Development of a predictive risk model for all-cause mortality in patients with diabetes in Hong Kong

Development and validation of a nomogram of all-cause mortality in adult Americans with diabetes

This study aimed to develop and validate a predictive model of all-cause mortality risk in American adults aged ≥ 18 years with diabetes. 7918 participants with diabetes were enrolled from the National Health and Nutrition Examination Survey (NHANES) 1999-2016 and followed for a median of 96 months. The primary study endpoint was the all-cause mortality. Predictors of all-cause mortality included age, Monocytes, Erythrocyte, creatinine, Nutrition Risk Index (NRI), neutrophils/lymphocytes (NLR), smoking habits, alcohol consumption, cardiovascular disease (CVD), urinary albumin excretion rate (UAE), and insulin use. The c-index was 0.790 (95% CI 0.779-0.801, P < 0.001) and 0.792 (95% CI: 0.776-0.808, P < 0.001) for the training and validation sets, respectively. The area under the ROC curve was 0.815, 0.814, 0.827 and 0.812, 0.818 and 0.829 for the training and validation sets at 3, 5, and 10 years of follow-up, respectively. Both calibration plots and DCA curves performed well. The model provides accurate predictions of the risk of death for American persons with diabetes and its scores can effectively determine the risk of death in outpatients, providing guidance for clinical decision-making and predicting prognosis for patients.

Role of angiotensin receptor-neprilysin inhibitor in diabetic complications

Diabetes mellitus is a prevalent disorder with multi-system manifestations, causing a significant burden in terms of disability and deaths globally. Angio-tensin receptor-neprilysin inhibitor (ARNI) belongs to a class of medications for treating heart failure, with the benefits of reducing hospitalization rates and mortality. This review mainly focuses on the clinical and basic investigations related to ARNI and diabetic complications, discussing possible physiological and molecular mechanisms, with insights for future applications.

Behind the Hospital Ward: In-Hospital Mortality of Type 2 Diabetes Mellitus Patients in Indonesia (Analysis of National Health Insurance Claim Sample Data)

The rising global prevalence of diabetes mellitus, a chronic metabolic disorder, poses significant challenges to healthcare systems worldwide. This study examined in-hospital mortality among patients diagnosed with non-insulin-dependent diabetes mellitus (NIDDM) of ICD-10, or Type 2 Diabetes Mellitus (T2DM), in Indonesia, utilizing hospital claims data spanning from 2017 to 2022 obtained from the Indonesia Health Social Security Agency or Badan Penyelenggara Jaminan Sosial (BPJS) Kesehatan. The analysis, which included 610,809 hospitalized T2DM patients, revealed an in-hospital mortality rate of 6.6%. Factors contributing to an elevated risk of mortality included advanced age, the presence of comorbidities, and severe complications. Additionally, patients receiving health subsidies and those treated in government hospitals were found to have higher mortality risks. Geographic disparities were observed, highlighting variations in healthcare outcomes across different regions. Notably, the complication of ketoacidosis emerged as the most significant risk factor for in-hospital mortality, with an odds ratio (OR) of 10.86, underscoring the critical need for prompt intervention and thorough management of complications to improve patient outcomes.

Healthcare Big Data in Hong Kong: Development and Implementation of Artificial Intelligence-Enhanced Predictive Models for Risk Stratification

Routinely collected electronic health records (EHRs) data contain a vast amount of valuable information for conducting epidemiological studies. With the right tools, we can gain insights into disease processes and development, identify the best treatment and develop accurate models for predicting outcomes. Our recent systematic review has found that the number of big data studies from Hong Kong has rapidly increased since 2015, with an increasingly common application of artificial intelligence (AI). The advantages of big data are that i) the models developed are highly generalisable to the population, ii) multiple outcomes can be determined simultaneously, iii) ease of cross-validation by for model training, development and calibration, iv) huge numbers of useful variables can be analyzed, v) static and dynamic variables can be analyzed, vi) non-linear and latent interactions between variables can be captured, vii) artificial intelligence approaches can enhance the performance of prediction models. In this paper, we will provide several examples (cardiovascular disease, diabetes mellitus, Brugada syndrome, long QT syndrome) to illustrate efforts from a multi-disciplinary team to identify data from different modalities to develop models using territory-wide datasets, with the possibility of real-time risk updates by using new data captured from patients. The benefit is that only routinely collected data are required for developing highly accurate and high-performance models. AI-driven models outperform traditional models in terms of sensitivity, specificity, accuracy, area under the receiver operating characteristic and precision-recall curve, and F1 score. Web and/or mobile versions of the risk models allow clinicians to risk stratify patients quickly in clinical settings, thereby enabling clinical decision-making. Efforts are required to identify the best ways of implementing AI algorithms on the web and mobile apps.

