Machine Learning Models for Prediction of Diabetic Microvascular Complications.

IMPORTANCE AND AIMS: Diabetic microvascular complications significantly impact morbidity and mortality. This review focuses on machine learning/artificial intelligence (ML/AI) in predicting diabetic retinopathy (DR), diabetic kidney disease (DKD), and diabetic neuropathy (DN). METHODS: A comprehensive PubMed search from 1990 to 2023 identified studies on ML/AI models for diabetic microvascular complications. The review analyzed study design, cohorts, predictors, ML techniques, prediction horizon, and performance metrics. RESULTS: Among the 74 identified studies, 256 featured internally validated ML models and 124 had externally validated models, with about half being retrospective. Since 2010, there has been a rise in the use of ML for predicting microvascular complications, mainly driven by DKD research across 27 countries. A more modest increase in ML research on DR and DN was observed, with publications from fewer countries. For all microvascular complications, predictive models achieved a mean (standard deviation) c-statistic of 0.79 (0.09) on internal validation and 0.72 (0.12) on external validation. Diabetic kidney disease models had the highest discrimination, with c-statistics of 0.81 (0.09) on internal validation and 0.74 (0.13) on external validation, respectively. Few studies externally validated prediction of DN. The prediction horizon, outcome definitions, number and type of predictors, and ML technique significantly influenced model performance. CONCLUSIONS AND RELEVANCE: There is growing global interest in using ML for predicting diabetic microvascular complications. Research on DKD is the most advanced in terms of publication volume and overall prediction performance. Both DR and DN require more research. External validation and adherence to recommended guidelines are crucial.

Precision prognostics for cardiovascular disease in Type 2 diabetes: a systematic review and meta-analysis.

BACKGROUND: Precision medicine has the potential to improve cardiovascular disease (CVD) risk prediction in individuals with Type 2 diabetes (T2D). METHODS: We conducted a systematic review and meta-analysis of longitudinal studies to identify potentially novel prognostic factors that may improve CVD risk prediction in T2D. Out of 9380 studies identified, 416 studies met inclusion criteria. Outcomes were reported for 321 biomarker studies, 48 genetic marker studies, and 47 risk score/model studies. RESULTS: Out of all evaluated biomarkers, only 13 showed improvement in prediction performance. Results of pooled meta-analyses, non-pooled analyses, and assessments of improvement in prediction performance and risk of bias, yielded the highest predictive utility for N-terminal pro b-type natriuretic peptide (NT-proBNP) (high-evidence), troponin-T (TnT) (moderate-evidence), triglyceride-glucose (TyG) index (moderate-evidence), Genetic Risk Score for Coronary Heart Disease (GRS-CHD) (moderate-evidence); moderate predictive utility for coronary computed tomography angiography (low-evidence), single-photon emission computed tomography (low-evidence), pulse wave velocity (moderate-evidence); and low predictive utility for C-reactive protein (moderate-evidence), coronary artery calcium score (low-evidence), galectin-3 (low-evidence), troponin-I (low-evidence), carotid plaque (low-evidence), and growth differentiation factor-15 (low-evidence). Risk scores showed modest discrimination, with lower performance in populations different from the original development cohort. CONCLUSIONS: Despite high interest in this topic, very few studies conducted rigorous analyses to demonstrate incremental predictive utility beyond established CVD risk factors for T2D. The most promising markers identified were NT-proBNP, TnT, TyG and GRS-CHD, with the highest strength of evidence for NT-proBNP. Further research is needed to determine their clinical utility in risk stratification and management of CVD in T2D.

People living with type 2 diabetes (T2D) are more likely to develop problems with their heart or blood circulation, known as cardiovascular disease (CVD), than people who do not have T2D. However, it can be difficult to predict which people with T2D are most likely to develop CVD. This is because current approaches, such as blood tests, do not identify all people with T2D who are at an increased risk of CVD. In this study we reviewed published papers that investigated the differences between people with T2D who experienced CVD compared to those who did not. We found some indicators that could potentially be used to determine which people with T2D are most likely to develop CVD. More studies are needed to determine how useful these are. However, they could potentially be used to enable clinicians to provide targeted advice and treatment to those people with T2D at most risk of developing CVD.

Sociodemographic characteristics and COVID-19 testing rates: spatiotemporal patterns and impact of test accessibility in Sweden.

