618 records from 364,863 unique older adults (age ≥65 years) with diabetes and at least 1 Hong Kong Hospital Authority attendance from 2013 to 2018. We used 258 predictors including demographics, admissions, diagnoses, medications, and routine laboratory tests in a one-year period to predict SH events requiring hospitalization in the following 12 months. The cohort was randomly split into training, testing, and internal validation sets in a 7:2:1 ratio. Six ML algorithms were evaluated including logistic-regression, random forest, gradient boost machine, deep neural network (DNN), XGBoost, and Rulefit. We tested our model in a temporal validation cohort in the Hong Kong Diabetes Register with predictors defined in 2018 and outcome events defined in 2019. Predictive performance was assessed using area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC) statistics, and positive predictive value (PPV). We identified 11,128 SH events requiring hospitalization during the observation periods. The XGBoost model yielded the best performance (AUROC = 0.978 [95% CI 0.972 to 0.984]; AUPRC = 0.670 [95% CI 0.652 to 0.688]; PPV = 0.721 [95% CI 0.703 to 0.739]). This was superior to an 11-variable conventional logistic-regression model comprised of age, sex, history of SH, hypertension, blood glucose, kidney function measurements, and use of oral glucose-lowering drugs (GLDs) (AUROC = 0.906; AUPRC = 0.085; PPV = 0.468). Top impactful predictors included non-use of lipid-regulating drugs, in-patient admission, urgent emergency triage, insulin use, and history of SH. External validation in the HKDR cohort yielded AUROC of 0.856 [95% CI 0.838 to 0.873]. Main limitations of this study included limited transportability of the model and lack of geographically independent validation. CONCLUSIONS: Our novel-ML model demonstrated good discrimination and high precision in predicting one-year risk of SH requiring hospitalization. This may be integrated into EHR decision support systems for preemptive intervention in older adults at highest risk.

Comparison of ß-Cell Function and Insulin Sensitivity Between Normal-Weight and Obese Chinese With Young-Onset Type 2 Diabetes.

Normal-weight individuals with usual-onset type 2 diabetes have reduced ß-cell function and greater insulin sensitivity compared with their obese counterparts. The relative contribution of ß-cell dysfunction and insulin resistance to young-onset type 2 diabetes (YOD) among normal-weight individuals is not well established. In 44 individuals with YOD (24 with normal weight and 20 with obesity) and 24 healthy control individuals with normoglycemia (12 with normal weight and 12 with obesity), we conducted 2-h 12 mmol/L hyperglycemic clamps to measure acute (0-10 min) and steady-state (100-120 min) insulin and C-peptide responses, as well as insulin sensitivity index. Normal-weight individuals with YOD had lower acute insulin response, steady-state insulin and C-peptide responses, and a higher insulin sensitivity index compared with their obese counterparts with YOD. Compared with BMI-matched healthy control individuals, normal-weight individuals with YOD had lower acute and steady-state insulin and C-peptide responses but a similar insulin sensitivity index. The impairment of steady-state ß-cell response relative to healthy control individuals was more pronounced in normal-weight versus obese individuals with YOD. In conclusion, normal-weight Chinese with YOD exhibited worse ß-cell function but preserved insulin sensitivity relative to obese individuals with YOD and BMI-matched healthy individuals with normoglycemia. The selection of glucose-lowering therapy should account for pathophysiological differences underlying YOD between normal-weight and obese individuals.

Circulating metabolomic markers linking diabetic kidney disease and incident cardiovascular disease in type 2 diabetes: analyses from the Hong Kong Diabetes Biobank.

AIMS/HYPOTHESIS: The aim of this study was to describe the metabolome in diabetic kidney disease (DKD) and its association with incident CVD in type 2 diabetes, and identify prognostic biomarkers. METHODS: From a prospective cohort of individuals with type 2 diabetes, baseline sera (N=1991) were quantified for 170 metabolites using NMR spectroscopy with median 5.2 years of follow-up. Associations of chronic kidney disease (CKD, eGFR<60 ml/min per 1.73 m2) or severely increased albuminuria with each metabolite were examined using linear regression, adjusted for confounders and multiplicity. Associations between DKD (CKD or severely increased albuminuria)-related metabolites and incident CVD were examined using Cox regressions. Metabolomic biomarkers were identified and assessed for CVD prediction and replicated in two independent cohorts. RESULTS: At false discovery rate (FDR)<0.05, 156 metabolites were associated with DKD (151 for CKD and 128 for severely increased albuminuria), including apolipoprotein B-containing lipoproteins, HDL, fatty acids, phenylalanine, tyrosine, albumin and glycoprotein acetyls. Over 5.2 years of follow-up, 75 metabolites were associated with incident CVD at FDR<0.05. A model comprising age, sex and three metabolites (albumin, triglycerides in large HDL and phospholipids in small LDL) performed comparably to conventional risk factors (C statistic 0.765 vs 0.762, p=0.893) and adding the three metabolites further improved CVD prediction (C statistic from 0.762 to 0.797, p=0.014) and improved discrimination and reclassification. The 3-metabolite score was validated in independent Chinese and Dutch cohorts. CONCLUSIONS/INTERPRETATION: Altered metabolomic signatures in DKD are associated with incident CVD and improve CVD risk stratification.

