Cardiometabolic syndrome among general adult population in Ghana: The role of lipid accumulation product, waist circumference-triglyceride index, and triglyceride-glucose index as surrogate indicators

Pathways to chronic disease detection and prediction: Mapping the potential of machine learning to the pathophysiological processes while navigating ethical challenges

Chronic diseases such as heart disease, cancer, and diabetes are leading drivers of mortality worldwide, underscoring the need for improved efforts around early detection and prediction. The pathophysiology and management of chronic diseases have benefitted from emerging fields in molecular biology like genomics, transcriptomics, proteomics, glycomics, and lipidomics. The complex biomarker and mechanistic data from these "omics" studies present analytical and interpretive challenges, especially for traditional statistical methods. Machine learning (ML) techniques offer considerable promise in unlocking new pathways for data-driven chronic disease risk assessment and prognosis. This review provides a comprehensive overview of state-of-the-art applications of ML algorithms for chronic disease detection and prediction across datasets, including medical imaging, genomics, wearables, and electronic health records. Specifically, we review and synthesize key studies leveraging major ML approaches ranging from traditional techniques such as logistic regression and random forests to modern deep learning neural network architectures. We consolidate existing literature to date around ML for chronic disease prediction to synthesize major trends and trajectories that may inform both future research and clinical translation efforts in this growing field. While highlighting the critical innovations and successes emerging in this space, we identify the key challenges and limitations that remain to be addressed. Finally, we discuss pathways forward toward scalable, equitable, and clinically implementable ML solutions for transforming chronic disease screening and prevention.

Differential risk of cardiovascular complications in patients with type-2 diabetes mellitus in Ghana: A hospital-based cross-sectional study

AIM

To characterize clinically relevant subgroups of patients with type-2 diabetes mellitus (T2DM) based on adiposity, insulin secretion, and resistance indices.

METHODS

A cross-sectional study was conducted at Eastern Regional Hospital in Ghana from July to October 2021 to investigate long-term patients with T2DM. To select participants, a systematic random sampling method was employed. Demographic data was collected using a structured questionnaire and fasting blood samples were taken to measure glycemic and lipid levels. Blood pressure and adiposity indices were measured during recruitment. The risk of cardiovascular disease (CVD) was defined using Framingham scores and standard low-density lipoprotein thresholds. To analyze the data, k-means clustering algorithms and regression analysis were used.

RESULTS

The study identified three groups in female patients according to body mass index, relative fat mass, glycated hemoglobin, and triglyceride-glucose index. These groups included the obesity-related phenotype, the severe insulin resistance phenotype, and the normal weight phenotype with improved insulin resistance. Among male patients with T2DM, two groups were identified, including the obesity-related phenotype with severe insulin resistance and the normal weight phenotype with improved insulin sensitivity. The severe insulin resistance phenotype in female patients was associated with an increased risk of high CVD (OR = 5.34, 95%CI:2.11-13.55) and metabolic syndrome (OR = 7.07; 95%CI:3.24-15.42). Among male patients, the obesity-related phenotype with severe insulin resistance was associated with an increased intermediate (OR = 21.78, 95%CI:4.17-113.78) and a high-risk CVD (OR = 6.84, 95%CI:1.45-32.12).

CONCLUSIONS

The findings highlight significant cardiometabolic heterogeneity among T2DM patients. The subgroups of T2DM patients characterized by obesity and/or severe insulin resistance with or without poor glycemic control, have increased risk of CVD. This underscores the importance of considering differences in adiposity, insulin secretion, and sensitivity indices when making clinical decisions for patients with T2DM.

Predictive modelling of metabolic syndrome in Ghanaian diabetic patients: an ensemble machine learning approach

Objectives

The burgeoning prevalence of cardiometabolic disorders, including type 2 diabetes mellitus (T2DM) and metabolic syndrome (MetS) within Africa is concerning. Machine learning (ML) techniques offer a unique opportunity to leverage data-driven insights and construct predictive models for MetS risk, thereby enhancing the implementation of personalised prevention strategies. In this work, we employed ML techniques to develop predictive models for pre-MetS and MetS among diabetic patients.

Methods

This multi-centre cross-sectional study comprised of 919 T2DM patients. Age, gender, novel anthropometric indices along with biochemical measures were analysed using BORUTA feature selection and an ensemble majority voting classification model, which included logistic regression, k-nearest neighbour, Gaussian Naive Bayes, Gradient boosting classification, and support vector machine.

Results

Distinct metabolic profiles and phenotype clusters were associated with MetS progression. The BORUTA algorithm identified 10 and 16 significant features for pre-MetS and MetS prediction, respectively. For pre-MetS, the top-ranked features were lipid accumulation product (LAP), triglyceride-glucose index adjusted for waist-to-height ratio (TyG-WHtR), coronary risk (CR), visceral adiposity index (VAI) and abdominal volume index (AVI). For MetS prediction, the most influential features were VAI, LAP, waist triglyceride index (WTI), Very low-density cholesterol (VLDLC) and TyG-WHtR. Majority voting ensemble classifier demonstrated superior performance in predicting pre-MetS (AUC = 0.79) and MetS (AUC = 0.87).