BACKGROUND: Diagnostic testing is essential for disease surveillance and test-trace-isolate efforts. We aimed to investigate if residential area sociodemographic characteristics and test accessibility were associated with Coronavirus Disease 2019 (COVID-19) testing rates. METHODS: We included 426 224 patient-initiated COVID-19 polymerase chain reaction tests from Uppsala County in Sweden from 24 June 2020 to 9 February 2022. Using Poisson regression analyses, we investigated if postal code area Care Need Index (CNI; median 1.0, IQR 0.8-1.4), a composite measure of sociodemographic factors used in Sweden to allocate primary healthcare resources, was associated with COVID-19 daily testing rates after adjustments for community transmission. We assessed if the distance to testing station influenced testing, and performed a difference-in-difference-analysis of a new testing station targeting a disadvantaged neighbourhood. RESULTS: We observed that CNI, i.e. primary healthcare need, was negatively associated with COVID-19 testing rates in inhabitants 5-69 years. More pronounced differences were noted across younger age groups and in Uppsala City, with test rate ratios in children (5-14 years) ranging from 0.56 (95% CI 0.47-0.67) to 0.87 (95% CI 0.80-0.93) across three pandemic waves. Longer distance to the nearest testing station was linked to lower testing rates, e.g. every additional 10 km was associated with a 10-18% decrease in inhabitants 15-29 years in Uppsala County. The opening of the targeted testing station was associated with increased testing, including twice as high testing rates in individuals aged 70-105, supporting an intervention effect. CONCLUSIONS: Ensuring accessible testing across all residential areas constitutes a promising tool to decrease inequalities in testing.

Precision Prognostics for Cardiovascular Disease in Type 2 Diabetes: A Systematic Review and Meta-analysis.

Background Precision medicine has the potential to improve cardiovascular disease (CVD) risk prediction in individuals with type 2 diabetes (T2D). Methods We conducted a systematic review and meta-analysis of longitudinal studies to identify potentially novel prognostic factors that may improve CVD risk prediction in T2D. Out of 9380 studies identified, 416 studies met inclusion criteria. Outcomes were reported for 321 biomarker studies, 48 genetic marker studies, and 47 risk score/model studies. Results Out of all evaluated biomarkers, only 13 showed improvement in prediction performance. Results of pooled meta-analyses, non-pooled analyses, and assessments of improvement in prediction performance and risk of bias, yielded the highest predictive utility for N-terminal pro b-type natriuretic peptide (NT-proBNP) (high-evidence), troponin-T (TnT) (moderate-evidence), triglyceride-glucose (TyG) index (moderate-evidence), Genetic Risk Score for Coronary Heart Disease (GRS-CHD) (moderate-evidence); moderate predictive utility for coronary computed tomography angiography (low-evidence), single-photon emission computed tomography (low-evidence), pulse wave velocity (moderate-evidence); and low predictive utility for C-reactive protein (moderate-evidence), coronary artery calcium score (low-evidence), galectin-3 (low-evidence), troponin-I (low-evidence), carotid plaque (low-evidence), and growth differentiation factor-15 (low-evidence). Risk scores showed modest discrimination, with lower performance in populations different from the original development cohort. Conclusions Despite high interest in this topic, very few studies conducted rigorous analyses to demonstrate incremental predictive utility beyond established CVD risk factors for T2D. The most promising markers identified were NT-proBNP, TnT, TyG and GRS-CHD, with the highest strength of evidence for NT-proBNP. Further research is needed to determine their clinical utility in risk stratification and management of CVD in T2D.

Identification of biomarkers for glycaemic deterioration in type 2 diabetes.

We identify biomarkers for disease progression in three type 2 diabetes cohorts encompassing 2,973 individuals across three molecular classes, metabolites, lipids and proteins. Homocitrulline, isoleucine and 2-aminoadipic acid, eight triacylglycerol species, and lowered sphingomyelin 42:2;2 levels are predictive of faster progression towards insulin requirement. Of ~1,300 proteins examined in two cohorts, levels of GDF15/MIC-1, IL-18Ra, CRELD1, NogoR, FAS, and ENPP7 are associated with faster progression, whilst SMAC/DIABLO, SPOCK1 and HEMK2 predict lower progression rates. In an external replication, proteins and lipids are associated with diabetes incidence and prevalence. NogoR/RTN4R injection improved glucose tolerance in high fat-fed male mice but impaired it in male db/db mice. High NogoR levels led to islet cell apoptosis, and IL-18R antagonised inflammatory IL-18 signalling towards nuclear factor kappa-B in vitro. This comprehensive, multi-disciplinary approach thus identifies biomarkers with potential prognostic utility, provides evidence for possible disease mechanisms, and identifies potential therapeutic avenues to slow diabetes progression.