Enhanced prediction of abnormal glucose tolerance using an extended non-invasive risk score incorporating routine renal biochemistry.

INTRODUCTION: Type 2 diabetes is preventable in subjects with impaired glucose tolerance based on 2-hour plasma glucose (2hPG) during 75 g oral glucose tolerance test (OGTT). We incorporated routine biochemistry to improve the performance of a non-invasive diabetes risk score to identify individuals with abnormal glucose tolerance (AGT) defined by 2hPG≥7.8 mmol/L during OGTT. RESEARCH DESIGN AND METHODS: We used baseline data of 1938 individuals from the community-based "Better Health for Better Hong Kong - Hong Kong Family Diabetes Study (BHBHK-HKFDS) Cohort" recruited in 1998-2003. We incorporated routine biochemistry in a validated non-invasive diabetes risk score, and evaluated its performance using area under receiver operating characteristics (AUROC) with internal and external validation. RESULTS: The AUROC of the original non-invasive risk score to predict AGT was 0.698 (95% CI, 0.662 to 0.733). Following additional inclusion of fasting plasma glucose, serum potassium, creatinine, and urea, the AUROC increased to 0.778 (95% CI, 0.744 to 0.809, p<0.001). Net reclassification improved by 31.9% (p<0.001) overall, by 30.8% among people with AGT and 1.1% among people without AGT. The extended model showed good calibration (χ2=11.315, p=0.1845) and performance on external validation using an independent data set (AUROC=0.722, 95% CI, 0.680 to 0.764). CONCLUSIONS: The extended risk score incorporating clinical and routine biochemistry can be integrated into an electronic health records system to select high-risk subjects for evaluation of AGT using OGTT for prevention of diabetes.

1-year weight change after diabetes diagnosis and long-term incidence and sustainability of remission of type 2 diabetes in real-world settings in Hong Kong: An observational cohort study.

BACKGROUND: Clinical trials have demonstrated that remission of type 2 diabetes can be achieved following sustained weight loss. However, the feasibility of achieving diabetes remission through weight management in real-world settings remains unclear. In this study, we aimed to examine the association of weight change at 1 year after diabetes diagnosis with long-term incidence and sustainability of type 2 diabetes remission in real-world settings in Hong Kong. METHODS AND FINDINGS: This was a population-based observational cohort study. The territory-wide Risk Assessment and Management Programme for Diabetes Mellitus (RAMP-DM) provides regular comprehensive assessments of metabolic control and complication screening for people with diabetes in Hong Kong. We included 37,326 people with newly diagnosed type 2 diabetes who were enrolled in the RAMP-DM between 2000 and 2017, followed until 2019. Diabetes remission was defined as 2 consecutive HbA1c <6.5% measurements at least 6 months apart in the absence of glucose-lowering drugs (GLDs) and with no record of GLDs at least 3 months before these measurements. During a median follow-up of 7.9 years, 6.1% (2,279) of people achieved diabetes remission, with an incidence rate of 7.8 (95% CI: 7.5, 8.1) per 1,000 person-years. After adjusting for age at diabetes diagnosis, sex, assessment year, body mass index, other metabolic indices, smoking, alcohol drinking, and medication use, the hazard ratio (HR) for diabetes remission was 3.28 (95% CI: 2.75, 3.92; p < 0.001) for people with ≥10% weight loss within 1 year of diagnosis, 2.29 (95% CI: 2.03, 2.59; p < 0.001) for those with 5% to 9.9% weight loss, and 1.34 (95% CI: 1.22, 1.47; p < 0.001) for those with 0% to 4.9% weight loss compared to people with weight gain. During a median follow-up of 3.1 years, 67.2% (1,531) of people who had achieved diabetes remission returned to hyperglycaemia, with an incidence rate of 184.8 (95% CI: 175.5, 194.0) per 1,000 person-years. The adjusted HR for returning to hyperglycaemia was 0.52 (95% CI: 0.41, 0.65; p < 0.001) for people with ≥10% weight loss, 0.78 (95% CI: 0.68, 0.92; p = 0.002) for those with 5% to 9.9% weight loss, and 0.90 (95% CI: 0.80, 1.01; p = 0.073) for those with 0% to 4.9% weight loss compared to people with weight gain. Diabetes remission was associated with a 31% (HR: 0.69, 95% CI: 0.52, 0.93; p = 0.014) decreased risk of all-cause mortality. The main limitation of the study is that the reliability of HbA1c used to define diabetes remission can be affected by other medical conditions. Furthermore, we did not have data on bariatric surgery. CONCLUSIONS: In this study, greater weight loss within the first year of diabetes diagnosis was associated with an increased likelihood of achieving diabetes remission and a decreased risk of returning to hyperglycaemia among those who had achieved diabetes remission. However, both the incidence of diabetes remission and the probability of its long-term sustainability were low with conventional management in real-world settings, in an era when the importance of weight loss was not fully appreciated. Our study provides evidence for policymakers to design and implement early weight management interventions and diabetes remission initiatives.