Developing a prediction model for all-cause mortality risk among patients with type 2 diabetes mellitus in Shanghai, China

BACKGROUND

All-cause mortality risk prediction models for patients with type 2 diabetes mellitus (T2DM) in mainland China have not been established. This study aimed to fill this gap.

METHODS

Based on the Shanghai Link Healthcare Database, patients diagnosed with T2DM and aged 40-99 years were identified between January 1, 2013 and December 31, 2016 and followed until December 31, 2021. All the patients were randomly allocated into training and validation sets at a 2:1 ratio. Cox proportional hazards models were used to develop the all-cause mortality risk prediction model. The model performance was evaluated by discrimination (Harrell C-index) and calibration (calibration plots).

RESULTS

A total of 399 784 patients with T2DM were eventually enrolled, with 68 318 deaths over a median follow-up of 6.93 years. The final prediction model included age, sex, heart failure, cerebrovascular disease, moderate or severe kidney disease, moderate or severe liver disease, cancer, insulin use, glycosylated hemoglobin, and high-density lipoprotein cholesterol. The model showed good discrimination and calibration in the validation sets: the mean C-index value was 0.8113 (range 0.8110-0.8115) and the predicted risks closely matched the observed risks in the calibration plots.

CONCLUSIONS

This study constructed the first 5-year all-cause mortality risk prediction model for patients with T2DM in south China, with good predictive performance.

Recent applications of machine learning and deep learning models in the prediction, diagnosis, and management of diabetes: a comprehensive review

Diabetes as a metabolic illness can be characterized by increased amounts of blood glucose. This abnormal increase can lead to critical detriment to the other organs such as the kidneys, eyes, heart, nerves, and blood vessels. Therefore, its prediction, prognosis, and management are essential to prevent harmful effects and also recommend more useful treatments. For these goals, machine learning algorithms have found considerable attention and have been developed successfully. This review surveys the recently proposed machine learning (ML) and deep learning (DL) models for the objectives mentioned earlier. The reported results disclose that the ML and DL algorithms are promising approaches for controlling blood glucose and diabetes. However, they should be improved and employed in large datasets to affirm their applicability.

Association Between Atherosclerosis-Related Cardiovascular Disease and Uveitis: A Systematic Review and Meta-Analysis

BACKGROUND

Uveitis is not only an intraocular inflammatory disease, but also an indicator of systemic inflammation. It is unclear whether uveitis can increase the risk of cardiovascular disease (CVD) through the atherosclerotic pathway.

METHODS

PubMed and Embase databases were searched until 5 September, 2022. Original studies investigating uveitis and cardiovascular events were selected. The random-effects model was used to calculate the difference of groups in pooled estimates.

RESULTS

A total of six observational studies that included mainly ankylosing spondylitis (AS) patients were included. Of these, three studies reported data on carotid plaques and carotid intima-media thickness (cIMT) and the other three studies provided data on atherosclerosis-related CVD. No significant difference was found in cIMT between uveitis and controls (MD = 0.01, 95% CI = -0.03-0.04, p = 0.66), consistent with the findings of carotid plaque incidence (OR = 1.30, 95% CI = 0.71-2.41, p = 0.39). However, uveitis was associated with a 1.49-fold increase in atherosclerosis-related CVD (HR = 1.49, 95% CI = 1.20-1.84, p = 0.0002).

CONCLUSIONS

Uveitis is a predictor of atherosclerosis-related CVD in AS patients. For autoimmune disease patients with uveitis, earlier screening of cardiovascular risk factors and the implementation of corresponding prevention strategies may be associated with a better prognosis.