A phenome-wide comparative analysis of genetic discordance between obesity and type 2 diabetes.

Obesity and type 2 diabetes are causally related, yet there is considerable heterogeneity in the consequences of both conditions and the mechanisms of action are poorly defined. Here we show a genetic-driven approach defining two obesity profiles that convey highly concordant and discordant diabetogenic effects. We annotate and then compare association signals for these profiles across clinical and molecular phenotypic layers. Key differences are identified in a wide range of traits, including cardiovascular mortality, fat distribution, liver metabolism, blood pressure, specific lipid fractions and blood levels of proteins involved in extracellular matrix remodelling. We find marginal differences in abundance of Bacteroidetes and Firmicutes bacteria in the gut. Instrumental analyses reveal prominent causal roles for waist-to-hip ratio, blood pressure and cholesterol content of high-density lipoprotein particles in the development of diabetes in obesity. We prioritize 17 genes from the discordant signature that convey protection against type 2 diabetes in obesity, which may represent logical targets for precision medicine approaches.

Ethnic, gender and other sociodemographic biases in genome-wide association studies for the most burdensome non-communicable diseases: 2005-2022.

INTRODUCTION: Since 2005, disease-related human genetic diversity has been intensively characterized using genome-wide association studies (GWAS). Understanding how and by whom this work was performed may yield valuable insights into the generalizability of GWAS discoveries to global populations and how high-impact genetics research can be equitably sustained in the future. MATERIALS AND METHODS: We mined the NHGRI-EBI GWAS Catalog (2005-2022) for the most burdensome non-communicable causes of death worldwide. We then compared (i) the geographic, ethnic and socioeconomic characteristics of study populations; (ii) the geographic and socioeconomic characteristics of the regions within which researchers were located and (iii) the extent to which male and female investigators undertook and led the research. RESULTS: The research institutions leading the work are often US-based (37%), while the origin of samples is more diverse, with the Nordic countries having contributed as much data to GWAS as the United States (~17% of data). The majority of first (60%), senior (75%) and all (66%) authors are male; although proportions vary by disease and leadership level, male co-authors are the ubiquitous majority. The vast majority (91%) of complex trait GWAS has been performed in European ancestry populations, with cohorts and scientists predominantly located in medium-to-high socioeconomically ranked countries; apart from East Asians (~5%), other ethnicities rarely feature in published GWAS. See: https://hugofitipaldi.shinyapps.io/gwas_results/ to browse all results. CONCLUSION: Most GWAS cohorts are of European ancestry residing outside the United States, with a smaller yet meaningful proportion of East Asian ancestry. Papers describing GWAS research are predominantly authored by male scientists based in medium-to-high income countries.

Investigating the causal relationships between excess adiposity and cardiometabolic health in men and women.

AIMS/HYPOTHESIS: Excess adiposity is differentially associated with increased risk of cardiometabolic disease in men and women, according to observational studies. Causal inference studies largely assume a linear relationship between BMI and cardiometabolic outcomes, which may not be the case. In this study, we investigated the shapes of the causal relationships between BMI and cardiometabolic diseases and risk factors. We further investigated sex differences within the causal framework. METHODS: To assess causal relationships between BMI and the outcomes, we used two-stage least-squares Mendelian randomisation (MR), with a polygenic risk score for BMI as the instrumental variable. To elucidate the shapes of the causal relationships, we used a non-linear MR fractional polynomial method, and used piecewise MR to investigate threshold relationships and confirm the shapes. RESULTS: BMI was associated with type 2 diabetes (OR 3.10; 95% CI 2.73, 3.53), hypertension (OR 1.53; 95% CI 1.44, 1.62) and coronary artery disease (OR 1.20; 95% CI 1.08, 1.33), but not chronic kidney disease (OR 1.08; 95% CI 0.67, 1.72) or stroke (OR 1.08; 95% CI 0.92, 1.28). The data suggest that these relationships are non-linear. For cardiometabolic risk factors, BMI was positively associated with glucose, HbA1c, triacylglycerol levels and both systolic and diastolic BP. BMI had an inverse causal relationship with total cholesterol, LDL-cholesterol and HDL-cholesterol. The data suggest a non-linear causal relationship between BMI and BP and other biomarkers (p<0.001) except lipoprotein A. The piecewise MR results were consistent with the fractional polynomial results. The causal effect of BMI on coronary artery disease, total cholesterol and LDL-cholesterol was different in men and women, but this sex difference was only significant for LDL-cholesterol after controlling for multiple testing (p<0.001). Further, the causal effect of BMI on coronary artery disease varied by menopause status in women. CONCLUSIONS/INTERPRETATION: We describe the shapes of causal effects of BMI on cardiometabolic diseases and risk factors, and report sex differences in the causal effects of BMI on LDL-cholesterol. We found evidence of non-linearity in the causal effect of BMI on diseases and risk factor biomarkers. Reducing excess adiposity is highly beneficial for health, but there is greater need to consider biological sex in the management of adiposity.