Conclusion

Identifying these risk factors reveals the complex interplay between visceral adiposity and metabolic dysregulation in African populations, enabling early detection and treatment. Ethical integration of ML algorithms in clinical decision-making can streamline identification of high-risk individuals, optimize resource allocation, and enable precise, tailored interventions.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40200-024-01491-7.

Multi-dimensional comparison of abdominal obesity indices and insulin resistance indicators for assessing NAFLD

BACKGROUND

The prevalence of non-alcoholic fatty liver disease (NAFLD) keeps increasing annually worldwide. Non-invasive assessment tools for evaluating the risk and severity of the disease are still limited. Insulin resistance (IR) and abdominal obesity (ABO) are closely related to NAFLD.

METHODS

A retrospective large-scale, population-based study was conducted based on the data from the 2017-2018 cycle of the National Health and Nutrition Examination Survey (NHANES). Three ABO indices, namely lipid accumulation product (LAP), visceral obesity index (VAI), waist circumference-triglyceride index (WTI), and three IR indices, including triglyceride glucose index (TyG), homeostasis model assessment of insulin resistance (HOMA-IR) and metabolic score for insulin resistance (METS-IR), were analyzed and compared for their relationships with NAFLD based on weighted multivariable logistic regression, spearman correlation heatmap, smooth curve fittings. The area under the curve (AUC) of receiver-operating characteristic (ROC) curve was used to evaluate the diagnostic capability of these indices for NAFLD. Differences among the AUCs were calculated and compared by Delong test.

RESULTS

In total, 3095 participants were included in our study among which 1368 adults were diagnosed with NAFLD. All six indices presented positive associations with NAFLD. There was a claw-shaped curve between HOMA-IR, VAI, LAP and NAFLD while a smooth semi-bell curve was observed in TyG, METS-IR and WTI. LAP and HOMA-IR had the best diagnostic capability for NAFLD (LAP: AUC = 0.8, Youden index = 0.48; HOMA-IR: AUC = 0.798, Youden index = 0.472) while VAI (AUC = 0.728, Youden index = 0.361) showed the lowest predictive value. The correlation heat map indicated positive correlations between all six indices and liver function, hepatic steatosis and fibrosis severity. In the NAFLD group, IR indicators presented a stronger association with alanine aminotransferase (ALT) compared with ABO indices.

CONCLUSIONS

All six indices can screen NAFLD withLAP and HOMA-IR being possibly optimal predictors. IR indices may be more sensitive to identify acute hepatic injury in NAFLD patients than ABO indices.

Unmasking the Risk Factors Associated with Undiagnosed Diabetes and Prediabetes in Ghana: Insights from Cardiometabolic Risk (CarMeR) Study-APTI Project

INTRODUCTION

Undiagnosed diabetes poses significant public health challenges in Ghana. Numerous factors may influence the prevalence of undiagnosed diabetes among adults, and therefore, using a model that takes into account the intricate network of these relationships should be considered. Our goal was to evaluate fasting plasma levels, a critical indicator of diabetes, and the associated direct and indirect associated or protective factors.

METHODS

This research employed a cross-sectional survey to sample 1200 adults aged 25-70 years who perceived themselves as healthy and had not been previously diagnosed with diabetes from 13 indigenous communities within the Cape Coast Metropolis, Ghana. Diabetes was diagnosed based on the American Diabetes Association (ADA) criteria for fasting plasma glucose, and lipid profiles were determined using Mindray equipment (August 2022, China). A stepwise WHO questionnaire was used to collect data on sociodemographic and lifestyle variables. We analyzed the associations among the exogenous, mediating, and endogenous variables using a generalized structural equation model (GSEM).

RESULTS

Overall, the prevalence of prediabetes and diabetes in the Cape Coast Metropolis was found to be 14.2% and 3.84%, respectively. In the sex domain, females had a higher prevalence of prediabetes (15.33%) and diabetes (5.15%) than males (12.62% and 1.24%, respectively). Rural areas had the highest prevalence, followed by peri-urban areas, whereas urban areas had the lowest prevalence. In the GSEM results, we found that body mass index (BMI), triglycerides (TG), systolic blood pressure (SBP), gamma-glutamyl transferase (GGT), and female sex were direct predictive factors for prediabetes and diabetes, based on fasting plasma glucose (FPG) levels. Indirect factors influencing diabetes and prediabetes through waist circumference (WC) included childhood overweight status, family history, age 35-55 and 56-70, and moderate and high socioeconomic status. High density lipoprotein (HDL) cholesterol, childhood overweight, low physical activity, female sex, moderate and high socioeconomic status, and market trading were also associated with high BMI, indirectly influencing prediabetes and diabetes. Total cholesterol, increased TG levels, WC, age, low physical activity, and rural dwellers were identified as indirectly associated factors with prediabetes and diabetes through SBP. Religion, male sex, and alcohol consumption were identified as predictive factors for GGT, indirectly influencing prediabetes and diabetes.