The Impact of Cardiac Comorbidity Sequence at Baseline and Mortality Risk in Type 2 Diabetes Mellitus: A Retrospective Population-Based Cohort Study

Introduction: The presence of multiple comorbidities increases the risk of all-cause mortality, but the effects of the comorbidity sequence before the baseline date on mortality remain unexplored. This study investigated the relationship between coronary heart disease (CHD), atrial fibrillation (AF) and heart failure (HF) through their sequence of development and the effect on all-cause mortality risk in type 2 diabetes mellitus. Methods: This study included patients with type 2 diabetes mellitus prescribed antidiabetic/cardiovascular medications in public hospitals of Hong Kong between 1 January 2009 and 31 December 2009, with follow-up until death or 31 December 2019. The Cox regression was used to identify comorbidity sequences predicting all-cause mortality in patients with different medication subgroups. Results: A total of 249,291 patients (age: 66.0 ± 12.4 years, 47.4% male) were included. At baseline, 7564, 10,900 and 25,589 patients had AF, HF and CHD, respectively. Over follow-up (3524 ± 1218 days), 85,870 patients died (mortality rate: 35.7 per 1000 person-years). Sulphonylurea users with CHD developing later and insulin users with CHD developing earlier in the disease course had lower mortality risks. Amongst insulin users with two of the three comorbidities, those with CHD with preceding AF (hazard ratio (HR): 3.06, 95% CI: [2.60−3.61], p < 0.001) or HF (HR: 3.84 [3.47−4.24], p < 0.001) had a higher mortality. In users of lipid-lowering agents with all three comorbidities, those with preceding AF had a higher risk of mortality (AF-CHD-HF: HR: 3.22, [2.24−4.61], p < 0.001; AF-HF-CHD: HR: 3.71, [2.66−5.16], p < 0.001). Conclusions: The sequence of comorbidity development affects the risk of all-cause mortality to varying degrees in diabetic patients on different antidiabetic/cardiovascular medications.

Development of an Electronic Frailty Index for Predicting Mortality and Complications Analysis in Pulmonary Hypertension Using Random Survival Forest Model

Background

The long-term prognosis of the cardio-metabolic and renal complications, in addition to mortality in patients with newly diagnosed pulmonary hypertension, are unclear. This study aims to develop a scalable predictive model in the form of an electronic frailty index (eFI) to predict different adverse outcomes.

Methods

This was a population-based cohort study of patients diagnosed with pulmonary hypertension between January 1st, 2000 and December 31st, 2017, in Hong Kong public hospitals. The primary outcomes were mortality, cardiovascular complications, renal diseases, and diabetes mellitus. The univariable and multivariable Cox regression analyses were applied to identify the significant risk factors, which were fed into the non-parametric random survival forest (RSF) model to develop an eFI.

Results

A total of 2,560 patients with a mean age of 63.4 years old (interquartile range: 38.0–79.0) were included. Over a follow-up, 1,347 died and 1,878, 437, and 684 patients developed cardiovascular complications, diabetes mellitus, and renal disease, respectively. The RSF-model-identified age, average readmission, anti-hypertensive drugs, cumulative length of stay, and total bilirubin were among the most important risk factors for predicting mortality. Pair-wise interactions of factors including diagnosis age, average readmission interval, and cumulative hospital stay were also crucial for the mortality prediction. Patients who developed all-cause mortality had higher values of the eFI compared to those who survived (P < 0.0001). An eFI ≥ 9.5 was associated with increased risks of mortality [hazard ratio (HR): 1.90; 95% confidence interval [CI]: 1.70–2.12; P < 0.0001]. The cumulative hazards were higher among patients who were 65 years old or above with eFI ≥ 9.5. Using the same cut-off point, the eFI predicted a long-term mortality over 10 years (HR: 1.71; 95% CI: 1.53–1.90; P < 0.0001). Compared to the multivariable Cox regression, the precision, recall, area under the curve (AUC), and C-index were significantly higher for RSF in the prediction of outcomes.

Conclusion

The RSF models identified the novel risk factors and interactions for the development of complications and mortality. The eFI constructed by RSF accurately predicts the complications and mortality of patients with pulmonary hypertension, especially among the elderly.

Adverse Cardiovascular Complications following prescription of programmed cell death 1 (PD-1) and programmed cell death ligand 1 (PD-L1) inhibitors: a propensity-score matched Cohort Study with competing risk analysis

Background

Programmed death-1 (PD-1) and programmed death- ligand 1 (PD-L1) inhibitors, such as pembrolizumab, nivolumab and atezolizumab, are major classes of immune checkpoint inhibitors that are increasingly used for cancer treatment. However, their use is associated with adverse cardiovascular events. We examined the incidence of new-onset cardiac complications in patients receiving PD-1 or PD-L1 inhibitors.