App-based COVID-19 syndromic surveillance and prediction of hospital admissions in COVID Symptom Study Sweden.

The app-based COVID Symptom Study was launched in Sweden in April 2020 to contribute to real-time COVID-19 surveillance. We enrolled 143,531 study participants (≥18 years) who contributed 10.6 million daily symptom reports between April 29, 2020 and February 10, 2021. Here, we include data from 19,161 self-reported PCR tests to create a symptom-based model to estimate the individual probability of symptomatic COVID-19, with an AUC of 0.78 (95% CI 0.74-0.83) in an external dataset. These individual probabilities are employed to estimate daily regional COVID-19 prevalence, which are in turn used together with current hospital data to predict next week COVID-19 hospital admissions. We show that this hospital prediction model demonstrates a lower median absolute percentage error (MdAPE: 25.9%) across the five most populated regions in Sweden during the first pandemic wave than a model based on case notifications (MdAPE: 30.3%). During the second wave, the error rates are similar. When we apply the same model to an English dataset, not including local COVID-19 test data, we observe MdAPEs of 22.3% and 19.0% during the first and second pandemic waves, respectively, highlighting the transferability of the prediction model.

Distinct Molecular Signatures of Clinical Clusters in People With Type 2 Diabetes: An IMI-RHAPSODY Study.

Type 2 diabetes is a multifactorial disease with multiple underlying aetiologies. To address this heterogeneity, investigators of a previous study clustered people with diabetes according to five diabetes subtypes. The aim of the current study is to investigate the etiology of these clusters by comparing their molecular signatures. In three independent cohorts, in total 15,940 individuals were clustered based on five clinical characteristics. In a subset, genetic (N = 12,828), metabolomic (N = 2,945), lipidomic (N = 2,593), and proteomic (N = 1,170) data were obtained in plasma. For each data type, each cluster was compared with the other four clusters as the reference. The insulin-resistant cluster showed the most distinct molecular signature, with higher branched-chain amino acid, diacylglycerol, and triacylglycerol levels and aberrant protein levels in plasma were enriched for proteins in the intracellular PI3K/Akt pathway. The obese cluster showed higher levels of cytokines. The mild diabetes cluster with high HDL showed the most beneficial molecular profile with effects opposite of those seen in the insulin-resistant cluster. This study shows that clustering people with type 2 diabetes can identify underlying molecular mechanisms related to pancreatic islets, liver, and adipose tissue metabolism. This provides novel biological insights into the diverse aetiological processes that would not be evident when type 2 diabetes is viewed as a homogeneous disease.

Replication and cross-validation of type 2 diabetes subtypes based on clinical variables: an IMI-RHAPSODY study.

AIMS/HYPOTHESIS: Five clusters based on clinical characteristics have been suggested as diabetes subtypes: one autoimmune and four subtypes of type 2 diabetes. In the current study we replicate and cross-validate these type 2 diabetes clusters in three large cohorts using variables readily measured in the clinic. METHODS: In three independent cohorts, in total 15,940 individuals were clustered based on age, BMI, HbA1c, random or fasting C-peptide, and HDL-cholesterol. Clusters were cross-validated against the original clusters based on HOMA measures. In addition, between cohorts, clusters were cross-validated by re-assigning people based on each cohort's cluster centres. Finally, we compared the time to insulin requirement for each cluster. RESULTS: Five distinct type 2 diabetes clusters were identified and mapped back to the original four All New Diabetics in Scania (ANDIS) clusters. Using C-peptide and HDL-cholesterol instead of HOMA2-B and HOMA2-IR, three of the clusters mapped with high sensitivity (80.6-90.7%) to the previously identified severe insulin-deficient diabetes (SIDD), severe insulin-resistant diabetes (SIRD) and mild obesity-related diabetes (MOD) clusters. The previously described ANDIS mild age-related diabetes (MARD) cluster could be mapped to the two milder groups in our study: one characterised by high HDL-cholesterol (mild diabetes with high HDL-cholesterol [MDH] cluster), and the other not having any extreme characteristic (mild diabetes [MD]). When these two milder groups were combined, they mapped well to the previously labelled MARD cluster (sensitivity 79.1%). In the cross-validation between cohorts, particularly the SIDD and MDH clusters cross-validated well, with sensitivities ranging from 73.3% to 97.1%. SIRD and MD showed a lower sensitivity, ranging from 36.1% to 92.3%, where individuals shifted from SIRD to MD and vice versa. People belonging to the SIDD cluster showed the fastest progression towards insulin requirement, while the MDH cluster showed the slowest progression. CONCLUSIONS/INTERPRETATION: Clusters based on C-peptide instead of HOMA2 measures resemble those based on HOMA2 measures, especially for SIDD, SIRD and MOD. By adding HDL-cholesterol, the MARD cluster based upon HOMA2 measures resulted in the current clustering into two clusters, with one cluster having high HDL levels. Cross-validation between cohorts showed generally a good resemblance between cohorts. Together, our results show that the clustering based on clinical variables readily measured in the clinic (age, HbA1c, HDL-cholesterol, BMI and C-peptide) results in informative clusters that are representative of the original ANDIS clusters and stable across cohorts. Adding HDL-cholesterol to the clustering resulted in the identification of a cluster with very slow glycaemic deterioration.