CONCLUSIONS

Diabetes in indigenous communities is directly influenced by blood lipid, BMI, SBP, and alcohol levels. Childhood obesity, physical inactivity, sex, socioeconomic status, and family history could indirectly influence diabetes development. These findings offer valuable insights for policymakers and health-sector stakeholders, enabling them to understand the factors associated with diabetes development and implement necessary public health interventions and personalized care strategies for prevention and management in Ghana.

Gender-specific accuracy of lipid accumulation product index for the screening of metabolic syndrome in general adults: a meta-analysis and comparative analysis with other adiposity indicators

BACKGROUND

Lipid accumulation product (LAP) is a novel predictor index of central lipid accumulation associated with metabolic and cardiovascular diseases. This study aims to investigate the accuracy of LAP for the screening of metabolic syndrome (MetS) in general adult males and females and its comparison with other lipid-related indicators.

METHODS

A systematic literature search was conducted in PubMed, Scopus, Web of Science, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and ProQuest for eligible studies up to May 8, 2024. Outcomes were pooled mean difference (MD), odds ratio (OR), and diagnostic accuracy parameters (sensitivity, specificity, and area under the summary receiver operating characteristic [AUSROC] curve). Comparative analysis was conducted using Z-test.

RESULTS

Forty-three studies involving 202,313 participants (98,164 males and 104,149 females) were included. Pooled MD analysis showed that LAP was 45.92 (P < 0.001) and 41.70 units (P < 0.001) higher in men and women with MetS, respectively. LAP was also significantly associated with MetS, with pooled ORs of 1.07 (P < 0.001) in men and 1.08 (P < 0.001) in women. In men, LAP could detect MetS with a pooled sensitivity of 85% (95% CI: 82%-87%), specificity of 81% (95% CI: 80%-83%), and AUSROC curve of 0.88 (95% CI: 0.85-0.90), while in women, LAP had a sensitivity of 83% (95% CI: 80%-86%), specificity of 80% (95% CI: 78%-82%), and AUSROC curve of 0.88 (95% CI: 0.85-0.91). LAP had a significantly higher AUSROC curve (P < 0.05) for detecting MetS compared to body mass index (BMI), waist-to-height ratio (WHtR), waist-to-hip ratio (WHR), body roundness index (BRI), a body shape index (ABSI), body adiposity index (BAI), conicity index (CI) in both genders, and waist circumference (WC) and abdominal volume index (AVI) in females.

CONCLUSION

LAP may serve as a simple, cost-effective, and more accurate screening tool for MetS in general adult male and female populations.

Diagnostic efficacy of the triglyceride-glucose index in the prediction of contrast-induced nephropathy following percutaneous coronary intervention

Introduction

Contrast-induced nephropathy (CIN) is a common complication of percutaneous coronary intervention (PCI). Identifying patients at high CIN risk remains challenging. The triglyceride-glucose (TyG) index may help predict CIN but evidence is limited. We conducted a meta-analysis to evaluate the diagnostic value of TyG index for CIN after PCI.

Methods

A systematic literature search was performed in MEDLINE, Cochrane, and EMBASE until August 2023 (PROSPERO registration: CRD42023452257). Observational studies examining TyG index for predicting CIN risk in PCI patients were included. This diagnostic meta-analysis aimed to evaluate the accuracy of the TyG index in predicting the likelihood of CIN. Secondary outcomes aimed to assess the pooled incidence of CIN and the association between an elevated TyG index and the risk of CIN.

Results

Five studies (Turkey, n=2; China, n=3) with 3518 patients (age range: 57.6 to 68.22 years) were included. The pooled incidence of CIN was 15.3% [95% confidence interval (CI) 11-20.8%]. A high TyG index associated with increased CIN risk (odds ratio: 2.25, 95% CI 1.82-2.77). Pooled sensitivity and specificity were 0.77 (95% CI 0.59-0.88) and 0.55 (95% CI 0.43-0.68) respectively. Analysis of the summary receiver operating characteristic (sROC) curve revealed an area under the curve of 0.69 (95% CI 0.65-0.73). There was a low risk of publication bias (p = 0.81).

Conclusion

The TyG index displayed a noteworthy correlation with the risk of CIN subsequent to PCI. However, its overall diagnostic accuracy was found to be moderate in nature. While promising, the TyG index should not be used in isolation for CIN screening given the heterogeneity between studies. In addition, the findings cannot be considered conclusive given the scarcity of data. Further large-scale studies are warranted to validate TyG cutoffs and determine how to optimally incorporate it into current risk prediction models.

Systematic Review Registration

https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023452257, identifier CRD42023452257.