Methods

Patients receiving PD-1 or PD-L1 inhibitors since their launch up to 31st December 2019 at publicly funded hospitals of Hong Kong, China, without pre-existing cardiac complications were included. The primary outcome was a composite of incident heart failure, acute myocardial infarction, atrial fibrillation, or atrial flutter with the last follow-up date of 31st December 2020. Propensity score matching between PD-L1 inhibitor use and PD-1 inhibitor use with a 1:2 ratio for patient demographics, past comorbidities and non-PD-1/PD-L1 medications was performed with nearest neighbour search strategy (0.1 caliper). Univariable and multivariable Cox regression analysis models were conducted. Competing risks models and multiple propensity matching approaches were considered for sensitivity analysis.

Results

A total of 1959 patients were included. Over a median follow-up of 247 days (interquartile range [IQR]: 72-506), 320 (incidence rate [IR]: 16.31%) patients met the primary outcome after PD-1/PD-L1 treatment: 244 (IR: 12.57%) with heart failure, 38 (IR: 1.93%) with acute myocardial infarction, 54 (IR: 2.75%) with atrial fibrillation, 6 (IR: 0.31%) with atrial flutter. Compared with PD-1 inhibitor treatment, PD-L1 inhibitor treatment was significantly associated with lower risks of the composite outcome both before (hazard ratio [HR]: 0.32, 95% CI: [0.18-0.59], P value=0.0002) and after matching (HR: 0.34, 95% CI: [0.18-0.65], P value=0.001), and lower all-cause mortality risks before matching (HR: 0.77, 95% CI: [0.64-0.93], P value=0.0078) and after matching (HR: 0.80, 95% CI: [0.65-1.00], P value=0.0463). Patients who developed cardiac complications had shorter average readmission intervals and a higher number of hospitalizations after treatment with PD-1/PD-L1 inhibitors in both the unmatched and matched cohorts (P value<0.0001). Multivariable Cox regression models, competing risk analysis with cause-specific and subdistribution hazard models, and multiple propensity approaches confirmed these observations.

Conclusions

Compared with PD-1 treatment, PD-L1 treatment was significantly associated with lower risk of new onset cardiac complications and all-cause mortality both before and after propensity score matching.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40959-021-00128-5.

Hip fractures risks in edoxaban versus warfarin users: A propensity score-matched population-based cohort study with competing risk analyses

OBJECTIVE

The three direct oral anticoagulants (DOAC), rivaroxaban, apixaban and dabigatran have been associated with lower risks of fractures compared to warfarin. However, no large scale studies have explored the associations with the newest DOAC, edoxaban, with fracture risk. The present study aims to elucidate the effects of edoxaban on the risk of hip fracture amongst elderly patients by comparing the incidence of new onset hip fracture between edoxaban and warfarin users in a Chinese population.

METHODS

This was a retrospective population-based cohort study of patients with edoxaban or warfarin use between January 1st, 2016 and December 31st, 2019 in Hong Kong, China. Patients with less than one-month exposure, medication switching between warfarin and edoxaban, those who died within 30 days after drug exposure, prior human immunodeficiency virus infection, age <50 years old, and those with prior hip fractures were excluded. Propensity score matching (1:2) between edoxaban and warfarin users using the nearest neighbour method was performed based on demographics, prior comorbidities, and use of different medications. The study outcomes were new onset hip fractures, medically attended falls and all-cause mortality.

RESULTS

A total of 5014 patients including 579 edoxaban users and 4435 warfarin users (median age: 70 years old [interquartile range (IQR): 62-79], 56.66% males) with a median follow-up of 637.5 (IQR: 320-1073) days were included. In the matched cohort, edoxaban users had significantly lower rates of new onset hip fractures, medically attended falls and all-cause mortality. The protective value of edoxaban use against new onset hip fracture (hazard ratio [HR]: 0.13, 95% confidence interval [CI]: [0.03-0.54], p = 0.0051), medically attended falls (HR: 0.47, [0.29-0.75], p = 0.0018) and all-cause mortality (HR: 0.61, [0.42-0.87], p = 0.0059) in comparison to warfarin use persisted after matching. The significant relationship between edoxaban use and lower fracture risk was preserved in all sensitivity analyses using different approaches using the propensity score.