Predicting and elucidating the etiology of fatty liver disease: A machine learning modeling and validation study in the IMI DIRECT cohorts.

BACKGROUND: Non-alcoholic fatty liver disease (NAFLD) is highly prevalent and causes serious health complications in individuals with and without type 2 diabetes (T2D). Early diagnosis of NAFLD is important, as this can help prevent irreversible damage to the liver and, ultimately, hepatocellular carcinomas. We sought to expand etiological understanding and develop a diagnostic tool for NAFLD using machine learning. METHODS AND FINDINGS: We utilized the baseline data from IMI DIRECT, a multicenter prospective cohort study of 3,029 European-ancestry adults recently diagnosed with T2D (n = 795) or at high risk of developing the disease (n = 2,234). Multi-omics (genetic, transcriptomic, proteomic, and metabolomic) and clinical (liver enzymes and other serological biomarkers, anthropometry, measures of beta-cell function, insulin sensitivity, and lifestyle) data comprised the key input variables. The models were trained on MRI-image-derived liver fat content (<5% or ≥5%) available for 1,514 participants. We applied LASSO (least absolute shrinkage and selection operator) to select features from the different layers of omics data and random forest analysis to develop the models. The prediction models included clinical and omics variables separately or in combination. A model including all omics and clinical variables yielded a cross-validated receiver operating characteristic area under the curve (ROCAUC) of 0.84 (95% CI 0.82, 0.86; p < 0.001), which compared with a ROCAUC of 0.82 (95% CI 0.81, 0.83; p < 0.001) for a model including 9 clinically accessible variables. The IMI DIRECT prediction models outperformed existing noninvasive NAFLD prediction tools. One limitation is that these analyses were performed in adults of European ancestry residing in northern Europe, and it is unknown how well these findings will translate to people of other ancestries and exposed to environmental risk factors that differ from those of the present cohort. Another key limitation of this study is that the prediction was done on a binary outcome of liver fat quantity (<5% or ≥5%) rather than a continuous one. CONCLUSIONS: In this study, we developed several models with different combinations of clinical and omics data and identified biological features that appear to be associated with liver fat accumulation. In general, the clinical variables showed better prediction ability than the complex omics variables. However, the combination of omics and clinical variables yielded the highest accuracy. We have incorporated the developed clinical models into a web interface (see: https://www.predictliverfat.org/) and made it available to the community. TRIAL REGISTRATION: ClinicalTrials.gov NCT03814915.

A Global Overview of Precision Medicine in Type 2 Diabetes.

The detailed characterization of human biology and behaviors is now possible at scale owing to innovations in biomarkers, bioimaging, and wearable technologies; "big data" from electronic medical records, health insurance databases, and other platforms becoming increasingly accessible; and rapidly evolving computational power and bioinformatics methods. Collectively, these advances are creating unprecedented opportunities to better understand diabetes and many other complex traits. Identifying hidden structures within these complex data sets and linking these structures to outcome data may yield unique insights into the risk factors and natural history of diabetes, which in turn may help optimize the prevention and management of the disease. This emerging area is broadly termed "precision medicine." In this Perspective, we give an overview of the evidence and barriers to the development and implementation of precision medicine in type 2 diabetes. We also discuss recently presented paradigms through which complex data might enhance our understanding of diabetes and ultimately our ability to tackle the disease more effectively than ever before.