CONCLUSIONS

Edoxaban use is associated with lower risks of new onset hip fractures, medically attended falls and mortality risks compared to warfarin after propensity score matching.

The effect of total cholesterol/high-density lipoprotein cholesterol ratio on mortality risk in the general population

BACKGROUND

The relationship between the total cholesterol/high-density lipoprotein cholesterol (TC/HDL-C) ratio and all-cause and cardiovascular mortality has not been elucidated. Herein, we intend to probe the effect of the TC/HDL-C ratio on all-cause and cardiovascular mortality in the general population.

METHODS

From the 1999-2014 National Health and Nutrition Examination Surveys (NHANES), a total of 32,405 health participants aged ≥18 years were included. The TC/HDL-C levels were divided into five groups: Q1: <2.86, Q2: 2.86-3.46, Q3: 3.46-4.12, Q4: 4.12-5.07, Q5: >5.07. Multivariate Cox regression models were used to explore the relationship between the TC/HDL-C ratio and cardiovascular and all-cause mortality. Two-piecewise linear regression models and restricted cubic spline regression were used to explore nonlinear and irregularly shaped relationships. Kaplan-Meier survival curve and subgroup analyses were conducted.

RESULTS

The population comprised 15,675 men and 16,730 women with a mean age of 43 years. During a median follow-up of 98 months (8.1 years), 2,859 mortality cases were recorded. The TC/HDL-C ratio and all-cause mortality showed a nonlinear association after adjusting for confounding variables in the restricted cubic spline analysis. Hazard ratios (HRs) of all-cause mortality were particularly positively related to the level of TC/HDL-C ratio in the higher range >5.07 and in the lower range <2.86 (HR 1.26; 95% CI 1.10, 1.45; HR 1.18; 95% CI 1.00, 1.38, respectively), although the HRs of cardiovascular disease mortality showed no difference among the five groups. In the two-piecewise linear regression model, a TC/HDL-C ratio range of ≥4.22 was positively correlated with cardiovascular mortality (HR 1.13; 95% CI 1.02, 1.25). In the subgroup analysis, a nonlinear association between TC/HDL-C and all-cause mortality was found in those aged <65 years, men, and the no lipid drug treatment population.

CONCLUSION

A nonlinear association between the TC/HDL-C ratio and all-cause mortality was found, indicating that a too-low or too-high TC/HDL-C ratio might increase all-cause mortality. However, for cardiovascular mortality, it does not seem so. The cutoff value was 4.22. The individuals had higher cardiovascular mortality with a TC/HDL-C ratio >4.22.

Risk stratification of cardiac arrhythmias and sudden cardiac death in type 2 diabetes mellitus patients receiving insulin therapy: A population-based cohort study

INTRODUCTION

Metabolic abnormalities may exacerbate the risk of adverse outcomes in patients with type 2 diabetes mellitus. The present study aims to assess the predictive value of HbA1c and lipid variability on the risks of sudden cardiac death (SCD) and incident atrial fibrillation (AF).

METHODS

The retrospective observational study consists of type 2 diabetic patients prescribed with insulin, who went to publicly funded clinics and hospitals in Hong Kong between January 1, 2009 and December 31, 2009. Variability in total cholesterol, low-density lipoprotein-cholesterol (LDL-C), high-density lipoprotein-cholesterol (HDL-C), triglyceride, and HbA1c were assessed through their SD and coefficient of variation. The primary outcomes were incident (1) ventricular tachycardia/ventricular fibrillation, actual or aborted SCD and (2) AF.

RESULTS

A total of 23 329 patients (mean ± SD age: 64 ± 14 years old; 51% male; mean HbA1c 8.6 ± 1.3%) were included. On multivariable analysis, HbA1c, total cholesterol, LDL-C and triglyceride variability were found to be predictors of SCD (p < .05).

CONCLUSION

HbA1c and lipid variability were predictive of SCD. Therefore, poor glucose control and variability in lipid parameters in diabetic patients are associated with aborted or actual SCD. These observations suggest the need to re-evaluate the extent of glycemic control required for outcome optimization.